Original Research

Droperidol is a butyrophenone antipsychotic approved by the US Food and Drug Administration (FDA) for use in postoperative nausea and vomiting (PONV). Off-label, it has also been utilized for its sedative, anxiolytic, and analgesic properties.¹ While its exact mechanism of action remains elusive, it is believed that binding to postsynaptic γ-aminobutyric acid receptors induces anxiolysis and sedation, while dopaminergic activity in the chemoreceptor trigger zone contributes to its antiemetic effects.² Since the introduction of droperidol in 1967, it has been widely used by emergency physicians, psychiatrists, and anesthesiologists globally.¹

Despite its therapeutic efficacy, use of droperidol has been tempered by concerns regarding its cardiovascular safety profile, specifically its potential to prolong the QT interval and precipitate cardiac arrhythmias. In 2001, the FDA placed a boxed warning on droperidol that mandated electrocardiogram (EKG) monitoring before and after treatment. This requirement has led to a widespread decrease in use, and the FDA decision sparked significant controversy among clinicians, with many organizations arguing that the evidence did not support this mandate.¹

Further review of the cases cited by the FDA revealed that there were 277 reported cases of droperidol-related adverse events (AEs), but many of these cases were duplicates and occurred outside the US.³ Additionally, the doses of droperidol used in these cases were significantly higher than the typical doses used in the emergency department (ED), ranging from 25 to 250 mg.⁴ Typical doses for PONV range from 0.625 to 2.5 mg intravenous (IV) or intramuscular (IM). Recommended doses for agitation typically range from 2.5 to 10 mg IV and 5 to 10 mg IM.⁵

There has been growing interest in reevaluating the risk-benefit profile of droperidol in the ED. Since the original decision by the FDA, multiple publications have challenged the idea that droperidol has significantly higher risks associated with its use. The 2014 review by the Clinical Guidelines Committee of the American Academy of Emergency Medicine did not find evidence that low-dose droperidol (< 2.5 is unsafe for use in the ED.⁶ A retrospective cohort study from 2020 found no fatalities in 5784 patients. Furthermore, a prospective observational study of 1009 patients at 6 EDs who received high-dose droperidol (≤ 20.0 mg) found no evidence of increased risk for QT prolongation.⁷ The evidence supports the safety of droperidol for use in prehospital and hospital settings as well as in pediatric, adult, and geriatric populations.^8-14 Droperidol was eventually reintroduced in 2019, which led to increased use.

The US Department of Veterans Affairs (VA) formulary has limited options (eg, haloperidol and olanzapine) that have robust evidence supporting their use to treat aggression or psychosis-related agitation. Ziprasidone injections are not on the formulary and require authorization for use, which may delay patient care and pose a safety risk. In 2021, VA Greater Los Angeles Healthcare System (VAGLAHS) received Pharmacy and Therapeutics Committee approval to use droperidol in the ED for agitation or nausea and vomiting. The purpose of this study was to evaluate safety outcomes for patients prescribed droperidol and the need for rescue medications (ie, effectiveness) in the VAGLAHS ED.

Methods

This retrospective chart review analyzed patients administered droperidol in the VAGLAHS ED from February 1, 2021, through April 30, 2023. A list of patients who had droperidol ordered in the VAGLAHS ED was obtained from the Veterans Health Information Systems and Technology Architecture. Charts were reviewed using the Computerized Patient Record System to confirm droperidol administration. Nurse documentation was reviewed to confirm the time, dose, and route of administration. In addition, droperidol dosages were categorized as < 5 mg, 5 to 10 mg, and > 10 mg to review outcomes based on the total amount administered to each patient.

Patients included in the study received droperidol in the ED within the study period, were aged ≥ 18 years, and received droperidol for acute agitation or antiemesis. Patients were excluded if they received droperidol for an indication other than agitation or antiemesis.

The study team reviewed the list of patients and audited the collected data. Reviewers were trained on the study protocols and variables identified. The following data were collected: patient demographics (age, sex, race, height, weight, allergies), Charlson Comorbidity Index (CCI) conditions, cardiac comorbidities, laboratory values at admission, basic metabolic panels, liver function tests, droperidol use (doses, indications, and documentation of safety), concomitant medications ordered with the initial droperidol order, AEs (arrhythmias, extrapyramidal symptoms [EPS], respiratory depression, mortality), medications used within 60 minutes of droperidol administration (rescue medications), other medications used within 24 hours after droperidol administration, and EKG/QTc (corrected QT interval) intervals. The data reviewed and recorded were from the date of the initial patient ED visit.

Outcomes

The primary outcome was all-cause mortality within 24 hours after droperidol administration. This outcome was measured in all patients included in this study. Secondary outcomes included rescue medications needed after droperidol administration, incidence of QT prolongation, incidence of EPS (defined as akathisia, dystonia, parkinsonism, or tardive dyskinesia), and incidence of respiratory depression. Clinically significant QTc was defined as an interval of ≥ 500 ms with incidence of arrhythmias, code blues, or intubations. Baseline risk factors for QTc prolongation were taken into consideration including electrolyte abnormalities, concomitant QT-prolonging medications, CCI score, and cardiac comorbidities. Incidence of EPS was counted if patients received medications such as diphenhydramine or benztropine after droperidol administration in addition to documentation of EPS signs and symptoms. Incidences of EPS findings were reviewed by emergency department physicians to confirm the diagnosis.

Safety was assessed by quantifying mortality rates 24 hours after droperidol administration along with incidence of AEs associated with droperidol use including QT prolongation, EPS, and respiratory depression.

The necessity of rescue medication use was assessed by nursing documentation, additional medications ordered, and/or no additional medications required for agitation within 60 minutes of droperidol administration. Sixty minutes was the chosen timeframe given that the onset of droperidol action is between 3 and 10 minutes and peaks in about 30 minutes. Medications that were considered rescue medications included diphenhydramine < 25 mg, diphenhydramine 25 to 50 mg, lorazepam < 1 mg, lorazepam 1 to 2 mg, diphenhydramine < 25 mg and lorazepam < 1 mg, diphenhydramine < 25 mg and lorazepam 1 to 2 mg, diphenhydramine 25 to 50 mg and lorazepam 1 to 2 mg, and other medications, the names and doses of which were manually documented by investigators.

Statistical Analysis

For all variables in the study, descriptive analysis was used to categorize findings. Microsoft Excel was used to calculate means, frequency counts, percentages, and categorize data.

Results

Between February 1, 2021, and April 16, 2023, 214 patients received droperidol in the VAGLAHS ED, and 207 patients were included in the study. Seven patients did not receive droperidol for the indications included (acute agitation or antiemesis). Most of the study population (89.4%) was male, and the mean age was 51.0 years. The mean CCI was 1.6. In the study, 183 (88.4%) patients received droperidol for agitation and 24 (11.6%) for nausea and vomiting (Table 1).

Primary Outcome

No deaths were observed in a 24-hour period after droperidol administration among the 207 patients included in the study. There were also no arrhythmias, code blues, or intubations observed with the administration of droperidol (Table 2).

Secondary Outcomes

A total of 144 patients (69.6%) received droperidol alone to resolve agitation or nausea and vomiting. In the remaining population, 63 (30.4%) patients were given medications concomitantly with droperidol.

Fifteen patients (7.2%) required rescue medications that were administered within 60 minutes of droperidol administration. Rescue medications were required for 7 patients (4.9%) who initially received droperidol alone compared with 8 patients (12.7%) who were administered concomitant medications with droperidol (Figure).

Extrapyramidal Symptoms

EPS occurred in 2 patients (1.0%). There was 1 incidence of tardive dyskinesia (TD) in which the patient received droperidol 2.5 mg IM for emesis. TD was resolved with diphenhydramine 50 mg. A second patient who experienced dystonia received droperidol 10 mg IM for agitation. Dystonia was resolved with benztropine 2 mg. Both patients had a CCI of 0, no cardiac comorbidities, and laboratory test results were within reference ranges. The second patient received olanzapine within 24 hours of droperidol administration; however, it was after the EPS event.

QTc Prolongation

Baseline EKGs (within 6 months prior to ED visit) were available for 102 patients (49.3%). Nine patients (8.8%) had a reported baseline QTc of ≥ 500 ms (Table 3). Of these patients, 6 had a repeat EKG and 5 had a repeat QTc < 500 ms. One patient had a baseline and repeated QTc of 512 ms with essentially no change after droperidol administration. Only 1 patient was on a potentially QTc-prolonging medication at home. None of the patients with baseline QTc > 500 ms experienced arrhythmias after droperidol administration.

We found that 59 patients (28.5%) had EKGs performed within 24 hours after droperidol administration. Five patients had documented QTc ≥ 500 ms, but no arrhythmias were observed in a 24-hour period. Table 4 describes the additional medications administered after the 60-minute window but within 24 hours after droperidol administration. Quetiapine 300 mg and metoclopramide 5 mg were the only medications documented that can potentially increase QTc. Patient adherence to home medications and the timing of the last dose prior to ED visit were unknown. However, no arrhythmias were noted in these patients with QT^c changes. No patients experienced respiratory depression within 24 hours of droperidol administration.

Older Adult Patients

Thirty-eight patients were aged ≥ 65 years with a mean age of 74.2 years. Thirty-four patients (89.5%) received droperidol for agitation and 4 (10.6%) for nausea and vomiting. Only 21 patients had a baseline EKG, and 4 had QTc ≤ 500 ms. At 24 hours, EKGs were performed for 18 patients and 3 had a QTc ≤ 500 ms. No mortality or arrhythmias were reported and there were no incidences of rescue medications, EPS, or respiratory depression.

Discussion

The study included 207 patients who received droperidol for either agitation or nausea/vomiting in the VAGLAHS ED. No mortality occurred within 24 hours of droperidol administration, which is consistent with recent studies.^8-14

Furthermore, 59 patients (28.5%) had an EKG performed within 24 hours of droperidol administration; 5 patients had documented QTc ≥ 500 ms. Only 3 of the patients with prolonged QTc had baseline readings for comparison. Only 2 patients had an increase in QTc interval. No arrhythmias were observed; however, the effects of observing QTc prolongation were limited due to the lack of post-EKG readings following droperidol administration. Because of the retrospective nature of the study, neither standardization of EKG at baseline nor 24-hour postadministration were possible. The study found that droperidol was effective with only 15 patients (7.3%) requiring rescue medications. In the patients who were given medications concomitantly with droperidol, it was not possible to conclude whether the patients would have required rescue medications to resolve their agitation or nausea/vomiting. Administration of concomitant medications with droperidol may be attributed to practice patterns associated with haloperidol use, which is frequently administered with concomitant medications such as diphenhydramine and/or a benzodiazepine.

AEs were rare with no documentation of respiratory depression and 2 cases (1.0%) of EPS. Both incidences of EPS resolved with diphenhydramine or benztropine. However, given the reliance on nursing documentation to capture AEs, the number of events may have been underreported.

Limitations

Standardization of dosing was a limiting factor that could affect the need for rescue medications. Another limitation was reliance on nursing reports of resolution of symptoms and comfort with agitated patients. Given the retrospective design and small sample size, this study may not have captured all potential AEs. However, the doses administered within this study population were consistent with what was expected based on other studies.^8-14

Conclusions

Droperidol, an antipsychotic, is currently approved for PONV, but is also used off-label for agitation. This study found no fatalities among patients who received droperidol in the ED. The findings suggest that droperidol used for agitation and as an antiemetic, despite its FDA boxed warning, appears to be safe and showed no evidence of mortality, arrhythmias, code blues, or intubations despite the lack of postdose EKG monitoring. Among the 38 patients aged ≥ 65 years, the use of droperidol revealed no increased risks. It should be noted that droperidol appeared safe and few patients required rescue medications within this study population.

Droperidol is a butyrophenone antipsychotic approved by the US Food and Drug Administration (FDA) for use in postoperative nausea and vomiting (PONV). Off-label, it has also been utilized for its sedative, anxiolytic, and analgesic properties.¹ While its exact mechanism of action remains elusive, it is believed that binding to postsynaptic γ-aminobutyric acid receptors induces anxiolysis and sedation, while dopaminergic activity in the chemoreceptor trigger zone contributes to its antiemetic effects.² Since the introduction of droperidol in 1967, it has been widely used by emergency physicians, psychiatrists, and anesthesiologists globally.¹

Despite its therapeutic efficacy, use of droperidol has been tempered by concerns regarding its cardiovascular safety profile, specifically its potential to prolong the QT interval and precipitate cardiac arrhythmias. In 2001, the FDA placed a boxed warning on droperidol that mandated electrocardiogram (EKG) monitoring before and after treatment. This requirement has led to a widespread decrease in use, and the FDA decision sparked significant controversy among clinicians, with many organizations arguing that the evidence did not support this mandate.¹

Further review of the cases cited by the FDA revealed that there were 277 reported cases of droperidol-related adverse events (AEs), but many of these cases were duplicates and occurred outside the US.³ Additionally, the doses of droperidol used in these cases were significantly higher than the typical doses used in the emergency department (ED), ranging from 25 to 250 mg.⁴ Typical doses for PONV range from 0.625 to 2.5 mg intravenous (IV) or intramuscular (IM). Recommended doses for agitation typically range from 2.5 to 10 mg IV and 5 to 10 mg IM.⁵

There has been growing interest in reevaluating the risk-benefit profile of droperidol in the ED. Since the original decision by the FDA, multiple publications have challenged the idea that droperidol has significantly higher risks associated with its use. The 2014 review by the Clinical Guidelines Committee of the American Academy of Emergency Medicine did not find evidence that low-dose droperidol (< 2.5 is unsafe for use in the ED.⁶ A retrospective cohort study from 2020 found no fatalities in 5784 patients. Furthermore, a prospective observational study of 1009 patients at 6 EDs who received high-dose droperidol (≤ 20.0 mg) found no evidence of increased risk for QT prolongation.⁷ The evidence supports the safety of droperidol for use in prehospital and hospital settings as well as in pediatric, adult, and geriatric populations.^8-14 Droperidol was eventually reintroduced in 2019, which led to increased use.

The US Department of Veterans Affairs (VA) formulary has limited options (eg, haloperidol and olanzapine) that have robust evidence supporting their use to treat aggression or psychosis-related agitation. Ziprasidone injections are not on the formulary and require authorization for use, which may delay patient care and pose a safety risk. In 2021, VA Greater Los Angeles Healthcare System (VAGLAHS) received Pharmacy and Therapeutics Committee approval to use droperidol in the ED for agitation or nausea and vomiting. The purpose of this study was to evaluate safety outcomes for patients prescribed droperidol and the need for rescue medications (ie, effectiveness) in the VAGLAHS ED.

Methods

This retrospective chart review analyzed patients administered droperidol in the VAGLAHS ED from February 1, 2021, through April 30, 2023. A list of patients who had droperidol ordered in the VAGLAHS ED was obtained from the Veterans Health Information Systems and Technology Architecture. Charts were reviewed using the Computerized Patient Record System to confirm droperidol administration. Nurse documentation was reviewed to confirm the time, dose, and route of administration. In addition, droperidol dosages were categorized as < 5 mg, 5 to 10 mg, and > 10 mg to review outcomes based on the total amount administered to each patient.

Patients included in the study received droperidol in the ED within the study period, were aged ≥ 18 years, and received droperidol for acute agitation or antiemesis. Patients were excluded if they received droperidol for an indication other than agitation or antiemesis.

The study team reviewed the list of patients and audited the collected data. Reviewers were trained on the study protocols and variables identified. The following data were collected: patient demographics (age, sex, race, height, weight, allergies), Charlson Comorbidity Index (CCI) conditions, cardiac comorbidities, laboratory values at admission, basic metabolic panels, liver function tests, droperidol use (doses, indications, and documentation of safety), concomitant medications ordered with the initial droperidol order, AEs (arrhythmias, extrapyramidal symptoms [EPS], respiratory depression, mortality), medications used within 60 minutes of droperidol administration (rescue medications), other medications used within 24 hours after droperidol administration, and EKG/QTc (corrected QT interval) intervals. The data reviewed and recorded were from the date of the initial patient ED visit.

Outcomes

The primary outcome was all-cause mortality within 24 hours after droperidol administration. This outcome was measured in all patients included in this study. Secondary outcomes included rescue medications needed after droperidol administration, incidence of QT prolongation, incidence of EPS (defined as akathisia, dystonia, parkinsonism, or tardive dyskinesia), and incidence of respiratory depression. Clinically significant QTc was defined as an interval of ≥ 500 ms with incidence of arrhythmias, code blues, or intubations. Baseline risk factors for QTc prolongation were taken into consideration including electrolyte abnormalities, concomitant QT-prolonging medications, CCI score, and cardiac comorbidities. Incidence of EPS was counted if patients received medications such as diphenhydramine or benztropine after droperidol administration in addition to documentation of EPS signs and symptoms. Incidences of EPS findings were reviewed by emergency department physicians to confirm the diagnosis.

Safety was assessed by quantifying mortality rates 24 hours after droperidol administration along with incidence of AEs associated with droperidol use including QT prolongation, EPS, and respiratory depression.

The necessity of rescue medication use was assessed by nursing documentation, additional medications ordered, and/or no additional medications required for agitation within 60 minutes of droperidol administration. Sixty minutes was the chosen timeframe given that the onset of droperidol action is between 3 and 10 minutes and peaks in about 30 minutes. Medications that were considered rescue medications included diphenhydramine < 25 mg, diphenhydramine 25 to 50 mg, lorazepam < 1 mg, lorazepam 1 to 2 mg, diphenhydramine < 25 mg and lorazepam < 1 mg, diphenhydramine < 25 mg and lorazepam 1 to 2 mg, diphenhydramine 25 to 50 mg and lorazepam 1 to 2 mg, and other medications, the names and doses of which were manually documented by investigators.

Statistical Analysis

For all variables in the study, descriptive analysis was used to categorize findings. Microsoft Excel was used to calculate means, frequency counts, percentages, and categorize data.

Results

Between February 1, 2021, and April 16, 2023, 214 patients received droperidol in the VAGLAHS ED, and 207 patients were included in the study. Seven patients did not receive droperidol for the indications included (acute agitation or antiemesis). Most of the study population (89.4%) was male, and the mean age was 51.0 years. The mean CCI was 1.6. In the study, 183 (88.4%) patients received droperidol for agitation and 24 (11.6%) for nausea and vomiting (Table 1).

Primary Outcome

No deaths were observed in a 24-hour period after droperidol administration among the 207 patients included in the study. There were also no arrhythmias, code blues, or intubations observed with the administration of droperidol (Table 2).

Secondary Outcomes

A total of 144 patients (69.6%) received droperidol alone to resolve agitation or nausea and vomiting. In the remaining population, 63 (30.4%) patients were given medications concomitantly with droperidol.

Fifteen patients (7.2%) required rescue medications that were administered within 60 minutes of droperidol administration. Rescue medications were required for 7 patients (4.9%) who initially received droperidol alone compared with 8 patients (12.7%) who were administered concomitant medications with droperidol (Figure).

Extrapyramidal Symptoms

EPS occurred in 2 patients (1.0%). There was 1 incidence of tardive dyskinesia (TD) in which the patient received droperidol 2.5 mg IM for emesis. TD was resolved with diphenhydramine 50 mg. A second patient who experienced dystonia received droperidol 10 mg IM for agitation. Dystonia was resolved with benztropine 2 mg. Both patients had a CCI of 0, no cardiac comorbidities, and laboratory test results were within reference ranges. The second patient received olanzapine within 24 hours of droperidol administration; however, it was after the EPS event.

QTc Prolongation

Baseline EKGs (within 6 months prior to ED visit) were available for 102 patients (49.3%). Nine patients (8.8%) had a reported baseline QTc of ≥ 500 ms (Table 3). Of these patients, 6 had a repeat EKG and 5 had a repeat QTc < 500 ms. One patient had a baseline and repeated QTc of 512 ms with essentially no change after droperidol administration. Only 1 patient was on a potentially QTc-prolonging medication at home. None of the patients with baseline QTc > 500 ms experienced arrhythmias after droperidol administration.

We found that 59 patients (28.5%) had EKGs performed within 24 hours after droperidol administration. Five patients had documented QTc ≥ 500 ms, but no arrhythmias were observed in a 24-hour period. Table 4 describes the additional medications administered after the 60-minute window but within 24 hours after droperidol administration. Quetiapine 300 mg and metoclopramide 5 mg were the only medications documented that can potentially increase QTc. Patient adherence to home medications and the timing of the last dose prior to ED visit were unknown. However, no arrhythmias were noted in these patients with QT^c changes. No patients experienced respiratory depression within 24 hours of droperidol administration.

Older Adult Patients

Thirty-eight patients were aged ≥ 65 years with a mean age of 74.2 years. Thirty-four patients (89.5%) received droperidol for agitation and 4 (10.6%) for nausea and vomiting. Only 21 patients had a baseline EKG, and 4 had QTc ≤ 500 ms. At 24 hours, EKGs were performed for 18 patients and 3 had a QTc ≤ 500 ms. No mortality or arrhythmias were reported and there were no incidences of rescue medications, EPS, or respiratory depression.

Discussion

The study included 207 patients who received droperidol for either agitation or nausea/vomiting in the VAGLAHS ED. No mortality occurred within 24 hours of droperidol administration, which is consistent with recent studies.^8-14

Furthermore, 59 patients (28.5%) had an EKG performed within 24 hours of droperidol administration; 5 patients had documented QTc ≥ 500 ms. Only 3 of the patients with prolonged QTc had baseline readings for comparison. Only 2 patients had an increase in QTc interval. No arrhythmias were observed; however, the effects of observing QTc prolongation were limited due to the lack of post-EKG readings following droperidol administration. Because of the retrospective nature of the study, neither standardization of EKG at baseline nor 24-hour postadministration were possible. The study found that droperidol was effective with only 15 patients (7.3%) requiring rescue medications. In the patients who were given medications concomitantly with droperidol, it was not possible to conclude whether the patients would have required rescue medications to resolve their agitation or nausea/vomiting. Administration of concomitant medications with droperidol may be attributed to practice patterns associated with haloperidol use, which is frequently administered with concomitant medications such as diphenhydramine and/or a benzodiazepine.

AEs were rare with no documentation of respiratory depression and 2 cases (1.0%) of EPS. Both incidences of EPS resolved with diphenhydramine or benztropine. However, given the reliance on nursing documentation to capture AEs, the number of events may have been underreported.

Limitations

Standardization of dosing was a limiting factor that could affect the need for rescue medications. Another limitation was reliance on nursing reports of resolution of symptoms and comfort with agitated patients. Given the retrospective design and small sample size, this study may not have captured all potential AEs. However, the doses administered within this study population were consistent with what was expected based on other studies.^8-14

Conclusions

Droperidol, an antipsychotic, is currently approved for PONV, but is also used off-label for agitation. This study found no fatalities among patients who received droperidol in the ED. The findings suggest that droperidol used for agitation and as an antiemetic, despite its FDA boxed warning, appears to be safe and showed no evidence of mortality, arrhythmias, code blues, or intubations despite the lack of postdose EKG monitoring. Among the 38 patients aged ≥ 65 years, the use of droperidol revealed no increased risks. It should be noted that droperidol appeared safe and few patients required rescue medications within this study population.

Droperidol is a butyrophenone antipsychotic approved by the US Food and Drug Administration (FDA) for use in postoperative nausea and vomiting (PONV). Off-label, it has also been utilized for its sedative, anxiolytic, and analgesic properties.¹ While its exact mechanism of action remains elusive, it is believed that binding to postsynaptic γ-aminobutyric acid receptors induces anxiolysis and sedation, while dopaminergic activity in the chemoreceptor trigger zone contributes to its antiemetic effects.² Since the introduction of droperidol in 1967, it has been widely used by emergency physicians, psychiatrists, and anesthesiologists globally.¹

Despite its therapeutic efficacy, use of droperidol has been tempered by concerns regarding its cardiovascular safety profile, specifically its potential to prolong the QT interval and precipitate cardiac arrhythmias. In 2001, the FDA placed a boxed warning on droperidol that mandated electrocardiogram (EKG) monitoring before and after treatment. This requirement has led to a widespread decrease in use, and the FDA decision sparked significant controversy among clinicians, with many organizations arguing that the evidence did not support this mandate.¹

Further review of the cases cited by the FDA revealed that there were 277 reported cases of droperidol-related adverse events (AEs), but many of these cases were duplicates and occurred outside the US.³ Additionally, the doses of droperidol used in these cases were significantly higher than the typical doses used in the emergency department (ED), ranging from 25 to 250 mg.⁴ Typical doses for PONV range from 0.625 to 2.5 mg intravenous (IV) or intramuscular (IM). Recommended doses for agitation typically range from 2.5 to 10 mg IV and 5 to 10 mg IM.⁵

There has been growing interest in reevaluating the risk-benefit profile of droperidol in the ED. Since the original decision by the FDA, multiple publications have challenged the idea that droperidol has significantly higher risks associated with its use. The 2014 review by the Clinical Guidelines Committee of the American Academy of Emergency Medicine did not find evidence that low-dose droperidol (< 2.5 is unsafe for use in the ED.⁶ A retrospective cohort study from 2020 found no fatalities in 5784 patients. Furthermore, a prospective observational study of 1009 patients at 6 EDs who received high-dose droperidol (≤ 20.0 mg) found no evidence of increased risk for QT prolongation.⁷ The evidence supports the safety of droperidol for use in prehospital and hospital settings as well as in pediatric, adult, and geriatric populations.^8-14 Droperidol was eventually reintroduced in 2019, which led to increased use.

The US Department of Veterans Affairs (VA) formulary has limited options (eg, haloperidol and olanzapine) that have robust evidence supporting their use to treat aggression or psychosis-related agitation. Ziprasidone injections are not on the formulary and require authorization for use, which may delay patient care and pose a safety risk. In 2021, VA Greater Los Angeles Healthcare System (VAGLAHS) received Pharmacy and Therapeutics Committee approval to use droperidol in the ED for agitation or nausea and vomiting. The purpose of this study was to evaluate safety outcomes for patients prescribed droperidol and the need for rescue medications (ie, effectiveness) in the VAGLAHS ED.

Methods

This retrospective chart review analyzed patients administered droperidol in the VAGLAHS ED from February 1, 2021, through April 30, 2023. A list of patients who had droperidol ordered in the VAGLAHS ED was obtained from the Veterans Health Information Systems and Technology Architecture. Charts were reviewed using the Computerized Patient Record System to confirm droperidol administration. Nurse documentation was reviewed to confirm the time, dose, and route of administration. In addition, droperidol dosages were categorized as < 5 mg, 5 to 10 mg, and > 10 mg to review outcomes based on the total amount administered to each patient.

Patients included in the study received droperidol in the ED within the study period, were aged ≥ 18 years, and received droperidol for acute agitation or antiemesis. Patients were excluded if they received droperidol for an indication other than agitation or antiemesis.

The study team reviewed the list of patients and audited the collected data. Reviewers were trained on the study protocols and variables identified. The following data were collected: patient demographics (age, sex, race, height, weight, allergies), Charlson Comorbidity Index (CCI) conditions, cardiac comorbidities, laboratory values at admission, basic metabolic panels, liver function tests, droperidol use (doses, indications, and documentation of safety), concomitant medications ordered with the initial droperidol order, AEs (arrhythmias, extrapyramidal symptoms [EPS], respiratory depression, mortality), medications used within 60 minutes of droperidol administration (rescue medications), other medications used within 24 hours after droperidol administration, and EKG/QTc (corrected QT interval) intervals. The data reviewed and recorded were from the date of the initial patient ED visit.

Outcomes

The primary outcome was all-cause mortality within 24 hours after droperidol administration. This outcome was measured in all patients included in this study. Secondary outcomes included rescue medications needed after droperidol administration, incidence of QT prolongation, incidence of EPS (defined as akathisia, dystonia, parkinsonism, or tardive dyskinesia), and incidence of respiratory depression. Clinically significant QTc was defined as an interval of ≥ 500 ms with incidence of arrhythmias, code blues, or intubations. Baseline risk factors for QTc prolongation were taken into consideration including electrolyte abnormalities, concomitant QT-prolonging medications, CCI score, and cardiac comorbidities. Incidence of EPS was counted if patients received medications such as diphenhydramine or benztropine after droperidol administration in addition to documentation of EPS signs and symptoms. Incidences of EPS findings were reviewed by emergency department physicians to confirm the diagnosis.

Safety was assessed by quantifying mortality rates 24 hours after droperidol administration along with incidence of AEs associated with droperidol use including QT prolongation, EPS, and respiratory depression.

The necessity of rescue medication use was assessed by nursing documentation, additional medications ordered, and/or no additional medications required for agitation within 60 minutes of droperidol administration. Sixty minutes was the chosen timeframe given that the onset of droperidol action is between 3 and 10 minutes and peaks in about 30 minutes. Medications that were considered rescue medications included diphenhydramine < 25 mg, diphenhydramine 25 to 50 mg, lorazepam < 1 mg, lorazepam 1 to 2 mg, diphenhydramine < 25 mg and lorazepam < 1 mg, diphenhydramine < 25 mg and lorazepam 1 to 2 mg, diphenhydramine 25 to 50 mg and lorazepam 1 to 2 mg, and other medications, the names and doses of which were manually documented by investigators.

Statistical Analysis

For all variables in the study, descriptive analysis was used to categorize findings. Microsoft Excel was used to calculate means, frequency counts, percentages, and categorize data.

Results

Between February 1, 2021, and April 16, 2023, 214 patients received droperidol in the VAGLAHS ED, and 207 patients were included in the study. Seven patients did not receive droperidol for the indications included (acute agitation or antiemesis). Most of the study population (89.4%) was male, and the mean age was 51.0 years. The mean CCI was 1.6. In the study, 183 (88.4%) patients received droperidol for agitation and 24 (11.6%) for nausea and vomiting (Table 1).

Primary Outcome

No deaths were observed in a 24-hour period after droperidol administration among the 207 patients included in the study. There were also no arrhythmias, code blues, or intubations observed with the administration of droperidol (Table 2).

Secondary Outcomes

A total of 144 patients (69.6%) received droperidol alone to resolve agitation or nausea and vomiting. In the remaining population, 63 (30.4%) patients were given medications concomitantly with droperidol.

Fifteen patients (7.2%) required rescue medications that were administered within 60 minutes of droperidol administration. Rescue medications were required for 7 patients (4.9%) who initially received droperidol alone compared with 8 patients (12.7%) who were administered concomitant medications with droperidol (Figure).

Extrapyramidal Symptoms

EPS occurred in 2 patients (1.0%). There was 1 incidence of tardive dyskinesia (TD) in which the patient received droperidol 2.5 mg IM for emesis. TD was resolved with diphenhydramine 50 mg. A second patient who experienced dystonia received droperidol 10 mg IM for agitation. Dystonia was resolved with benztropine 2 mg. Both patients had a CCI of 0, no cardiac comorbidities, and laboratory test results were within reference ranges. The second patient received olanzapine within 24 hours of droperidol administration; however, it was after the EPS event.

QTc Prolongation

Baseline EKGs (within 6 months prior to ED visit) were available for 102 patients (49.3%). Nine patients (8.8%) had a reported baseline QTc of ≥ 500 ms (Table 3). Of these patients, 6 had a repeat EKG and 5 had a repeat QTc < 500 ms. One patient had a baseline and repeated QTc of 512 ms with essentially no change after droperidol administration. Only 1 patient was on a potentially QTc-prolonging medication at home. None of the patients with baseline QTc > 500 ms experienced arrhythmias after droperidol administration.

We found that 59 patients (28.5%) had EKGs performed within 24 hours after droperidol administration. Five patients had documented QTc ≥ 500 ms, but no arrhythmias were observed in a 24-hour period. Table 4 describes the additional medications administered after the 60-minute window but within 24 hours after droperidol administration. Quetiapine 300 mg and metoclopramide 5 mg were the only medications documented that can potentially increase QTc. Patient adherence to home medications and the timing of the last dose prior to ED visit were unknown. However, no arrhythmias were noted in these patients with QT^c changes. No patients experienced respiratory depression within 24 hours of droperidol administration.

Older Adult Patients

Thirty-eight patients were aged ≥ 65 years with a mean age of 74.2 years. Thirty-four patients (89.5%) received droperidol for agitation and 4 (10.6%) for nausea and vomiting. Only 21 patients had a baseline EKG, and 4 had QTc ≤ 500 ms. At 24 hours, EKGs were performed for 18 patients and 3 had a QTc ≤ 500 ms. No mortality or arrhythmias were reported and there were no incidences of rescue medications, EPS, or respiratory depression.

Discussion

The study included 207 patients who received droperidol for either agitation or nausea/vomiting in the VAGLAHS ED. No mortality occurred within 24 hours of droperidol administration, which is consistent with recent studies.^8-14

Furthermore, 59 patients (28.5%) had an EKG performed within 24 hours of droperidol administration; 5 patients had documented QTc ≥ 500 ms. Only 3 of the patients with prolonged QTc had baseline readings for comparison. Only 2 patients had an increase in QTc interval. No arrhythmias were observed; however, the effects of observing QTc prolongation were limited due to the lack of post-EKG readings following droperidol administration. Because of the retrospective nature of the study, neither standardization of EKG at baseline nor 24-hour postadministration were possible. The study found that droperidol was effective with only 15 patients (7.3%) requiring rescue medications. In the patients who were given medications concomitantly with droperidol, it was not possible to conclude whether the patients would have required rescue medications to resolve their agitation or nausea/vomiting. Administration of concomitant medications with droperidol may be attributed to practice patterns associated with haloperidol use, which is frequently administered with concomitant medications such as diphenhydramine and/or a benzodiazepine.

AEs were rare with no documentation of respiratory depression and 2 cases (1.0%) of EPS. Both incidences of EPS resolved with diphenhydramine or benztropine. However, given the reliance on nursing documentation to capture AEs, the number of events may have been underreported.

Limitations

Standardization of dosing was a limiting factor that could affect the need for rescue medications. Another limitation was reliance on nursing reports of resolution of symptoms and comfort with agitated patients. Given the retrospective design and small sample size, this study may not have captured all potential AEs. However, the doses administered within this study population were consistent with what was expected based on other studies.^8-14

Conclusions

Droperidol, an antipsychotic, is currently approved for PONV, but is also used off-label for agitation. This study found no fatalities among patients who received droperidol in the ED. The findings suggest that droperidol used for agitation and as an antiemetic, despite its FDA boxed warning, appears to be safe and showed no evidence of mortality, arrhythmias, code blues, or intubations despite the lack of postdose EKG monitoring. Among the 38 patients aged ≥ 65 years, the use of droperidol revealed no increased risks. It should be noted that droperidol appeared safe and few patients required rescue medications within this study population.

Calcitonin gene-related peptide (CGRP) is a neuropeptide that plays a key role in migraine pathophysiology by promoting the dilation of cerebral blood vessels and transmitting pain signals.¹ CGRP has generated interest for the prevention and acute treatment of migraine. Since 2018, 8 novel CGRP-targeting therapies have been approved by the US Food and Drug Administration (FDA) for the management of migraines.^2,3 For migraine prevention, there are 4 injectable monoclonal antibodies (mAbs) directed against the CGRP receptor (erenumab) or the CGRP ligand (fremanezumab, galcanezumab, and eptinezumab). There are also 2 oral small-molecule CGRP receptor antagonists, termed gepants, that also are approved for migraine prevention (atogepant and rimegepant). Three gepants are approved for acute migraine treatment and are administered orally (rimegepant and ubrogepant) or intranasally (zavegepant) (Table 1).

CGRP-targeting therapies have received attention for their role in vasodilation within the cerebral, coronary, and renal vasculature.⁴ CGRP-mediated vasodilatory effects cause systemic regulation of blood pressure (BP) and play a protective role in hypertension.² Some studies, particularly with erenumab, have shown that the inhibitory role of the agent leads to an increase in BP, as well as gastrointestinal issues such as constipation.^2,5 The FDA recently updated monitoring recommendations for all CGRP-targeting therapies to include the potential for BP elevations and hypertension. Outside of this, there is no definitive evidence linking dual CGRP-targeted therapy to higher cardiovascular or gastrointestinal risks and prescribing information does not carry contraindications.⁶

In a 2021 consensus statement, the American Headache Society (AHS) recommended CGRP-targeting therapies for migraine prevention after inability to tolerate or inadequate response to an 8-week trial of ≥ 2 drug classes including antihypertensives, antiseizure medications, antidepressants, and onabotulinumtoxinA.⁷ For acute treatment, AHS recommended gepant use after contraindication to or inadequate response to ≥ 2 triptans. Guidance on combination CGRP-targeting therapies for both prevention and acute treatment was not provided.⁷ More recently, the AHS published a position statement noting substantial efficacy and safety data for CGRP-targeting therapies and suggested its consideration as a first-line option for migraine prevention, though use for acute treatment or combination CGRP-targeting therapies for both prevention and acute treatment were not addressed.⁸

The International Headache Society guidelines for the acute treatment of migraines recommend nonopioid analgesics as first-line therapy for mild migraine attacks. For moderate to severe attacks, triptans with or without a nonopioid analgesic were recommended as first-line therapy, prior to consideration of CGRP-targeted therapy.⁹ The increased use of this new drug class has also led to combination use of CGRP-targeting therapies for migraine prevention and acute treatment as seen in clinical practice and reflected by some case reports, case series, and small studies describing such use.^10-14 In light of the similar mechanism of action of these therapies and the physiologic role of CGRP, there have been calls for safety evaluation.¹⁵

To our knowledge, no studies have evaluated dual CGRP-targeting regimens for migraine in the veteran population. In 2023, the US Department of Veterans Affairs (VA) and US Department of Defense (DoD) updated their clinical practice guidelines for the management of headache.³ For migraine prevention, the VA/DoD guidelines include a strong recommendation for the use of erenumab, fremanezumab, and galcanezumab; a weak recommendation for the use of atogepant; and a recommendation neither for nor against the use of rimegepant. For acute treatment, the guidelines assign a weak recommendation for the use of rimegepant and ubrogepant. Combination use was not addressed.³

Prior to the VA/DoD guidelines, the Veterans Health Administration restricted the dual use of CGRP-targeting therapies for both preventive and acute migraine treatment. However, the VA Pharmacy Benefit Management Service removed the restriction in the Criteria for Use documents, allowing broader access to these medications for veterans.^16-22 This change permits the use of CGRP-targeting drugs for both acute and preventive migraine treatment after initial data reflecting real-world case reports and open-label studies suggested possible efficacy without a clear safety concern.^11,12 This study aims to fill the gap in the literature by evaluating the safety, efficacy, and overall outcomes of combination CGRP-targeting treatment for migraine prevention and acute treatment in a veteran population.

Methods

This single-center, retrospective, medication use evaluation at the Ralph H. Johnson VA Medical Center (RHJVAMC) was reviewed by the RHJVAMC Research and Development Committee and Quality Improvement Program Evaluation Self Certification Tool, which both determined that institutional review board approval was not required because it was considered part of routine care and quality improvement. Computerized Patient Record System (CPRS) data were reviewed between April 1, 2023 (after the Criteria for Use for CGRP-targeting therapies was updated), through January 31, 2025. Patients were included if they had a confirmed diagnosis of migraine using the International Classification of Headache Disorders, 3rd edition criteria and had concomitant active prescriptions for both a preventive and acute treatment CGRP-targeting agent during the project period.²³ Only patients receiving care from the RHJVAMC neurology department were included.

The primary objective was to assess the safety of dual CGRP-targeting therapies for migraine treatment. Key safety endpoints included effects on liver function, kidney function, and BP. Safety outcomes were graded using Common Terminology Criteria for Adverse Events.²⁴ Changes in liver function were categorized as grade 1, 2, or 3 elevations: grade 1 (aspartate aminotransferase [AST]/alanine aminotransferase [ALT] up to 3x the upper limit of normal [ULN] or bilirubin > 1.5 x ULN); grade 2 (AST/ALT 3-6 x ULN or bilirubin 1.5-3 x ULN); and grade 3 (AST/ALT 5-10 x ULN or bilirubin 3-10 x ULN). Kidney function changes were assessed by serum creatinine levels using a similar grading system: Grade 1 (≤ 1.5 x ULN); grade 2 (1.5-3 x baseline of normal); and grade 3 (3-6 x ULN or baseline). Changes in BP were monitored from baseline to the time of the first neurology follow-up. Elevations were grouped into 2 categories, defined as BP ≥ 140 mm Hg systolic and/or 90 mm Hg diastolic (category 1) and ≥ 160 mm Hg systolic and/or 100 mm Hg diastolic (category 2). Neurology documentation was also reviewed in CPRS for individual patient-reported adverse effects (AEs). Safety endpoints were tracked for any occurrence during the project period.

The secondary objective was to describe the patient-reported efficacy of adding a gepant for acute migraine treatment to existing CGRP-targeting therapies for migraine prevention, in those patients who were stable for ≥ 12 weeks on the preventive therapy. Neurology documentation of headache characteristics, including headache severity as rated on a numerical pain score from 0 (no pain) to 10 (worst pain), and duration of headaches (in hours) were recorded during the project period. Changes in headache characteristics were tracked from baseline (ie, the neurology visit when the gepant was first requested) to the first neurology follow-up within 6 months of initiating gepant for acute treatment. If ranges were provided within documentation, a mean was calculated and used for data collection. Neurology documentation was also reviewed for any patient report of overall effectiveness with the added gepant, and categorized as symptoms improved, worsened, or did not change based on subjective report. Descriptive statistics were used for data analysis. A 1-sample Wilcoxon signed rank test was performed as an exploratory analysis for change in headache characteristics from baseline to first neurology follow-up within 6 months. Each individual CGRP regimen was counted as a unique data point to adequately describe changes associated with each new medication and/or dose adjustment. Therefore, patients could be included more than once to account for each distinct treatment regimen.

Results

From April 1, 2023, to January 31, 2025, 96 patients were identified with active prescriptions for dual CGRP-targeting therapies. Of the 96 patients, 89 were included in the final analysis; 1 patient lacked a migraine diagnosis and 6 did not have a concomitant dual CGRP-targeted regimen and were excluded. The mean age of patients was 46.8 years and 54 (61%) were female. The most common migraine diagnosis was chronic migraine in 68 patients (76%). Triptans, ibuprofen, and acetaminophen were the most commonly used acute treatment medications (Table 2).

Safety Assessment

Many of the 89 unique patients trialed > 1 regimen. Thus, for the safety analysis, we analyzed 149 patients on unique dual CGRP-targeting regimens (Table 3). Ubrogepant was used by 126 patients (84.6%) for acute treatment. For preventive therapy, 63 patients (42.3%) used erenumab injections and 55 (36.9%) used fremanezumab injections. Seven patients (4.7%) reported AEs (Table 4). Five of the 7 AEs were noted in the package inserts.^25-32 One patient taking both atogepant and ubrogepant reported brain fog that resolved after a dose reduction of atogepant to every other day dosing. A patient taking fremanezumab and rimegepant reported myalgia/joint pain after the first fremanezumab injection, which resolved after a few days and did not recur during the study period.

Nine of 149 patient regimens (6.0%) were associated with changes in liver function tests or serum creatinine, though all but 1 were grade 1 (1 patient had a grade 2 ALT elevation). Twenty-five patients (16.8%) experienced changes in BP, most of which were category 1 elevations. Four patients had systolic or diastolic BP ≥ 160 mm Hg or 100 mm Hg, respectively (Table 5).

Efficacy Assessment

Of the 149 unique dual CGRP regimens, 59 were eligible for the exploratory efficacy analysis. Data were excluded from the efficacy analysis if patients had not been on a stable CGRP preventive migraine regimen for ≥ 12 weeks prior to the addition of a gepant. Fourteen regimens were excluded due to a lack of clear documentation on efficacy, leaving 45 analyzed regimens. Of the 45 regimens, 34 were from unique patients. There was no median change in migraine intensity or duration found in the efficacy analysis (0.0, P = .18, and 0.0, P = .92, respectively). Ten patients on dual CGRP therapy reported that the addition of a gepant for acute treatment improved their symptoms, 20 reported that their symptoms were unchanged and/or worsened, and 29 lacked documentation.

Discussion

This study aimed to describe the safety and efficacy of concomitant CGRP regimens for migraine prevention and acute treatment. To our knowledge, this was the first descriptive study of these agents in a veteran population. The potential for increased AEs with concomitant use of CGRP antagonists is due to the similarities in the mechanism of action between the agents, which both target the same receptor/ligand pathway. Given CGRP activity in both the gastrointestinal and cardiovascular systems, the potential for related AEs is speculative. Patient-reported AEs occurred in 7 of 149 unique treatment regimens reviewed for an incidence rate < 5%. All AEs were nonserious and self-limiting.

Our findings are consistent with available research. A 2024 retrospective, exploratory real-world study evaluating the safety and tolerability of combining CGRP-targeting mAbs with gepants reported findings consistent with our results. This analysis included adult patients treated with ≥ 1 previous anti-CGRP mAb and found that 234 of 516 patients included received a combination of a gepant in addition to a CGRP-targeting mAb. Of these 234 patients, 1.3% reported nonserious AEs.³³ Similarly, in a multicenter, open-label, long-term safety study in adults experiencing multiple monthly migraine attacks, a subgroup of 13 participants taking a stable dose of an anti-CGRP mAb also took rimegepant 75 mg as needed for acute treatment for 12 weeks. These patients experienced no serious AEs or any AEs leading to discontinuation.¹⁴ A study evaluating the drug-drug interaction, safety, and tolerability of dual therapy (atogepant 60 mg daily and ubrogepant 100 mg every 3 days) in 26 patients found no serious AEs, including no significant changes from baseline in laboratory results, vital signs, or safety-related 12-lead electrocardiogram parameters.¹⁵The TANDEM real-world, open-label, prospective study demonstrated similar results. It evaluated the safety and tolerability of concomitant use of ubrogepant and atogepant in patients with episodic migraines and found no increase in AEs when comparing atogepant alone with combination therapy. Twenty-six patients (9.9%) discontinued treatment due to AEs. The most common treatment-related AEs were constipation, nausea, decreased appetite, and fatigue. Efficacy data were also noted to be an exploratory endpoint in the TANDEM study; however, results have not been published.¹²

Within this safety analysis, new onset gastrointestinal AEs, specifically nausea, only occurred in 1 patient. Hypertension occurred in 25 treatment regimens (16.8%) for 21 unique patients (4 BP elevations occurred in 1 patient on 4 different regimens). However, the retrospective nature of reporting may limit accurate assessment. A closer analysis determined that elevated BP readings correlated with elevated pain scores at the time of the readings, which could have factored into the BP elevations. However, ongoing monitoring is needed due to an increased risk of hypertension, particularly given recent FDA labeling updates for CGRP-targeting therapies including gepants. In light of this, and the overall low incidence of hypertension reported, no new safety concerns were identified.

Limitations

Efficacy data in this project were exploratory. This evaluation did not show a significant difference in migraine intensity or duration after adding a gepant for acute treatment. The study was not powered to detect a significant difference. Limited data exist assessing efficacy outcomes with dual CGRP-targeting treatment regimens. The COURAGE study assessed the real-world effectiveness of ubrogepant and CGRP mAbs with or without the addition of onabotulinumtoxinA. The final analysis of the ubrogepant and CGRP mAb arm included 245 total patients and assessed meaningful migraine pain relief, restoration of normal function after a migraine, and treatment satisfaction. By hour 2, 61.6% of patients reported achieving migraine pain relief, rising to 80.4% by hour 4. Return to normal function occurred in 34.7% at hour 2 and 55.5% by hour 4.¹³ The long-term safety and efficacy of combining erenumab and rimegepant were described in a case series involving 2 patients. Both patients reported that the concomitant CGRP-targeted therapies were effective and reported no AEs.¹⁴

The retrospective design of this study meant that there was potential for limited documentation and introduction of bias into the results. Data were collected at a single VA health care system, and thus, results may not be generalizable to a broader population. However, the study population was consistent with the higher incidence of migraine expected in females in the general population. The sample size was limited, particularly in the exploratory efficacy endpoint assessment.

Limitations were observed due to inconsistent documentation regarding headache characteristics, making it challenging to draw meaningful conclusions from this data set. Additional confounding factors, including polypharmacy, nonadherence to medications, and comorbidities, may have skewed results. For example, while our study design required that the preventive CGRP-targeting medication be stable for 12 weeks for inclusion in further efficacy analysis, other medications commonly used for migraine prevention may have been adjusted (which was not accounted for in this analysis). Given this, more large-scale, placebo-controlled, randomized studies are needed to continue to assess the safety and efficacy of these combination treatment regimens.

Conclusions

Few AEs or safety events were reported with combination CGRP-targeting treatment for acute and preventive treatment of migraine. Those that were identified were considered mild. Efficacy data were limited, and further studies are needed to fully assess outcomes.

Calcitonin gene-related peptide (CGRP) is a neuropeptide that plays a key role in migraine pathophysiology by promoting the dilation of cerebral blood vessels and transmitting pain signals.¹ CGRP has generated interest for the prevention and acute treatment of migraine. Since 2018, 8 novel CGRP-targeting therapies have been approved by the US Food and Drug Administration (FDA) for the management of migraines.^2,3 For migraine prevention, there are 4 injectable monoclonal antibodies (mAbs) directed against the CGRP receptor (erenumab) or the CGRP ligand (fremanezumab, galcanezumab, and eptinezumab). There are also 2 oral small-molecule CGRP receptor antagonists, termed gepants, that also are approved for migraine prevention (atogepant and rimegepant). Three gepants are approved for acute migraine treatment and are administered orally (rimegepant and ubrogepant) or intranasally (zavegepant) (Table 1).

CGRP-targeting therapies have received attention for their role in vasodilation within the cerebral, coronary, and renal vasculature.⁴ CGRP-mediated vasodilatory effects cause systemic regulation of blood pressure (BP) and play a protective role in hypertension.² Some studies, particularly with erenumab, have shown that the inhibitory role of the agent leads to an increase in BP, as well as gastrointestinal issues such as constipation.^2,5 The FDA recently updated monitoring recommendations for all CGRP-targeting therapies to include the potential for BP elevations and hypertension. Outside of this, there is no definitive evidence linking dual CGRP-targeted therapy to higher cardiovascular or gastrointestinal risks and prescribing information does not carry contraindications.⁶

In a 2021 consensus statement, the American Headache Society (AHS) recommended CGRP-targeting therapies for migraine prevention after inability to tolerate or inadequate response to an 8-week trial of ≥ 2 drug classes including antihypertensives, antiseizure medications, antidepressants, and onabotulinumtoxinA.⁷ For acute treatment, AHS recommended gepant use after contraindication to or inadequate response to ≥ 2 triptans. Guidance on combination CGRP-targeting therapies for both prevention and acute treatment was not provided.⁷ More recently, the AHS published a position statement noting substantial efficacy and safety data for CGRP-targeting therapies and suggested its consideration as a first-line option for migraine prevention, though use for acute treatment or combination CGRP-targeting therapies for both prevention and acute treatment were not addressed.⁸

The International Headache Society guidelines for the acute treatment of migraines recommend nonopioid analgesics as first-line therapy for mild migraine attacks. For moderate to severe attacks, triptans with or without a nonopioid analgesic were recommended as first-line therapy, prior to consideration of CGRP-targeted therapy.⁹ The increased use of this new drug class has also led to combination use of CGRP-targeting therapies for migraine prevention and acute treatment as seen in clinical practice and reflected by some case reports, case series, and small studies describing such use.^10-14 In light of the similar mechanism of action of these therapies and the physiologic role of CGRP, there have been calls for safety evaluation.¹⁵

To our knowledge, no studies have evaluated dual CGRP-targeting regimens for migraine in the veteran population. In 2023, the US Department of Veterans Affairs (VA) and US Department of Defense (DoD) updated their clinical practice guidelines for the management of headache.³ For migraine prevention, the VA/DoD guidelines include a strong recommendation for the use of erenumab, fremanezumab, and galcanezumab; a weak recommendation for the use of atogepant; and a recommendation neither for nor against the use of rimegepant. For acute treatment, the guidelines assign a weak recommendation for the use of rimegepant and ubrogepant. Combination use was not addressed.³

Prior to the VA/DoD guidelines, the Veterans Health Administration restricted the dual use of CGRP-targeting therapies for both preventive and acute migraine treatment. However, the VA Pharmacy Benefit Management Service removed the restriction in the Criteria for Use documents, allowing broader access to these medications for veterans.^16-22 This change permits the use of CGRP-targeting drugs for both acute and preventive migraine treatment after initial data reflecting real-world case reports and open-label studies suggested possible efficacy without a clear safety concern.^11,12 This study aims to fill the gap in the literature by evaluating the safety, efficacy, and overall outcomes of combination CGRP-targeting treatment for migraine prevention and acute treatment in a veteran population.

Methods

This single-center, retrospective, medication use evaluation at the Ralph H. Johnson VA Medical Center (RHJVAMC) was reviewed by the RHJVAMC Research and Development Committee and Quality Improvement Program Evaluation Self Certification Tool, which both determined that institutional review board approval was not required because it was considered part of routine care and quality improvement. Computerized Patient Record System (CPRS) data were reviewed between April 1, 2023 (after the Criteria for Use for CGRP-targeting therapies was updated), through January 31, 2025. Patients were included if they had a confirmed diagnosis of migraine using the International Classification of Headache Disorders, 3rd edition criteria and had concomitant active prescriptions for both a preventive and acute treatment CGRP-targeting agent during the project period.²³ Only patients receiving care from the RHJVAMC neurology department were included.

The primary objective was to assess the safety of dual CGRP-targeting therapies for migraine treatment. Key safety endpoints included effects on liver function, kidney function, and BP. Safety outcomes were graded using Common Terminology Criteria for Adverse Events.²⁴ Changes in liver function were categorized as grade 1, 2, or 3 elevations: grade 1 (aspartate aminotransferase [AST]/alanine aminotransferase [ALT] up to 3x the upper limit of normal [ULN] or bilirubin > 1.5 x ULN); grade 2 (AST/ALT 3-6 x ULN or bilirubin 1.5-3 x ULN); and grade 3 (AST/ALT 5-10 x ULN or bilirubin 3-10 x ULN). Kidney function changes were assessed by serum creatinine levels using a similar grading system: Grade 1 (≤ 1.5 x ULN); grade 2 (1.5-3 x baseline of normal); and grade 3 (3-6 x ULN or baseline). Changes in BP were monitored from baseline to the time of the first neurology follow-up. Elevations were grouped into 2 categories, defined as BP ≥ 140 mm Hg systolic and/or 90 mm Hg diastolic (category 1) and ≥ 160 mm Hg systolic and/or 100 mm Hg diastolic (category 2). Neurology documentation was also reviewed in CPRS for individual patient-reported adverse effects (AEs). Safety endpoints were tracked for any occurrence during the project period.

The secondary objective was to describe the patient-reported efficacy of adding a gepant for acute migraine treatment to existing CGRP-targeting therapies for migraine prevention, in those patients who were stable for ≥ 12 weeks on the preventive therapy. Neurology documentation of headache characteristics, including headache severity as rated on a numerical pain score from 0 (no pain) to 10 (worst pain), and duration of headaches (in hours) were recorded during the project period. Changes in headache characteristics were tracked from baseline (ie, the neurology visit when the gepant was first requested) to the first neurology follow-up within 6 months of initiating gepant for acute treatment. If ranges were provided within documentation, a mean was calculated and used for data collection. Neurology documentation was also reviewed for any patient report of overall effectiveness with the added gepant, and categorized as symptoms improved, worsened, or did not change based on subjective report. Descriptive statistics were used for data analysis. A 1-sample Wilcoxon signed rank test was performed as an exploratory analysis for change in headache characteristics from baseline to first neurology follow-up within 6 months. Each individual CGRP regimen was counted as a unique data point to adequately describe changes associated with each new medication and/or dose adjustment. Therefore, patients could be included more than once to account for each distinct treatment regimen.

Results

From April 1, 2023, to January 31, 2025, 96 patients were identified with active prescriptions for dual CGRP-targeting therapies. Of the 96 patients, 89 were included in the final analysis; 1 patient lacked a migraine diagnosis and 6 did not have a concomitant dual CGRP-targeted regimen and were excluded. The mean age of patients was 46.8 years and 54 (61%) were female. The most common migraine diagnosis was chronic migraine in 68 patients (76%). Triptans, ibuprofen, and acetaminophen were the most commonly used acute treatment medications (Table 2).

Safety Assessment

Many of the 89 unique patients trialed > 1 regimen. Thus, for the safety analysis, we analyzed 149 patients on unique dual CGRP-targeting regimens (Table 3). Ubrogepant was used by 126 patients (84.6%) for acute treatment. For preventive therapy, 63 patients (42.3%) used erenumab injections and 55 (36.9%) used fremanezumab injections. Seven patients (4.7%) reported AEs (Table 4). Five of the 7 AEs were noted in the package inserts.^25-32 One patient taking both atogepant and ubrogepant reported brain fog that resolved after a dose reduction of atogepant to every other day dosing. A patient taking fremanezumab and rimegepant reported myalgia/joint pain after the first fremanezumab injection, which resolved after a few days and did not recur during the study period.

Nine of 149 patient regimens (6.0%) were associated with changes in liver function tests or serum creatinine, though all but 1 were grade 1 (1 patient had a grade 2 ALT elevation). Twenty-five patients (16.8%) experienced changes in BP, most of which were category 1 elevations. Four patients had systolic or diastolic BP ≥ 160 mm Hg or 100 mm Hg, respectively (Table 5).

Efficacy Assessment

Of the 149 unique dual CGRP regimens, 59 were eligible for the exploratory efficacy analysis. Data were excluded from the efficacy analysis if patients had not been on a stable CGRP preventive migraine regimen for ≥ 12 weeks prior to the addition of a gepant. Fourteen regimens were excluded due to a lack of clear documentation on efficacy, leaving 45 analyzed regimens. Of the 45 regimens, 34 were from unique patients. There was no median change in migraine intensity or duration found in the efficacy analysis (0.0, P = .18, and 0.0, P = .92, respectively). Ten patients on dual CGRP therapy reported that the addition of a gepant for acute treatment improved their symptoms, 20 reported that their symptoms were unchanged and/or worsened, and 29 lacked documentation.

Discussion

This study aimed to describe the safety and efficacy of concomitant CGRP regimens for migraine prevention and acute treatment. To our knowledge, this was the first descriptive study of these agents in a veteran population. The potential for increased AEs with concomitant use of CGRP antagonists is due to the similarities in the mechanism of action between the agents, which both target the same receptor/ligand pathway. Given CGRP activity in both the gastrointestinal and cardiovascular systems, the potential for related AEs is speculative. Patient-reported AEs occurred in 7 of 149 unique treatment regimens reviewed for an incidence rate < 5%. All AEs were nonserious and self-limiting.

Our findings are consistent with available research. A 2024 retrospective, exploratory real-world study evaluating the safety and tolerability of combining CGRP-targeting mAbs with gepants reported findings consistent with our results. This analysis included adult patients treated with ≥ 1 previous anti-CGRP mAb and found that 234 of 516 patients included received a combination of a gepant in addition to a CGRP-targeting mAb. Of these 234 patients, 1.3% reported nonserious AEs.³³ Similarly, in a multicenter, open-label, long-term safety study in adults experiencing multiple monthly migraine attacks, a subgroup of 13 participants taking a stable dose of an anti-CGRP mAb also took rimegepant 75 mg as needed for acute treatment for 12 weeks. These patients experienced no serious AEs or any AEs leading to discontinuation.¹⁴ A study evaluating the drug-drug interaction, safety, and tolerability of dual therapy (atogepant 60 mg daily and ubrogepant 100 mg every 3 days) in 26 patients found no serious AEs, including no significant changes from baseline in laboratory results, vital signs, or safety-related 12-lead electrocardiogram parameters.¹⁵The TANDEM real-world, open-label, prospective study demonstrated similar results. It evaluated the safety and tolerability of concomitant use of ubrogepant and atogepant in patients with episodic migraines and found no increase in AEs when comparing atogepant alone with combination therapy. Twenty-six patients (9.9%) discontinued treatment due to AEs. The most common treatment-related AEs were constipation, nausea, decreased appetite, and fatigue. Efficacy data were also noted to be an exploratory endpoint in the TANDEM study; however, results have not been published.¹²

Within this safety analysis, new onset gastrointestinal AEs, specifically nausea, only occurred in 1 patient. Hypertension occurred in 25 treatment regimens (16.8%) for 21 unique patients (4 BP elevations occurred in 1 patient on 4 different regimens). However, the retrospective nature of reporting may limit accurate assessment. A closer analysis determined that elevated BP readings correlated with elevated pain scores at the time of the readings, which could have factored into the BP elevations. However, ongoing monitoring is needed due to an increased risk of hypertension, particularly given recent FDA labeling updates for CGRP-targeting therapies including gepants. In light of this, and the overall low incidence of hypertension reported, no new safety concerns were identified.

Limitations

Efficacy data in this project were exploratory. This evaluation did not show a significant difference in migraine intensity or duration after adding a gepant for acute treatment. The study was not powered to detect a significant difference. Limited data exist assessing efficacy outcomes with dual CGRP-targeting treatment regimens. The COURAGE study assessed the real-world effectiveness of ubrogepant and CGRP mAbs with or without the addition of onabotulinumtoxinA. The final analysis of the ubrogepant and CGRP mAb arm included 245 total patients and assessed meaningful migraine pain relief, restoration of normal function after a migraine, and treatment satisfaction. By hour 2, 61.6% of patients reported achieving migraine pain relief, rising to 80.4% by hour 4. Return to normal function occurred in 34.7% at hour 2 and 55.5% by hour 4.¹³ The long-term safety and efficacy of combining erenumab and rimegepant were described in a case series involving 2 patients. Both patients reported that the concomitant CGRP-targeted therapies were effective and reported no AEs.¹⁴

The retrospective design of this study meant that there was potential for limited documentation and introduction of bias into the results. Data were collected at a single VA health care system, and thus, results may not be generalizable to a broader population. However, the study population was consistent with the higher incidence of migraine expected in females in the general population. The sample size was limited, particularly in the exploratory efficacy endpoint assessment.

Limitations were observed due to inconsistent documentation regarding headache characteristics, making it challenging to draw meaningful conclusions from this data set. Additional confounding factors, including polypharmacy, nonadherence to medications, and comorbidities, may have skewed results. For example, while our study design required that the preventive CGRP-targeting medication be stable for 12 weeks for inclusion in further efficacy analysis, other medications commonly used for migraine prevention may have been adjusted (which was not accounted for in this analysis). Given this, more large-scale, placebo-controlled, randomized studies are needed to continue to assess the safety and efficacy of these combination treatment regimens.

Conclusions

Few AEs or safety events were reported with combination CGRP-targeting treatment for acute and preventive treatment of migraine. Those that were identified were considered mild. Efficacy data were limited, and further studies are needed to fully assess outcomes.

Calcitonin gene-related peptide (CGRP) is a neuropeptide that plays a key role in migraine pathophysiology by promoting the dilation of cerebral blood vessels and transmitting pain signals.¹ CGRP has generated interest for the prevention and acute treatment of migraine. Since 2018, 8 novel CGRP-targeting therapies have been approved by the US Food and Drug Administration (FDA) for the management of migraines.^2,3 For migraine prevention, there are 4 injectable monoclonal antibodies (mAbs) directed against the CGRP receptor (erenumab) or the CGRP ligand (fremanezumab, galcanezumab, and eptinezumab). There are also 2 oral small-molecule CGRP receptor antagonists, termed gepants, that also are approved for migraine prevention (atogepant and rimegepant). Three gepants are approved for acute migraine treatment and are administered orally (rimegepant and ubrogepant) or intranasally (zavegepant) (Table 1).

CGRP-targeting therapies have received attention for their role in vasodilation within the cerebral, coronary, and renal vasculature.⁴ CGRP-mediated vasodilatory effects cause systemic regulation of blood pressure (BP) and play a protective role in hypertension.² Some studies, particularly with erenumab, have shown that the inhibitory role of the agent leads to an increase in BP, as well as gastrointestinal issues such as constipation.^2,5 The FDA recently updated monitoring recommendations for all CGRP-targeting therapies to include the potential for BP elevations and hypertension. Outside of this, there is no definitive evidence linking dual CGRP-targeted therapy to higher cardiovascular or gastrointestinal risks and prescribing information does not carry contraindications.⁶

In a 2021 consensus statement, the American Headache Society (AHS) recommended CGRP-targeting therapies for migraine prevention after inability to tolerate or inadequate response to an 8-week trial of ≥ 2 drug classes including antihypertensives, antiseizure medications, antidepressants, and onabotulinumtoxinA.⁷ For acute treatment, AHS recommended gepant use after contraindication to or inadequate response to ≥ 2 triptans. Guidance on combination CGRP-targeting therapies for both prevention and acute treatment was not provided.⁷ More recently, the AHS published a position statement noting substantial efficacy and safety data for CGRP-targeting therapies and suggested its consideration as a first-line option for migraine prevention, though use for acute treatment or combination CGRP-targeting therapies for both prevention and acute treatment were not addressed.⁸

The International Headache Society guidelines for the acute treatment of migraines recommend nonopioid analgesics as first-line therapy for mild migraine attacks. For moderate to severe attacks, triptans with or without a nonopioid analgesic were recommended as first-line therapy, prior to consideration of CGRP-targeted therapy.⁹ The increased use of this new drug class has also led to combination use of CGRP-targeting therapies for migraine prevention and acute treatment as seen in clinical practice and reflected by some case reports, case series, and small studies describing such use.^10-14 In light of the similar mechanism of action of these therapies and the physiologic role of CGRP, there have been calls for safety evaluation.¹⁵

To our knowledge, no studies have evaluated dual CGRP-targeting regimens for migraine in the veteran population. In 2023, the US Department of Veterans Affairs (VA) and US Department of Defense (DoD) updated their clinical practice guidelines for the management of headache.³ For migraine prevention, the VA/DoD guidelines include a strong recommendation for the use of erenumab, fremanezumab, and galcanezumab; a weak recommendation for the use of atogepant; and a recommendation neither for nor against the use of rimegepant. For acute treatment, the guidelines assign a weak recommendation for the use of rimegepant and ubrogepant. Combination use was not addressed.³

Prior to the VA/DoD guidelines, the Veterans Health Administration restricted the dual use of CGRP-targeting therapies for both preventive and acute migraine treatment. However, the VA Pharmacy Benefit Management Service removed the restriction in the Criteria for Use documents, allowing broader access to these medications for veterans.^16-22 This change permits the use of CGRP-targeting drugs for both acute and preventive migraine treatment after initial data reflecting real-world case reports and open-label studies suggested possible efficacy without a clear safety concern.^11,12 This study aims to fill the gap in the literature by evaluating the safety, efficacy, and overall outcomes of combination CGRP-targeting treatment for migraine prevention and acute treatment in a veteran population.

Methods

This single-center, retrospective, medication use evaluation at the Ralph H. Johnson VA Medical Center (RHJVAMC) was reviewed by the RHJVAMC Research and Development Committee and Quality Improvement Program Evaluation Self Certification Tool, which both determined that institutional review board approval was not required because it was considered part of routine care and quality improvement. Computerized Patient Record System (CPRS) data were reviewed between April 1, 2023 (after the Criteria for Use for CGRP-targeting therapies was updated), through January 31, 2025. Patients were included if they had a confirmed diagnosis of migraine using the International Classification of Headache Disorders, 3rd edition criteria and had concomitant active prescriptions for both a preventive and acute treatment CGRP-targeting agent during the project period.²³ Only patients receiving care from the RHJVAMC neurology department were included.

The primary objective was to assess the safety of dual CGRP-targeting therapies for migraine treatment. Key safety endpoints included effects on liver function, kidney function, and BP. Safety outcomes were graded using Common Terminology Criteria for Adverse Events.²⁴ Changes in liver function were categorized as grade 1, 2, or 3 elevations: grade 1 (aspartate aminotransferase [AST]/alanine aminotransferase [ALT] up to 3x the upper limit of normal [ULN] or bilirubin > 1.5 x ULN); grade 2 (AST/ALT 3-6 x ULN or bilirubin 1.5-3 x ULN); and grade 3 (AST/ALT 5-10 x ULN or bilirubin 3-10 x ULN). Kidney function changes were assessed by serum creatinine levels using a similar grading system: Grade 1 (≤ 1.5 x ULN); grade 2 (1.5-3 x baseline of normal); and grade 3 (3-6 x ULN or baseline). Changes in BP were monitored from baseline to the time of the first neurology follow-up. Elevations were grouped into 2 categories, defined as BP ≥ 140 mm Hg systolic and/or 90 mm Hg diastolic (category 1) and ≥ 160 mm Hg systolic and/or 100 mm Hg diastolic (category 2). Neurology documentation was also reviewed in CPRS for individual patient-reported adverse effects (AEs). Safety endpoints were tracked for any occurrence during the project period.

The secondary objective was to describe the patient-reported efficacy of adding a gepant for acute migraine treatment to existing CGRP-targeting therapies for migraine prevention, in those patients who were stable for ≥ 12 weeks on the preventive therapy. Neurology documentation of headache characteristics, including headache severity as rated on a numerical pain score from 0 (no pain) to 10 (worst pain), and duration of headaches (in hours) were recorded during the project period. Changes in headache characteristics were tracked from baseline (ie, the neurology visit when the gepant was first requested) to the first neurology follow-up within 6 months of initiating gepant for acute treatment. If ranges were provided within documentation, a mean was calculated and used for data collection. Neurology documentation was also reviewed for any patient report of overall effectiveness with the added gepant, and categorized as symptoms improved, worsened, or did not change based on subjective report. Descriptive statistics were used for data analysis. A 1-sample Wilcoxon signed rank test was performed as an exploratory analysis for change in headache characteristics from baseline to first neurology follow-up within 6 months. Each individual CGRP regimen was counted as a unique data point to adequately describe changes associated with each new medication and/or dose adjustment. Therefore, patients could be included more than once to account for each distinct treatment regimen.

Results

From April 1, 2023, to January 31, 2025, 96 patients were identified with active prescriptions for dual CGRP-targeting therapies. Of the 96 patients, 89 were included in the final analysis; 1 patient lacked a migraine diagnosis and 6 did not have a concomitant dual CGRP-targeted regimen and were excluded. The mean age of patients was 46.8 years and 54 (61%) were female. The most common migraine diagnosis was chronic migraine in 68 patients (76%). Triptans, ibuprofen, and acetaminophen were the most commonly used acute treatment medications (Table 2).

Safety Assessment

Many of the 89 unique patients trialed > 1 regimen. Thus, for the safety analysis, we analyzed 149 patients on unique dual CGRP-targeting regimens (Table 3). Ubrogepant was used by 126 patients (84.6%) for acute treatment. For preventive therapy, 63 patients (42.3%) used erenumab injections and 55 (36.9%) used fremanezumab injections. Seven patients (4.7%) reported AEs (Table 4). Five of the 7 AEs were noted in the package inserts.^25-32 One patient taking both atogepant and ubrogepant reported brain fog that resolved after a dose reduction of atogepant to every other day dosing. A patient taking fremanezumab and rimegepant reported myalgia/joint pain after the first fremanezumab injection, which resolved after a few days and did not recur during the study period.

Nine of 149 patient regimens (6.0%) were associated with changes in liver function tests or serum creatinine, though all but 1 were grade 1 (1 patient had a grade 2 ALT elevation). Twenty-five patients (16.8%) experienced changes in BP, most of which were category 1 elevations. Four patients had systolic or diastolic BP ≥ 160 mm Hg or 100 mm Hg, respectively (Table 5).

Efficacy Assessment

Of the 149 unique dual CGRP regimens, 59 were eligible for the exploratory efficacy analysis. Data were excluded from the efficacy analysis if patients had not been on a stable CGRP preventive migraine regimen for ≥ 12 weeks prior to the addition of a gepant. Fourteen regimens were excluded due to a lack of clear documentation on efficacy, leaving 45 analyzed regimens. Of the 45 regimens, 34 were from unique patients. There was no median change in migraine intensity or duration found in the efficacy analysis (0.0, P = .18, and 0.0, P = .92, respectively). Ten patients on dual CGRP therapy reported that the addition of a gepant for acute treatment improved their symptoms, 20 reported that their symptoms were unchanged and/or worsened, and 29 lacked documentation.

Discussion

This study aimed to describe the safety and efficacy of concomitant CGRP regimens for migraine prevention and acute treatment. To our knowledge, this was the first descriptive study of these agents in a veteran population. The potential for increased AEs with concomitant use of CGRP antagonists is due to the similarities in the mechanism of action between the agents, which both target the same receptor/ligand pathway. Given CGRP activity in both the gastrointestinal and cardiovascular systems, the potential for related AEs is speculative. Patient-reported AEs occurred in 7 of 149 unique treatment regimens reviewed for an incidence rate < 5%. All AEs were nonserious and self-limiting.

Our findings are consistent with available research. A 2024 retrospective, exploratory real-world study evaluating the safety and tolerability of combining CGRP-targeting mAbs with gepants reported findings consistent with our results. This analysis included adult patients treated with ≥ 1 previous anti-CGRP mAb and found that 234 of 516 patients included received a combination of a gepant in addition to a CGRP-targeting mAb. Of these 234 patients, 1.3% reported nonserious AEs.³³ Similarly, in a multicenter, open-label, long-term safety study in adults experiencing multiple monthly migraine attacks, a subgroup of 13 participants taking a stable dose of an anti-CGRP mAb also took rimegepant 75 mg as needed for acute treatment for 12 weeks. These patients experienced no serious AEs or any AEs leading to discontinuation.¹⁴ A study evaluating the drug-drug interaction, safety, and tolerability of dual therapy (atogepant 60 mg daily and ubrogepant 100 mg every 3 days) in 26 patients found no serious AEs, including no significant changes from baseline in laboratory results, vital signs, or safety-related 12-lead electrocardiogram parameters.¹⁵The TANDEM real-world, open-label, prospective study demonstrated similar results. It evaluated the safety and tolerability of concomitant use of ubrogepant and atogepant in patients with episodic migraines and found no increase in AEs when comparing atogepant alone with combination therapy. Twenty-six patients (9.9%) discontinued treatment due to AEs. The most common treatment-related AEs were constipation, nausea, decreased appetite, and fatigue. Efficacy data were also noted to be an exploratory endpoint in the TANDEM study; however, results have not been published.¹²

Within this safety analysis, new onset gastrointestinal AEs, specifically nausea, only occurred in 1 patient. Hypertension occurred in 25 treatment regimens (16.8%) for 21 unique patients (4 BP elevations occurred in 1 patient on 4 different regimens). However, the retrospective nature of reporting may limit accurate assessment. A closer analysis determined that elevated BP readings correlated with elevated pain scores at the time of the readings, which could have factored into the BP elevations. However, ongoing monitoring is needed due to an increased risk of hypertension, particularly given recent FDA labeling updates for CGRP-targeting therapies including gepants. In light of this, and the overall low incidence of hypertension reported, no new safety concerns were identified.

Limitations

Efficacy data in this project were exploratory. This evaluation did not show a significant difference in migraine intensity or duration after adding a gepant for acute treatment. The study was not powered to detect a significant difference. Limited data exist assessing efficacy outcomes with dual CGRP-targeting treatment regimens. The COURAGE study assessed the real-world effectiveness of ubrogepant and CGRP mAbs with or without the addition of onabotulinumtoxinA. The final analysis of the ubrogepant and CGRP mAb arm included 245 total patients and assessed meaningful migraine pain relief, restoration of normal function after a migraine, and treatment satisfaction. By hour 2, 61.6% of patients reported achieving migraine pain relief, rising to 80.4% by hour 4. Return to normal function occurred in 34.7% at hour 2 and 55.5% by hour 4.¹³ The long-term safety and efficacy of combining erenumab and rimegepant were described in a case series involving 2 patients. Both patients reported that the concomitant CGRP-targeted therapies were effective and reported no AEs.¹⁴

The retrospective design of this study meant that there was potential for limited documentation and introduction of bias into the results. Data were collected at a single VA health care system, and thus, results may not be generalizable to a broader population. However, the study population was consistent with the higher incidence of migraine expected in females in the general population. The sample size was limited, particularly in the exploratory efficacy endpoint assessment.

Limitations were observed due to inconsistent documentation regarding headache characteristics, making it challenging to draw meaningful conclusions from this data set. Additional confounding factors, including polypharmacy, nonadherence to medications, and comorbidities, may have skewed results. For example, while our study design required that the preventive CGRP-targeting medication be stable for 12 weeks for inclusion in further efficacy analysis, other medications commonly used for migraine prevention may have been adjusted (which was not accounted for in this analysis). Given this, more large-scale, placebo-controlled, randomized studies are needed to continue to assess the safety and efficacy of these combination treatment regimens.

Conclusions

Few AEs or safety events were reported with combination CGRP-targeting treatment for acute and preventive treatment of migraine. Those that were identified were considered mild. Efficacy data were limited, and further studies are needed to fully assess outcomes.

Heart failure (HF) is a prevalent disease in the United States affecting > 6.5 million adults and contributing to significant morbidity and mortality.¹ The disease course associated with HF includes potential symptom improvement with intermittent periods of decompensation and possible clinical deterioration. Multiple therapies have been developed to improve outcomes in people with HF—to palliate HF symptoms, prevent hospitalizations, and reduce mortality.² However, the risks of decompensation and hospitalization remain. HF decompensation development may precede clear actionable symptoms such as worsening dyspnea, noticeable edema, or weight gain. Tools to identify patient deterioration and trigger interventions to prevent HF admissions are clinically attractive compared with reliance on subjective factors alone.

Cardiac resynchronization therapy (CRT) and implantable cardioverter-defibrillator (ICD) devices made by Boston Scientific include the HeartLogic monitoring feature. Five main sensors produce an index risk score; an index score > 16 warns clinicians that the patient is at an increased risk for a HF event.³ The 5 sensors are thoracic impedance, first (S1) and third heart sounds (S3), night heart rate (NHR), respiratory rate (RR), and activity. Each sensor can draw attention to the primary driver of the alert and guide health care practitioners (HCPs) to the appropriate interventions.³ A HeartLogic alert example is shown in Figure 1.

The S3 occurs during the early diastolic phase when blood moves into the ventricles. As HF worsens, with a combination of elevated filling pressures and reduced cardiac muscle compliance, S3 can become more pronounced.⁴ The S1 is correlated with the contractility of the left ventricle and will be reduced in patients at risk for HF events.⁵ Physical activity is a long-term prognostic marker in patients with HF; reduced activity is associated with mortality and increased risk of an HF event.⁶ Thoracic impedance is a sensor used to identify pulmonary congestion, pocket infections, pleural/pericardial effusion, and respiratory infections. The accumulation of intrathoracic fluid during pulmonary congestion increases conductance, causing a decrease in impedance.⁷ RR will increase as patients experience dyspnea with a more rapid, shallow breath and may trigger alerts closer to the actual HF event than other sensors. Nearly 90% of patients hospitalized for HF experience shortness of breath.^8,9 NHR is used as a surrogate for resting heart rate (HR). A high resting HR is correlated with the progression of coronary atherosclerosis, harmful effects on left ventricular function, and increased risk of myocardial ischemia and ventricular arrhythmias.¹⁰

One of the challenges with preventing hospitalizations may be the lack of patient reported symptoms leading up to the event. The purpose of the sensors and HeartLogic index is to identify patients a median of 34 days before an HF event (HF admission or unscheduled intervention with intravenous treatment) with a sensitivity rate of 70%.³ According to real-word experience data, alerts have been found to precede HF symptoms by a median of 12 days and HF events such as hospitalizations by a median of 38 days, with an overall 67% reduction in HF hospitalizations when integrated into clinical care.^11,12

MANAGE-HF evaluated 191 patients with HF with reduced ejection fraction (HFrEF) (< 35%), New York Heart Association class II-III symptoms, and who had an implanted CRT and/or ICD to develop an alert management guide to optimize medical treatment.¹² It aimed to adjust patient regimen within 6 days of an elevated Heart- Logic index by either initiation, escalation, or maintenance of HF treatment depending on the index trend after the initial alert. This trial found that by focusing on such optimization, HF treatment was augmented during 74% of the 585 alert cases and during 54% of 3290 weekly alerts.

Initiation and uptitration of the 4 primary components of guideline-directed medical therapy (GDMT) are recommended by the 2022 Heart Failure Guidelines to reduce mortality and morbidity in patients with HFrEF.² The 4 pillars of GDMT consist of -blockers (BB), sodium-glucose cotransporter type 2 inhibitors (SGLT2i), mineralocorticoid receptor antagonists (MRA), and renin-angiotensin-system inhibitors (RASi) including angiotensin II receptor blocker/neprilysin inhibitors (ARNi), angiotensin-converting enzyme inhibitors (ACEi) and angiotensin II receptor blockers (ARB) (Appendix 1). Obtaining and titrating to target doses wherever possible is recommended, as those were the doses that established safety and efficacy in patients with HFrEF in clinical trials.² Pharmacists are adequately equipped to optimize HF GDMT and appropriately monitor drug response.

Through the use of HeartLogic in clinical practice, patients with HF have been shown to have improved clinical outcomes and are more likely to receive effective care; 80% of alerts were shown to provide new information to clinicians.¹³ This project sought to quantify the total number and types of pharmacist interventions driven by integration of HeartLogic index monitoring into practice.

Methods

The West Palm Beach Veterans Affairs Medical Center (WPBVAMC) Research Program Office approved this project and determined it was exempt from institutional review board oversight. Patients were screened retrospectively and prospectively from May 26, 2022, through December 31, 2022, by a cardiology clinical pharmacist practitioner (CPP) and a cardiology pharmacy resident using the local monitoring suite for the HeartLogic-compatible device, LATITUDE NXT. Read-only access to the local monitoring suit was granted by the National Cardiac Device Surveillance Program. Training for HeartLogic was completed through continuing education courses provided by Boston Scientific. Additional information was provided by Boston Scientific representative presentations and collaboration with WPBVAMC pacemaker clinic HCPs.

Individuals included were patients with HeartLogic-capable ICDs. A HeartLogic alert had to be present at initial patient contact. Patients were also contacted as part of routine clinical practice, but no formal number or frequency of calls to patients was required. The initial contact must be with a pharmacist for the patient to be included, but subsequent contact by other HCPs was included. Patients in the cardiology clinic are required to meet with a cardiologist at least annually; however, interim visits can be completed by advanced practice registered nurse practitioners, physicians assistants, or CPPs.

Patients in alert status were contacted by telephone and appropriate modifications of HF therapy were made by the CPP based on score metrics, medical record review, and patient interview. Information surrounding the initial alert, baseline patient data, medication and monitoring interventions made, and clinical outcomes such as hospitalization, symptom improvement, follow-up, and mortality were collected. Information for each encounter was collected until 42 days from the initial date of pharmacist contact.

Clinically successful tolerability of intervention implementation was defined as tolerability, adherence, and lack of adverse effects (AEs) per patient report at follow-up or within 42 days from initial alert (Appendix 2). A decrease in dose was not counted as intolerance. A single patient may have been counted as multiple encounters if the original intervention resulted in treatment intolerance and the patient remained in alert or if an additional alert occurred after 42 days of the initial alert. There were no specific time criteria for follow-up, which occurred at the CPP’s discretion.

There was no mandated algorithm used to alter medications based on the Heart- Logic score, nor were there required minimum or maximum numbers of interventions after an alert. Patient contact by telephone initiated an encounter. The types of interventions included medication increases, decreases, initiation, discontinuation, or no medication change. Each medication change and rationale, if applicable, was recorded for the encounter ≤ 42 days after the initial contact date. If a medication with required monitoring parameters was augmented, the pharmacist was responsible for ordering laboratory testing and follow-up. Most interventions were completed by telephone; however, some patients had in-person visits in the HF CPP clinic.

Outcomes

The primary outcome was the number of pharmacist interventions made to optimize GDMT, defined as either an initiation or dose increase. Key intervention analysis included the use and dosing of the 4 primary components of HF GDMT: BB, SGLT2i, MRA, and ARNi/ARB/ACEi. In addition to the 4 primary components of GDMT, loop diuretic changes were also recorded and analyzed. Secondary endpoints were the number of HF hospitalizations ≤ 42 days after the initial alert, and the effect of medication interventions on device metrics, patient symptoms, and tolerability. Successful tolerability was defined as continued use of augmented GDMT without intolerance or discontinuation. The primary analysis was analyzed through descriptive statistics. Median changes in HeartLogic scores and metrics from baseline were analyzed using a paired, 2-sided t test with an α of .05 to detect significance.

Results

There were 39 WPBVAMC patients with a HeartLogic-capable device. Twenty-one alert encounters were analyzed in 16 patients (41%) over 31 weeks of data collection. The 16 patients at baseline had a mean age of 74 years, all were male, and 12 (75%) were White. Eight patients (50%) had a recent ejection fraction (EF) between 30% and 40%. Three patients had an EF ≥ 40%. At the time of alert, 15 patients used BB (94%), 10 used loop diuretics (63%), and 9 used ARNi (56%) (Table 1).

There were 23 medication changes made during initial contact. The most common change was starting an SGLT2i (30%; n = 7), followed by starting an MRA (22%; n = 5), and increasing the ARNi dose (22%; n = 5). At the initial contact, ≥ 1 medication optimization occurred in 95% (n = 20) of encounters. The CPP contacted patients a mean of 4.8 days after the initial alert.

Patients were taking a mean of 2.6 primary GDMT medications at baseline and 3.0 at 42 days. CPP encounters led to a mean of 1.8 medication changes over the 6-week period (range, 0-5). Seventeen medications were started, 13 medications were increased, 3 medications were decreased, and 4 medications were stopped (Table 2). One ACEi and 1 ARB were switched as a therapy escalation to an ARNi. One patient was on 1 of 4 primary GDMTs at baseline, which increased to 4 GDMT agents at 42 days.

SGLT2 inhibitors were added most often at initial contact (54%) and throughout the 42-day period (41%). The most common successfully tolerated optimizations were RASi, followed by MRA, SGLT2 inhibitors, BB, and loop diuretics with 11, 6, 5, 3, and 2 patients, respectively. Interventions were tolerated by 90% of patients, and no HF hospitalization occurred during follow-up. All possible rationales for patients with the same or reduced number of GDMT at 42 days compared with baseline are shown in Appendix 2.

Device Metrics

During initial contact, the most common HeartLogic metric category that was predominantly worsening were heart sounds (S1, S3, and S3/S1 ratio), followed by compensatory mechanism sensors (NHR and RR) and congestion (impedance) at rates of 61.9%, 23.8%, and 14.3%, respectively (Figure 2).

The median HeartLogic index score was 18 at baseline and 5 at the end of the follow-up period (P < .001). The changes in score and metrics were compared with the type of successfully tolerated GDMT optimization made (Table 3). The GDMT optimization analysis included SGLT2i, RASi, MRA, BB, and loop diuretics. All interventions reduced the overall HeartLogic index score, ranging from a 9.5-point reduction (loop diuretics) to a 16-point reduction (SGLT2i and BB). Optimization of SGLT2i, RASi, and loop diuretics had a positive impact on S1 score. For S3 score, SGLT2i, MRA, and BB had a positive impact. All medications, except for SGLT2i therapy, reduced the NHR score. Optimization of MRA, SGLT2i, and BB had positive impacts on the impedance score. All medications reduced RR from baseline. Only SGLT2i and loop diuretics had positive impacts on the activity score.

Clinical Outcomes and Adverse Effects

Within 42 days of contact, 17 encounters (81%) had ≥ 1 follow-up appointment with a CPP and all 21 patients had ≥  1 follow-up health care team member. One patient had a HF-related hospitalization within 42 days of contact; however, that individual refused the recommended medication intervention. There were 13 encounters (62%) with reported symptoms at the time of the initial alert and 10 (77%) had subjective symptom improvement at 42 days (Appendix 3).

Of 30 medication optimizations, 27 primary GDMT medications were tolerated. Two medication intolerances led to discontinuation (1 SGLT2i and 1 loop diuretic) and 1 patient never started the SGLT2i (Table 4). There was only 1 known patient who did not follow the directions to adjust their medications. That individual was included because the patient agreed to the change during the CPP visit but later reported that he had never started the SGLT2i.

Discussion

The HeartLogic tool created a bridge for patients with HF to work with CPPs as soon as possible to optimize medication therapy to reduce HF events. This study highlights an additional area of expertise and service that CPPs may offer to their specialty HF clinic team. Over 31 weeks, 21 encounters and 30 medication optimizations were completed. These interventions led to significant reductions in HeartLogic scores, improvements in symptoms, and optimization of HF care, most of which were well tolerated.

Additional hemodynamic monitoring devices are available. Similar to HeartLogic, OptiVol is a tool embedded in select Medtronic implantable devices that monitors fluid status. ¹⁴ CardioMEMS is an implantable pulmonary artery pressure sensor used as a presymptomatic data point to alert clinicians when HF is worsening. In the CHAMPION trial, the use of CardioMEMS showed a 28% reduction in HF-related hospitalization at 6 months.¹⁵ Conversely, in the GUIDE-HF trial, monitoring with CardioMEMS did not significantly reduce the composite endpoint of mortality and total HF events.¹⁶ Therefore, remote hemodynamic monitoring has variable results and the use of these tools remains uncertain per the clinical guidelines.²

The MANAGE-HF study that contributed to the validation of the HeartLogic tool may provide a comparison with this smaller single-center project. The time to follow-up within 7 days of alert was noted in only 54% of the patients in MANAGE-HF.¹² In this study, 86% of patients received follow-up within 7 days, with a mean of 4.8 days. The quick turnaround from the time of alert to intervention portrays pharmacists as readily available HCPs.

In MANAGE-HF, 89% of medication augmentation involved loop diuretics or thiazides; in our project, loop diuretics were the least frequently changed medication. Most optimizations in this project included ARNi, SGLT2i, BB, and MRA, which have been shown to reduce morbidity and mortality.² Our project included use of SGLT2i therapy to affect HeartLogic metrics, which has not been evaluated previously. SGLT2i were the most commonly initiated medication after an alert. Of the 5 tolerated SGLT2i optimization encounters, 4 were out of alert at 42 days.

SGLT2i resulted in a significant decrease in HeartLogic index score from baseline and were the only class of medication that did not produce a negative change in any metric. In this study, CPPs utilizing and acting on HeartLogic alerts led to 1 (4.8%) hospitalization with HF as the primary reason for admission and no hospitalizations as a secondary cause in 42 days, compared to 37% and 7.9% in the MANAGE-HF in 1 year, respectively. An additional screening 1 year after the initial alert found that 2 (12.5%) of 12 patients had been admitted with 1 HF hospitalization each.

A strength of this study was the ability to use HeartLogic to identify high-risk patients, provide a source of patient contact and monitoring, interpret 5 cardiac sensors, and optimize all HF GDMT, not just volume management. By focusing efforts on making patient contact and pharmacotherapy interventions with morbidity and mortality benefit, remote hemodynamic monitoring may show a clear clinical benefit and become a vital part of HF care.

Limitations

Checking for adherence and tolerance to medications were mainly patient reported if there was a CPP follow-up within 42 days, or potentially through refill history when unclear. However, this limitation is reflective of current practice where patients may have multiple clinicians working to optimize HF care and where there is reliance on patients in order to guide continued therapy. Although unable to explicitly show a reduction in HF events given lack of comparator group, the interventions made are associated with improved outcomes and thus would be expected to improve patient outcomes. Changes in vital signs were not tracked as part of this project, however the main rationale for changes made were to optimize GDMT therapy, not specifically to impact vital sign measures.

HeartLogic alerts prompted identification of high-risk patients with HF, pharmacist evaluation and outreach, patient-focused pharmacotherapy care, and beneficial patient outcomes. With only 2 cardiology CPPs checking alerts once weekly, future studies may be needed with larger samples to create algorithms and protocols to increase the clinical utility of this tool on a greater scale.

Conclusions

Cardiology CPP-led HF interventions triggered by HeartLogic alerts lead to effective patient identification, increased access to care, reductions in HeartLogic scores, improvements in symptoms, and optimization of HF care. This project demonstrates the practical utility of the HeartLogic suite in conjunction with CPP care to prioritize treatment for highrisk patients with HF in an efficient manner. The data highlight the potential value of the HeartLogic tool and a CPP in HF care to facilitate initiation and optimization of GDMT to ultimately improve the morbidity and mortality in patients with HF.

Heart failure (HF) is a prevalent disease in the United States affecting > 6.5 million adults and contributing to significant morbidity and mortality.¹ The disease course associated with HF includes potential symptom improvement with intermittent periods of decompensation and possible clinical deterioration. Multiple therapies have been developed to improve outcomes in people with HF—to palliate HF symptoms, prevent hospitalizations, and reduce mortality.² However, the risks of decompensation and hospitalization remain. HF decompensation development may precede clear actionable symptoms such as worsening dyspnea, noticeable edema, or weight gain. Tools to identify patient deterioration and trigger interventions to prevent HF admissions are clinically attractive compared with reliance on subjective factors alone.

Cardiac resynchronization therapy (CRT) and implantable cardioverter-defibrillator (ICD) devices made by Boston Scientific include the HeartLogic monitoring feature. Five main sensors produce an index risk score; an index score > 16 warns clinicians that the patient is at an increased risk for a HF event.³ The 5 sensors are thoracic impedance, first (S1) and third heart sounds (S3), night heart rate (NHR), respiratory rate (RR), and activity. Each sensor can draw attention to the primary driver of the alert and guide health care practitioners (HCPs) to the appropriate interventions.³ A HeartLogic alert example is shown in Figure 1.

The S3 occurs during the early diastolic phase when blood moves into the ventricles. As HF worsens, with a combination of elevated filling pressures and reduced cardiac muscle compliance, S3 can become more pronounced.⁴ The S1 is correlated with the contractility of the left ventricle and will be reduced in patients at risk for HF events.⁵ Physical activity is a long-term prognostic marker in patients with HF; reduced activity is associated with mortality and increased risk of an HF event.⁶ Thoracic impedance is a sensor used to identify pulmonary congestion, pocket infections, pleural/pericardial effusion, and respiratory infections. The accumulation of intrathoracic fluid during pulmonary congestion increases conductance, causing a decrease in impedance.⁷ RR will increase as patients experience dyspnea with a more rapid, shallow breath and may trigger alerts closer to the actual HF event than other sensors. Nearly 90% of patients hospitalized for HF experience shortness of breath.^8,9 NHR is used as a surrogate for resting heart rate (HR). A high resting HR is correlated with the progression of coronary atherosclerosis, harmful effects on left ventricular function, and increased risk of myocardial ischemia and ventricular arrhythmias.¹⁰

One of the challenges with preventing hospitalizations may be the lack of patient reported symptoms leading up to the event. The purpose of the sensors and HeartLogic index is to identify patients a median of 34 days before an HF event (HF admission or unscheduled intervention with intravenous treatment) with a sensitivity rate of 70%.³ According to real-word experience data, alerts have been found to precede HF symptoms by a median of 12 days and HF events such as hospitalizations by a median of 38 days, with an overall 67% reduction in HF hospitalizations when integrated into clinical care.^11,12

MANAGE-HF evaluated 191 patients with HF with reduced ejection fraction (HFrEF) (< 35%), New York Heart Association class II-III symptoms, and who had an implanted CRT and/or ICD to develop an alert management guide to optimize medical treatment.¹² It aimed to adjust patient regimen within 6 days of an elevated Heart- Logic index by either initiation, escalation, or maintenance of HF treatment depending on the index trend after the initial alert. This trial found that by focusing on such optimization, HF treatment was augmented during 74% of the 585 alert cases and during 54% of 3290 weekly alerts.

Initiation and uptitration of the 4 primary components of guideline-directed medical therapy (GDMT) are recommended by the 2022 Heart Failure Guidelines to reduce mortality and morbidity in patients with HFrEF.² The 4 pillars of GDMT consist of -blockers (BB), sodium-glucose cotransporter type 2 inhibitors (SGLT2i), mineralocorticoid receptor antagonists (MRA), and renin-angiotensin-system inhibitors (RASi) including angiotensin II receptor blocker/neprilysin inhibitors (ARNi), angiotensin-converting enzyme inhibitors (ACEi) and angiotensin II receptor blockers (ARB) (Appendix 1). Obtaining and titrating to target doses wherever possible is recommended, as those were the doses that established safety and efficacy in patients with HFrEF in clinical trials.² Pharmacists are adequately equipped to optimize HF GDMT and appropriately monitor drug response.

Through the use of HeartLogic in clinical practice, patients with HF have been shown to have improved clinical outcomes and are more likely to receive effective care; 80% of alerts were shown to provide new information to clinicians.¹³ This project sought to quantify the total number and types of pharmacist interventions driven by integration of HeartLogic index monitoring into practice.

Methods

The West Palm Beach Veterans Affairs Medical Center (WPBVAMC) Research Program Office approved this project and determined it was exempt from institutional review board oversight. Patients were screened retrospectively and prospectively from May 26, 2022, through December 31, 2022, by a cardiology clinical pharmacist practitioner (CPP) and a cardiology pharmacy resident using the local monitoring suite for the HeartLogic-compatible device, LATITUDE NXT. Read-only access to the local monitoring suit was granted by the National Cardiac Device Surveillance Program. Training for HeartLogic was completed through continuing education courses provided by Boston Scientific. Additional information was provided by Boston Scientific representative presentations and collaboration with WPBVAMC pacemaker clinic HCPs.

Individuals included were patients with HeartLogic-capable ICDs. A HeartLogic alert had to be present at initial patient contact. Patients were also contacted as part of routine clinical practice, but no formal number or frequency of calls to patients was required. The initial contact must be with a pharmacist for the patient to be included, but subsequent contact by other HCPs was included. Patients in the cardiology clinic are required to meet with a cardiologist at least annually; however, interim visits can be completed by advanced practice registered nurse practitioners, physicians assistants, or CPPs.

Patients in alert status were contacted by telephone and appropriate modifications of HF therapy were made by the CPP based on score metrics, medical record review, and patient interview. Information surrounding the initial alert, baseline patient data, medication and monitoring interventions made, and clinical outcomes such as hospitalization, symptom improvement, follow-up, and mortality were collected. Information for each encounter was collected until 42 days from the initial date of pharmacist contact.

Clinically successful tolerability of intervention implementation was defined as tolerability, adherence, and lack of adverse effects (AEs) per patient report at follow-up or within 42 days from initial alert (Appendix 2). A decrease in dose was not counted as intolerance. A single patient may have been counted as multiple encounters if the original intervention resulted in treatment intolerance and the patient remained in alert or if an additional alert occurred after 42 days of the initial alert. There were no specific time criteria for follow-up, which occurred at the CPP’s discretion.

There was no mandated algorithm used to alter medications based on the Heart- Logic score, nor were there required minimum or maximum numbers of interventions after an alert. Patient contact by telephone initiated an encounter. The types of interventions included medication increases, decreases, initiation, discontinuation, or no medication change. Each medication change and rationale, if applicable, was recorded for the encounter ≤ 42 days after the initial contact date. If a medication with required monitoring parameters was augmented, the pharmacist was responsible for ordering laboratory testing and follow-up. Most interventions were completed by telephone; however, some patients had in-person visits in the HF CPP clinic.

Outcomes

The primary outcome was the number of pharmacist interventions made to optimize GDMT, defined as either an initiation or dose increase. Key intervention analysis included the use and dosing of the 4 primary components of HF GDMT: BB, SGLT2i, MRA, and ARNi/ARB/ACEi. In addition to the 4 primary components of GDMT, loop diuretic changes were also recorded and analyzed. Secondary endpoints were the number of HF hospitalizations ≤ 42 days after the initial alert, and the effect of medication interventions on device metrics, patient symptoms, and tolerability. Successful tolerability was defined as continued use of augmented GDMT without intolerance or discontinuation. The primary analysis was analyzed through descriptive statistics. Median changes in HeartLogic scores and metrics from baseline were analyzed using a paired, 2-sided t test with an α of .05 to detect significance.

Results

There were 39 WPBVAMC patients with a HeartLogic-capable device. Twenty-one alert encounters were analyzed in 16 patients (41%) over 31 weeks of data collection. The 16 patients at baseline had a mean age of 74 years, all were male, and 12 (75%) were White. Eight patients (50%) had a recent ejection fraction (EF) between 30% and 40%. Three patients had an EF ≥ 40%. At the time of alert, 15 patients used BB (94%), 10 used loop diuretics (63%), and 9 used ARNi (56%) (Table 1).

There were 23 medication changes made during initial contact. The most common change was starting an SGLT2i (30%; n = 7), followed by starting an MRA (22%; n = 5), and increasing the ARNi dose (22%; n = 5). At the initial contact, ≥ 1 medication optimization occurred in 95% (n = 20) of encounters. The CPP contacted patients a mean of 4.8 days after the initial alert.

Patients were taking a mean of 2.6 primary GDMT medications at baseline and 3.0 at 42 days. CPP encounters led to a mean of 1.8 medication changes over the 6-week period (range, 0-5). Seventeen medications were started, 13 medications were increased, 3 medications were decreased, and 4 medications were stopped (Table 2). One ACEi and 1 ARB were switched as a therapy escalation to an ARNi. One patient was on 1 of 4 primary GDMTs at baseline, which increased to 4 GDMT agents at 42 days.

SGLT2 inhibitors were added most often at initial contact (54%) and throughout the 42-day period (41%). The most common successfully tolerated optimizations were RASi, followed by MRA, SGLT2 inhibitors, BB, and loop diuretics with 11, 6, 5, 3, and 2 patients, respectively. Interventions were tolerated by 90% of patients, and no HF hospitalization occurred during follow-up. All possible rationales for patients with the same or reduced number of GDMT at 42 days compared with baseline are shown in Appendix 2.

Device Metrics

During initial contact, the most common HeartLogic metric category that was predominantly worsening were heart sounds (S1, S3, and S3/S1 ratio), followed by compensatory mechanism sensors (NHR and RR) and congestion (impedance) at rates of 61.9%, 23.8%, and 14.3%, respectively (Figure 2).

The median HeartLogic index score was 18 at baseline and 5 at the end of the follow-up period (P < .001). The changes in score and metrics were compared with the type of successfully tolerated GDMT optimization made (Table 3). The GDMT optimization analysis included SGLT2i, RASi, MRA, BB, and loop diuretics. All interventions reduced the overall HeartLogic index score, ranging from a 9.5-point reduction (loop diuretics) to a 16-point reduction (SGLT2i and BB). Optimization of SGLT2i, RASi, and loop diuretics had a positive impact on S1 score. For S3 score, SGLT2i, MRA, and BB had a positive impact. All medications, except for SGLT2i therapy, reduced the NHR score. Optimization of MRA, SGLT2i, and BB had positive impacts on the impedance score. All medications reduced RR from baseline. Only SGLT2i and loop diuretics had positive impacts on the activity score.

Clinical Outcomes and Adverse Effects

Within 42 days of contact, 17 encounters (81%) had ≥ 1 follow-up appointment with a CPP and all 21 patients had ≥  1 follow-up health care team member. One patient had a HF-related hospitalization within 42 days of contact; however, that individual refused the recommended medication intervention. There were 13 encounters (62%) with reported symptoms at the time of the initial alert and 10 (77%) had subjective symptom improvement at 42 days (Appendix 3).

Of 30 medication optimizations, 27 primary GDMT medications were tolerated. Two medication intolerances led to discontinuation (1 SGLT2i and 1 loop diuretic) and 1 patient never started the SGLT2i (Table 4). There was only 1 known patient who did not follow the directions to adjust their medications. That individual was included because the patient agreed to the change during the CPP visit but later reported that he had never started the SGLT2i.

Discussion

The HeartLogic tool created a bridge for patients with HF to work with CPPs as soon as possible to optimize medication therapy to reduce HF events. This study highlights an additional area of expertise and service that CPPs may offer to their specialty HF clinic team. Over 31 weeks, 21 encounters and 30 medication optimizations were completed. These interventions led to significant reductions in HeartLogic scores, improvements in symptoms, and optimization of HF care, most of which were well tolerated.

Additional hemodynamic monitoring devices are available. Similar to HeartLogic, OptiVol is a tool embedded in select Medtronic implantable devices that monitors fluid status. ¹⁴ CardioMEMS is an implantable pulmonary artery pressure sensor used as a presymptomatic data point to alert clinicians when HF is worsening. In the CHAMPION trial, the use of CardioMEMS showed a 28% reduction in HF-related hospitalization at 6 months.¹⁵ Conversely, in the GUIDE-HF trial, monitoring with CardioMEMS did not significantly reduce the composite endpoint of mortality and total HF events.¹⁶ Therefore, remote hemodynamic monitoring has variable results and the use of these tools remains uncertain per the clinical guidelines.²

The MANAGE-HF study that contributed to the validation of the HeartLogic tool may provide a comparison with this smaller single-center project. The time to follow-up within 7 days of alert was noted in only 54% of the patients in MANAGE-HF.¹² In this study, 86% of patients received follow-up within 7 days, with a mean of 4.8 days. The quick turnaround from the time of alert to intervention portrays pharmacists as readily available HCPs.

In MANAGE-HF, 89% of medication augmentation involved loop diuretics or thiazides; in our project, loop diuretics were the least frequently changed medication. Most optimizations in this project included ARNi, SGLT2i, BB, and MRA, which have been shown to reduce morbidity and mortality.² Our project included use of SGLT2i therapy to affect HeartLogic metrics, which has not been evaluated previously. SGLT2i were the most commonly initiated medication after an alert. Of the 5 tolerated SGLT2i optimization encounters, 4 were out of alert at 42 days.

SGLT2i resulted in a significant decrease in HeartLogic index score from baseline and were the only class of medication that did not produce a negative change in any metric. In this study, CPPs utilizing and acting on HeartLogic alerts led to 1 (4.8%) hospitalization with HF as the primary reason for admission and no hospitalizations as a secondary cause in 42 days, compared to 37% and 7.9% in the MANAGE-HF in 1 year, respectively. An additional screening 1 year after the initial alert found that 2 (12.5%) of 12 patients had been admitted with 1 HF hospitalization each.

A strength of this study was the ability to use HeartLogic to identify high-risk patients, provide a source of patient contact and monitoring, interpret 5 cardiac sensors, and optimize all HF GDMT, not just volume management. By focusing efforts on making patient contact and pharmacotherapy interventions with morbidity and mortality benefit, remote hemodynamic monitoring may show a clear clinical benefit and become a vital part of HF care.

Limitations

Checking for adherence and tolerance to medications were mainly patient reported if there was a CPP follow-up within 42 days, or potentially through refill history when unclear. However, this limitation is reflective of current practice where patients may have multiple clinicians working to optimize HF care and where there is reliance on patients in order to guide continued therapy. Although unable to explicitly show a reduction in HF events given lack of comparator group, the interventions made are associated with improved outcomes and thus would be expected to improve patient outcomes. Changes in vital signs were not tracked as part of this project, however the main rationale for changes made were to optimize GDMT therapy, not specifically to impact vital sign measures.

HeartLogic alerts prompted identification of high-risk patients with HF, pharmacist evaluation and outreach, patient-focused pharmacotherapy care, and beneficial patient outcomes. With only 2 cardiology CPPs checking alerts once weekly, future studies may be needed with larger samples to create algorithms and protocols to increase the clinical utility of this tool on a greater scale.

Conclusions

Cardiology CPP-led HF interventions triggered by HeartLogic alerts lead to effective patient identification, increased access to care, reductions in HeartLogic scores, improvements in symptoms, and optimization of HF care. This project demonstrates the practical utility of the HeartLogic suite in conjunction with CPP care to prioritize treatment for highrisk patients with HF in an efficient manner. The data highlight the potential value of the HeartLogic tool and a CPP in HF care to facilitate initiation and optimization of GDMT to ultimately improve the morbidity and mortality in patients with HF.

Heart failure (HF) is a prevalent disease in the United States affecting > 6.5 million adults and contributing to significant morbidity and mortality.¹ The disease course associated with HF includes potential symptom improvement with intermittent periods of decompensation and possible clinical deterioration. Multiple therapies have been developed to improve outcomes in people with HF—to palliate HF symptoms, prevent hospitalizations, and reduce mortality.² However, the risks of decompensation and hospitalization remain. HF decompensation development may precede clear actionable symptoms such as worsening dyspnea, noticeable edema, or weight gain. Tools to identify patient deterioration and trigger interventions to prevent HF admissions are clinically attractive compared with reliance on subjective factors alone.

Cardiac resynchronization therapy (CRT) and implantable cardioverter-defibrillator (ICD) devices made by Boston Scientific include the HeartLogic monitoring feature. Five main sensors produce an index risk score; an index score > 16 warns clinicians that the patient is at an increased risk for a HF event.³ The 5 sensors are thoracic impedance, first (S1) and third heart sounds (S3), night heart rate (NHR), respiratory rate (RR), and activity. Each sensor can draw attention to the primary driver of the alert and guide health care practitioners (HCPs) to the appropriate interventions.³ A HeartLogic alert example is shown in Figure 1.

The S3 occurs during the early diastolic phase when blood moves into the ventricles. As HF worsens, with a combination of elevated filling pressures and reduced cardiac muscle compliance, S3 can become more pronounced.⁴ The S1 is correlated with the contractility of the left ventricle and will be reduced in patients at risk for HF events.⁵ Physical activity is a long-term prognostic marker in patients with HF; reduced activity is associated with mortality and increased risk of an HF event.⁶ Thoracic impedance is a sensor used to identify pulmonary congestion, pocket infections, pleural/pericardial effusion, and respiratory infections. The accumulation of intrathoracic fluid during pulmonary congestion increases conductance, causing a decrease in impedance.⁷ RR will increase as patients experience dyspnea with a more rapid, shallow breath and may trigger alerts closer to the actual HF event than other sensors. Nearly 90% of patients hospitalized for HF experience shortness of breath.^8,9 NHR is used as a surrogate for resting heart rate (HR). A high resting HR is correlated with the progression of coronary atherosclerosis, harmful effects on left ventricular function, and increased risk of myocardial ischemia and ventricular arrhythmias.¹⁰

One of the challenges with preventing hospitalizations may be the lack of patient reported symptoms leading up to the event. The purpose of the sensors and HeartLogic index is to identify patients a median of 34 days before an HF event (HF admission or unscheduled intervention with intravenous treatment) with a sensitivity rate of 70%.³ According to real-word experience data, alerts have been found to precede HF symptoms by a median of 12 days and HF events such as hospitalizations by a median of 38 days, with an overall 67% reduction in HF hospitalizations when integrated into clinical care.^11,12

MANAGE-HF evaluated 191 patients with HF with reduced ejection fraction (HFrEF) (< 35%), New York Heart Association class II-III symptoms, and who had an implanted CRT and/or ICD to develop an alert management guide to optimize medical treatment.¹² It aimed to adjust patient regimen within 6 days of an elevated Heart- Logic index by either initiation, escalation, or maintenance of HF treatment depending on the index trend after the initial alert. This trial found that by focusing on such optimization, HF treatment was augmented during 74% of the 585 alert cases and during 54% of 3290 weekly alerts.

Initiation and uptitration of the 4 primary components of guideline-directed medical therapy (GDMT) are recommended by the 2022 Heart Failure Guidelines to reduce mortality and morbidity in patients with HFrEF.² The 4 pillars of GDMT consist of -blockers (BB), sodium-glucose cotransporter type 2 inhibitors (SGLT2i), mineralocorticoid receptor antagonists (MRA), and renin-angiotensin-system inhibitors (RASi) including angiotensin II receptor blocker/neprilysin inhibitors (ARNi), angiotensin-converting enzyme inhibitors (ACEi) and angiotensin II receptor blockers (ARB) (Appendix 1). Obtaining and titrating to target doses wherever possible is recommended, as those were the doses that established safety and efficacy in patients with HFrEF in clinical trials.² Pharmacists are adequately equipped to optimize HF GDMT and appropriately monitor drug response.

Through the use of HeartLogic in clinical practice, patients with HF have been shown to have improved clinical outcomes and are more likely to receive effective care; 80% of alerts were shown to provide new information to clinicians.¹³ This project sought to quantify the total number and types of pharmacist interventions driven by integration of HeartLogic index monitoring into practice.

Methods

The West Palm Beach Veterans Affairs Medical Center (WPBVAMC) Research Program Office approved this project and determined it was exempt from institutional review board oversight. Patients were screened retrospectively and prospectively from May 26, 2022, through December 31, 2022, by a cardiology clinical pharmacist practitioner (CPP) and a cardiology pharmacy resident using the local monitoring suite for the HeartLogic-compatible device, LATITUDE NXT. Read-only access to the local monitoring suit was granted by the National Cardiac Device Surveillance Program. Training for HeartLogic was completed through continuing education courses provided by Boston Scientific. Additional information was provided by Boston Scientific representative presentations and collaboration with WPBVAMC pacemaker clinic HCPs.

Individuals included were patients with HeartLogic-capable ICDs. A HeartLogic alert had to be present at initial patient contact. Patients were also contacted as part of routine clinical practice, but no formal number or frequency of calls to patients was required. The initial contact must be with a pharmacist for the patient to be included, but subsequent contact by other HCPs was included. Patients in the cardiology clinic are required to meet with a cardiologist at least annually; however, interim visits can be completed by advanced practice registered nurse practitioners, physicians assistants, or CPPs.

Patients in alert status were contacted by telephone and appropriate modifications of HF therapy were made by the CPP based on score metrics, medical record review, and patient interview. Information surrounding the initial alert, baseline patient data, medication and monitoring interventions made, and clinical outcomes such as hospitalization, symptom improvement, follow-up, and mortality were collected. Information for each encounter was collected until 42 days from the initial date of pharmacist contact.

Clinically successful tolerability of intervention implementation was defined as tolerability, adherence, and lack of adverse effects (AEs) per patient report at follow-up or within 42 days from initial alert (Appendix 2). A decrease in dose was not counted as intolerance. A single patient may have been counted as multiple encounters if the original intervention resulted in treatment intolerance and the patient remained in alert or if an additional alert occurred after 42 days of the initial alert. There were no specific time criteria for follow-up, which occurred at the CPP’s discretion.

There was no mandated algorithm used to alter medications based on the Heart- Logic score, nor were there required minimum or maximum numbers of interventions after an alert. Patient contact by telephone initiated an encounter. The types of interventions included medication increases, decreases, initiation, discontinuation, or no medication change. Each medication change and rationale, if applicable, was recorded for the encounter ≤ 42 days after the initial contact date. If a medication with required monitoring parameters was augmented, the pharmacist was responsible for ordering laboratory testing and follow-up. Most interventions were completed by telephone; however, some patients had in-person visits in the HF CPP clinic.

Outcomes

The primary outcome was the number of pharmacist interventions made to optimize GDMT, defined as either an initiation or dose increase. Key intervention analysis included the use and dosing of the 4 primary components of HF GDMT: BB, SGLT2i, MRA, and ARNi/ARB/ACEi. In addition to the 4 primary components of GDMT, loop diuretic changes were also recorded and analyzed. Secondary endpoints were the number of HF hospitalizations ≤ 42 days after the initial alert, and the effect of medication interventions on device metrics, patient symptoms, and tolerability. Successful tolerability was defined as continued use of augmented GDMT without intolerance or discontinuation. The primary analysis was analyzed through descriptive statistics. Median changes in HeartLogic scores and metrics from baseline were analyzed using a paired, 2-sided t test with an α of .05 to detect significance.

Results

There were 39 WPBVAMC patients with a HeartLogic-capable device. Twenty-one alert encounters were analyzed in 16 patients (41%) over 31 weeks of data collection. The 16 patients at baseline had a mean age of 74 years, all were male, and 12 (75%) were White. Eight patients (50%) had a recent ejection fraction (EF) between 30% and 40%. Three patients had an EF ≥ 40%. At the time of alert, 15 patients used BB (94%), 10 used loop diuretics (63%), and 9 used ARNi (56%) (Table 1).

There were 23 medication changes made during initial contact. The most common change was starting an SGLT2i (30%; n = 7), followed by starting an MRA (22%; n = 5), and increasing the ARNi dose (22%; n = 5). At the initial contact, ≥ 1 medication optimization occurred in 95% (n = 20) of encounters. The CPP contacted patients a mean of 4.8 days after the initial alert.

Patients were taking a mean of 2.6 primary GDMT medications at baseline and 3.0 at 42 days. CPP encounters led to a mean of 1.8 medication changes over the 6-week period (range, 0-5). Seventeen medications were started, 13 medications were increased, 3 medications were decreased, and 4 medications were stopped (Table 2). One ACEi and 1 ARB were switched as a therapy escalation to an ARNi. One patient was on 1 of 4 primary GDMTs at baseline, which increased to 4 GDMT agents at 42 days.

SGLT2 inhibitors were added most often at initial contact (54%) and throughout the 42-day period (41%). The most common successfully tolerated optimizations were RASi, followed by MRA, SGLT2 inhibitors, BB, and loop diuretics with 11, 6, 5, 3, and 2 patients, respectively. Interventions were tolerated by 90% of patients, and no HF hospitalization occurred during follow-up. All possible rationales for patients with the same or reduced number of GDMT at 42 days compared with baseline are shown in Appendix 2.

Device Metrics

During initial contact, the most common HeartLogic metric category that was predominantly worsening were heart sounds (S1, S3, and S3/S1 ratio), followed by compensatory mechanism sensors (NHR and RR) and congestion (impedance) at rates of 61.9%, 23.8%, and 14.3%, respectively (Figure 2).

The median HeartLogic index score was 18 at baseline and 5 at the end of the follow-up period (P < .001). The changes in score and metrics were compared with the type of successfully tolerated GDMT optimization made (Table 3). The GDMT optimization analysis included SGLT2i, RASi, MRA, BB, and loop diuretics. All interventions reduced the overall HeartLogic index score, ranging from a 9.5-point reduction (loop diuretics) to a 16-point reduction (SGLT2i and BB). Optimization of SGLT2i, RASi, and loop diuretics had a positive impact on S1 score. For S3 score, SGLT2i, MRA, and BB had a positive impact. All medications, except for SGLT2i therapy, reduced the NHR score. Optimization of MRA, SGLT2i, and BB had positive impacts on the impedance score. All medications reduced RR from baseline. Only SGLT2i and loop diuretics had positive impacts on the activity score.

Clinical Outcomes and Adverse Effects

Within 42 days of contact, 17 encounters (81%) had ≥ 1 follow-up appointment with a CPP and all 21 patients had ≥  1 follow-up health care team member. One patient had a HF-related hospitalization within 42 days of contact; however, that individual refused the recommended medication intervention. There were 13 encounters (62%) with reported symptoms at the time of the initial alert and 10 (77%) had subjective symptom improvement at 42 days (Appendix 3).

Of 30 medication optimizations, 27 primary GDMT medications were tolerated. Two medication intolerances led to discontinuation (1 SGLT2i and 1 loop diuretic) and 1 patient never started the SGLT2i (Table 4). There was only 1 known patient who did not follow the directions to adjust their medications. That individual was included because the patient agreed to the change during the CPP visit but later reported that he had never started the SGLT2i.

Discussion

The HeartLogic tool created a bridge for patients with HF to work with CPPs as soon as possible to optimize medication therapy to reduce HF events. This study highlights an additional area of expertise and service that CPPs may offer to their specialty HF clinic team. Over 31 weeks, 21 encounters and 30 medication optimizations were completed. These interventions led to significant reductions in HeartLogic scores, improvements in symptoms, and optimization of HF care, most of which were well tolerated.

Additional hemodynamic monitoring devices are available. Similar to HeartLogic, OptiVol is a tool embedded in select Medtronic implantable devices that monitors fluid status. ¹⁴ CardioMEMS is an implantable pulmonary artery pressure sensor used as a presymptomatic data point to alert clinicians when HF is worsening. In the CHAMPION trial, the use of CardioMEMS showed a 28% reduction in HF-related hospitalization at 6 months.¹⁵ Conversely, in the GUIDE-HF trial, monitoring with CardioMEMS did not significantly reduce the composite endpoint of mortality and total HF events.¹⁶ Therefore, remote hemodynamic monitoring has variable results and the use of these tools remains uncertain per the clinical guidelines.²

The MANAGE-HF study that contributed to the validation of the HeartLogic tool may provide a comparison with this smaller single-center project. The time to follow-up within 7 days of alert was noted in only 54% of the patients in MANAGE-HF.¹² In this study, 86% of patients received follow-up within 7 days, with a mean of 4.8 days. The quick turnaround from the time of alert to intervention portrays pharmacists as readily available HCPs.

In MANAGE-HF, 89% of medication augmentation involved loop diuretics or thiazides; in our project, loop diuretics were the least frequently changed medication. Most optimizations in this project included ARNi, SGLT2i, BB, and MRA, which have been shown to reduce morbidity and mortality.² Our project included use of SGLT2i therapy to affect HeartLogic metrics, which has not been evaluated previously. SGLT2i were the most commonly initiated medication after an alert. Of the 5 tolerated SGLT2i optimization encounters, 4 were out of alert at 42 days.

SGLT2i resulted in a significant decrease in HeartLogic index score from baseline and were the only class of medication that did not produce a negative change in any metric. In this study, CPPs utilizing and acting on HeartLogic alerts led to 1 (4.8%) hospitalization with HF as the primary reason for admission and no hospitalizations as a secondary cause in 42 days, compared to 37% and 7.9% in the MANAGE-HF in 1 year, respectively. An additional screening 1 year after the initial alert found that 2 (12.5%) of 12 patients had been admitted with 1 HF hospitalization each.

A strength of this study was the ability to use HeartLogic to identify high-risk patients, provide a source of patient contact and monitoring, interpret 5 cardiac sensors, and optimize all HF GDMT, not just volume management. By focusing efforts on making patient contact and pharmacotherapy interventions with morbidity and mortality benefit, remote hemodynamic monitoring may show a clear clinical benefit and become a vital part of HF care.

Limitations

Checking for adherence and tolerance to medications were mainly patient reported if there was a CPP follow-up within 42 days, or potentially through refill history when unclear. However, this limitation is reflective of current practice where patients may have multiple clinicians working to optimize HF care and where there is reliance on patients in order to guide continued therapy. Although unable to explicitly show a reduction in HF events given lack of comparator group, the interventions made are associated with improved outcomes and thus would be expected to improve patient outcomes. Changes in vital signs were not tracked as part of this project, however the main rationale for changes made were to optimize GDMT therapy, not specifically to impact vital sign measures.

HeartLogic alerts prompted identification of high-risk patients with HF, pharmacist evaluation and outreach, patient-focused pharmacotherapy care, and beneficial patient outcomes. With only 2 cardiology CPPs checking alerts once weekly, future studies may be needed with larger samples to create algorithms and protocols to increase the clinical utility of this tool on a greater scale.

Conclusions

Cardiology CPP-led HF interventions triggered by HeartLogic alerts lead to effective patient identification, increased access to care, reductions in HeartLogic scores, improvements in symptoms, and optimization of HF care. This project demonstrates the practical utility of the HeartLogic suite in conjunction with CPP care to prioritize treatment for highrisk patients with HF in an efficient manner. The data highlight the potential value of the HeartLogic tool and a CPP in HF care to facilitate initiation and optimization of GDMT to ultimately improve the morbidity and mortality in patients with HF.

Hypertension remains one of the most important modifiable risk factors for the prevention of cardiovascular (CV) events. According to a population-based study, 25% of CV events (CV death, heart disease, coronary revascularization, stroke, or heart failure) are attributable to hypertension.¹ Recent guidelines have emphasized the importance of accurate blood pressure (BP) measurement in facilitating appropriate hypertension diagnosis and management.^2-4

Currently, there are different BP measurement methods endorsed by practice guidelines. These include conventional in-office measurement, 24-hour ambulatory BP monitoring (ABPM), home BP monitoring (HBPM), and automated office BP (AOBP) measurement.^2-4 AOBP device protocols vary but generally involve devices automatically taking multiple BP measurements while the patient is unattended. These measurements are then presented as a single averaged reading, with individual BP values available for review by the clinician.

Researchers have found that AOBP measurements have a greater association with ABPM values and can mitigate the white coat effect observed in a substantial proportion of patients during in-clinic BP measurement.⁵ A meta-analysis found that the use of AOBP was associated with a 10.5 mm Hg reduction in systolic BP (SBP) compared with traditional office-based BP assessments.⁵ Similarly, a separate meta-analysis found that AOBP SBP measures were on average 14.5 mm Hg lower than routine office or research setting values.⁶ In addition, CV risk outcomes data support the use of AOBP to screen and manage patients with hypertension. The Cardiovascular Health Awareness Program (CHAP) study used AOBP values to determine the risk for CV events (myocardial infarction, congestive heart failure, and stroke) in community-based patients aged ≥ 65 years.⁷ The study showed a significantly higher risk of CV events in patients with an SBP of 135 to 144 mm Hg and a diastolic BP (DBP) of 80 to 89 mm Hg. Therefore, the CHAP study researchers suggested an AOBP target of < 135/85 mm Hg to decrease the risk of CV events.⁷The landmark SPRINT trial, which was a major contributor to the development of BP target recommendations in guidelines, utilized AOBP to classify hypertension and guide management.^2-4,8 SPRINT ultimately showed that intensive BP-lowering treatment (to SBP < 120 mm Hg) was associated with a 25% reduction in major CV events and a 27% reduction in all-cause mortality.⁸ Other evaluations found a close association between AOBP values and left ventricular mass index and carotid artery wall thickness as surrogate markers for end-organ damage.^9,10 These data show AOBP as a reliable method to guide antihypertensive therapy interventions in the clinical setting.

Considering these proposed advantages, the 2017 Canadian guidelines for hypertension management recommend AOBP as the preferred method for clinic-based BP measurement, and the 2018 European Society of Cardiology/European Society of Hypertension blood pressure guidelines recommend the use of AOBP when feasible.^3,4 The 2017 American College of Cardiology/American Heart Association Guideline for the Prevention, Detection, Evaluation, and Management of High Blood Pressure in Adults also discusses AOBP as a method to minimize potential confounders in BP values.²

This study evaluated the difference between AOBP and conventional in-office BP measurements obtained during cardiology clinic visits at the West Palm Beach Veterans Affairs Medical Center (WPBVAMC).

METHODS

A retrospective review of AOBP measurements was performed at the WPBVAMC cardiology clinic between May 26, 2017, and February 19, 2019. These AOBP measurements were taken at the discretion of a nurse or other clinician after initial, conventional BP measurements had been taken as part of clinic check-in procedures. No formal protocols dictated the use or timing of AOBP measurements. Similarly, the AOBP results were factored into clinical care decisions.

Clinicians at the cardiology clinic used AOBP averages that were derived using the BpTRU BPM-100 (BpTRU Medical Devices) meter, which averaged 5 BP readings taken at 1-minute intervals. Clinicians selected cuff size based on manufacturer recommendations. The testing was done with the patient seated alone in either a nursing triage area or a clinic office.

Data collected during the retrospective review included the clinician associated with the visit, the patient’s physical location and accompaniment status during AOBP measurement, conventionally measured BP and heart rates, and AOBP-derived BP and heart rate averages. Differences in BP values were compared with the paired t test, while binary comparisons were conducted through the McNemar test. Data collection and analysis were performed using Microsoft Excel.

During data collection, all information was stored in a secure drive accessible only to the investigators. The project was approved by the West Palm Beach Veterans Affairs Healthcare System Research and Development Committee as a nonresearch activity in accordance with Veterans Health Administration Handbook 1058.05; thus, institutional review board approval was not required.

RESULTS

Ninety-five nonconsecutive patients were included in the analysis. AOBP measurements were taken with the patient sitting alone in either a clinic office (n = 83) or nursing triage area (n = 12). Most patients were coming in for follow-up appointments; 13 patients (14%) had appointments related to a 24-hour ABPM session.

The mean SBP and DBP values were lower for the AOBP measurements vs the conventional BP measurements (mean SBP difference, 14.6 mm Hg; P < .001; mean DBP difference, 3.5 mm Hg; P = .0002) (Table). There were no appreciable differences in heart rates. The white coat effect was suggested based on an SBP reduction of > 20 mm Hg from conventional to AOBP measurements in 22 patients (23%), a DBP reduction of > 10 mm Hg in 21 patients (22%), and a reduction in both values in 8 patients (8%).

A controlled BP (< 130/80 mm Hg) was more common in the AOBP group than in the conventional group (22% vs 7%, respectively; P =.001).² Review of conventional BP measurements indicated that 11 patients had systolic readings ≥ 180 mm Hg, 2 had diastolic readings ≥ 110 mm Hg, and 1 had a reading that was ≥ 180/110 mm Hg. AOBP measurements indicated that these 14 patients had SBP readings < 180 mm Hg and DBP readings < 110 mm Hg. The use of AOBP measurements may have mitigated unnecessary emergency room visits for these patients.

On review of clinic notes and actions associated with episodes of AOBP testing during routine follow-up clinic appointments, AOBP was determined to be useful with regard to clinical decision-making for 65 (79%) patients. Impacts of AOBP inclusion vs conventional BP assessments included clinician notation of AOBP, support for making changes that would have been considered based on conventional BP assessment. AOBP results gave support to forgoing a therapeutic intervention (ie, therapy addition or intensification) that may have been pursued based on conventional BP measurements in 25 of 82 patients (30%). These data suggest that AOBP readings can be useful and actionable by clinicians.

DISCUSSION

The findings of this study add to the growing evidence regarding AOBP use, application, and advantages in clinical practice. In this evaluation, the mean difference in SBP and DBP was 14.6 mm Hg and 3.5 mm Hg, respectively, from the conventional office measurements to the AOBP measurements. This difference is similar to that reported by the CAMBO trial and other evaluations, where the use of AOBP measurements corresponded to a reduction in SBP of between 10 and 20 mm Hg vs conventional measures.^5,11-18

These findings showed a significantly higher percentage of controlled BP values (< 130/80 mm Hg) with AOBP values compared with conventional office measurements. The data supported the decision to defer antihypertensive therapy intervention in 30% of patients. Without AOBP data, patients may have been classified as uncontrolled, prompting therapy addition or intensification that could increase the risk of adverse events. Additionally, 14 patients would have met the criteria for hypertensive urgency under the guidelines at that time.² With the use of AOBP readings, none of these patients were identified as having a hypertensive urgency, and they avoided an acute care referral or urgent intervention.

The discrepancy between AOBP and conventional office BP measurements suggested a white coat effect based on SBP and DBP readings in 22 (23%) and 21 (22%) patients, respectively. Practice guidelines recommend ABPM to mitigate a potential white coat effect.^2-4 However, ABPM can be inconvenient for patients, as they need to travel to and from the clinic for fitting and removal (assuming that a facility has the device available for patient use). In addition, some patients may find it uncomfortable. Based on the correlation between AOBP and awake ABPM values, AOBP represents a feasible way to identify a white coat effect.

AOBP monitoring does not appear to be affected by the type of practice setting, as it has been evaluated in a variety of locations, including community-based pharmacies, primary care offices, and waiting rooms.^12,19-22 However, potential AOBP implementation challenges may include office space constraints, clinician perception that it will delay workflow, and device cost. Costs associated with an AOBP meter vary widely based on device and procurement source, but have been estimated to range from $650 to > $2000.²³ Published reports have described how to overcome AOBP implementation barriers.^24,25

Limitations

The results of this evaluation should be interpreted cautiously due to several limitations. First, the retrospective study was conducted at a single clinic that may not be representative of other Veterans Health Administration or community-based populations. In addition, patient data such as age, sex, and body mass index were not available. AOBP measurements were obtained at the discretion of the clinician and not according to a prespecified protocol.

Conclusions

This analysis showed AOBP measurement leads to a greater percentage of controlled BP values compared with conventional office BP measurement, positioning it as a way to reduce BP misclassification, prevent potentially unnecessary therapeutic interventions, and mitigate the white coat effect.

Hypertension remains one of the most important modifiable risk factors for the prevention of cardiovascular (CV) events. According to a population-based study, 25% of CV events (CV death, heart disease, coronary revascularization, stroke, or heart failure) are attributable to hypertension.¹ Recent guidelines have emphasized the importance of accurate blood pressure (BP) measurement in facilitating appropriate hypertension diagnosis and management.^2-4

Currently, there are different BP measurement methods endorsed by practice guidelines. These include conventional in-office measurement, 24-hour ambulatory BP monitoring (ABPM), home BP monitoring (HBPM), and automated office BP (AOBP) measurement.^2-4 AOBP device protocols vary but generally involve devices automatically taking multiple BP measurements while the patient is unattended. These measurements are then presented as a single averaged reading, with individual BP values available for review by the clinician.

Researchers have found that AOBP measurements have a greater association with ABPM values and can mitigate the white coat effect observed in a substantial proportion of patients during in-clinic BP measurement.⁵ A meta-analysis found that the use of AOBP was associated with a 10.5 mm Hg reduction in systolic BP (SBP) compared with traditional office-based BP assessments.⁵ Similarly, a separate meta-analysis found that AOBP SBP measures were on average 14.5 mm Hg lower than routine office or research setting values.⁶ In addition, CV risk outcomes data support the use of AOBP to screen and manage patients with hypertension. The Cardiovascular Health Awareness Program (CHAP) study used AOBP values to determine the risk for CV events (myocardial infarction, congestive heart failure, and stroke) in community-based patients aged ≥ 65 years.⁷ The study showed a significantly higher risk of CV events in patients with an SBP of 135 to 144 mm Hg and a diastolic BP (DBP) of 80 to 89 mm Hg. Therefore, the CHAP study researchers suggested an AOBP target of < 135/85 mm Hg to decrease the risk of CV events.⁷The landmark SPRINT trial, which was a major contributor to the development of BP target recommendations in guidelines, utilized AOBP to classify hypertension and guide management.^2-4,8 SPRINT ultimately showed that intensive BP-lowering treatment (to SBP < 120 mm Hg) was associated with a 25% reduction in major CV events and a 27% reduction in all-cause mortality.⁸ Other evaluations found a close association between AOBP values and left ventricular mass index and carotid artery wall thickness as surrogate markers for end-organ damage.^9,10 These data show AOBP as a reliable method to guide antihypertensive therapy interventions in the clinical setting.

Considering these proposed advantages, the 2017 Canadian guidelines for hypertension management recommend AOBP as the preferred method for clinic-based BP measurement, and the 2018 European Society of Cardiology/European Society of Hypertension blood pressure guidelines recommend the use of AOBP when feasible.^3,4 The 2017 American College of Cardiology/American Heart Association Guideline for the Prevention, Detection, Evaluation, and Management of High Blood Pressure in Adults also discusses AOBP as a method to minimize potential confounders in BP values.²

This study evaluated the difference between AOBP and conventional in-office BP measurements obtained during cardiology clinic visits at the West Palm Beach Veterans Affairs Medical Center (WPBVAMC).

METHODS

A retrospective review of AOBP measurements was performed at the WPBVAMC cardiology clinic between May 26, 2017, and February 19, 2019. These AOBP measurements were taken at the discretion of a nurse or other clinician after initial, conventional BP measurements had been taken as part of clinic check-in procedures. No formal protocols dictated the use or timing of AOBP measurements. Similarly, the AOBP results were factored into clinical care decisions.

Clinicians at the cardiology clinic used AOBP averages that were derived using the BpTRU BPM-100 (BpTRU Medical Devices) meter, which averaged 5 BP readings taken at 1-minute intervals. Clinicians selected cuff size based on manufacturer recommendations. The testing was done with the patient seated alone in either a nursing triage area or a clinic office.

Data collected during the retrospective review included the clinician associated with the visit, the patient’s physical location and accompaniment status during AOBP measurement, conventionally measured BP and heart rates, and AOBP-derived BP and heart rate averages. Differences in BP values were compared with the paired t test, while binary comparisons were conducted through the McNemar test. Data collection and analysis were performed using Microsoft Excel.

During data collection, all information was stored in a secure drive accessible only to the investigators. The project was approved by the West Palm Beach Veterans Affairs Healthcare System Research and Development Committee as a nonresearch activity in accordance with Veterans Health Administration Handbook 1058.05; thus, institutional review board approval was not required.

RESULTS

Ninety-five nonconsecutive patients were included in the analysis. AOBP measurements were taken with the patient sitting alone in either a clinic office (n = 83) or nursing triage area (n = 12). Most patients were coming in for follow-up appointments; 13 patients (14%) had appointments related to a 24-hour ABPM session.

The mean SBP and DBP values were lower for the AOBP measurements vs the conventional BP measurements (mean SBP difference, 14.6 mm Hg; P < .001; mean DBP difference, 3.5 mm Hg; P = .0002) (Table). There were no appreciable differences in heart rates. The white coat effect was suggested based on an SBP reduction of > 20 mm Hg from conventional to AOBP measurements in 22 patients (23%), a DBP reduction of > 10 mm Hg in 21 patients (22%), and a reduction in both values in 8 patients (8%).

A controlled BP (< 130/80 mm Hg) was more common in the AOBP group than in the conventional group (22% vs 7%, respectively; P =.001).² Review of conventional BP measurements indicated that 11 patients had systolic readings ≥ 180 mm Hg, 2 had diastolic readings ≥ 110 mm Hg, and 1 had a reading that was ≥ 180/110 mm Hg. AOBP measurements indicated that these 14 patients had SBP readings < 180 mm Hg and DBP readings < 110 mm Hg. The use of AOBP measurements may have mitigated unnecessary emergency room visits for these patients.

On review of clinic notes and actions associated with episodes of AOBP testing during routine follow-up clinic appointments, AOBP was determined to be useful with regard to clinical decision-making for 65 (79%) patients. Impacts of AOBP inclusion vs conventional BP assessments included clinician notation of AOBP, support for making changes that would have been considered based on conventional BP assessment. AOBP results gave support to forgoing a therapeutic intervention (ie, therapy addition or intensification) that may have been pursued based on conventional BP measurements in 25 of 82 patients (30%). These data suggest that AOBP readings can be useful and actionable by clinicians.

DISCUSSION

The findings of this study add to the growing evidence regarding AOBP use, application, and advantages in clinical practice. In this evaluation, the mean difference in SBP and DBP was 14.6 mm Hg and 3.5 mm Hg, respectively, from the conventional office measurements to the AOBP measurements. This difference is similar to that reported by the CAMBO trial and other evaluations, where the use of AOBP measurements corresponded to a reduction in SBP of between 10 and 20 mm Hg vs conventional measures.^5,11-18

These findings showed a significantly higher percentage of controlled BP values (< 130/80 mm Hg) with AOBP values compared with conventional office measurements. The data supported the decision to defer antihypertensive therapy intervention in 30% of patients. Without AOBP data, patients may have been classified as uncontrolled, prompting therapy addition or intensification that could increase the risk of adverse events. Additionally, 14 patients would have met the criteria for hypertensive urgency under the guidelines at that time.² With the use of AOBP readings, none of these patients were identified as having a hypertensive urgency, and they avoided an acute care referral or urgent intervention.

The discrepancy between AOBP and conventional office BP measurements suggested a white coat effect based on SBP and DBP readings in 22 (23%) and 21 (22%) patients, respectively. Practice guidelines recommend ABPM to mitigate a potential white coat effect.^2-4 However, ABPM can be inconvenient for patients, as they need to travel to and from the clinic for fitting and removal (assuming that a facility has the device available for patient use). In addition, some patients may find it uncomfortable. Based on the correlation between AOBP and awake ABPM values, AOBP represents a feasible way to identify a white coat effect.

AOBP monitoring does not appear to be affected by the type of practice setting, as it has been evaluated in a variety of locations, including community-based pharmacies, primary care offices, and waiting rooms.^12,19-22 However, potential AOBP implementation challenges may include office space constraints, clinician perception that it will delay workflow, and device cost. Costs associated with an AOBP meter vary widely based on device and procurement source, but have been estimated to range from $650 to > $2000.²³ Published reports have described how to overcome AOBP implementation barriers.^24,25

Limitations

The results of this evaluation should be interpreted cautiously due to several limitations. First, the retrospective study was conducted at a single clinic that may not be representative of other Veterans Health Administration or community-based populations. In addition, patient data such as age, sex, and body mass index were not available. AOBP measurements were obtained at the discretion of the clinician and not according to a prespecified protocol.

Conclusions

This analysis showed AOBP measurement leads to a greater percentage of controlled BP values compared with conventional office BP measurement, positioning it as a way to reduce BP misclassification, prevent potentially unnecessary therapeutic interventions, and mitigate the white coat effect.

Hypertension remains one of the most important modifiable risk factors for the prevention of cardiovascular (CV) events. According to a population-based study, 25% of CV events (CV death, heart disease, coronary revascularization, stroke, or heart failure) are attributable to hypertension.¹ Recent guidelines have emphasized the importance of accurate blood pressure (BP) measurement in facilitating appropriate hypertension diagnosis and management.^2-4

Currently, there are different BP measurement methods endorsed by practice guidelines. These include conventional in-office measurement, 24-hour ambulatory BP monitoring (ABPM), home BP monitoring (HBPM), and automated office BP (AOBP) measurement.^2-4 AOBP device protocols vary but generally involve devices automatically taking multiple BP measurements while the patient is unattended. These measurements are then presented as a single averaged reading, with individual BP values available for review by the clinician.

Researchers have found that AOBP measurements have a greater association with ABPM values and can mitigate the white coat effect observed in a substantial proportion of patients during in-clinic BP measurement.⁵ A meta-analysis found that the use of AOBP was associated with a 10.5 mm Hg reduction in systolic BP (SBP) compared with traditional office-based BP assessments.⁵ Similarly, a separate meta-analysis found that AOBP SBP measures were on average 14.5 mm Hg lower than routine office or research setting values.⁶ In addition, CV risk outcomes data support the use of AOBP to screen and manage patients with hypertension. The Cardiovascular Health Awareness Program (CHAP) study used AOBP values to determine the risk for CV events (myocardial infarction, congestive heart failure, and stroke) in community-based patients aged ≥ 65 years.⁷ The study showed a significantly higher risk of CV events in patients with an SBP of 135 to 144 mm Hg and a diastolic BP (DBP) of 80 to 89 mm Hg. Therefore, the CHAP study researchers suggested an AOBP target of < 135/85 mm Hg to decrease the risk of CV events.⁷The landmark SPRINT trial, which was a major contributor to the development of BP target recommendations in guidelines, utilized AOBP to classify hypertension and guide management.^2-4,8 SPRINT ultimately showed that intensive BP-lowering treatment (to SBP < 120 mm Hg) was associated with a 25% reduction in major CV events and a 27% reduction in all-cause mortality.⁸ Other evaluations found a close association between AOBP values and left ventricular mass index and carotid artery wall thickness as surrogate markers for end-organ damage.^9,10 These data show AOBP as a reliable method to guide antihypertensive therapy interventions in the clinical setting.

Considering these proposed advantages, the 2017 Canadian guidelines for hypertension management recommend AOBP as the preferred method for clinic-based BP measurement, and the 2018 European Society of Cardiology/European Society of Hypertension blood pressure guidelines recommend the use of AOBP when feasible.^3,4 The 2017 American College of Cardiology/American Heart Association Guideline for the Prevention, Detection, Evaluation, and Management of High Blood Pressure in Adults also discusses AOBP as a method to minimize potential confounders in BP values.²

This study evaluated the difference between AOBP and conventional in-office BP measurements obtained during cardiology clinic visits at the West Palm Beach Veterans Affairs Medical Center (WPBVAMC).

METHODS

A retrospective review of AOBP measurements was performed at the WPBVAMC cardiology clinic between May 26, 2017, and February 19, 2019. These AOBP measurements were taken at the discretion of a nurse or other clinician after initial, conventional BP measurements had been taken as part of clinic check-in procedures. No formal protocols dictated the use or timing of AOBP measurements. Similarly, the AOBP results were factored into clinical care decisions.

Clinicians at the cardiology clinic used AOBP averages that were derived using the BpTRU BPM-100 (BpTRU Medical Devices) meter, which averaged 5 BP readings taken at 1-minute intervals. Clinicians selected cuff size based on manufacturer recommendations. The testing was done with the patient seated alone in either a nursing triage area or a clinic office.

Data collected during the retrospective review included the clinician associated with the visit, the patient’s physical location and accompaniment status during AOBP measurement, conventionally measured BP and heart rates, and AOBP-derived BP and heart rate averages. Differences in BP values were compared with the paired t test, while binary comparisons were conducted through the McNemar test. Data collection and analysis were performed using Microsoft Excel.

During data collection, all information was stored in a secure drive accessible only to the investigators. The project was approved by the West Palm Beach Veterans Affairs Healthcare System Research and Development Committee as a nonresearch activity in accordance with Veterans Health Administration Handbook 1058.05; thus, institutional review board approval was not required.

RESULTS

Ninety-five nonconsecutive patients were included in the analysis. AOBP measurements were taken with the patient sitting alone in either a clinic office (n = 83) or nursing triage area (n = 12). Most patients were coming in for follow-up appointments; 13 patients (14%) had appointments related to a 24-hour ABPM session.

The mean SBP and DBP values were lower for the AOBP measurements vs the conventional BP measurements (mean SBP difference, 14.6 mm Hg; P < .001; mean DBP difference, 3.5 mm Hg; P = .0002) (Table). There were no appreciable differences in heart rates. The white coat effect was suggested based on an SBP reduction of > 20 mm Hg from conventional to AOBP measurements in 22 patients (23%), a DBP reduction of > 10 mm Hg in 21 patients (22%), and a reduction in both values in 8 patients (8%).

A controlled BP (< 130/80 mm Hg) was more common in the AOBP group than in the conventional group (22% vs 7%, respectively; P =.001).² Review of conventional BP measurements indicated that 11 patients had systolic readings ≥ 180 mm Hg, 2 had diastolic readings ≥ 110 mm Hg, and 1 had a reading that was ≥ 180/110 mm Hg. AOBP measurements indicated that these 14 patients had SBP readings < 180 mm Hg and DBP readings < 110 mm Hg. The use of AOBP measurements may have mitigated unnecessary emergency room visits for these patients.

On review of clinic notes and actions associated with episodes of AOBP testing during routine follow-up clinic appointments, AOBP was determined to be useful with regard to clinical decision-making for 65 (79%) patients. Impacts of AOBP inclusion vs conventional BP assessments included clinician notation of AOBP, support for making changes that would have been considered based on conventional BP assessment. AOBP results gave support to forgoing a therapeutic intervention (ie, therapy addition or intensification) that may have been pursued based on conventional BP measurements in 25 of 82 patients (30%). These data suggest that AOBP readings can be useful and actionable by clinicians.

DISCUSSION

The findings of this study add to the growing evidence regarding AOBP use, application, and advantages in clinical practice. In this evaluation, the mean difference in SBP and DBP was 14.6 mm Hg and 3.5 mm Hg, respectively, from the conventional office measurements to the AOBP measurements. This difference is similar to that reported by the CAMBO trial and other evaluations, where the use of AOBP measurements corresponded to a reduction in SBP of between 10 and 20 mm Hg vs conventional measures.^5,11-18

These findings showed a significantly higher percentage of controlled BP values (< 130/80 mm Hg) with AOBP values compared with conventional office measurements. The data supported the decision to defer antihypertensive therapy intervention in 30% of patients. Without AOBP data, patients may have been classified as uncontrolled, prompting therapy addition or intensification that could increase the risk of adverse events. Additionally, 14 patients would have met the criteria for hypertensive urgency under the guidelines at that time.² With the use of AOBP readings, none of these patients were identified as having a hypertensive urgency, and they avoided an acute care referral or urgent intervention.

The discrepancy between AOBP and conventional office BP measurements suggested a white coat effect based on SBP and DBP readings in 22 (23%) and 21 (22%) patients, respectively. Practice guidelines recommend ABPM to mitigate a potential white coat effect.^2-4 However, ABPM can be inconvenient for patients, as they need to travel to and from the clinic for fitting and removal (assuming that a facility has the device available for patient use). In addition, some patients may find it uncomfortable. Based on the correlation between AOBP and awake ABPM values, AOBP represents a feasible way to identify a white coat effect.

AOBP monitoring does not appear to be affected by the type of practice setting, as it has been evaluated in a variety of locations, including community-based pharmacies, primary care offices, and waiting rooms.^12,19-22 However, potential AOBP implementation challenges may include office space constraints, clinician perception that it will delay workflow, and device cost. Costs associated with an AOBP meter vary widely based on device and procurement source, but have been estimated to range from $650 to > $2000.²³ Published reports have described how to overcome AOBP implementation barriers.^24,25

Limitations

The results of this evaluation should be interpreted cautiously due to several limitations. First, the retrospective study was conducted at a single clinic that may not be representative of other Veterans Health Administration or community-based populations. In addition, patient data such as age, sex, and body mass index were not available. AOBP measurements were obtained at the discretion of the clinician and not according to a prespecified protocol.

Conclusions

This analysis showed AOBP measurement leads to a greater percentage of controlled BP values compared with conventional office BP measurement, positioning it as a way to reduce BP misclassification, prevent potentially unnecessary therapeutic interventions, and mitigate the white coat effect.

Due to the underlying mechanism of atrial fibrillation (Afib), clots can form within the left atrial appendage. Clots that become dislodged may lead to ischemic stroke and possibly death. The 2023 guidelines for atrial fibrillation from the American College of Cardiology and American Heart Association recommend anticoagulation therapy for patients with an Afib diagnosis and a CHA₂DS₂-VASc (congestive heart failure, hypertension, age ≥ 75 years, diabetes, stroke/vascular disease, age 65 to 74 years, and female sex) score pertinent for ≥ 1 non–sex-related factor (score ≥ 2 for women; ≥ 1 for men) to prevent stroke-related complications. The CHA₂DS₂-VASc score is a 9-point scoring tool based on comorbidities and conditions that increase risk of stroke in patients with Afib. Each value correlates to an annualized stroke risk percentage that increases as the score increases.

In clinical practice, patients meeting these thresholds are indicated for anticoagulation and are considered for indefinite use unless ≥ 1 of the following conditions are present: bleeding risk outweighs the stroke prevention benefit, Afib is episodic (< 48 hours) or a nonpharmacologic intervention, such as a left atrial appendage occlusion (LAAO) device is present.¹

In patients with a diagnosed venous thromboembolism (VTE), such as deep vein thrombosis or pulmonary embolism, anticoagulation is used to treat the current thrombosis and prevent embolization that can ultimately lead to death. The 2021 guideline for VTE from the American College of Chest Physicians identifies certain risk factors that increase risk for VTE and categorizes them as transient or persistent. Transient risk factors include hospitalization > 3 days, major trauma, surgery, cast immobilization, hormone therapy, pregnancy, or prolonged travel > 8 hours. Persistent risk factors include malignancy, thrombophilia, and certain medications.

The guideline recommends therapy durations based on event frequency, the presence and classification of provoking risk factors, and bleeding risk. As the risk of recurrent thrombosis and other potential complications is greatest in the first 3 to 6 months after a diagnosed event, at least 3 months anticoagulation therapy is recommended following VTE diagnosis. At the 3-month mark, all regimens are suggested to be re-evaluated and considered for extended treatment duration if the event was unprovoked, recurrent, secondary to a persistent risk factor, or low bleed risk.²Anticoagulation is an important guideline-recommended pharmacologic intervention for various disease states, although its use is not without risks. The Institute for Safe Medication Practices has classified oral anticoagulants as high-alert medications. This designation was made because anticoagulant medications have the potential to cause harm when used or omitted in error and lead to life-threatening bleed or thrombotic complications.³Anticoagulation stewardship ensures that anticoagulation therapy is appropriately initiated, maintained, and discontinued when indicated. Because of the potential for harm, anticoagulation stewardship is an important part of Afib and VTE management. Pharmacists can help verify and evaluate anticoagulation therapies. Research suggests that pharmacist-led anticoagulation stewardship efforts may play a role in ensuring safer patient outcomes.⁴The purpose of this quality improvement (QI) study was to implement pharmacist-led anticoagulation stewardship practices at Veterans Affairs Phoenix Health Care System (VAPHCS) to identify veterans with Afib not currently on anticoagulation, as well as to identify veterans with a history of VTE events who have completed a sufficient treatment duration.

Methods

Anticoagulation stewardship efforts were implemented in 2 cohorts of patients: those with Afib who may be indicated to initiate anticoagulation, and those with a history of VTE events who may be indicated to consider anticoagulation discontinuation. Patient records were reviewed using a standardized note template, and recommendations to either initiate or discontinue anticoagulation therapy were documented. The VAPHCS Research Service reviewed this study and determined that it was not research and was exempt from institutional review board review.

Atrial Fibrillation Cohort

A population health dashboard created by the Stroke Prevention in Atrial Fibrillation/Flutter Targeting the uNTreated: a focus on health care disparities (SPAFF-TNT-D) national VA study team was used to identify veterans at VAPHCS with a diagnosis of Afib without an active VA prescription for an anticoagulant. The dashboard filtered and produced data points from the medical record that correlated to the components of the CHA₂DS₂-VASc score. All veterans identified by the dashboard with scores of 7 or 8 were included. No patients had a score of 9. Comprehensive chart reviews of available VA and non–VA-provided care records were conducted by the investigators, and a standardized note template designed by the SPAFF-TNT-D team (eAppendix 1) was used to document findings within the electronic health record (EHR). If anticoagulation was deemed to be indicated, the assigned primary care practitioner (PCP) as listed in the EHR was alerted to the note by the investigators for further evaluation and consideration of prescribing anticoagulation.

Venous Thromboembolism Cohort

VAPHCS pharmacy informatics pulled data that included veterans with documented VTE and an active VA anticoagulant prescription between November 2022 and November 2023. Veterans were reviewed in chronological order based on when the anticoagulant prescription was written. All veterans were included until an equal number of charts were reviewed in both the Afib and VTE cohorts. Comprehensive chart review of available VA- and non–VA-provided care records was conducted by the investigators, and a standardized note template as designed by the investigators (eAppendix 2) was used to document findings within the EHR. If the duration of anticoagulation therapy was deemed sufficient, the assigned anticoagulation clinical pharmacist practitioner (CPP) was alerted to the note by the investigators for further evaluation and consideration of discontinuing anticoagulation.

EHR reviews were conducted in October and November 2023 and lasted about 10 to 20 minutes per patient. To evaluate completeness and accuracy of the documented findings within the EHR, both investigators reviewed and cosigned the completed note template and verified the correct PCP was alerted to the recommendation for appropriate continuity of care. Results were reviewed in March 2024.

Outcomes

Atrial fibrillation cohort. The primary outcome was the number of veterans with Afib who were recommended to start anticoagulation therapy. Additional outcomes evaluated included the number of interventions completed, action taken by PCPs in response to the provided recommendation, and reasons provided by the investigators for not recommending initiation of anticoagulation therapy in specific veteran cases.

Venous thromboembolism cohort. The primary outcome was the number of veterans with a history of VTE events recommended to discontinue anticoagulation therapy. Additional outcomes included number of interventions completed, action taken by the anticoagulation CPP in response to the provided recommendation, and reasons provided by the investigators for not recommending discontinuation of anticoagulation therapy in specific veteran cases.

Analysis

Sample size was determined by the inclusion criteria and was not designed to attain statistical power. Data embedded in the Afib cohort standardized note template, also known as health factors, were later used for data analysis. Recommendations in the VTE cohort were manually tracked and recorded by the investigators. Results for this study were analyzed using descriptive statistics.

Results

A total of 114 veterans were reviewed and included in this study: 57 in each cohort. Seven recommendations were made regarding anticoagulation initiation for patients with Afib and 7 were made for anticoagulation discontinuation for patients with VTE (Table 1).

In the Afib cohort, 1 veteran was successfully initiated on anticoagulation therapy and 1 veteran was deemed appropriate for initiation of anticoagulation but was not reachable. Of the 5 recommendations with no action taken, 4 PCPs acknowledged the alert with no further documentation, and 1 PCP deferred the decision to cardiology with no further documentation. In the VTE cohort, 3 veterans successfully discontinued anticoagulation therapy and 2 veterans were further evaluated by the anticoagulation CPP and deemed appropriate to continue therapy based on potential for malignancy. Of the 2 recommendations with no action taken, 1 anticoagulation CPP acknowledged the alert with no further documentation and 1 anticoagulation CPP suggested further evaluation by PCP with no further documentation.

In the Afib cohort, a nonpharmacologic approach was defined as documentation of a LAAO device. An inaccurate diagnosis was defined as an Afib diagnosis being used in a previous visit, although there was no further confirmation of diagnosis via chart review. Veterans classified as already being on anticoagulation had documentation of non–VA-written anticoagulant prescriptions or receiving a supply of anticoagulants from a facility such as a nursing home. Anticoagulation was defined as unfavorable if a documented risk/benefit conversation was found via EHR review. Anticoagulation was defined as not indicated if the Afib was documented as transient, episodic, or historical (Table 2).

In the VTE cohort, no recommendations for discontinuation were made for veterans indicated to continue anticoagulation due to a concurrent Afib diagnosis. Chronic or recurrent events were defined as documentation of multiple VTE events and associated dates in the EHR. Persistent risk factors included malignancy or medications contributing to hypercoagulable states. Thrombophilia was defined as having documentation of a diagnosis in the EHR. An unprovoked event was defined as VTE without any documented transient risk factors (eg, hospitalization, trauma, surgery, cast immobilization, hormone therapy, pregnancy, or prolonged travel). Anticoagulation had already been discontinued in 1 veteran after the data were collected but before chart review occurred (Table 3).

Discussion

Pharmacy-led indication reviews resulted in appropriate recommendations for anticoagulation use in veterans with Afib and a history of VTE events. Overall, 12.3% of chart reviews in each cohort resulted in a recommendation being made, which was similar to the rate found by Koolian et al.⁵ In that study, 10% of recommendations were related to initiation or interruption of anticoagulation. This recommendation category consisted of several subcategories, including “suggesting therapeutic anticoagulation when none is currently ordered” and “suggesting anticoagulation cessation if no longer indicated,” but specific numerical prevalence was not provided.⁵

Online dashboard use allowed for greater population health management and identification of veterans with Afib who were not on active anticoagulation, providing opportunities to prevent stroke-related complications. Wang et al completed a similarly designed study that included a population health tool to identify patients with Afib who were not on anticoagulation and implemented pharmacist-led chart review and facilitation of recommendations to the responsible clinician. This study reviewed 1727 patients and recommended initiation of anticoagulation therapy for 75 (4.3%).⁶ The current study had a higher percentage of patients with recommendations for changes despite its smaller size.

Evaluating the duration of therapy for anticoagulation in veterans with a history of VTE events provided an opportunity to reduce unnecessary exposure to anticoagulation and minimize bleeding risks. Using a chart review process and standardized note template enabled the documentation of pertinent information that could be readily reviewed by the PCP. This process is a step toward ensuring VAPHCS PCPs provide guideline-recommended care and actively prevent stroke and bleeding complications. Adoption of this process into the current VAPHCS Anticoagulation Clinic workflow for review of veterans with either Afib or VTE could lead to more EHRs being reviewed and recommendations made, ultimately improving patient outcomes. 

Therapeutic interventions based on the recommendations were completed for 1 of 7 veterans (14%) and 3 of 7 veterans (43%) in the Afib and VTE cohorts, respectively. The prevalence of completed interventions in this anticoagulation stewardship study was higher than those in Wang et al, who found only 9% of their recommendations resulted in PCPs considering action related to anticoagulation, and only 4% were successfully initiated.⁶

In the Afib cohort, veterans identified by the dashboard with a CHA₂DS₂-VASc of 7 or 8 were prioritized for review. Reviewing these veterans ensured that patients with the highest stroke risk were sufficiently evaluated and started on anticoagulation as needed to reduce stroke-related complications. In contrast, because these veterans had higher CHA₂DS₂-VASc scores, they may have already been evaluated for anticoagulation in the past and had a documented rationale for not being placed on anticoagulation (LAAO device placement was the most common rationale). Focusing on veterans with a lower CHA₂DS₂-VASc score such as 1 for men or 2 for women could potentially include more opportunities for recommendations. Although stroke risk may be lower in this population compared with those with higher CHA₂DS₂-VASc scores, guideline-recommended anticoagulation use may be missed for these patients. 

In the VTE cohort, veterans with an anticoagulant prescription written 12 months before data collection were prioritized for review. Reviewing these veterans ensured that anticoagulation therapy met guideline recommendations of at least 3 months, with potential for extended duration upon further evaluation by a provider at that time. Based on collected results, most veterans were already reevaluated and had documented reasons why anticoagulation was still indicated; concurrent Afib was most common followed by chronic or recurrent VTE. Reviewing veterans with more recent prescriptions just over the recommended 3-month duration could potentially include more opportunities for recommendations to be made. It is more likely that by 3 months another PCP had not already weighed in on the duration of therapy, and the anticoagulation CPP could ensure a thorough review is conducted with guideline-based recommendations.

Most published literature on anticoagulation stewardship efforts is focused on inpatient management and policy changes, or concentrate on attributes of therapy such as appropriate dosing and drug interactions. This study highlighted that gaps in care related to anticoagulation use and discontinuation are present in the VAPHCS population and can be appropriately addressed via pharmacist-led indication reviews. Future studies designed to focus on initiating anticoagulation where appropriate, and discontinuing where a sufficient treatment period has been completed, are warranted to minimize this gap in care and allow health systems to work toward process changes to ensure safe and optimized care is provided for the patients they serve.

Limitations

In the Afib cohort, 5 of 7 recommendations (71%) had no further action taken by the PCP, which may represent a barrier to care. In contrast, 2 of 7 recommendations (29%) had no further action in the VTE cohort. It is possible that the difference can be attributed to the anticoagulation CPP receiving VTE alerts and PCPs receiving Afib alerts. The anticoagulation CPP was familiar with this QI study and may have better understood the purpose of the chart review and the need to provide a timely response. PCPs may have been less likely to take action because they were unfamiliar with the anticoagulation stewardship initiative and standardized note template or overwhelmed by too many EHR alerts.

The lack of PCP response to a virtual alert or message also was observed by Wang et al, whereas Koolian et al reported higher intervention completion rates, with verbal recommendations being made to the responsible clinicians. To further ensure these pertinent recommendations for anticoagulation initiation in veterans with Afib are properly reviewed and evaluated, future research could include intentional follow-up with the PCP regarding the alert, PCP-specific education about the anticoagulation stewardship initiative and the role of the standardized note template, and collaboration with PCPs to identify alternative ways to relay recommendations in a way that would ensure the completion of appropriate and timely review.

Conclusions

This study identified gaps in care related to anticoagulation needs in the VAPHCS veteran population. Utilizing a standardized indication review process allows pharmacists to evaluate anticoagulant use for both appropriate indication and duration of therapy. Providing recommendations via chart review notes and alerting respective PCPs and CPPs results in veterans receiving safe and optimized care regarding their anticoagulation needs.

Due to the underlying mechanism of atrial fibrillation (Afib), clots can form within the left atrial appendage. Clots that become dislodged may lead to ischemic stroke and possibly death. The 2023 guidelines for atrial fibrillation from the American College of Cardiology and American Heart Association recommend anticoagulation therapy for patients with an Afib diagnosis and a CHA₂DS₂-VASc (congestive heart failure, hypertension, age ≥ 75 years, diabetes, stroke/vascular disease, age 65 to 74 years, and female sex) score pertinent for ≥ 1 non–sex-related factor (score ≥ 2 for women; ≥ 1 for men) to prevent stroke-related complications. The CHA₂DS₂-VASc score is a 9-point scoring tool based on comorbidities and conditions that increase risk of stroke in patients with Afib. Each value correlates to an annualized stroke risk percentage that increases as the score increases.

In clinical practice, patients meeting these thresholds are indicated for anticoagulation and are considered for indefinite use unless ≥ 1 of the following conditions are present: bleeding risk outweighs the stroke prevention benefit, Afib is episodic (< 48 hours) or a nonpharmacologic intervention, such as a left atrial appendage occlusion (LAAO) device is present.¹

In patients with a diagnosed venous thromboembolism (VTE), such as deep vein thrombosis or pulmonary embolism, anticoagulation is used to treat the current thrombosis and prevent embolization that can ultimately lead to death. The 2021 guideline for VTE from the American College of Chest Physicians identifies certain risk factors that increase risk for VTE and categorizes them as transient or persistent. Transient risk factors include hospitalization > 3 days, major trauma, surgery, cast immobilization, hormone therapy, pregnancy, or prolonged travel > 8 hours. Persistent risk factors include malignancy, thrombophilia, and certain medications.

The guideline recommends therapy durations based on event frequency, the presence and classification of provoking risk factors, and bleeding risk. As the risk of recurrent thrombosis and other potential complications is greatest in the first 3 to 6 months after a diagnosed event, at least 3 months anticoagulation therapy is recommended following VTE diagnosis. At the 3-month mark, all regimens are suggested to be re-evaluated and considered for extended treatment duration if the event was unprovoked, recurrent, secondary to a persistent risk factor, or low bleed risk.²Anticoagulation is an important guideline-recommended pharmacologic intervention for various disease states, although its use is not without risks. The Institute for Safe Medication Practices has classified oral anticoagulants as high-alert medications. This designation was made because anticoagulant medications have the potential to cause harm when used or omitted in error and lead to life-threatening bleed or thrombotic complications.³Anticoagulation stewardship ensures that anticoagulation therapy is appropriately initiated, maintained, and discontinued when indicated. Because of the potential for harm, anticoagulation stewardship is an important part of Afib and VTE management. Pharmacists can help verify and evaluate anticoagulation therapies. Research suggests that pharmacist-led anticoagulation stewardship efforts may play a role in ensuring safer patient outcomes.⁴The purpose of this quality improvement (QI) study was to implement pharmacist-led anticoagulation stewardship practices at Veterans Affairs Phoenix Health Care System (VAPHCS) to identify veterans with Afib not currently on anticoagulation, as well as to identify veterans with a history of VTE events who have completed a sufficient treatment duration.

Methods

Anticoagulation stewardship efforts were implemented in 2 cohorts of patients: those with Afib who may be indicated to initiate anticoagulation, and those with a history of VTE events who may be indicated to consider anticoagulation discontinuation. Patient records were reviewed using a standardized note template, and recommendations to either initiate or discontinue anticoagulation therapy were documented. The VAPHCS Research Service reviewed this study and determined that it was not research and was exempt from institutional review board review.

Atrial Fibrillation Cohort

A population health dashboard created by the Stroke Prevention in Atrial Fibrillation/Flutter Targeting the uNTreated: a focus on health care disparities (SPAFF-TNT-D) national VA study team was used to identify veterans at VAPHCS with a diagnosis of Afib without an active VA prescription for an anticoagulant. The dashboard filtered and produced data points from the medical record that correlated to the components of the CHA₂DS₂-VASc score. All veterans identified by the dashboard with scores of 7 or 8 were included. No patients had a score of 9. Comprehensive chart reviews of available VA and non–VA-provided care records were conducted by the investigators, and a standardized note template designed by the SPAFF-TNT-D team (eAppendix 1) was used to document findings within the electronic health record (EHR). If anticoagulation was deemed to be indicated, the assigned primary care practitioner (PCP) as listed in the EHR was alerted to the note by the investigators for further evaluation and consideration of prescribing anticoagulation.

Venous Thromboembolism Cohort

VAPHCS pharmacy informatics pulled data that included veterans with documented VTE and an active VA anticoagulant prescription between November 2022 and November 2023. Veterans were reviewed in chronological order based on when the anticoagulant prescription was written. All veterans were included until an equal number of charts were reviewed in both the Afib and VTE cohorts. Comprehensive chart review of available VA- and non–VA-provided care records was conducted by the investigators, and a standardized note template as designed by the investigators (eAppendix 2) was used to document findings within the EHR. If the duration of anticoagulation therapy was deemed sufficient, the assigned anticoagulation clinical pharmacist practitioner (CPP) was alerted to the note by the investigators for further evaluation and consideration of discontinuing anticoagulation.

EHR reviews were conducted in October and November 2023 and lasted about 10 to 20 minutes per patient. To evaluate completeness and accuracy of the documented findings within the EHR, both investigators reviewed and cosigned the completed note template and verified the correct PCP was alerted to the recommendation for appropriate continuity of care. Results were reviewed in March 2024.

Outcomes

Atrial fibrillation cohort. The primary outcome was the number of veterans with Afib who were recommended to start anticoagulation therapy. Additional outcomes evaluated included the number of interventions completed, action taken by PCPs in response to the provided recommendation, and reasons provided by the investigators for not recommending initiation of anticoagulation therapy in specific veteran cases.

Venous thromboembolism cohort. The primary outcome was the number of veterans with a history of VTE events recommended to discontinue anticoagulation therapy. Additional outcomes included number of interventions completed, action taken by the anticoagulation CPP in response to the provided recommendation, and reasons provided by the investigators for not recommending discontinuation of anticoagulation therapy in specific veteran cases.

Analysis

Sample size was determined by the inclusion criteria and was not designed to attain statistical power. Data embedded in the Afib cohort standardized note template, also known as health factors, were later used for data analysis. Recommendations in the VTE cohort were manually tracked and recorded by the investigators. Results for this study were analyzed using descriptive statistics.

Results

A total of 114 veterans were reviewed and included in this study: 57 in each cohort. Seven recommendations were made regarding anticoagulation initiation for patients with Afib and 7 were made for anticoagulation discontinuation for patients with VTE (Table 1).

In the Afib cohort, 1 veteran was successfully initiated on anticoagulation therapy and 1 veteran was deemed appropriate for initiation of anticoagulation but was not reachable. Of the 5 recommendations with no action taken, 4 PCPs acknowledged the alert with no further documentation, and 1 PCP deferred the decision to cardiology with no further documentation. In the VTE cohort, 3 veterans successfully discontinued anticoagulation therapy and 2 veterans were further evaluated by the anticoagulation CPP and deemed appropriate to continue therapy based on potential for malignancy. Of the 2 recommendations with no action taken, 1 anticoagulation CPP acknowledged the alert with no further documentation and 1 anticoagulation CPP suggested further evaluation by PCP with no further documentation.

In the Afib cohort, a nonpharmacologic approach was defined as documentation of a LAAO device. An inaccurate diagnosis was defined as an Afib diagnosis being used in a previous visit, although there was no further confirmation of diagnosis via chart review. Veterans classified as already being on anticoagulation had documentation of non–VA-written anticoagulant prescriptions or receiving a supply of anticoagulants from a facility such as a nursing home. Anticoagulation was defined as unfavorable if a documented risk/benefit conversation was found via EHR review. Anticoagulation was defined as not indicated if the Afib was documented as transient, episodic, or historical (Table 2).

In the VTE cohort, no recommendations for discontinuation were made for veterans indicated to continue anticoagulation due to a concurrent Afib diagnosis. Chronic or recurrent events were defined as documentation of multiple VTE events and associated dates in the EHR. Persistent risk factors included malignancy or medications contributing to hypercoagulable states. Thrombophilia was defined as having documentation of a diagnosis in the EHR. An unprovoked event was defined as VTE without any documented transient risk factors (eg, hospitalization, trauma, surgery, cast immobilization, hormone therapy, pregnancy, or prolonged travel). Anticoagulation had already been discontinued in 1 veteran after the data were collected but before chart review occurred (Table 3).

Discussion

Pharmacy-led indication reviews resulted in appropriate recommendations for anticoagulation use in veterans with Afib and a history of VTE events. Overall, 12.3% of chart reviews in each cohort resulted in a recommendation being made, which was similar to the rate found by Koolian et al.⁵ In that study, 10% of recommendations were related to initiation or interruption of anticoagulation. This recommendation category consisted of several subcategories, including “suggesting therapeutic anticoagulation when none is currently ordered” and “suggesting anticoagulation cessation if no longer indicated,” but specific numerical prevalence was not provided.⁵

Online dashboard use allowed for greater population health management and identification of veterans with Afib who were not on active anticoagulation, providing opportunities to prevent stroke-related complications. Wang et al completed a similarly designed study that included a population health tool to identify patients with Afib who were not on anticoagulation and implemented pharmacist-led chart review and facilitation of recommendations to the responsible clinician. This study reviewed 1727 patients and recommended initiation of anticoagulation therapy for 75 (4.3%).⁶ The current study had a higher percentage of patients with recommendations for changes despite its smaller size.

Evaluating the duration of therapy for anticoagulation in veterans with a history of VTE events provided an opportunity to reduce unnecessary exposure to anticoagulation and minimize bleeding risks. Using a chart review process and standardized note template enabled the documentation of pertinent information that could be readily reviewed by the PCP. This process is a step toward ensuring VAPHCS PCPs provide guideline-recommended care and actively prevent stroke and bleeding complications. Adoption of this process into the current VAPHCS Anticoagulation Clinic workflow for review of veterans with either Afib or VTE could lead to more EHRs being reviewed and recommendations made, ultimately improving patient outcomes. 

Therapeutic interventions based on the recommendations were completed for 1 of 7 veterans (14%) and 3 of 7 veterans (43%) in the Afib and VTE cohorts, respectively. The prevalence of completed interventions in this anticoagulation stewardship study was higher than those in Wang et al, who found only 9% of their recommendations resulted in PCPs considering action related to anticoagulation, and only 4% were successfully initiated.⁶

In the Afib cohort, veterans identified by the dashboard with a CHA₂DS₂-VASc of 7 or 8 were prioritized for review. Reviewing these veterans ensured that patients with the highest stroke risk were sufficiently evaluated and started on anticoagulation as needed to reduce stroke-related complications. In contrast, because these veterans had higher CHA₂DS₂-VASc scores, they may have already been evaluated for anticoagulation in the past and had a documented rationale for not being placed on anticoagulation (LAAO device placement was the most common rationale). Focusing on veterans with a lower CHA₂DS₂-VASc score such as 1 for men or 2 for women could potentially include more opportunities for recommendations. Although stroke risk may be lower in this population compared with those with higher CHA₂DS₂-VASc scores, guideline-recommended anticoagulation use may be missed for these patients. 

In the VTE cohort, veterans with an anticoagulant prescription written 12 months before data collection were prioritized for review. Reviewing these veterans ensured that anticoagulation therapy met guideline recommendations of at least 3 months, with potential for extended duration upon further evaluation by a provider at that time. Based on collected results, most veterans were already reevaluated and had documented reasons why anticoagulation was still indicated; concurrent Afib was most common followed by chronic or recurrent VTE. Reviewing veterans with more recent prescriptions just over the recommended 3-month duration could potentially include more opportunities for recommendations to be made. It is more likely that by 3 months another PCP had not already weighed in on the duration of therapy, and the anticoagulation CPP could ensure a thorough review is conducted with guideline-based recommendations.

Most published literature on anticoagulation stewardship efforts is focused on inpatient management and policy changes, or concentrate on attributes of therapy such as appropriate dosing and drug interactions. This study highlighted that gaps in care related to anticoagulation use and discontinuation are present in the VAPHCS population and can be appropriately addressed via pharmacist-led indication reviews. Future studies designed to focus on initiating anticoagulation where appropriate, and discontinuing where a sufficient treatment period has been completed, are warranted to minimize this gap in care and allow health systems to work toward process changes to ensure safe and optimized care is provided for the patients they serve.

Limitations

In the Afib cohort, 5 of 7 recommendations (71%) had no further action taken by the PCP, which may represent a barrier to care. In contrast, 2 of 7 recommendations (29%) had no further action in the VTE cohort. It is possible that the difference can be attributed to the anticoagulation CPP receiving VTE alerts and PCPs receiving Afib alerts. The anticoagulation CPP was familiar with this QI study and may have better understood the purpose of the chart review and the need to provide a timely response. PCPs may have been less likely to take action because they were unfamiliar with the anticoagulation stewardship initiative and standardized note template or overwhelmed by too many EHR alerts.

The lack of PCP response to a virtual alert or message also was observed by Wang et al, whereas Koolian et al reported higher intervention completion rates, with verbal recommendations being made to the responsible clinicians. To further ensure these pertinent recommendations for anticoagulation initiation in veterans with Afib are properly reviewed and evaluated, future research could include intentional follow-up with the PCP regarding the alert, PCP-specific education about the anticoagulation stewardship initiative and the role of the standardized note template, and collaboration with PCPs to identify alternative ways to relay recommendations in a way that would ensure the completion of appropriate and timely review.

Conclusions

This study identified gaps in care related to anticoagulation needs in the VAPHCS veteran population. Utilizing a standardized indication review process allows pharmacists to evaluate anticoagulant use for both appropriate indication and duration of therapy. Providing recommendations via chart review notes and alerting respective PCPs and CPPs results in veterans receiving safe and optimized care regarding their anticoagulation needs.

Due to the underlying mechanism of atrial fibrillation (Afib), clots can form within the left atrial appendage. Clots that become dislodged may lead to ischemic stroke and possibly death. The 2023 guidelines for atrial fibrillation from the American College of Cardiology and American Heart Association recommend anticoagulation therapy for patients with an Afib diagnosis and a CHA₂DS₂-VASc (congestive heart failure, hypertension, age ≥ 75 years, diabetes, stroke/vascular disease, age 65 to 74 years, and female sex) score pertinent for ≥ 1 non–sex-related factor (score ≥ 2 for women; ≥ 1 for men) to prevent stroke-related complications. The CHA₂DS₂-VASc score is a 9-point scoring tool based on comorbidities and conditions that increase risk of stroke in patients with Afib. Each value correlates to an annualized stroke risk percentage that increases as the score increases.

In clinical practice, patients meeting these thresholds are indicated for anticoagulation and are considered for indefinite use unless ≥ 1 of the following conditions are present: bleeding risk outweighs the stroke prevention benefit, Afib is episodic (< 48 hours) or a nonpharmacologic intervention, such as a left atrial appendage occlusion (LAAO) device is present.¹

In patients with a diagnosed venous thromboembolism (VTE), such as deep vein thrombosis or pulmonary embolism, anticoagulation is used to treat the current thrombosis and prevent embolization that can ultimately lead to death. The 2021 guideline for VTE from the American College of Chest Physicians identifies certain risk factors that increase risk for VTE and categorizes them as transient or persistent. Transient risk factors include hospitalization > 3 days, major trauma, surgery, cast immobilization, hormone therapy, pregnancy, or prolonged travel > 8 hours. Persistent risk factors include malignancy, thrombophilia, and certain medications.

The guideline recommends therapy durations based on event frequency, the presence and classification of provoking risk factors, and bleeding risk. As the risk of recurrent thrombosis and other potential complications is greatest in the first 3 to 6 months after a diagnosed event, at least 3 months anticoagulation therapy is recommended following VTE diagnosis. At the 3-month mark, all regimens are suggested to be re-evaluated and considered for extended treatment duration if the event was unprovoked, recurrent, secondary to a persistent risk factor, or low bleed risk.²Anticoagulation is an important guideline-recommended pharmacologic intervention for various disease states, although its use is not without risks. The Institute for Safe Medication Practices has classified oral anticoagulants as high-alert medications. This designation was made because anticoagulant medications have the potential to cause harm when used or omitted in error and lead to life-threatening bleed or thrombotic complications.³Anticoagulation stewardship ensures that anticoagulation therapy is appropriately initiated, maintained, and discontinued when indicated. Because of the potential for harm, anticoagulation stewardship is an important part of Afib and VTE management. Pharmacists can help verify and evaluate anticoagulation therapies. Research suggests that pharmacist-led anticoagulation stewardship efforts may play a role in ensuring safer patient outcomes.⁴The purpose of this quality improvement (QI) study was to implement pharmacist-led anticoagulation stewardship practices at Veterans Affairs Phoenix Health Care System (VAPHCS) to identify veterans with Afib not currently on anticoagulation, as well as to identify veterans with a history of VTE events who have completed a sufficient treatment duration.

Methods

Anticoagulation stewardship efforts were implemented in 2 cohorts of patients: those with Afib who may be indicated to initiate anticoagulation, and those with a history of VTE events who may be indicated to consider anticoagulation discontinuation. Patient records were reviewed using a standardized note template, and recommendations to either initiate or discontinue anticoagulation therapy were documented. The VAPHCS Research Service reviewed this study and determined that it was not research and was exempt from institutional review board review.

Atrial Fibrillation Cohort

A population health dashboard created by the Stroke Prevention in Atrial Fibrillation/Flutter Targeting the uNTreated: a focus on health care disparities (SPAFF-TNT-D) national VA study team was used to identify veterans at VAPHCS with a diagnosis of Afib without an active VA prescription for an anticoagulant. The dashboard filtered and produced data points from the medical record that correlated to the components of the CHA₂DS₂-VASc score. All veterans identified by the dashboard with scores of 7 or 8 were included. No patients had a score of 9. Comprehensive chart reviews of available VA and non–VA-provided care records were conducted by the investigators, and a standardized note template designed by the SPAFF-TNT-D team (eAppendix 1) was used to document findings within the electronic health record (EHR). If anticoagulation was deemed to be indicated, the assigned primary care practitioner (PCP) as listed in the EHR was alerted to the note by the investigators for further evaluation and consideration of prescribing anticoagulation.

Venous Thromboembolism Cohort

VAPHCS pharmacy informatics pulled data that included veterans with documented VTE and an active VA anticoagulant prescription between November 2022 and November 2023. Veterans were reviewed in chronological order based on when the anticoagulant prescription was written. All veterans were included until an equal number of charts were reviewed in both the Afib and VTE cohorts. Comprehensive chart review of available VA- and non–VA-provided care records was conducted by the investigators, and a standardized note template as designed by the investigators (eAppendix 2) was used to document findings within the EHR. If the duration of anticoagulation therapy was deemed sufficient, the assigned anticoagulation clinical pharmacist practitioner (CPP) was alerted to the note by the investigators for further evaluation and consideration of discontinuing anticoagulation.

EHR reviews were conducted in October and November 2023 and lasted about 10 to 20 minutes per patient. To evaluate completeness and accuracy of the documented findings within the EHR, both investigators reviewed and cosigned the completed note template and verified the correct PCP was alerted to the recommendation for appropriate continuity of care. Results were reviewed in March 2024.

Outcomes

Atrial fibrillation cohort. The primary outcome was the number of veterans with Afib who were recommended to start anticoagulation therapy. Additional outcomes evaluated included the number of interventions completed, action taken by PCPs in response to the provided recommendation, and reasons provided by the investigators for not recommending initiation of anticoagulation therapy in specific veteran cases.

Venous thromboembolism cohort. The primary outcome was the number of veterans with a history of VTE events recommended to discontinue anticoagulation therapy. Additional outcomes included number of interventions completed, action taken by the anticoagulation CPP in response to the provided recommendation, and reasons provided by the investigators for not recommending discontinuation of anticoagulation therapy in specific veteran cases.

Analysis

Sample size was determined by the inclusion criteria and was not designed to attain statistical power. Data embedded in the Afib cohort standardized note template, also known as health factors, were later used for data analysis. Recommendations in the VTE cohort were manually tracked and recorded by the investigators. Results for this study were analyzed using descriptive statistics.

Results

A total of 114 veterans were reviewed and included in this study: 57 in each cohort. Seven recommendations were made regarding anticoagulation initiation for patients with Afib and 7 were made for anticoagulation discontinuation for patients with VTE (Table 1).

In the Afib cohort, 1 veteran was successfully initiated on anticoagulation therapy and 1 veteran was deemed appropriate for initiation of anticoagulation but was not reachable. Of the 5 recommendations with no action taken, 4 PCPs acknowledged the alert with no further documentation, and 1 PCP deferred the decision to cardiology with no further documentation. In the VTE cohort, 3 veterans successfully discontinued anticoagulation therapy and 2 veterans were further evaluated by the anticoagulation CPP and deemed appropriate to continue therapy based on potential for malignancy. Of the 2 recommendations with no action taken, 1 anticoagulation CPP acknowledged the alert with no further documentation and 1 anticoagulation CPP suggested further evaluation by PCP with no further documentation.

In the Afib cohort, a nonpharmacologic approach was defined as documentation of a LAAO device. An inaccurate diagnosis was defined as an Afib diagnosis being used in a previous visit, although there was no further confirmation of diagnosis via chart review. Veterans classified as already being on anticoagulation had documentation of non–VA-written anticoagulant prescriptions or receiving a supply of anticoagulants from a facility such as a nursing home. Anticoagulation was defined as unfavorable if a documented risk/benefit conversation was found via EHR review. Anticoagulation was defined as not indicated if the Afib was documented as transient, episodic, or historical (Table 2).

In the VTE cohort, no recommendations for discontinuation were made for veterans indicated to continue anticoagulation due to a concurrent Afib diagnosis. Chronic or recurrent events were defined as documentation of multiple VTE events and associated dates in the EHR. Persistent risk factors included malignancy or medications contributing to hypercoagulable states. Thrombophilia was defined as having documentation of a diagnosis in the EHR. An unprovoked event was defined as VTE without any documented transient risk factors (eg, hospitalization, trauma, surgery, cast immobilization, hormone therapy, pregnancy, or prolonged travel). Anticoagulation had already been discontinued in 1 veteran after the data were collected but before chart review occurred (Table 3).

Discussion

Pharmacy-led indication reviews resulted in appropriate recommendations for anticoagulation use in veterans with Afib and a history of VTE events. Overall, 12.3% of chart reviews in each cohort resulted in a recommendation being made, which was similar to the rate found by Koolian et al.⁵ In that study, 10% of recommendations were related to initiation or interruption of anticoagulation. This recommendation category consisted of several subcategories, including “suggesting therapeutic anticoagulation when none is currently ordered” and “suggesting anticoagulation cessation if no longer indicated,” but specific numerical prevalence was not provided.⁵

Online dashboard use allowed for greater population health management and identification of veterans with Afib who were not on active anticoagulation, providing opportunities to prevent stroke-related complications. Wang et al completed a similarly designed study that included a population health tool to identify patients with Afib who were not on anticoagulation and implemented pharmacist-led chart review and facilitation of recommendations to the responsible clinician. This study reviewed 1727 patients and recommended initiation of anticoagulation therapy for 75 (4.3%).⁶ The current study had a higher percentage of patients with recommendations for changes despite its smaller size.

Evaluating the duration of therapy for anticoagulation in veterans with a history of VTE events provided an opportunity to reduce unnecessary exposure to anticoagulation and minimize bleeding risks. Using a chart review process and standardized note template enabled the documentation of pertinent information that could be readily reviewed by the PCP. This process is a step toward ensuring VAPHCS PCPs provide guideline-recommended care and actively prevent stroke and bleeding complications. Adoption of this process into the current VAPHCS Anticoagulation Clinic workflow for review of veterans with either Afib or VTE could lead to more EHRs being reviewed and recommendations made, ultimately improving patient outcomes. 

Therapeutic interventions based on the recommendations were completed for 1 of 7 veterans (14%) and 3 of 7 veterans (43%) in the Afib and VTE cohorts, respectively. The prevalence of completed interventions in this anticoagulation stewardship study was higher than those in Wang et al, who found only 9% of their recommendations resulted in PCPs considering action related to anticoagulation, and only 4% were successfully initiated.⁶

In the Afib cohort, veterans identified by the dashboard with a CHA₂DS₂-VASc of 7 or 8 were prioritized for review. Reviewing these veterans ensured that patients with the highest stroke risk were sufficiently evaluated and started on anticoagulation as needed to reduce stroke-related complications. In contrast, because these veterans had higher CHA₂DS₂-VASc scores, they may have already been evaluated for anticoagulation in the past and had a documented rationale for not being placed on anticoagulation (LAAO device placement was the most common rationale). Focusing on veterans with a lower CHA₂DS₂-VASc score such as 1 for men or 2 for women could potentially include more opportunities for recommendations. Although stroke risk may be lower in this population compared with those with higher CHA₂DS₂-VASc scores, guideline-recommended anticoagulation use may be missed for these patients. 

In the VTE cohort, veterans with an anticoagulant prescription written 12 months before data collection were prioritized for review. Reviewing these veterans ensured that anticoagulation therapy met guideline recommendations of at least 3 months, with potential for extended duration upon further evaluation by a provider at that time. Based on collected results, most veterans were already reevaluated and had documented reasons why anticoagulation was still indicated; concurrent Afib was most common followed by chronic or recurrent VTE. Reviewing veterans with more recent prescriptions just over the recommended 3-month duration could potentially include more opportunities for recommendations to be made. It is more likely that by 3 months another PCP had not already weighed in on the duration of therapy, and the anticoagulation CPP could ensure a thorough review is conducted with guideline-based recommendations.

Most published literature on anticoagulation stewardship efforts is focused on inpatient management and policy changes, or concentrate on attributes of therapy such as appropriate dosing and drug interactions. This study highlighted that gaps in care related to anticoagulation use and discontinuation are present in the VAPHCS population and can be appropriately addressed via pharmacist-led indication reviews. Future studies designed to focus on initiating anticoagulation where appropriate, and discontinuing where a sufficient treatment period has been completed, are warranted to minimize this gap in care and allow health systems to work toward process changes to ensure safe and optimized care is provided for the patients they serve.

Limitations

In the Afib cohort, 5 of 7 recommendations (71%) had no further action taken by the PCP, which may represent a barrier to care. In contrast, 2 of 7 recommendations (29%) had no further action in the VTE cohort. It is possible that the difference can be attributed to the anticoagulation CPP receiving VTE alerts and PCPs receiving Afib alerts. The anticoagulation CPP was familiar with this QI study and may have better understood the purpose of the chart review and the need to provide a timely response. PCPs may have been less likely to take action because they were unfamiliar with the anticoagulation stewardship initiative and standardized note template or overwhelmed by too many EHR alerts.

The lack of PCP response to a virtual alert or message also was observed by Wang et al, whereas Koolian et al reported higher intervention completion rates, with verbal recommendations being made to the responsible clinicians. To further ensure these pertinent recommendations for anticoagulation initiation in veterans with Afib are properly reviewed and evaluated, future research could include intentional follow-up with the PCP regarding the alert, PCP-specific education about the anticoagulation stewardship initiative and the role of the standardized note template, and collaboration with PCPs to identify alternative ways to relay recommendations in a way that would ensure the completion of appropriate and timely review.

Conclusions

This study identified gaps in care related to anticoagulation needs in the VAPHCS veteran population. Utilizing a standardized indication review process allows pharmacists to evaluate anticoagulant use for both appropriate indication and duration of therapy. Providing recommendations via chart review notes and alerting respective PCPs and CPPs results in veterans receiving safe and optimized care regarding their anticoagulation needs.

The veteran population, with its unique and diverse types of exposure and military service experiences, faces distinct health factors compared with the general population. These factors can be categorized into exposures during military service and those occurring postservice. While the latter phase incorporates psychological issues that may arise while transitioning to civilian life, the service period is associated with major physical, chemical, and psychological exposures that can impact veterans’ health. Carcinogenesis related to military exposures is concerning, and different types of malignancies have been associated with military exposures.¹ The 2022 introduction of the Cancer Moonshot initiative served as a breeding ground for multiple projects aimed at investigation of exposure-related carcinogenesis, prompting increased attention and efforts to linking specific exposures to specific malignancies.²

Melanoma is the deadliest skin cancer, accounting for 1.3% of all cancer deaths.³ Although it may only account for 1% to 5% of skin cancer diagnoses, its incidence in the United States’ population has been increasing.^4,5 There were 97,610 estimated new cases of melanoma in 2023, according to the National Cancer Institute.⁶

The incidence of melanoma may be higher in the military population compared with the general population.⁷ Melanoma is the fourth-most common cancer diagnosed in veterans.⁸

Several demographic characteristics of the US military population are associated with higher melanoma incidence and poorer prognosis, including male sex, older age, and White race. Apart from sun exposure—a known risk factor for melanoma development—other factors, such as service branch, seem to contribute to risk, with the highest melanoma rates noted in the Air Force.⁹ According to a study by Chang et al, veterans have a higher risk of stage III (18%) or stage IV (13%) melanoma at initial diagnosis.⁸

Molecular testing of metastatic melanoma is currently the standard of care for guiding the use of US Food and Drug Administration-approved targeted therapies such as BRAF, MEK, and KIT inhibitors. This comparative analysis details the melanoma comprehensive genomic profiles observed at a large US Department of Veterans Affairs (VA) medical center (VAMC) and those reported in reference databases.

Methods

A query to select all metastatic melanomas sent for comprehensive genomic profiling from the Kansas City VAMC (KCVAMC), identified 35 cases from 2019 through 2023 as the study population. The health records of these patients were reviewed to collect demographic information, military service history, melanoma history, other medical, social, and family histories. The comprehensive genomic profiling reports were reviewed to collect the reported pathogenic variants, microsatellite instability (MSI) status, and tumor mutational burden (TMB) for each case.

The Catalogue of Somatic Mutations in Cancer (COSMIC) was used to identify the most commonly mutated genes in melanomas from The Cancer Genome Atlas for the general population.^4,5 The literature was consulted to determine the MSI status and TMB in melanomas from The Cancer Genome Atlas for separate reference populations.^6,7 The frequency of MSI-high (MSI-H) status, TMB ≥ 10 mutations/megabase (mut/Mb), and mutations in each of the 20 most commonly mutated genes was determined and compared between melanomas from The Cancer Genome Atlas and KCVAMC cases. Corresponding P  values were calculated to identify significant differences. Values were calculated for the entire sample as well as a subgroup with Agent Orange (AO) exposure. The study was approved by the KCVAMC Institutional Review Board.

Results

The mean (SD) age of study participants was 72.9 (9.4) years (range, 39-90 years). The mean (SD) duration of military service was 1654 (1421) days (about 4 years, 6 months, and 10 days). Of the 35 patients included, 22 (63%) served during the Vietnam era (November 1, 1965, to April 30, 1975) and 2 (6%) served during the Persian Gulf War era (August 2, 1990, to February 28, 1991). Seventeen veterans (49%) served in the Army, 9 in the Navy (26%), 5 in the Air Force (14%), and 4 in the Marine Corps (11%). Definitive AO exposure was noted in 13 patients (37%) (Table 1).

Of the 35 patients, 24 (69%) had metastatic disease and the primary site of melanoma was unknown in 14 patients (40%). One patient (Patient 32) had an intraocular melanoma. The primary site was the trunk for 11 patients (31%), the face/head for 7 patients (20%) and extremities for 3 patients (9%). Eight patients (23%) were pT3 stage (thickness > 2 mm but < 4 mm), 7 patients (20%) were pT4 stage (thickness > 4 mm), and 5 patients (14%) were pT1 (thickness ≥ 1 mm). One patient had a primary lesion at pT2 stage, and 1 had a Tis stage lesion. Three patients (9%) had a family history of melanoma in a first-degree relative.

The list of genes mutated in melanoma cells in the study population is provided in the eAppendix.^10,11 Twenty-seven patients (77%) had mutations in TERT promoter, 15 (43%) in CDKN2A/B, 13 (37%) in BRAF, 11 (31%) in NF1, 9 (26%) in TP53, and 8 (23%) in NRAS (Table 2). The majority of mutations in TERT promoter were c.- 146C>T (18 of 27 patients [67%]), whereas c.-124C>T was the second-most common (8 of 27 patients [30%]). The 2 observed mutations in the 13 patients with BRAF mutations were V600E and V600K, with almost equal distribution (54% and 46%, respectively). The mean (SD) TMB was 33.2 (39) mut/Mb (range, 1-203 mut/Mb). Ten patients (29%) had a TMB < 10 mut/Mb, whereas 24 (69%) had a TMB > 10 mut/Mb. The TMB could not be determined in 1 case. The frequency of TMB-high tumors in the study population compared with frequency in the reference population is shown in Table 3.¹² Only 3 patients (0.64%) in the reference population had MSI-H tumors, and the microsatellite status could not be determined in those tumors (Table 4).¹³ Table 5 outlines statistically significant findings.

Agent Orange Subgroup

AO was a tactical herbicide used by the US military, named for the orange band around the storage barrels. Possible mutagenic properties of AO have been attributed to its byproduct, dioxin. Among the most common cancers known to be associated with AO exposure are bladder and prostate carcinoma and hematopoietic neoplasms. The association between genetic alterations and AO exposure was studied in veterans with prostate cancer.¹⁴ However, to our knowledge, insufficient information is available to determine whether an association exists between exposure to herbicides used in Vietnam or the contaminant dioxin and melanoma. Because a significant proportion of this study population had a well-documented history of AO exposure (37.1%), we were able to analyze them as a subgroup and to separately compare their mutation frequency with the general population.

Results were notable for different distributions of the most frequently mutated genes in the AO subgroup compared with the whole study population. As such, TERT promoter remained the most frequently mutated gene (92%), followed by CDKN2A/B (46%); however, frequency of mutations in NF1 (46%) outnumbered those of BRAF (31%), the fourth-most common mutation. Moreover, when compared with the general melanoma population, a significantly higher frequency of mutations in the NF1 gene was observed in the AO subgroup—not the entire study population.

Discussion

Given that veterans constitute a distinct population, there is reasonable interest in investigating characteristic health issues related to military service. Skin cancer—melanoma in particular—has been researched recently in a veteran population. The differences in demographics, tumor characteristics, and melanoma- specific survival in veterans compared with the general population have already been assessed. According to Chang et al, compared with the general population, veterans are more likely to present with metastatic disease and have lower 5-year survival rates.⁸

Melanoma is one of the most highly mutated malignancies.¹⁵ Fortunately, the most common mutation in melanoma, BRAF V600E, is now considered therapeutically targetable. However, there are still many mutations that are less often discussed and not well understood. Regardless of therapeutic implications, all mutations observed in melanoma are worth investigating because a tumor’s genomic profile also can provide prognostic and etiologic information. Developing comprehensive descriptions of melanoma mutational profiles in specific populations is critical to advancing etiologic understanding and informing prevention strategies.

Our results demonstrate the high prevalence of TERT promoter mutations with characteristic ultraviolet signature (C>T) in the study population. This aligns with general evidence that TERT promoter mutations are common in cutaneous melanomas: 77% of this study sample and up to 86% of all mutations are TERT promoter mutations, according to Davis et al.¹⁵ TERT promoter mutations are positively associated with the initiation, invasion, and metastasis of melanoma. In certain subtypes, there is evidence that the presence of TERT promoter mutations is significantly associated with risk for extranodal metastasis and death.¹⁶ The second-most common mutated gene in the veteran study population was CDKN2A/B (43%), and the third-most mutated gene was BRAF (37%).

In chronically sun-exposed skin NF1, NRAS, and occasionally BRAF V600K mutations tend to predominate. BRAF V600E mutations, on the other hand, are rare in these melanomas.15 In our study population, the most prevalent melanoma site was the trunk (31%), which is considered a location with an intermittent pattern of sun exposure.¹⁷

This study population also had a higher frequency of CDKN2A/B mutations. High frequencies of CDKN2A/B mutations have been reported in familial melanomas, but only 1 patient with CDKN2A/B mutations had a known family history of melanoma.¹⁵ Tumors in the study population showed significantly lower frequency of mutations in ROS1, GRIN2A, KDR, KMT2C (MLL3), KMT2D (MLL2), LRP1B, PTPRT, PTCH1, FAT4, and PREX2 (P < .05).

In this study the subgroup of veterans with AO exposure differed from the whole study population. As such, CDKN2A/B mutations were observed with the same frequency as NF1 mutations (46% each); however, BRAF mutations constituted only 31% of the mutations. In addition, the frequency of NF1 mutations was significantly higher in the AO subgroup compared with the general population, but not in the whole study population.

Our sample also differed from the reference population by showing a significantly higher frequency of TMB-high (ie, ≥ 10 mut/Mb) tumors (71% vs 49%; P = .01).¹² Interestingly, no significant difference in the frequency of TMB-high tumors was observed between the AO subgroup and the reference population (69% vs 49%; P = .16). There also was no statistically significant difference between the frequency of MSI-H tumors in our study population and the reference population (P = .64).¹³

One patient in the study population had uveal melanoma. Mutations encountered in this patient’s tumor differed from the general mutational profile of tumors. None of the 21 mutations depicted in Table 2 were present in this sample.^10,11 On the other hand, those mutations frequently observed in intraocular melanomas, BAP1 and GNA11, were present in this patient.¹⁸ Additionally, this particular melanoma possessed mutations in genes RICTOR, RAD21, and PIK3R1.

Limitations

This study population consisted exclusively of male patients, introducing sex as a potential confounder in analyzing differences between the study population and the general population. As noted in a 2020 systematic review, there were no sex-based differences in the frequency of mutations in BRAF, NRAS, and KIT genes.¹⁹

Regarding NF1 mutations, only NF1-mutated acral and mucosal melanomas were more frequently observed in female patients, whereas nonacral NF1-mutated melanomas were more frequently observed in male patients.²⁰ However, there is currently no clear evidence of whether the mutational landscapes of cutaneous melanoma differ by sex.²¹ Among the 11 cases with NF1-mutatation, site of origin was known in 6, 5 of which originated at nonacral sites. Although the AO subgroup also consisted entirely of male patients, this does not explain the observed increased frequency of NF1 mutations relative to the general population. No such difference was observed between the whole study population, which also consisted exclusively of male patients, and the general population. The similar frequencies of nonacral location in the whole study population (3 acral, 18 nonacral, 14 unknown site of origin) and AO subgroup (1 acral, 7 nonacral, 5 unknown site of origin) preclude location as an explanation.

The Cancer Genome Atlas Network proposed a framework for genomic classification of melanoma into 4 subtypes based on the pattern of the most prevalent significantly mutated genes: mutant BRAF, mutant RAS, mutant NF1, and triple–wild-type. According to that study, BRAF mutations were indeed associated with younger age, in contrast to the NF1-mutant genomic subtype, which was more prevalent in older individuals with higher TMB.²² This emphasizes the need to interpret the potential association of AO exposure and NF1 mutation in melanoma with caution, although additional studies are required to observe the difference between the veteran population and age-matched general population.

On the other hand, Yu et al reported no significant differences of TMB values between patients aged < 60 and ≥ 60 years with melanoma.²³ In short, the observed differences we report in our limited study warrant additional investigation with larger sample sizes, sex-matched controlling, and age-matched controlling. The study was limited by its small sample size and the single location.

Conclusion

The genomic profile of melanomas in the veteran population appears to be similar to that of the general population with a few possible differences. Melanomas in the veteran study population showed a higher frequency of CDKN2A/B mutations; lower frequency of ROS1, GRIN2A, KDR, KMT2C (MLL3), KMT2D (MLL2), LRP1B, PTPRT, PTCH1, FAT4, and PREX2 mutations; and higher TMB. In addition, melanomas in the AO subgroup showed higher frequencies of NF1 mutations. The significance of such findings remains to be determined by further investigation.

The veteran population, with its unique and diverse types of exposure and military service experiences, faces distinct health factors compared with the general population. These factors can be categorized into exposures during military service and those occurring postservice. While the latter phase incorporates psychological issues that may arise while transitioning to civilian life, the service period is associated with major physical, chemical, and psychological exposures that can impact veterans’ health. Carcinogenesis related to military exposures is concerning, and different types of malignancies have been associated with military exposures.¹ The 2022 introduction of the Cancer Moonshot initiative served as a breeding ground for multiple projects aimed at investigation of exposure-related carcinogenesis, prompting increased attention and efforts to linking specific exposures to specific malignancies.²

Melanoma is the deadliest skin cancer, accounting for 1.3% of all cancer deaths.³ Although it may only account for 1% to 5% of skin cancer diagnoses, its incidence in the United States’ population has been increasing.^4,5 There were 97,610 estimated new cases of melanoma in 2023, according to the National Cancer Institute.⁶

The incidence of melanoma may be higher in the military population compared with the general population.⁷ Melanoma is the fourth-most common cancer diagnosed in veterans.⁸

Several demographic characteristics of the US military population are associated with higher melanoma incidence and poorer prognosis, including male sex, older age, and White race. Apart from sun exposure—a known risk factor for melanoma development—other factors, such as service branch, seem to contribute to risk, with the highest melanoma rates noted in the Air Force.⁹ According to a study by Chang et al, veterans have a higher risk of stage III (18%) or stage IV (13%) melanoma at initial diagnosis.⁸

Molecular testing of metastatic melanoma is currently the standard of care for guiding the use of US Food and Drug Administration-approved targeted therapies such as BRAF, MEK, and KIT inhibitors. This comparative analysis details the melanoma comprehensive genomic profiles observed at a large US Department of Veterans Affairs (VA) medical center (VAMC) and those reported in reference databases.

Methods

A query to select all metastatic melanomas sent for comprehensive genomic profiling from the Kansas City VAMC (KCVAMC), identified 35 cases from 2019 through 2023 as the study population. The health records of these patients were reviewed to collect demographic information, military service history, melanoma history, other medical, social, and family histories. The comprehensive genomic profiling reports were reviewed to collect the reported pathogenic variants, microsatellite instability (MSI) status, and tumor mutational burden (TMB) for each case.

The Catalogue of Somatic Mutations in Cancer (COSMIC) was used to identify the most commonly mutated genes in melanomas from The Cancer Genome Atlas for the general population.^4,5 The literature was consulted to determine the MSI status and TMB in melanomas from The Cancer Genome Atlas for separate reference populations.^6,7 The frequency of MSI-high (MSI-H) status, TMB ≥ 10 mutations/megabase (mut/Mb), and mutations in each of the 20 most commonly mutated genes was determined and compared between melanomas from The Cancer Genome Atlas and KCVAMC cases. Corresponding P  values were calculated to identify significant differences. Values were calculated for the entire sample as well as a subgroup with Agent Orange (AO) exposure. The study was approved by the KCVAMC Institutional Review Board.

Results

The mean (SD) age of study participants was 72.9 (9.4) years (range, 39-90 years). The mean (SD) duration of military service was 1654 (1421) days (about 4 years, 6 months, and 10 days). Of the 35 patients included, 22 (63%) served during the Vietnam era (November 1, 1965, to April 30, 1975) and 2 (6%) served during the Persian Gulf War era (August 2, 1990, to February 28, 1991). Seventeen veterans (49%) served in the Army, 9 in the Navy (26%), 5 in the Air Force (14%), and 4 in the Marine Corps (11%). Definitive AO exposure was noted in 13 patients (37%) (Table 1).

Of the 35 patients, 24 (69%) had metastatic disease and the primary site of melanoma was unknown in 14 patients (40%). One patient (Patient 32) had an intraocular melanoma. The primary site was the trunk for 11 patients (31%), the face/head for 7 patients (20%) and extremities for 3 patients (9%). Eight patients (23%) were pT3 stage (thickness > 2 mm but < 4 mm), 7 patients (20%) were pT4 stage (thickness > 4 mm), and 5 patients (14%) were pT1 (thickness ≥ 1 mm). One patient had a primary lesion at pT2 stage, and 1 had a Tis stage lesion. Three patients (9%) had a family history of melanoma in a first-degree relative.

The list of genes mutated in melanoma cells in the study population is provided in the eAppendix.^10,11 Twenty-seven patients (77%) had mutations in TERT promoter, 15 (43%) in CDKN2A/B, 13 (37%) in BRAF, 11 (31%) in NF1, 9 (26%) in TP53, and 8 (23%) in NRAS (Table 2). The majority of mutations in TERT promoter were c.- 146C>T (18 of 27 patients [67%]), whereas c.-124C>T was the second-most common (8 of 27 patients [30%]). The 2 observed mutations in the 13 patients with BRAF mutations were V600E and V600K, with almost equal distribution (54% and 46%, respectively). The mean (SD) TMB was 33.2 (39) mut/Mb (range, 1-203 mut/Mb). Ten patients (29%) had a TMB < 10 mut/Mb, whereas 24 (69%) had a TMB > 10 mut/Mb. The TMB could not be determined in 1 case. The frequency of TMB-high tumors in the study population compared with frequency in the reference population is shown in Table 3.¹² Only 3 patients (0.64%) in the reference population had MSI-H tumors, and the microsatellite status could not be determined in those tumors (Table 4).¹³ Table 5 outlines statistically significant findings.

Agent Orange Subgroup

AO was a tactical herbicide used by the US military, named for the orange band around the storage barrels. Possible mutagenic properties of AO have been attributed to its byproduct, dioxin. Among the most common cancers known to be associated with AO exposure are bladder and prostate carcinoma and hematopoietic neoplasms. The association between genetic alterations and AO exposure was studied in veterans with prostate cancer.¹⁴ However, to our knowledge, insufficient information is available to determine whether an association exists between exposure to herbicides used in Vietnam or the contaminant dioxin and melanoma. Because a significant proportion of this study population had a well-documented history of AO exposure (37.1%), we were able to analyze them as a subgroup and to separately compare their mutation frequency with the general population.

Results were notable for different distributions of the most frequently mutated genes in the AO subgroup compared with the whole study population. As such, TERT promoter remained the most frequently mutated gene (92%), followed by CDKN2A/B (46%); however, frequency of mutations in NF1 (46%) outnumbered those of BRAF (31%), the fourth-most common mutation. Moreover, when compared with the general melanoma population, a significantly higher frequency of mutations in the NF1 gene was observed in the AO subgroup—not the entire study population.

Discussion

Given that veterans constitute a distinct population, there is reasonable interest in investigating characteristic health issues related to military service. Skin cancer—melanoma in particular—has been researched recently in a veteran population. The differences in demographics, tumor characteristics, and melanoma- specific survival in veterans compared with the general population have already been assessed. According to Chang et al, compared with the general population, veterans are more likely to present with metastatic disease and have lower 5-year survival rates.⁸

Melanoma is one of the most highly mutated malignancies.¹⁵ Fortunately, the most common mutation in melanoma, BRAF V600E, is now considered therapeutically targetable. However, there are still many mutations that are less often discussed and not well understood. Regardless of therapeutic implications, all mutations observed in melanoma are worth investigating because a tumor’s genomic profile also can provide prognostic and etiologic information. Developing comprehensive descriptions of melanoma mutational profiles in specific populations is critical to advancing etiologic understanding and informing prevention strategies.

Our results demonstrate the high prevalence of TERT promoter mutations with characteristic ultraviolet signature (C>T) in the study population. This aligns with general evidence that TERT promoter mutations are common in cutaneous melanomas: 77% of this study sample and up to 86% of all mutations are TERT promoter mutations, according to Davis et al.¹⁵ TERT promoter mutations are positively associated with the initiation, invasion, and metastasis of melanoma. In certain subtypes, there is evidence that the presence of TERT promoter mutations is significantly associated with risk for extranodal metastasis and death.¹⁶ The second-most common mutated gene in the veteran study population was CDKN2A/B (43%), and the third-most mutated gene was BRAF (37%).

In chronically sun-exposed skin NF1, NRAS, and occasionally BRAF V600K mutations tend to predominate. BRAF V600E mutations, on the other hand, are rare in these melanomas.15 In our study population, the most prevalent melanoma site was the trunk (31%), which is considered a location with an intermittent pattern of sun exposure.¹⁷

This study population also had a higher frequency of CDKN2A/B mutations. High frequencies of CDKN2A/B mutations have been reported in familial melanomas, but only 1 patient with CDKN2A/B mutations had a known family history of melanoma.¹⁵ Tumors in the study population showed significantly lower frequency of mutations in ROS1, GRIN2A, KDR, KMT2C (MLL3), KMT2D (MLL2), LRP1B, PTPRT, PTCH1, FAT4, and PREX2 (P < .05).

In this study the subgroup of veterans with AO exposure differed from the whole study population. As such, CDKN2A/B mutations were observed with the same frequency as NF1 mutations (46% each); however, BRAF mutations constituted only 31% of the mutations. In addition, the frequency of NF1 mutations was significantly higher in the AO subgroup compared with the general population, but not in the whole study population.

Our sample also differed from the reference population by showing a significantly higher frequency of TMB-high (ie, ≥ 10 mut/Mb) tumors (71% vs 49%; P = .01).¹² Interestingly, no significant difference in the frequency of TMB-high tumors was observed between the AO subgroup and the reference population (69% vs 49%; P = .16). There also was no statistically significant difference between the frequency of MSI-H tumors in our study population and the reference population (P = .64).¹³

One patient in the study population had uveal melanoma. Mutations encountered in this patient’s tumor differed from the general mutational profile of tumors. None of the 21 mutations depicted in Table 2 were present in this sample.^10,11 On the other hand, those mutations frequently observed in intraocular melanomas, BAP1 and GNA11, were present in this patient.¹⁸ Additionally, this particular melanoma possessed mutations in genes RICTOR, RAD21, and PIK3R1.

Limitations

This study population consisted exclusively of male patients, introducing sex as a potential confounder in analyzing differences between the study population and the general population. As noted in a 2020 systematic review, there were no sex-based differences in the frequency of mutations in BRAF, NRAS, and KIT genes.¹⁹

Regarding NF1 mutations, only NF1-mutated acral and mucosal melanomas were more frequently observed in female patients, whereas nonacral NF1-mutated melanomas were more frequently observed in male patients.²⁰ However, there is currently no clear evidence of whether the mutational landscapes of cutaneous melanoma differ by sex.²¹ Among the 11 cases with NF1-mutatation, site of origin was known in 6, 5 of which originated at nonacral sites. Although the AO subgroup also consisted entirely of male patients, this does not explain the observed increased frequency of NF1 mutations relative to the general population. No such difference was observed between the whole study population, which also consisted exclusively of male patients, and the general population. The similar frequencies of nonacral location in the whole study population (3 acral, 18 nonacral, 14 unknown site of origin) and AO subgroup (1 acral, 7 nonacral, 5 unknown site of origin) preclude location as an explanation.

The Cancer Genome Atlas Network proposed a framework for genomic classification of melanoma into 4 subtypes based on the pattern of the most prevalent significantly mutated genes: mutant BRAF, mutant RAS, mutant NF1, and triple–wild-type. According to that study, BRAF mutations were indeed associated with younger age, in contrast to the NF1-mutant genomic subtype, which was more prevalent in older individuals with higher TMB.²² This emphasizes the need to interpret the potential association of AO exposure and NF1 mutation in melanoma with caution, although additional studies are required to observe the difference between the veteran population and age-matched general population.

On the other hand, Yu et al reported no significant differences of TMB values between patients aged < 60 and ≥ 60 years with melanoma.²³ In short, the observed differences we report in our limited study warrant additional investigation with larger sample sizes, sex-matched controlling, and age-matched controlling. The study was limited by its small sample size and the single location.

Conclusion

The genomic profile of melanomas in the veteran population appears to be similar to that of the general population with a few possible differences. Melanomas in the veteran study population showed a higher frequency of CDKN2A/B mutations; lower frequency of ROS1, GRIN2A, KDR, KMT2C (MLL3), KMT2D (MLL2), LRP1B, PTPRT, PTCH1, FAT4, and PREX2 mutations; and higher TMB. In addition, melanomas in the AO subgroup showed higher frequencies of NF1 mutations. The significance of such findings remains to be determined by further investigation.

The veteran population, with its unique and diverse types of exposure and military service experiences, faces distinct health factors compared with the general population. These factors can be categorized into exposures during military service and those occurring postservice. While the latter phase incorporates psychological issues that may arise while transitioning to civilian life, the service period is associated with major physical, chemical, and psychological exposures that can impact veterans’ health. Carcinogenesis related to military exposures is concerning, and different types of malignancies have been associated with military exposures.¹ The 2022 introduction of the Cancer Moonshot initiative served as a breeding ground for multiple projects aimed at investigation of exposure-related carcinogenesis, prompting increased attention and efforts to linking specific exposures to specific malignancies.²

Melanoma is the deadliest skin cancer, accounting for 1.3% of all cancer deaths.³ Although it may only account for 1% to 5% of skin cancer diagnoses, its incidence in the United States’ population has been increasing.^4,5 There were 97,610 estimated new cases of melanoma in 2023, according to the National Cancer Institute.⁶

The incidence of melanoma may be higher in the military population compared with the general population.⁷ Melanoma is the fourth-most common cancer diagnosed in veterans.⁸

Several demographic characteristics of the US military population are associated with higher melanoma incidence and poorer prognosis, including male sex, older age, and White race. Apart from sun exposure—a known risk factor for melanoma development—other factors, such as service branch, seem to contribute to risk, with the highest melanoma rates noted in the Air Force.⁹ According to a study by Chang et al, veterans have a higher risk of stage III (18%) or stage IV (13%) melanoma at initial diagnosis.⁸

Molecular testing of metastatic melanoma is currently the standard of care for guiding the use of US Food and Drug Administration-approved targeted therapies such as BRAF, MEK, and KIT inhibitors. This comparative analysis details the melanoma comprehensive genomic profiles observed at a large US Department of Veterans Affairs (VA) medical center (VAMC) and those reported in reference databases.

Methods

A query to select all metastatic melanomas sent for comprehensive genomic profiling from the Kansas City VAMC (KCVAMC), identified 35 cases from 2019 through 2023 as the study population. The health records of these patients were reviewed to collect demographic information, military service history, melanoma history, other medical, social, and family histories. The comprehensive genomic profiling reports were reviewed to collect the reported pathogenic variants, microsatellite instability (MSI) status, and tumor mutational burden (TMB) for each case.

The Catalogue of Somatic Mutations in Cancer (COSMIC) was used to identify the most commonly mutated genes in melanomas from The Cancer Genome Atlas for the general population.^4,5 The literature was consulted to determine the MSI status and TMB in melanomas from The Cancer Genome Atlas for separate reference populations.^6,7 The frequency of MSI-high (MSI-H) status, TMB ≥ 10 mutations/megabase (mut/Mb), and mutations in each of the 20 most commonly mutated genes was determined and compared between melanomas from The Cancer Genome Atlas and KCVAMC cases. Corresponding P  values were calculated to identify significant differences. Values were calculated for the entire sample as well as a subgroup with Agent Orange (AO) exposure. The study was approved by the KCVAMC Institutional Review Board.

Results

The mean (SD) age of study participants was 72.9 (9.4) years (range, 39-90 years). The mean (SD) duration of military service was 1654 (1421) days (about 4 years, 6 months, and 10 days). Of the 35 patients included, 22 (63%) served during the Vietnam era (November 1, 1965, to April 30, 1975) and 2 (6%) served during the Persian Gulf War era (August 2, 1990, to February 28, 1991). Seventeen veterans (49%) served in the Army, 9 in the Navy (26%), 5 in the Air Force (14%), and 4 in the Marine Corps (11%). Definitive AO exposure was noted in 13 patients (37%) (Table 1).

Of the 35 patients, 24 (69%) had metastatic disease and the primary site of melanoma was unknown in 14 patients (40%). One patient (Patient 32) had an intraocular melanoma. The primary site was the trunk for 11 patients (31%), the face/head for 7 patients (20%) and extremities for 3 patients (9%). Eight patients (23%) were pT3 stage (thickness > 2 mm but < 4 mm), 7 patients (20%) were pT4 stage (thickness > 4 mm), and 5 patients (14%) were pT1 (thickness ≥ 1 mm). One patient had a primary lesion at pT2 stage, and 1 had a Tis stage lesion. Three patients (9%) had a family history of melanoma in a first-degree relative.

The list of genes mutated in melanoma cells in the study population is provided in the eAppendix.^10,11 Twenty-seven patients (77%) had mutations in TERT promoter, 15 (43%) in CDKN2A/B, 13 (37%) in BRAF, 11 (31%) in NF1, 9 (26%) in TP53, and 8 (23%) in NRAS (Table 2). The majority of mutations in TERT promoter were c.- 146C>T (18 of 27 patients [67%]), whereas c.-124C>T was the second-most common (8 of 27 patients [30%]). The 2 observed mutations in the 13 patients with BRAF mutations were V600E and V600K, with almost equal distribution (54% and 46%, respectively). The mean (SD) TMB was 33.2 (39) mut/Mb (range, 1-203 mut/Mb). Ten patients (29%) had a TMB < 10 mut/Mb, whereas 24 (69%) had a TMB > 10 mut/Mb. The TMB could not be determined in 1 case. The frequency of TMB-high tumors in the study population compared with frequency in the reference population is shown in Table 3.¹² Only 3 patients (0.64%) in the reference population had MSI-H tumors, and the microsatellite status could not be determined in those tumors (Table 4).¹³ Table 5 outlines statistically significant findings.

Agent Orange Subgroup

AO was a tactical herbicide used by the US military, named for the orange band around the storage barrels. Possible mutagenic properties of AO have been attributed to its byproduct, dioxin. Among the most common cancers known to be associated with AO exposure are bladder and prostate carcinoma and hematopoietic neoplasms. The association between genetic alterations and AO exposure was studied in veterans with prostate cancer.¹⁴ However, to our knowledge, insufficient information is available to determine whether an association exists between exposure to herbicides used in Vietnam or the contaminant dioxin and melanoma. Because a significant proportion of this study population had a well-documented history of AO exposure (37.1%), we were able to analyze them as a subgroup and to separately compare their mutation frequency with the general population.

Results were notable for different distributions of the most frequently mutated genes in the AO subgroup compared with the whole study population. As such, TERT promoter remained the most frequently mutated gene (92%), followed by CDKN2A/B (46%); however, frequency of mutations in NF1 (46%) outnumbered those of BRAF (31%), the fourth-most common mutation. Moreover, when compared with the general melanoma population, a significantly higher frequency of mutations in the NF1 gene was observed in the AO subgroup—not the entire study population.

Discussion

Given that veterans constitute a distinct population, there is reasonable interest in investigating characteristic health issues related to military service. Skin cancer—melanoma in particular—has been researched recently in a veteran population. The differences in demographics, tumor characteristics, and melanoma- specific survival in veterans compared with the general population have already been assessed. According to Chang et al, compared with the general population, veterans are more likely to present with metastatic disease and have lower 5-year survival rates.⁸

Melanoma is one of the most highly mutated malignancies.¹⁵ Fortunately, the most common mutation in melanoma, BRAF V600E, is now considered therapeutically targetable. However, there are still many mutations that are less often discussed and not well understood. Regardless of therapeutic implications, all mutations observed in melanoma are worth investigating because a tumor’s genomic profile also can provide prognostic and etiologic information. Developing comprehensive descriptions of melanoma mutational profiles in specific populations is critical to advancing etiologic understanding and informing prevention strategies.

Our results demonstrate the high prevalence of TERT promoter mutations with characteristic ultraviolet signature (C>T) in the study population. This aligns with general evidence that TERT promoter mutations are common in cutaneous melanomas: 77% of this study sample and up to 86% of all mutations are TERT promoter mutations, according to Davis et al.¹⁵ TERT promoter mutations are positively associated with the initiation, invasion, and metastasis of melanoma. In certain subtypes, there is evidence that the presence of TERT promoter mutations is significantly associated with risk for extranodal metastasis and death.¹⁶ The second-most common mutated gene in the veteran study population was CDKN2A/B (43%), and the third-most mutated gene was BRAF (37%).

In chronically sun-exposed skin NF1, NRAS, and occasionally BRAF V600K mutations tend to predominate. BRAF V600E mutations, on the other hand, are rare in these melanomas.15 In our study population, the most prevalent melanoma site was the trunk (31%), which is considered a location with an intermittent pattern of sun exposure.¹⁷

This study population also had a higher frequency of CDKN2A/B mutations. High frequencies of CDKN2A/B mutations have been reported in familial melanomas, but only 1 patient with CDKN2A/B mutations had a known family history of melanoma.¹⁵ Tumors in the study population showed significantly lower frequency of mutations in ROS1, GRIN2A, KDR, KMT2C (MLL3), KMT2D (MLL2), LRP1B, PTPRT, PTCH1, FAT4, and PREX2 (P < .05).

In this study the subgroup of veterans with AO exposure differed from the whole study population. As such, CDKN2A/B mutations were observed with the same frequency as NF1 mutations (46% each); however, BRAF mutations constituted only 31% of the mutations. In addition, the frequency of NF1 mutations was significantly higher in the AO subgroup compared with the general population, but not in the whole study population.

Our sample also differed from the reference population by showing a significantly higher frequency of TMB-high (ie, ≥ 10 mut/Mb) tumors (71% vs 49%; P = .01).¹² Interestingly, no significant difference in the frequency of TMB-high tumors was observed between the AO subgroup and the reference population (69% vs 49%; P = .16). There also was no statistically significant difference between the frequency of MSI-H tumors in our study population and the reference population (P = .64).¹³

One patient in the study population had uveal melanoma. Mutations encountered in this patient’s tumor differed from the general mutational profile of tumors. None of the 21 mutations depicted in Table 2 were present in this sample.^10,11 On the other hand, those mutations frequently observed in intraocular melanomas, BAP1 and GNA11, were present in this patient.¹⁸ Additionally, this particular melanoma possessed mutations in genes RICTOR, RAD21, and PIK3R1.

Limitations

This study population consisted exclusively of male patients, introducing sex as a potential confounder in analyzing differences between the study population and the general population. As noted in a 2020 systematic review, there were no sex-based differences in the frequency of mutations in BRAF, NRAS, and KIT genes.¹⁹

Regarding NF1 mutations, only NF1-mutated acral and mucosal melanomas were more frequently observed in female patients, whereas nonacral NF1-mutated melanomas were more frequently observed in male patients.²⁰ However, there is currently no clear evidence of whether the mutational landscapes of cutaneous melanoma differ by sex.²¹ Among the 11 cases with NF1-mutatation, site of origin was known in 6, 5 of which originated at nonacral sites. Although the AO subgroup also consisted entirely of male patients, this does not explain the observed increased frequency of NF1 mutations relative to the general population. No such difference was observed between the whole study population, which also consisted exclusively of male patients, and the general population. The similar frequencies of nonacral location in the whole study population (3 acral, 18 nonacral, 14 unknown site of origin) and AO subgroup (1 acral, 7 nonacral, 5 unknown site of origin) preclude location as an explanation.

The Cancer Genome Atlas Network proposed a framework for genomic classification of melanoma into 4 subtypes based on the pattern of the most prevalent significantly mutated genes: mutant BRAF, mutant RAS, mutant NF1, and triple–wild-type. According to that study, BRAF mutations were indeed associated with younger age, in contrast to the NF1-mutant genomic subtype, which was more prevalent in older individuals with higher TMB.²² This emphasizes the need to interpret the potential association of AO exposure and NF1 mutation in melanoma with caution, although additional studies are required to observe the difference between the veteran population and age-matched general population.

On the other hand, Yu et al reported no significant differences of TMB values between patients aged < 60 and ≥ 60 years with melanoma.²³ In short, the observed differences we report in our limited study warrant additional investigation with larger sample sizes, sex-matched controlling, and age-matched controlling. The study was limited by its small sample size and the single location.

Conclusion

The genomic profile of melanomas in the veteran population appears to be similar to that of the general population with a few possible differences. Melanomas in the veteran study population showed a higher frequency of CDKN2A/B mutations; lower frequency of ROS1, GRIN2A, KDR, KMT2C (MLL3), KMT2D (MLL2), LRP1B, PTPRT, PTCH1, FAT4, and PREX2 mutations; and higher TMB. In addition, melanomas in the AO subgroup showed higher frequencies of NF1 mutations. The significance of such findings remains to be determined by further investigation.

Colorectal cancer (CRC) is the second-leading cause of cancer-related deaths in the United States, with an estimated 52,550 deaths in 2023.¹ However, the disease burden varies among different segments of the population.² While both CRC incidence and mortality have been decreasing due to screening and advances in treatment, there are disparities in incidence and mortality across the sociodemographic spectrum including race, ethnicity, education, and income.^1-4 While CRC incidence is decreasing for older adults, it is increasing among those aged < 55 years.⁵ The incidence of CRC in adults aged 40 to 54 years has increased by 0.5% to 1.3% annually since the mid-1990s.⁶ The US Preventive Services Task Force now recommends starting CRC screening at age 45 years for asymptomatic adults with average risk.⁷

Disparities also exist across geographical boundaries and living environment. Rural Americans faces additional challenges in health and lifestyle that can affect CRC outcomes. Compared to their urban counterparts, rural residents are more likely to be older, have lower levels of education, higher levels of poverty, lack health insurance, and less access to health care practitioners (HCPs).^8-10 Geographic proximity, defined as travel time or physical distance to a health facility, has been recognized as a predictor of inferior outcomes.¹¹ These aspects of rural living may pose challenges for accessing care for CRC screening and treatment.^11-13 National and local studies have shown disparities in CRC screening rates, incidence, and mortality between rural and urban populations.^14-16

It is unclear whether rural/urban disparities persist under the Veterans Health Administration (VHA) health care delivery model. This study examined differences in baseline characteristics and mortality between rural and urban veterans newly diagnosed with CRC. We also focused on a subpopulation aged ≤ 45 years.

Methods

This study extracted national data from the US Department of Veterans Affairs (VA) Corporate Data Warehouse (CDW) hosted in the VA Informatics and Computing Infrastructure (VINCI) environment. VINCI is an initiative to improve access to VA data and facilitate the analysis of these data while ensuring veterans’ privacy and data security.¹⁷ CDW is the VHA business intelligence information repository, which extracts data from clinical and nonclinical sources following prescribed and validated protocols. Data extracted included demographics, diagnosis, and procedure codes for both inpatient and outpatient encounters, vital signs, and vital status. This study used data previously extracted from a national cohort of veterans that encompassed all patients who received a group of commonly prescribed medications, such as statins, proton pump inhibitors, histamine-2 blockers, acetaminophen-containing products, and hydrocortisone-containing skin applications. This cohort encompassed 8,648,754 veterans, from whom 2,460,727 had encounters during fiscal years (FY) 2016 to 2021 (study period). The cohort was used to ensure that subjects were VHA patients, allowing them to adequately capture their clinical profiles.

Patients were identified as rural or urban based on their residence address at the date of their first diagnosis of CRC. The Geospatial Service Support Center (GSSC) aggregates and updates veterans’ residence address records for all enrolled veterans from the National Change of Address database. The data contain 1 record per enrollee. GSSC Geocoded Enrollee File contains enrollee addresses and their rurality indicators, categorized as urban, rural, or highly rural.¹⁸ Rurality is defined by the Rural Urban Commuting Area (RUCA) categories developed by the Department of Agriculture and the Health Resources and Services Administration of the US Department of Health and Human Services.¹⁹ Urban areas had RUCA codes of 1.0 to 1.1, and highly rural areas had RUCA scores of 10.0. All other areas were classified as rural. Since the proportion of veterans from highly rural areas was small, we included residents from highly rural areas in the rural residents’ group.

Inclusion and Exclusion Criteria

All veterans newly diagnosed with CRC from FY 2016 to 2021 were included. We used the ninth and tenth clinical modification revisions of the International Classification of Diseases (ICD-9-CM and ICD-10-CM) to define CRC diagnosis (Supplemental materials).^4,20 To ensure that patients were newly diagnosed with CRC, this study excluded patients with a previous ICD-9-CM code for CRC diagnosis since FY 2003.

Comorbidities were identified using diagnosis and procedure codes from inpatient and outpatient encounters, which were used to calculate the Charlson Comorbidity Index (CCI) at the time of CRC diagnosis using the weighted method described by Schneeweiss et al.²¹ We defined CRC high-risk conditions and CRC screening tests, including flexible sigmoidoscopy and stool tests, as described in previous studies (Supplemental materials).²⁰

The main outcome was total mortality. The date of death was extracted from the VHA Death Ascertainment File, which contains mortality data from the Master Person Index file in CDW and the Social Security Administration Death Master File. We used the date of death from any cause, as cause of death was not available.

A propensity score (PS) was created to match rural (including highly rural) and urban residents at a ratio of 1:1. Using a standard procedure described in prior publications, multivariable logistic regression used all baseline characteristics to estimate the PS and perform nearest-number matching without replacement.^22,23 A caliper of 0.01 maximized the matched cohort size and achieved balance (Supplemental materials). We then examined the balance of baseline characteristics between PS-matched groups.

Analyses

Cox proportional hazards regression analysis estimated the hazard ratio (HR) of death in rural residents compared to urban residents in the PS-matched cohort. The outcome event was the date of death during the study’s follow-up period (defined as period from first CRC diagnosis to death or study end), with censoring at the study’s end date (September 30, 2021). The proportional hazards assumption was assessed by inspecting the Kaplan-Meier curves. Multiple analyses examined the HR of total mortality in the PS-matched cohort, stratified by sex, race, and ethnicity. We also examined the HR of total mortality stratified by duration of follow-up.

Another PS-matching analysis among veterans aged ≤ 45 years was performed using the same techniques described earlier in this article. We performed a Cox proportional hazards regression analysis to compare mortality in PS-matched urban and rural veterans aged ≤ 45 years. The HR of death in all veterans aged ≤ 45 years (before PS-matching) was estimated using Cox proportional hazard regression analysis, adjusting for PS.

Dichotomous variables were compared using X² tests and continuous variables were compared using t tests. Baseline characteristics with missing values were converted into categorical variables and the proportion of subjects with missing values was equalized between treatment groups after PS-matching. For subgroup analysis, we examined the HR of total mortality in each subgroup using separate Cox proportional hazards regression models similar to the primary analysis but adjusted for PS. Due to multiple comparisons in the subgroup analysis, the findings should be considered exploratory. Statistical tests were 2-tailed, and significance was defined as P < .05. Data management and statistical analyses were conducted from June 2022 to January 2023 using STATA, Version 17. The VA Orlando Healthcare System Institutional Review Board approved the study and waived requirements for informed consent because only deidentified data were used.

Results

After excluding 49 patients (Supplemental materials, available at doi:10.12788/fp.0560), we identified 30,219 veterans with newly diagnosed CRC between FY 2016 to 2021 (Table 1). Of these, 19,422 (64.3%) resided in urban areas and 10,797 (35.7%) resided in rural areas (Table 2). The mean (SD) duration from the first CRC diagnosis to death or study end was 832 (640) days, and the median (IQR) was 723 (246–1330) days. Overall, incident CRC diagnoses were numerically highest in FY 2016 and lowest in FY 2020 (Figure 1). Patients with CRC in rural areas vs urban areas were significantly older (mean, 71.2 years vs 70.8 years, respectively; P < .001), more likely to be male (96.7% vs 95.7%, respectively; P < .001), more likely to be White (83.6% vs 67.8%, respectively; P < .001) and more likely to be non-Hispanic (92.2% vs 87.5%, respectively; P < .001). In terms of general health, rural veterans with CRC were more likely to be overweight or obese (81.5% rural vs 78.5% urban; P < .001) but had fewer mean comorbidities as measured by CCI (5.66 rural vs 5.90 urban; P < .001). A higher proportion of rural veterans with CRC had received stool-based (fecal occult blood test or fecal immunochemical test) CRC screening tests (61.6% rural vs 57.2% urban; P < .001). Fewer rural patients presented with systemic symptoms or signs within 1 year of CRC diagnosis (54.4% rural vs 57.5% urban, P < .001). Among urban patients with CRC, 6959 (35.8%) deaths were observed, compared with 3766 (34.9%) among rural patients (P = .10).

There were 21,568 PS-matched veterans: 10,784 in each group. In the PS-matched cohort, baseline characteristics were similar between veterans in urban and rural communities, including age, sex, race/ethnicity, body mass index, and comorbidities. Among rural patients with CRC, 3763 deaths (34.9%) were observed compared with 3702 (34.3%) among urban veterans. There was no significant difference in the HR of mortality between rural and urban CRC residents (HR, 1.01; 95% CI, 0.97-1.06; P = .53) (Figure 2).

Among veterans aged ≤ 45 years, 551 were diagnosed with CRC (391 urban and 160 rural). We PS-matched 142 pairs of urban and rural veterans without residual differences in baseline characteristics (eAppendix 1). There was no significant difference in the HR of mortality between rural and urban veterans aged ≤ 45 years (HR, 0.97; 95% CI, 0.57-1.63; P = .90) (Figure 2). Similarly, no difference in mortality was observed adjusting for PS between all rural and urban veterans aged ≤ 45 years (HR, 1.03; 95% CI, 0.67-1.59; P = .88).

There was no difference in total mortality between rural and urban veterans in any subgroup except for American Indian or Alaska Native veterans (HR, 2.41; 95% CI, 1.29-4.50; P = .006) (eAppendix 2).

Discussion

This study examined characteristics of patients with CRC between urban and rural areas among veterans who were VHA patients. Similar to other studies, rural veterans with CRC were older, more likely to be White, and were obese, but exhibited fewer comorbidities (lower CCI and lower incidence of congestive heart failure, dementia, hemiplegia, kidney diseases, liver diseases and AIDS, but higher incidence of chronic obstructive lung disease).^8,16 The incidence of CRC in this study population was lowest in FY 2020, which was reported by the Centers for Disease Control and Prevention and is attributed to COVID-19 pandemic disruption of health services.²⁴ The overall mortality in this study was similar to rates reported in other studies from the VA Central Cancer Registry.⁴ In the PS-matched cohort, where baseline characteristics were similar between urban and rural patients with CRC, we found no disparities in CRC-specific mortality between veterans in rural and urban areas. Additionally, when analysis was restricted to veterans aged ≤ 45 years, the results remained consistent.

Subgroup analyses showed no significant difference in mortality between rural and urban areas by sex, race or ethnicity, except rural American Indian or Alaska Native veterans who had double the mortality of their urban counterparts (HR, 2.41; 95% CI, 1.29-4.50; P = .006). This finding is difficult to interpret due to the small number of events and the wide CI. While with a Bonferroni correction the adjusted P value was .08, which is not statistically significant, a previous study found that although CRC incidence was lower overall in American Indian or Alaska Native populations compared to non-Hispanic White populations, CRC incidence was higher among American Indian or Alaska Native individuals in some areas such as Alaska and the Northern Plains.^25,26 Studies have noted that rural American Indian/Alaska Native populations experience greater poverty, less access to broadband internet, and limited access to care, contributing to poorer cancer outcomes and lower survival.²⁷ Thus, the finding of disparity in mortality between rural and urban American Indian or Alaska Native veterans warrants further study.

Other studies have raised concerns that CRC disproportionately affects adults in rural areas with higher mortality rates.^14-16 These disparities arise from sociodemographic factors and modifiable risk factors, including physical activity, dietary patterns, access to cancer screening, and gaps in quality treatment resources.^16,28 These factors operate at multiple levels: from individual, local health system, to community and policy.^2,27 For example, a South Carolina study (1996–2016) found that residents in rural areas were more likely to be diagnosed with advanced CRC, possibly indicating lower rates of CRC screening in rural areas. They also had higher likelihood of death from CRC.¹⁵ However, the study did not include any clinical parameters, such as comorbidities or obesity. A statewide, population-based study in Utah showed that rural men experienced a lower CRC survival in their unadjusted analysis.¹⁶ However, the study was small, with only 3948 urban and 712 rural residents. Additionally, there was no difference in total mortality in the whole cohort (HR, 0.96; 95% CI, 0.86-1.07) or in CRC-specific death (HR, 0.93; 95% CI, 0.81-1.08). A nationwide study also showed that CRC mortality rates were 8% higher in nonmetropolitan or rural areas than in the most urbanized areas containing large metropolitan counties.²⁹ However, this study did not include descriptions of clinical confounders, such as comorbidities, making it difficult to ascertain whether the difference in CRC mortality was due to rurality or differences in baseline risk characteristics.

In this study, the lack of CRC-specific mortality disparities may be attributed to the structures and practices of VHA health care. Recent studies have noted that mortality of several chronic medical conditions treated at the VHA was lower than at non-VHA hospitals.^30,31 One study that measured the quality of nonmetastatic CRC care based on National Comprehensive Cancer Network guidelines showed that > 72% of VHA patients received guideline-concordant care for each diagnostic and therapeutic measure, except for follow-up colonoscopy timing, which appear to be similar or superior to that of the private sector.^30,32,33 Some of the VA initiative for CRC screening may bypass the urban-rurality divide such as the mailed fecal immunochemical test program for CRC. This program was implemented at the onset of the COVID-19 pandemic to avoid disruptions of medical care.³⁴ Rural patients are more likely to undergo fecal immunochemical testing when compared to urban patients in this data. Beyond clinical care, the VHA uses processes to tackle social determinants of health such as housing, food security, and transportation, promoting equal access to health care, and promoting cultural competency among HCPs.^35-37

The results suggest that solutions to CRC disparities between rural and urban areas need to consider known barriers to rural health care, including transportation, diminished rural health care workforce, and other social determinants of health.^9,10,27,38 VHA makes considerable efforts to provide equitable care to all enrolled veterans, including specific programs for rural veterans, including ongoing outreach.³⁹ This study demonstrated lack of disparity in CRC-specific mortality in veterans receiving VHA care, highlighting the importance of these efforts.

Strengths and Limitations

This study used the VHA cohort to compare patient characteristics and mortality between patients with CRC residing in rural and urban areas. The study provides nationwide perspectives on CRC across the geographical spectrum and used a longitudinal cohort with prolonged follow-up to account for comorbidities.

However, the study compared a cohort of rural and urban veterans enrolled in the VHA; hence, the results may not reflect CRC outcomes in veterans without access to VHA care. Rurality has been independently associated with decreased likelihood of meeting CRC screening guidelines among veterans and military service members.³⁸ This study lacked sufficient information to compare CRC staging or treatment modalities among veterans. Although the data cannot identify CRC stage, the proportions of patients with metastatic CRC at diagnosis and CRC location were similar between groups. The study did not have information on their care outside of VHA setting.

This study could not ascertain whether disparities existed in CRC treatment modality since rural residence may result in referral to community-based CRC care, which did not appear in the data. To address these limitations, we used death from any cause as the primary outcome, since death is a hard outcome and is not subject to ascertainment bias. The relatively short follow-up time is another limitation, though subgroup analysis by follow-up did not show significant differences. Despite PS matching, residual unmeasured confounding may exist between urban and rural groups. The predominantly White, male VHA population with high CCI may limit the generalizability of the results.

Conclusions

Rural VHA enrollees had similar survival rates after CRC diagnosis compared to their urban counterparts in a PS-matched analysis. The VHA models of care—including mailed CRC screening tools, several socioeconomic determinants of health (housing, food security, and transportation), and promoting equal access to health care, as well as cultural competency among HCPs—HCPs—may help alleviate disparities across the rural-urban spectrum. The VHA should continue efforts to enroll veterans and provide comprehensive coordinated care in community partnerships.

Colorectal cancer (CRC) is the second-leading cause of cancer-related deaths in the United States, with an estimated 52,550 deaths in 2023.¹ However, the disease burden varies among different segments of the population.² While both CRC incidence and mortality have been decreasing due to screening and advances in treatment, there are disparities in incidence and mortality across the sociodemographic spectrum including race, ethnicity, education, and income.^1-4 While CRC incidence is decreasing for older adults, it is increasing among those aged < 55 years.⁵ The incidence of CRC in adults aged 40 to 54 years has increased by 0.5% to 1.3% annually since the mid-1990s.⁶ The US Preventive Services Task Force now recommends starting CRC screening at age 45 years for asymptomatic adults with average risk.⁷

Disparities also exist across geographical boundaries and living environment. Rural Americans faces additional challenges in health and lifestyle that can affect CRC outcomes. Compared to their urban counterparts, rural residents are more likely to be older, have lower levels of education, higher levels of poverty, lack health insurance, and less access to health care practitioners (HCPs).^8-10 Geographic proximity, defined as travel time or physical distance to a health facility, has been recognized as a predictor of inferior outcomes.¹¹ These aspects of rural living may pose challenges for accessing care for CRC screening and treatment.^11-13 National and local studies have shown disparities in CRC screening rates, incidence, and mortality between rural and urban populations.^14-16

It is unclear whether rural/urban disparities persist under the Veterans Health Administration (VHA) health care delivery model. This study examined differences in baseline characteristics and mortality between rural and urban veterans newly diagnosed with CRC. We also focused on a subpopulation aged ≤ 45 years.

Methods

This study extracted national data from the US Department of Veterans Affairs (VA) Corporate Data Warehouse (CDW) hosted in the VA Informatics and Computing Infrastructure (VINCI) environment. VINCI is an initiative to improve access to VA data and facilitate the analysis of these data while ensuring veterans’ privacy and data security.¹⁷ CDW is the VHA business intelligence information repository, which extracts data from clinical and nonclinical sources following prescribed and validated protocols. Data extracted included demographics, diagnosis, and procedure codes for both inpatient and outpatient encounters, vital signs, and vital status. This study used data previously extracted from a national cohort of veterans that encompassed all patients who received a group of commonly prescribed medications, such as statins, proton pump inhibitors, histamine-2 blockers, acetaminophen-containing products, and hydrocortisone-containing skin applications. This cohort encompassed 8,648,754 veterans, from whom 2,460,727 had encounters during fiscal years (FY) 2016 to 2021 (study period). The cohort was used to ensure that subjects were VHA patients, allowing them to adequately capture their clinical profiles.

Patients were identified as rural or urban based on their residence address at the date of their first diagnosis of CRC. The Geospatial Service Support Center (GSSC) aggregates and updates veterans’ residence address records for all enrolled veterans from the National Change of Address database. The data contain 1 record per enrollee. GSSC Geocoded Enrollee File contains enrollee addresses and their rurality indicators, categorized as urban, rural, or highly rural.¹⁸ Rurality is defined by the Rural Urban Commuting Area (RUCA) categories developed by the Department of Agriculture and the Health Resources and Services Administration of the US Department of Health and Human Services.¹⁹ Urban areas had RUCA codes of 1.0 to 1.1, and highly rural areas had RUCA scores of 10.0. All other areas were classified as rural. Since the proportion of veterans from highly rural areas was small, we included residents from highly rural areas in the rural residents’ group.

Inclusion and Exclusion Criteria

All veterans newly diagnosed with CRC from FY 2016 to 2021 were included. We used the ninth and tenth clinical modification revisions of the International Classification of Diseases (ICD-9-CM and ICD-10-CM) to define CRC diagnosis (Supplemental materials).^4,20 To ensure that patients were newly diagnosed with CRC, this study excluded patients with a previous ICD-9-CM code for CRC diagnosis since FY 2003.

Comorbidities were identified using diagnosis and procedure codes from inpatient and outpatient encounters, which were used to calculate the Charlson Comorbidity Index (CCI) at the time of CRC diagnosis using the weighted method described by Schneeweiss et al.²¹ We defined CRC high-risk conditions and CRC screening tests, including flexible sigmoidoscopy and stool tests, as described in previous studies (Supplemental materials).²⁰

The main outcome was total mortality. The date of death was extracted from the VHA Death Ascertainment File, which contains mortality data from the Master Person Index file in CDW and the Social Security Administration Death Master File. We used the date of death from any cause, as cause of death was not available.

A propensity score (PS) was created to match rural (including highly rural) and urban residents at a ratio of 1:1. Using a standard procedure described in prior publications, multivariable logistic regression used all baseline characteristics to estimate the PS and perform nearest-number matching without replacement.^22,23 A caliper of 0.01 maximized the matched cohort size and achieved balance (Supplemental materials). We then examined the balance of baseline characteristics between PS-matched groups.

Analyses

Cox proportional hazards regression analysis estimated the hazard ratio (HR) of death in rural residents compared to urban residents in the PS-matched cohort. The outcome event was the date of death during the study’s follow-up period (defined as period from first CRC diagnosis to death or study end), with censoring at the study’s end date (September 30, 2021). The proportional hazards assumption was assessed by inspecting the Kaplan-Meier curves. Multiple analyses examined the HR of total mortality in the PS-matched cohort, stratified by sex, race, and ethnicity. We also examined the HR of total mortality stratified by duration of follow-up.

Another PS-matching analysis among veterans aged ≤ 45 years was performed using the same techniques described earlier in this article. We performed a Cox proportional hazards regression analysis to compare mortality in PS-matched urban and rural veterans aged ≤ 45 years. The HR of death in all veterans aged ≤ 45 years (before PS-matching) was estimated using Cox proportional hazard regression analysis, adjusting for PS.

Dichotomous variables were compared using X² tests and continuous variables were compared using t tests. Baseline characteristics with missing values were converted into categorical variables and the proportion of subjects with missing values was equalized between treatment groups after PS-matching. For subgroup analysis, we examined the HR of total mortality in each subgroup using separate Cox proportional hazards regression models similar to the primary analysis but adjusted for PS. Due to multiple comparisons in the subgroup analysis, the findings should be considered exploratory. Statistical tests were 2-tailed, and significance was defined as P < .05. Data management and statistical analyses were conducted from June 2022 to January 2023 using STATA, Version 17. The VA Orlando Healthcare System Institutional Review Board approved the study and waived requirements for informed consent because only deidentified data were used.

Results

After excluding 49 patients (Supplemental materials, available at doi:10.12788/fp.0560), we identified 30,219 veterans with newly diagnosed CRC between FY 2016 to 2021 (Table 1). Of these, 19,422 (64.3%) resided in urban areas and 10,797 (35.7%) resided in rural areas (Table 2). The mean (SD) duration from the first CRC diagnosis to death or study end was 832 (640) days, and the median (IQR) was 723 (246–1330) days. Overall, incident CRC diagnoses were numerically highest in FY 2016 and lowest in FY 2020 (Figure 1). Patients with CRC in rural areas vs urban areas were significantly older (mean, 71.2 years vs 70.8 years, respectively; P < .001), more likely to be male (96.7% vs 95.7%, respectively; P < .001), more likely to be White (83.6% vs 67.8%, respectively; P < .001) and more likely to be non-Hispanic (92.2% vs 87.5%, respectively; P < .001). In terms of general health, rural veterans with CRC were more likely to be overweight or obese (81.5% rural vs 78.5% urban; P < .001) but had fewer mean comorbidities as measured by CCI (5.66 rural vs 5.90 urban; P < .001). A higher proportion of rural veterans with CRC had received stool-based (fecal occult blood test or fecal immunochemical test) CRC screening tests (61.6% rural vs 57.2% urban; P < .001). Fewer rural patients presented with systemic symptoms or signs within 1 year of CRC diagnosis (54.4% rural vs 57.5% urban, P < .001). Among urban patients with CRC, 6959 (35.8%) deaths were observed, compared with 3766 (34.9%) among rural patients (P = .10).

There were 21,568 PS-matched veterans: 10,784 in each group. In the PS-matched cohort, baseline characteristics were similar between veterans in urban and rural communities, including age, sex, race/ethnicity, body mass index, and comorbidities. Among rural patients with CRC, 3763 deaths (34.9%) were observed compared with 3702 (34.3%) among urban veterans. There was no significant difference in the HR of mortality between rural and urban CRC residents (HR, 1.01; 95% CI, 0.97-1.06; P = .53) (Figure 2).

Among veterans aged ≤ 45 years, 551 were diagnosed with CRC (391 urban and 160 rural). We PS-matched 142 pairs of urban and rural veterans without residual differences in baseline characteristics (eAppendix 1). There was no significant difference in the HR of mortality between rural and urban veterans aged ≤ 45 years (HR, 0.97; 95% CI, 0.57-1.63; P = .90) (Figure 2). Similarly, no difference in mortality was observed adjusting for PS between all rural and urban veterans aged ≤ 45 years (HR, 1.03; 95% CI, 0.67-1.59; P = .88).

There was no difference in total mortality between rural and urban veterans in any subgroup except for American Indian or Alaska Native veterans (HR, 2.41; 95% CI, 1.29-4.50; P = .006) (eAppendix 2).

Discussion

This study examined characteristics of patients with CRC between urban and rural areas among veterans who were VHA patients. Similar to other studies, rural veterans with CRC were older, more likely to be White, and were obese, but exhibited fewer comorbidities (lower CCI and lower incidence of congestive heart failure, dementia, hemiplegia, kidney diseases, liver diseases and AIDS, but higher incidence of chronic obstructive lung disease).^8,16 The incidence of CRC in this study population was lowest in FY 2020, which was reported by the Centers for Disease Control and Prevention and is attributed to COVID-19 pandemic disruption of health services.²⁴ The overall mortality in this study was similar to rates reported in other studies from the VA Central Cancer Registry.⁴ In the PS-matched cohort, where baseline characteristics were similar between urban and rural patients with CRC, we found no disparities in CRC-specific mortality between veterans in rural and urban areas. Additionally, when analysis was restricted to veterans aged ≤ 45 years, the results remained consistent.

Subgroup analyses showed no significant difference in mortality between rural and urban areas by sex, race or ethnicity, except rural American Indian or Alaska Native veterans who had double the mortality of their urban counterparts (HR, 2.41; 95% CI, 1.29-4.50; P = .006). This finding is difficult to interpret due to the small number of events and the wide CI. While with a Bonferroni correction the adjusted P value was .08, which is not statistically significant, a previous study found that although CRC incidence was lower overall in American Indian or Alaska Native populations compared to non-Hispanic White populations, CRC incidence was higher among American Indian or Alaska Native individuals in some areas such as Alaska and the Northern Plains.^25,26 Studies have noted that rural American Indian/Alaska Native populations experience greater poverty, less access to broadband internet, and limited access to care, contributing to poorer cancer outcomes and lower survival.²⁷ Thus, the finding of disparity in mortality between rural and urban American Indian or Alaska Native veterans warrants further study.

Other studies have raised concerns that CRC disproportionately affects adults in rural areas with higher mortality rates.^14-16 These disparities arise from sociodemographic factors and modifiable risk factors, including physical activity, dietary patterns, access to cancer screening, and gaps in quality treatment resources.^16,28 These factors operate at multiple levels: from individual, local health system, to community and policy.^2,27 For example, a South Carolina study (1996–2016) found that residents in rural areas were more likely to be diagnosed with advanced CRC, possibly indicating lower rates of CRC screening in rural areas. They also had higher likelihood of death from CRC.¹⁵ However, the study did not include any clinical parameters, such as comorbidities or obesity. A statewide, population-based study in Utah showed that rural men experienced a lower CRC survival in their unadjusted analysis.¹⁶ However, the study was small, with only 3948 urban and 712 rural residents. Additionally, there was no difference in total mortality in the whole cohort (HR, 0.96; 95% CI, 0.86-1.07) or in CRC-specific death (HR, 0.93; 95% CI, 0.81-1.08). A nationwide study also showed that CRC mortality rates were 8% higher in nonmetropolitan or rural areas than in the most urbanized areas containing large metropolitan counties.²⁹ However, this study did not include descriptions of clinical confounders, such as comorbidities, making it difficult to ascertain whether the difference in CRC mortality was due to rurality or differences in baseline risk characteristics.

In this study, the lack of CRC-specific mortality disparities may be attributed to the structures and practices of VHA health care. Recent studies have noted that mortality of several chronic medical conditions treated at the VHA was lower than at non-VHA hospitals.^30,31 One study that measured the quality of nonmetastatic CRC care based on National Comprehensive Cancer Network guidelines showed that > 72% of VHA patients received guideline-concordant care for each diagnostic and therapeutic measure, except for follow-up colonoscopy timing, which appear to be similar or superior to that of the private sector.^30,32,33 Some of the VA initiative for CRC screening may bypass the urban-rurality divide such as the mailed fecal immunochemical test program for CRC. This program was implemented at the onset of the COVID-19 pandemic to avoid disruptions of medical care.³⁴ Rural patients are more likely to undergo fecal immunochemical testing when compared to urban patients in this data. Beyond clinical care, the VHA uses processes to tackle social determinants of health such as housing, food security, and transportation, promoting equal access to health care, and promoting cultural competency among HCPs.^35-37

The results suggest that solutions to CRC disparities between rural and urban areas need to consider known barriers to rural health care, including transportation, diminished rural health care workforce, and other social determinants of health.^9,10,27,38 VHA makes considerable efforts to provide equitable care to all enrolled veterans, including specific programs for rural veterans, including ongoing outreach.³⁹ This study demonstrated lack of disparity in CRC-specific mortality in veterans receiving VHA care, highlighting the importance of these efforts.

Strengths and Limitations

This study used the VHA cohort to compare patient characteristics and mortality between patients with CRC residing in rural and urban areas. The study provides nationwide perspectives on CRC across the geographical spectrum and used a longitudinal cohort with prolonged follow-up to account for comorbidities.

However, the study compared a cohort of rural and urban veterans enrolled in the VHA; hence, the results may not reflect CRC outcomes in veterans without access to VHA care. Rurality has been independently associated with decreased likelihood of meeting CRC screening guidelines among veterans and military service members.³⁸ This study lacked sufficient information to compare CRC staging or treatment modalities among veterans. Although the data cannot identify CRC stage, the proportions of patients with metastatic CRC at diagnosis and CRC location were similar between groups. The study did not have information on their care outside of VHA setting.

This study could not ascertain whether disparities existed in CRC treatment modality since rural residence may result in referral to community-based CRC care, which did not appear in the data. To address these limitations, we used death from any cause as the primary outcome, since death is a hard outcome and is not subject to ascertainment bias. The relatively short follow-up time is another limitation, though subgroup analysis by follow-up did not show significant differences. Despite PS matching, residual unmeasured confounding may exist between urban and rural groups. The predominantly White, male VHA population with high CCI may limit the generalizability of the results.

Conclusions

Rural VHA enrollees had similar survival rates after CRC diagnosis compared to their urban counterparts in a PS-matched analysis. The VHA models of care—including mailed CRC screening tools, several socioeconomic determinants of health (housing, food security, and transportation), and promoting equal access to health care, as well as cultural competency among HCPs—HCPs—may help alleviate disparities across the rural-urban spectrum. The VHA should continue efforts to enroll veterans and provide comprehensive coordinated care in community partnerships.

Colorectal cancer (CRC) is the second-leading cause of cancer-related deaths in the United States, with an estimated 52,550 deaths in 2023.¹ However, the disease burden varies among different segments of the population.² While both CRC incidence and mortality have been decreasing due to screening and advances in treatment, there are disparities in incidence and mortality across the sociodemographic spectrum including race, ethnicity, education, and income.^1-4 While CRC incidence is decreasing for older adults, it is increasing among those aged < 55 years.⁵ The incidence of CRC in adults aged 40 to 54 years has increased by 0.5% to 1.3% annually since the mid-1990s.⁶ The US Preventive Services Task Force now recommends starting CRC screening at age 45 years for asymptomatic adults with average risk.⁷

Disparities also exist across geographical boundaries and living environment. Rural Americans faces additional challenges in health and lifestyle that can affect CRC outcomes. Compared to their urban counterparts, rural residents are more likely to be older, have lower levels of education, higher levels of poverty, lack health insurance, and less access to health care practitioners (HCPs).^8-10 Geographic proximity, defined as travel time or physical distance to a health facility, has been recognized as a predictor of inferior outcomes.¹¹ These aspects of rural living may pose challenges for accessing care for CRC screening and treatment.^11-13 National and local studies have shown disparities in CRC screening rates, incidence, and mortality between rural and urban populations.^14-16

It is unclear whether rural/urban disparities persist under the Veterans Health Administration (VHA) health care delivery model. This study examined differences in baseline characteristics and mortality between rural and urban veterans newly diagnosed with CRC. We also focused on a subpopulation aged ≤ 45 years.

Methods

This study extracted national data from the US Department of Veterans Affairs (VA) Corporate Data Warehouse (CDW) hosted in the VA Informatics and Computing Infrastructure (VINCI) environment. VINCI is an initiative to improve access to VA data and facilitate the analysis of these data while ensuring veterans’ privacy and data security.¹⁷ CDW is the VHA business intelligence information repository, which extracts data from clinical and nonclinical sources following prescribed and validated protocols. Data extracted included demographics, diagnosis, and procedure codes for both inpatient and outpatient encounters, vital signs, and vital status. This study used data previously extracted from a national cohort of veterans that encompassed all patients who received a group of commonly prescribed medications, such as statins, proton pump inhibitors, histamine-2 blockers, acetaminophen-containing products, and hydrocortisone-containing skin applications. This cohort encompassed 8,648,754 veterans, from whom 2,460,727 had encounters during fiscal years (FY) 2016 to 2021 (study period). The cohort was used to ensure that subjects were VHA patients, allowing them to adequately capture their clinical profiles.

Patients were identified as rural or urban based on their residence address at the date of their first diagnosis of CRC. The Geospatial Service Support Center (GSSC) aggregates and updates veterans’ residence address records for all enrolled veterans from the National Change of Address database. The data contain 1 record per enrollee. GSSC Geocoded Enrollee File contains enrollee addresses and their rurality indicators, categorized as urban, rural, or highly rural.¹⁸ Rurality is defined by the Rural Urban Commuting Area (RUCA) categories developed by the Department of Agriculture and the Health Resources and Services Administration of the US Department of Health and Human Services.¹⁹ Urban areas had RUCA codes of 1.0 to 1.1, and highly rural areas had RUCA scores of 10.0. All other areas were classified as rural. Since the proportion of veterans from highly rural areas was small, we included residents from highly rural areas in the rural residents’ group.

Inclusion and Exclusion Criteria

All veterans newly diagnosed with CRC from FY 2016 to 2021 were included. We used the ninth and tenth clinical modification revisions of the International Classification of Diseases (ICD-9-CM and ICD-10-CM) to define CRC diagnosis (Supplemental materials).^4,20 To ensure that patients were newly diagnosed with CRC, this study excluded patients with a previous ICD-9-CM code for CRC diagnosis since FY 2003.

Comorbidities were identified using diagnosis and procedure codes from inpatient and outpatient encounters, which were used to calculate the Charlson Comorbidity Index (CCI) at the time of CRC diagnosis using the weighted method described by Schneeweiss et al.²¹ We defined CRC high-risk conditions and CRC screening tests, including flexible sigmoidoscopy and stool tests, as described in previous studies (Supplemental materials).²⁰

The main outcome was total mortality. The date of death was extracted from the VHA Death Ascertainment File, which contains mortality data from the Master Person Index file in CDW and the Social Security Administration Death Master File. We used the date of death from any cause, as cause of death was not available.

A propensity score (PS) was created to match rural (including highly rural) and urban residents at a ratio of 1:1. Using a standard procedure described in prior publications, multivariable logistic regression used all baseline characteristics to estimate the PS and perform nearest-number matching without replacement.^22,23 A caliper of 0.01 maximized the matched cohort size and achieved balance (Supplemental materials). We then examined the balance of baseline characteristics between PS-matched groups.

Analyses

Cox proportional hazards regression analysis estimated the hazard ratio (HR) of death in rural residents compared to urban residents in the PS-matched cohort. The outcome event was the date of death during the study’s follow-up period (defined as period from first CRC diagnosis to death or study end), with censoring at the study’s end date (September 30, 2021). The proportional hazards assumption was assessed by inspecting the Kaplan-Meier curves. Multiple analyses examined the HR of total mortality in the PS-matched cohort, stratified by sex, race, and ethnicity. We also examined the HR of total mortality stratified by duration of follow-up.

Another PS-matching analysis among veterans aged ≤ 45 years was performed using the same techniques described earlier in this article. We performed a Cox proportional hazards regression analysis to compare mortality in PS-matched urban and rural veterans aged ≤ 45 years. The HR of death in all veterans aged ≤ 45 years (before PS-matching) was estimated using Cox proportional hazard regression analysis, adjusting for PS.

Dichotomous variables were compared using X² tests and continuous variables were compared using t tests. Baseline characteristics with missing values were converted into categorical variables and the proportion of subjects with missing values was equalized between treatment groups after PS-matching. For subgroup analysis, we examined the HR of total mortality in each subgroup using separate Cox proportional hazards regression models similar to the primary analysis but adjusted for PS. Due to multiple comparisons in the subgroup analysis, the findings should be considered exploratory. Statistical tests were 2-tailed, and significance was defined as P < .05. Data management and statistical analyses were conducted from June 2022 to January 2023 using STATA, Version 17. The VA Orlando Healthcare System Institutional Review Board approved the study and waived requirements for informed consent because only deidentified data were used.

Results

After excluding 49 patients (Supplemental materials, available at doi:10.12788/fp.0560), we identified 30,219 veterans with newly diagnosed CRC between FY 2016 to 2021 (Table 1). Of these, 19,422 (64.3%) resided in urban areas and 10,797 (35.7%) resided in rural areas (Table 2). The mean (SD) duration from the first CRC diagnosis to death or study end was 832 (640) days, and the median (IQR) was 723 (246–1330) days. Overall, incident CRC diagnoses were numerically highest in FY 2016 and lowest in FY 2020 (Figure 1). Patients with CRC in rural areas vs urban areas were significantly older (mean, 71.2 years vs 70.8 years, respectively; P < .001), more likely to be male (96.7% vs 95.7%, respectively; P < .001), more likely to be White (83.6% vs 67.8%, respectively; P < .001) and more likely to be non-Hispanic (92.2% vs 87.5%, respectively; P < .001). In terms of general health, rural veterans with CRC were more likely to be overweight or obese (81.5% rural vs 78.5% urban; P < .001) but had fewer mean comorbidities as measured by CCI (5.66 rural vs 5.90 urban; P < .001). A higher proportion of rural veterans with CRC had received stool-based (fecal occult blood test or fecal immunochemical test) CRC screening tests (61.6% rural vs 57.2% urban; P < .001). Fewer rural patients presented with systemic symptoms or signs within 1 year of CRC diagnosis (54.4% rural vs 57.5% urban, P < .001). Among urban patients with CRC, 6959 (35.8%) deaths were observed, compared with 3766 (34.9%) among rural patients (P = .10).

There were 21,568 PS-matched veterans: 10,784 in each group. In the PS-matched cohort, baseline characteristics were similar between veterans in urban and rural communities, including age, sex, race/ethnicity, body mass index, and comorbidities. Among rural patients with CRC, 3763 deaths (34.9%) were observed compared with 3702 (34.3%) among urban veterans. There was no significant difference in the HR of mortality between rural and urban CRC residents (HR, 1.01; 95% CI, 0.97-1.06; P = .53) (Figure 2).

Among veterans aged ≤ 45 years, 551 were diagnosed with CRC (391 urban and 160 rural). We PS-matched 142 pairs of urban and rural veterans without residual differences in baseline characteristics (eAppendix 1). There was no significant difference in the HR of mortality between rural and urban veterans aged ≤ 45 years (HR, 0.97; 95% CI, 0.57-1.63; P = .90) (Figure 2). Similarly, no difference in mortality was observed adjusting for PS between all rural and urban veterans aged ≤ 45 years (HR, 1.03; 95% CI, 0.67-1.59; P = .88).

There was no difference in total mortality between rural and urban veterans in any subgroup except for American Indian or Alaska Native veterans (HR, 2.41; 95% CI, 1.29-4.50; P = .006) (eAppendix 2).

Discussion

This study examined characteristics of patients with CRC between urban and rural areas among veterans who were VHA patients. Similar to other studies, rural veterans with CRC were older, more likely to be White, and were obese, but exhibited fewer comorbidities (lower CCI and lower incidence of congestive heart failure, dementia, hemiplegia, kidney diseases, liver diseases and AIDS, but higher incidence of chronic obstructive lung disease).^8,16 The incidence of CRC in this study population was lowest in FY 2020, which was reported by the Centers for Disease Control and Prevention and is attributed to COVID-19 pandemic disruption of health services.²⁴ The overall mortality in this study was similar to rates reported in other studies from the VA Central Cancer Registry.⁴ In the PS-matched cohort, where baseline characteristics were similar between urban and rural patients with CRC, we found no disparities in CRC-specific mortality between veterans in rural and urban areas. Additionally, when analysis was restricted to veterans aged ≤ 45 years, the results remained consistent.

Subgroup analyses showed no significant difference in mortality between rural and urban areas by sex, race or ethnicity, except rural American Indian or Alaska Native veterans who had double the mortality of their urban counterparts (HR, 2.41; 95% CI, 1.29-4.50; P = .006). This finding is difficult to interpret due to the small number of events and the wide CI. While with a Bonferroni correction the adjusted P value was .08, which is not statistically significant, a previous study found that although CRC incidence was lower overall in American Indian or Alaska Native populations compared to non-Hispanic White populations, CRC incidence was higher among American Indian or Alaska Native individuals in some areas such as Alaska and the Northern Plains.^25,26 Studies have noted that rural American Indian/Alaska Native populations experience greater poverty, less access to broadband internet, and limited access to care, contributing to poorer cancer outcomes and lower survival.²⁷ Thus, the finding of disparity in mortality between rural and urban American Indian or Alaska Native veterans warrants further study.

Other studies have raised concerns that CRC disproportionately affects adults in rural areas with higher mortality rates.^14-16 These disparities arise from sociodemographic factors and modifiable risk factors, including physical activity, dietary patterns, access to cancer screening, and gaps in quality treatment resources.^16,28 These factors operate at multiple levels: from individual, local health system, to community and policy.^2,27 For example, a South Carolina study (1996–2016) found that residents in rural areas were more likely to be diagnosed with advanced CRC, possibly indicating lower rates of CRC screening in rural areas. They also had higher likelihood of death from CRC.¹⁵ However, the study did not include any clinical parameters, such as comorbidities or obesity. A statewide, population-based study in Utah showed that rural men experienced a lower CRC survival in their unadjusted analysis.¹⁶ However, the study was small, with only 3948 urban and 712 rural residents. Additionally, there was no difference in total mortality in the whole cohort (HR, 0.96; 95% CI, 0.86-1.07) or in CRC-specific death (HR, 0.93; 95% CI, 0.81-1.08). A nationwide study also showed that CRC mortality rates were 8% higher in nonmetropolitan or rural areas than in the most urbanized areas containing large metropolitan counties.²⁹ However, this study did not include descriptions of clinical confounders, such as comorbidities, making it difficult to ascertain whether the difference in CRC mortality was due to rurality or differences in baseline risk characteristics.

In this study, the lack of CRC-specific mortality disparities may be attributed to the structures and practices of VHA health care. Recent studies have noted that mortality of several chronic medical conditions treated at the VHA was lower than at non-VHA hospitals.^30,31 One study that measured the quality of nonmetastatic CRC care based on National Comprehensive Cancer Network guidelines showed that > 72% of VHA patients received guideline-concordant care for each diagnostic and therapeutic measure, except for follow-up colonoscopy timing, which appear to be similar or superior to that of the private sector.^30,32,33 Some of the VA initiative for CRC screening may bypass the urban-rurality divide such as the mailed fecal immunochemical test program for CRC. This program was implemented at the onset of the COVID-19 pandemic to avoid disruptions of medical care.³⁴ Rural patients are more likely to undergo fecal immunochemical testing when compared to urban patients in this data. Beyond clinical care, the VHA uses processes to tackle social determinants of health such as housing, food security, and transportation, promoting equal access to health care, and promoting cultural competency among HCPs.^35-37

The results suggest that solutions to CRC disparities between rural and urban areas need to consider known barriers to rural health care, including transportation, diminished rural health care workforce, and other social determinants of health.^9,10,27,38 VHA makes considerable efforts to provide equitable care to all enrolled veterans, including specific programs for rural veterans, including ongoing outreach.³⁹ This study demonstrated lack of disparity in CRC-specific mortality in veterans receiving VHA care, highlighting the importance of these efforts.

Strengths and Limitations

This study used the VHA cohort to compare patient characteristics and mortality between patients with CRC residing in rural and urban areas. The study provides nationwide perspectives on CRC across the geographical spectrum and used a longitudinal cohort with prolonged follow-up to account for comorbidities.

However, the study compared a cohort of rural and urban veterans enrolled in the VHA; hence, the results may not reflect CRC outcomes in veterans without access to VHA care. Rurality has been independently associated with decreased likelihood of meeting CRC screening guidelines among veterans and military service members.³⁸ This study lacked sufficient information to compare CRC staging or treatment modalities among veterans. Although the data cannot identify CRC stage, the proportions of patients with metastatic CRC at diagnosis and CRC location were similar between groups. The study did not have information on their care outside of VHA setting.

This study could not ascertain whether disparities existed in CRC treatment modality since rural residence may result in referral to community-based CRC care, which did not appear in the data. To address these limitations, we used death from any cause as the primary outcome, since death is a hard outcome and is not subject to ascertainment bias. The relatively short follow-up time is another limitation, though subgroup analysis by follow-up did not show significant differences. Despite PS matching, residual unmeasured confounding may exist between urban and rural groups. The predominantly White, male VHA population with high CCI may limit the generalizability of the results.

Conclusions

Rural VHA enrollees had similar survival rates after CRC diagnosis compared to their urban counterparts in a PS-matched analysis. The VHA models of care—including mailed CRC screening tools, several socioeconomic determinants of health (housing, food security, and transportation), and promoting equal access to health care, as well as cultural competency among HCPs—HCPs—may help alleviate disparities across the rural-urban spectrum. The VHA should continue efforts to enroll veterans and provide comprehensive coordinated care in community partnerships.

In the United States, 1 in 4 veterans lives with type 2 diabetes mellitus (T2DM), double the rate of the general population.¹ Medications are important for the treatment of T2DM and preventing complications that may develop if not properly managed. Common classes of medications for diabetes include biguanides, sodiumglucose cotransporter-2 (SGLT-2) inhibitors, glucagon-like peptide-1 (GLP-1) receptor agonists, dipeptidyl peptidase-4 inhibitors, thiazolidinediones, sulfonylureas, and insulin. The selection of treatment depends on patient-specific factors including hemoglobin A_1c (HbA_1c) goal, potential effects on weight, risk of hypoglycemia, and comorbidities such as atherosclerotic cardiovascular disease, heart failure, or chronic kidney disease.²

HbA_1c level reflects the mean blood glucose over the previous 3 months and serves as an indication of diabetes control. In patients with diabetes, it is recommended that HbA_1c is checked ≥ 2 times annually for those meeting treatment goals, or more often if the patient needs to adjust medications to reach their HbA_1c goal. The goal HbA_1c level for most adults with diabetes is < 7%.³ This target can be adjusted based on age, comorbidities, or other patient factors. It is generally recommended that frequent glucose monitoring is not needed for patients with T2DM who are only taking oral agents and/or noninsulin injectables. However, for those on insulin regimens, it is advised to monitor glucose closely, with even more frequent testing for those with an intensive insulin regimen.³

Most patients with diabetes use fingerstick testing to self-monitor their blood glucose. However, continuous glucose monitors (CGMs) are becoming widely available and offer a solution to those who do not have the ability to check their glucose multiple times a day and throughout the night. The American Diabetes Association recommends that the frequency and timing of blood glucose monitoring, or the consideration of CGM use, should be based on the specific needs and goals of each patient.³ Guidelines also encourage those on intensive insulin regimens to check glucose levels when fasting, before and after meals, prior to exercise, and when hypoglycemia or hyperglycemia is suspected. Frequent testing can become a burden for patients, whereas once a CGM sensor is placed, it can be worn for 10 to 14 days. CGMs are also capable of transmitting glucose readings every 1 to 15 minutes to a receiver or mobile phone, allowing for further adaptability to a patient’s lifestyle.³

CGMs work by measuring the interstitial glucose with a small filament sensor and have demonstrated accuracy when compared to blood glucose readings. The ability of a CGM to accurately reflect HbA_1c levels is a potential benefit, reducing the need for frequent testing to determine whether patients have achieved glycemic control.⁴ Another benefit of a CGM is the ease of sharing data; patient accounts can be linked with a health care site, allowing clinicians to access glucose data even if the patient is not able to be seen in clinic. This allows health care practitioners (HCPs) to more efficiently tailor medications and optimize regimens based on patient-specific data that was not available by fingerstick testing alone.

Vigersky and colleagues provided one of the few studies on the long-term effects of CGM in patients managing T2DM through diet and exercise alone, oral medications, or basal insulin and found significant improvement in HbA_1c after only 3 months of CGM use.⁵

An important aspect of CGM use is the ability to alert the patient to low blood glucose readings, which can be dangerous for those unaware of hypoglycemia. Many studies have investigated the association between CGM use and acute metabolic events, demonstrating the potential for CGMs to prevent these emergencies. Karter and colleagues found a reduction in emergency department visits and hospitalizations for hypoglycemia associated with the use of CGMs in patients with type 1 DM (T1DM) and T2DM.⁶

There have been few studies on the use of CGM in veterans. Langford and colleagues found a reduction of HbA_1c among veterans with T2DM using CGMs. However, > 50% of the patients in the study were not receiving insulin therapy, which currently is a US Department of Veterans Affairs (VA) CGM criteria for use.⁷ While current studies provide evidence that supports improvement in HbA_1c levels with the use of CGMs, data are lacking for veterans with T2DM taking insulin. There is also minimal research that indicates which patients should be offered a CGM. The objective of this study was to evaluate glycemic control in veterans with T2DM on insulin using a CGM who were previously monitoring blood glucose with fingerstick testing. Secondary endpoints were explored to identify subgroups that may benefit from a CGM and other potential advantages of CGMs.

Methods

This was a retrospective study of veterans who transitioned from fingerstick testing to CGM for glucose monitoring. Each veteran served as their own control to limit confounding variables when comparing HbA_1c levels. Veterans with an active or suspended CGM order were identified by reviewing outpatient prescription data. All data collection and analysis were done within the Veterans Affairs Sioux Falls Health Care System.

The primary objective of this study was to assess glycemic control from the use of a CGM by evaluating the change in HbA_1c after transitioning to a CGM compared to the change in HbA_1c with standard fingerstick monitoring. Three HbA_1c values were collected for each veteran: before starting CGM, at initiation, and following CGM initiation (Figure 1). CGM start date was the date the CGM prescription order was placed. The pre-CGM HbA_1c level was ≥ 1 year prior to the CGM start date or the HbA_1c closest to 1 year. The start CGM HbA_1c level was within 3 months before or 1 month after the CGM start date. The post-CGM HbA_1c level was the most recent time of data collection and at least 6 months after CGM initiation. The change in HbA_1c from fingerstick glucose monitoring was the difference between the pre-CGM and start CGM values. The change in HbA_1c from use of a CGM was the difference between start CGM and post-CGM values, which were compared to determine HbA_1c reduction from CGM use.

This study also explored secondary outcomes including changes in HbA_1c by prescriber type, differences in HbA_1c reduction based on age, and changes in diabetes medications, including total daily insulin doses. For secondary outcomes, diabetes medication information and the total daily dose of insulin were gathered at the start of CGM use and at the time of data collection. The most recent CGM order prescribed was also collected.

Veterans were included if they were aged ≥ 18 years, had an active order for a CGM, T2DM diagnosis, an insulin prescription, and previously used test strips for glucose monitoring. Patients with T1DM, those who accessed CGMs or care in the community, and patients without HbA_1c values pre-CGM, were excluded.

Statistical Analysis

The primary endpoint of change in HbA_1c level before and after CGM use was compared using a paired t test. A 0.5% change in HbA_1c was considered clinically significant, as suggested in other studies.^8,9 P < .05 was considered statistically significant. Analysis for continuous baseline characteristics, including age and total daily insulin, were reported as mean values. Nominal characteristics including sex, race, diabetes medications, and prescriber type are reported as percentages.

Results

A total of 402 veterans were identified with an active CGM at the time of initial data collection in January 2024 and 175 met inclusion criteria. Sixty patients were excluded due to diabetes managed through a community HCP, 38 had T1DM, and 129 lacked HbA_1c within all specified time periods. The 175 veterans were randomized, and 150 were selected to perform a chart review for data collection. The mean age was 70 years, most were male and identified as White (Table 1). The majority of patients were managed by endocrinology (53.3%), followed by primary care (24.0%), and pharmacy (22.7%) (Table 2). The mean baseline HbA_1c was 8.6%.

The difference in HbA_1c before and after use of CGM was -0.97% (P = .0001). Prior to use of a CGM the change in HbA_1c was minimal, with an increase of 0.003% with the use of selfmonitoring glucose. After use of a CGM, HbA_1c decreased by 0.971%. This reduction in HbA_1c would also be considered clinically significant as the change was > 0.5%. The mean pre-, at start, and post-CGM HbA_1c levels were 8.6%, 8.6%, and 7.6%, respectively (Figure 2). Pharmacy prescribers had a 0.7% reduction in HbA_1c post-CGM, the least of all prescribers. While most age groups saw a reduction in HbA_1c, those aged ≥ 80 years had an increase of 0.18% (Table 3). There was an overall mean reduction in insulin of 22 units, which was similar between all prescribers.

Discussion

The primary endpoint of difference in change of HbA_1c before and after CGM use was found to be statistically and clinically significant, with a nearly 1% reduction in HbA_1c, which was similar to the reduction found by Vigersky and colleagues. ⁵ Across all prescribers, post-CGM HbA_1c levels were similar; however, patients with CGM prescribed by pharmacists had the smallest change in HbA_1c. VA pharmacists primarily assess veterans taking insulin who have HbA_1c levels that are below the goal with the aim of decreasing insulin to reduce the risk of hypoglycemia, which could result in increased HbA1c levels. This may also explain the observed increase in post-CGM HbA_1c levels in patients aged ≥ 80 years. Patients under the care of pharmacists also had baseline mean HbA1c levels that were lower than primary care and endocrinology prescribers and were closer to their HbA_1c goal at baseline, which likely was reflected in the smaller reduction in post-CGM HbA_1c level.

While there was a decrease in HbA_1c levels with CGM use, there were also changes to medications during this timeframe that also may have impacted HbA_1c levels. The most common diabetes medications started during CGM use were GLP-1 agonists and SGLT2-inhibitors. Additionally, there was a reduction in the total daily dose of insulin in the study population. These results demonstrate the potential benefits of CGMs for prescribers who take advantage of the CGM glucose data available to assist with medication adjustments. Another consideration for differences in changes of HbA_1c among prescriber types is the opportunity for more frequent follow- up visits with pharmacy or endocrinology compared with primary care. If veterans are followed more closely, it may be associated with improved HbA_1c control. Further research investigating changes in HbA_1c levels based on followup frequency may be useful.

Strengths and Limitations

The crossover design was a strength of this study. This design reduced confounding variables by having veterans serve as their own controls. In addition, the collection of multiple secondary outcomes adds to the knowledge base for future studies. This study focused on a unique population of veterans with T2DM who were taking insulin, an area that previously had very little data available to determine the benefits of CGM use.

Although the use of a CGM showed statistical significance in lowering HbA_1c, many veterans were started on new diabetes medication during the period of CGM use, which also likely contributed to the reduction in HbA_1c and may have confounded the results. The study was limited by its small population size due to time constraints of chart reviews and the limited generalizability of results outside of the VA system. The majority of patients were from a single site, male and identified as White, which may not be reflective of other VA and community health care systems. It was also noted that the time from the initiation of CGM use to the most recent HbA_1c level varied from 6 months to several years. Additionally, veterans managed by community-based HCPs with complex diabetes cases were excluded.

Conclusions

This study demonstrated a clinically and statistically significant reduction in HbA_1c with the use of a CGM compared to fingerstick monitoring in veterans with T2DM who were being treated with insulin. The change in post-CGM HbA_1c levels across prescribers was similar. In the subgroup analysis of change in HbA_1c among age groups, there was a lower HbA_1c reduction in individuals aged ≥ 80 years. The results from this study support the idea that CGM use may be beneficial for patients who require a reduction in HbA_1c by allowing more precise adjustments to medications and optimization of therapy, as well as the potential to reduce insulin requirements, which is especially valuable in the older adult veteran population.

In the United States, 1 in 4 veterans lives with type 2 diabetes mellitus (T2DM), double the rate of the general population.¹ Medications are important for the treatment of T2DM and preventing complications that may develop if not properly managed. Common classes of medications for diabetes include biguanides, sodiumglucose cotransporter-2 (SGLT-2) inhibitors, glucagon-like peptide-1 (GLP-1) receptor agonists, dipeptidyl peptidase-4 inhibitors, thiazolidinediones, sulfonylureas, and insulin. The selection of treatment depends on patient-specific factors including hemoglobin A_1c (HbA_1c) goal, potential effects on weight, risk of hypoglycemia, and comorbidities such as atherosclerotic cardiovascular disease, heart failure, or chronic kidney disease.²

HbA_1c level reflects the mean blood glucose over the previous 3 months and serves as an indication of diabetes control. In patients with diabetes, it is recommended that HbA_1c is checked ≥ 2 times annually for those meeting treatment goals, or more often if the patient needs to adjust medications to reach their HbA_1c goal. The goal HbA_1c level for most adults with diabetes is < 7%.³ This target can be adjusted based on age, comorbidities, or other patient factors. It is generally recommended that frequent glucose monitoring is not needed for patients with T2DM who are only taking oral agents and/or noninsulin injectables. However, for those on insulin regimens, it is advised to monitor glucose closely, with even more frequent testing for those with an intensive insulin regimen.³

Most patients with diabetes use fingerstick testing to self-monitor their blood glucose. However, continuous glucose monitors (CGMs) are becoming widely available and offer a solution to those who do not have the ability to check their glucose multiple times a day and throughout the night. The American Diabetes Association recommends that the frequency and timing of blood glucose monitoring, or the consideration of CGM use, should be based on the specific needs and goals of each patient.³ Guidelines also encourage those on intensive insulin regimens to check glucose levels when fasting, before and after meals, prior to exercise, and when hypoglycemia or hyperglycemia is suspected. Frequent testing can become a burden for patients, whereas once a CGM sensor is placed, it can be worn for 10 to 14 days. CGMs are also capable of transmitting glucose readings every 1 to 15 minutes to a receiver or mobile phone, allowing for further adaptability to a patient’s lifestyle.³

CGMs work by measuring the interstitial glucose with a small filament sensor and have demonstrated accuracy when compared to blood glucose readings. The ability of a CGM to accurately reflect HbA_1c levels is a potential benefit, reducing the need for frequent testing to determine whether patients have achieved glycemic control.⁴ Another benefit of a CGM is the ease of sharing data; patient accounts can be linked with a health care site, allowing clinicians to access glucose data even if the patient is not able to be seen in clinic. This allows health care practitioners (HCPs) to more efficiently tailor medications and optimize regimens based on patient-specific data that was not available by fingerstick testing alone.

Vigersky and colleagues provided one of the few studies on the long-term effects of CGM in patients managing T2DM through diet and exercise alone, oral medications, or basal insulin and found significant improvement in HbA_1c after only 3 months of CGM use.⁵

An important aspect of CGM use is the ability to alert the patient to low blood glucose readings, which can be dangerous for those unaware of hypoglycemia. Many studies have investigated the association between CGM use and acute metabolic events, demonstrating the potential for CGMs to prevent these emergencies. Karter and colleagues found a reduction in emergency department visits and hospitalizations for hypoglycemia associated with the use of CGMs in patients with type 1 DM (T1DM) and T2DM.⁶

There have been few studies on the use of CGM in veterans. Langford and colleagues found a reduction of HbA_1c among veterans with T2DM using CGMs. However, > 50% of the patients in the study were not receiving insulin therapy, which currently is a US Department of Veterans Affairs (VA) CGM criteria for use.⁷ While current studies provide evidence that supports improvement in HbA_1c levels with the use of CGMs, data are lacking for veterans with T2DM taking insulin. There is also minimal research that indicates which patients should be offered a CGM. The objective of this study was to evaluate glycemic control in veterans with T2DM on insulin using a CGM who were previously monitoring blood glucose with fingerstick testing. Secondary endpoints were explored to identify subgroups that may benefit from a CGM and other potential advantages of CGMs.

Methods

This was a retrospective study of veterans who transitioned from fingerstick testing to CGM for glucose monitoring. Each veteran served as their own control to limit confounding variables when comparing HbA_1c levels. Veterans with an active or suspended CGM order were identified by reviewing outpatient prescription data. All data collection and analysis were done within the Veterans Affairs Sioux Falls Health Care System.

The primary objective of this study was to assess glycemic control from the use of a CGM by evaluating the change in HbA_1c after transitioning to a CGM compared to the change in HbA_1c with standard fingerstick monitoring. Three HbA_1c values were collected for each veteran: before starting CGM, at initiation, and following CGM initiation (Figure 1). CGM start date was the date the CGM prescription order was placed. The pre-CGM HbA_1c level was ≥ 1 year prior to the CGM start date or the HbA_1c closest to 1 year. The start CGM HbA_1c level was within 3 months before or 1 month after the CGM start date. The post-CGM HbA_1c level was the most recent time of data collection and at least 6 months after CGM initiation. The change in HbA_1c from fingerstick glucose monitoring was the difference between the pre-CGM and start CGM values. The change in HbA_1c from use of a CGM was the difference between start CGM and post-CGM values, which were compared to determine HbA_1c reduction from CGM use.

This study also explored secondary outcomes including changes in HbA_1c by prescriber type, differences in HbA_1c reduction based on age, and changes in diabetes medications, including total daily insulin doses. For secondary outcomes, diabetes medication information and the total daily dose of insulin were gathered at the start of CGM use and at the time of data collection. The most recent CGM order prescribed was also collected.

Veterans were included if they were aged ≥ 18 years, had an active order for a CGM, T2DM diagnosis, an insulin prescription, and previously used test strips for glucose monitoring. Patients with T1DM, those who accessed CGMs or care in the community, and patients without HbA_1c values pre-CGM, were excluded.

Statistical Analysis

The primary endpoint of change in HbA_1c level before and after CGM use was compared using a paired t test. A 0.5% change in HbA_1c was considered clinically significant, as suggested in other studies.^8,9 P < .05 was considered statistically significant. Analysis for continuous baseline characteristics, including age and total daily insulin, were reported as mean values. Nominal characteristics including sex, race, diabetes medications, and prescriber type are reported as percentages.

Results

A total of 402 veterans were identified with an active CGM at the time of initial data collection in January 2024 and 175 met inclusion criteria. Sixty patients were excluded due to diabetes managed through a community HCP, 38 had T1DM, and 129 lacked HbA_1c within all specified time periods. The 175 veterans were randomized, and 150 were selected to perform a chart review for data collection. The mean age was 70 years, most were male and identified as White (Table 1). The majority of patients were managed by endocrinology (53.3%), followed by primary care (24.0%), and pharmacy (22.7%) (Table 2). The mean baseline HbA_1c was 8.6%.

The difference in HbA_1c before and after use of CGM was -0.97% (P = .0001). Prior to use of a CGM the change in HbA_1c was minimal, with an increase of 0.003% with the use of selfmonitoring glucose. After use of a CGM, HbA_1c decreased by 0.971%. This reduction in HbA_1c would also be considered clinically significant as the change was > 0.5%. The mean pre-, at start, and post-CGM HbA_1c levels were 8.6%, 8.6%, and 7.6%, respectively (Figure 2). Pharmacy prescribers had a 0.7% reduction in HbA_1c post-CGM, the least of all prescribers. While most age groups saw a reduction in HbA_1c, those aged ≥ 80 years had an increase of 0.18% (Table 3). There was an overall mean reduction in insulin of 22 units, which was similar between all prescribers.

Discussion

The primary endpoint of difference in change of HbA_1c before and after CGM use was found to be statistically and clinically significant, with a nearly 1% reduction in HbA_1c, which was similar to the reduction found by Vigersky and colleagues. ⁵ Across all prescribers, post-CGM HbA_1c levels were similar; however, patients with CGM prescribed by pharmacists had the smallest change in HbA_1c. VA pharmacists primarily assess veterans taking insulin who have HbA_1c levels that are below the goal with the aim of decreasing insulin to reduce the risk of hypoglycemia, which could result in increased HbA1c levels. This may also explain the observed increase in post-CGM HbA_1c levels in patients aged ≥ 80 years. Patients under the care of pharmacists also had baseline mean HbA1c levels that were lower than primary care and endocrinology prescribers and were closer to their HbA_1c goal at baseline, which likely was reflected in the smaller reduction in post-CGM HbA_1c level.

While there was a decrease in HbA_1c levels with CGM use, there were also changes to medications during this timeframe that also may have impacted HbA_1c levels. The most common diabetes medications started during CGM use were GLP-1 agonists and SGLT2-inhibitors. Additionally, there was a reduction in the total daily dose of insulin in the study population. These results demonstrate the potential benefits of CGMs for prescribers who take advantage of the CGM glucose data available to assist with medication adjustments. Another consideration for differences in changes of HbA_1c among prescriber types is the opportunity for more frequent follow- up visits with pharmacy or endocrinology compared with primary care. If veterans are followed more closely, it may be associated with improved HbA_1c control. Further research investigating changes in HbA_1c levels based on followup frequency may be useful.

Strengths and Limitations

The crossover design was a strength of this study. This design reduced confounding variables by having veterans serve as their own controls. In addition, the collection of multiple secondary outcomes adds to the knowledge base for future studies. This study focused on a unique population of veterans with T2DM who were taking insulin, an area that previously had very little data available to determine the benefits of CGM use.

Although the use of a CGM showed statistical significance in lowering HbA_1c, many veterans were started on new diabetes medication during the period of CGM use, which also likely contributed to the reduction in HbA_1c and may have confounded the results. The study was limited by its small population size due to time constraints of chart reviews and the limited generalizability of results outside of the VA system. The majority of patients were from a single site, male and identified as White, which may not be reflective of other VA and community health care systems. It was also noted that the time from the initiation of CGM use to the most recent HbA_1c level varied from 6 months to several years. Additionally, veterans managed by community-based HCPs with complex diabetes cases were excluded.

Conclusions

This study demonstrated a clinically and statistically significant reduction in HbA_1c with the use of a CGM compared to fingerstick monitoring in veterans with T2DM who were being treated with insulin. The change in post-CGM HbA_1c levels across prescribers was similar. In the subgroup analysis of change in HbA_1c among age groups, there was a lower HbA_1c reduction in individuals aged ≥ 80 years. The results from this study support the idea that CGM use may be beneficial for patients who require a reduction in HbA_1c by allowing more precise adjustments to medications and optimization of therapy, as well as the potential to reduce insulin requirements, which is especially valuable in the older adult veteran population.

In the United States, 1 in 4 veterans lives with type 2 diabetes mellitus (T2DM), double the rate of the general population.¹ Medications are important for the treatment of T2DM and preventing complications that may develop if not properly managed. Common classes of medications for diabetes include biguanides, sodiumglucose cotransporter-2 (SGLT-2) inhibitors, glucagon-like peptide-1 (GLP-1) receptor agonists, dipeptidyl peptidase-4 inhibitors, thiazolidinediones, sulfonylureas, and insulin. The selection of treatment depends on patient-specific factors including hemoglobin A_1c (HbA_1c) goal, potential effects on weight, risk of hypoglycemia, and comorbidities such as atherosclerotic cardiovascular disease, heart failure, or chronic kidney disease.²

HbA_1c level reflects the mean blood glucose over the previous 3 months and serves as an indication of diabetes control. In patients with diabetes, it is recommended that HbA_1c is checked ≥ 2 times annually for those meeting treatment goals, or more often if the patient needs to adjust medications to reach their HbA_1c goal. The goal HbA_1c level for most adults with diabetes is < 7%.³ This target can be adjusted based on age, comorbidities, or other patient factors. It is generally recommended that frequent glucose monitoring is not needed for patients with T2DM who are only taking oral agents and/or noninsulin injectables. However, for those on insulin regimens, it is advised to monitor glucose closely, with even more frequent testing for those with an intensive insulin regimen.³

Most patients with diabetes use fingerstick testing to self-monitor their blood glucose. However, continuous glucose monitors (CGMs) are becoming widely available and offer a solution to those who do not have the ability to check their glucose multiple times a day and throughout the night. The American Diabetes Association recommends that the frequency and timing of blood glucose monitoring, or the consideration of CGM use, should be based on the specific needs and goals of each patient.³ Guidelines also encourage those on intensive insulin regimens to check glucose levels when fasting, before and after meals, prior to exercise, and when hypoglycemia or hyperglycemia is suspected. Frequent testing can become a burden for patients, whereas once a CGM sensor is placed, it can be worn for 10 to 14 days. CGMs are also capable of transmitting glucose readings every 1 to 15 minutes to a receiver or mobile phone, allowing for further adaptability to a patient’s lifestyle.³

CGMs work by measuring the interstitial glucose with a small filament sensor and have demonstrated accuracy when compared to blood glucose readings. The ability of a CGM to accurately reflect HbA_1c levels is a potential benefit, reducing the need for frequent testing to determine whether patients have achieved glycemic control.⁴ Another benefit of a CGM is the ease of sharing data; patient accounts can be linked with a health care site, allowing clinicians to access glucose data even if the patient is not able to be seen in clinic. This allows health care practitioners (HCPs) to more efficiently tailor medications and optimize regimens based on patient-specific data that was not available by fingerstick testing alone.

Vigersky and colleagues provided one of the few studies on the long-term effects of CGM in patients managing T2DM through diet and exercise alone, oral medications, or basal insulin and found significant improvement in HbA_1c after only 3 months of CGM use.⁵

An important aspect of CGM use is the ability to alert the patient to low blood glucose readings, which can be dangerous for those unaware of hypoglycemia. Many studies have investigated the association between CGM use and acute metabolic events, demonstrating the potential for CGMs to prevent these emergencies. Karter and colleagues found a reduction in emergency department visits and hospitalizations for hypoglycemia associated with the use of CGMs in patients with type 1 DM (T1DM) and T2DM.⁶

There have been few studies on the use of CGM in veterans. Langford and colleagues found a reduction of HbA_1c among veterans with T2DM using CGMs. However, > 50% of the patients in the study were not receiving insulin therapy, which currently is a US Department of Veterans Affairs (VA) CGM criteria for use.⁷ While current studies provide evidence that supports improvement in HbA_1c levels with the use of CGMs, data are lacking for veterans with T2DM taking insulin. There is also minimal research that indicates which patients should be offered a CGM. The objective of this study was to evaluate glycemic control in veterans with T2DM on insulin using a CGM who were previously monitoring blood glucose with fingerstick testing. Secondary endpoints were explored to identify subgroups that may benefit from a CGM and other potential advantages of CGMs.

Methods

This was a retrospective study of veterans who transitioned from fingerstick testing to CGM for glucose monitoring. Each veteran served as their own control to limit confounding variables when comparing HbA_1c levels. Veterans with an active or suspended CGM order were identified by reviewing outpatient prescription data. All data collection and analysis were done within the Veterans Affairs Sioux Falls Health Care System.

The primary objective of this study was to assess glycemic control from the use of a CGM by evaluating the change in HbA_1c after transitioning to a CGM compared to the change in HbA_1c with standard fingerstick monitoring. Three HbA_1c values were collected for each veteran: before starting CGM, at initiation, and following CGM initiation (Figure 1). CGM start date was the date the CGM prescription order was placed. The pre-CGM HbA_1c level was ≥ 1 year prior to the CGM start date or the HbA_1c closest to 1 year. The start CGM HbA_1c level was within 3 months before or 1 month after the CGM start date. The post-CGM HbA_1c level was the most recent time of data collection and at least 6 months after CGM initiation. The change in HbA_1c from fingerstick glucose monitoring was the difference between the pre-CGM and start CGM values. The change in HbA_1c from use of a CGM was the difference between start CGM and post-CGM values, which were compared to determine HbA_1c reduction from CGM use.

This study also explored secondary outcomes including changes in HbA_1c by prescriber type, differences in HbA_1c reduction based on age, and changes in diabetes medications, including total daily insulin doses. For secondary outcomes, diabetes medication information and the total daily dose of insulin were gathered at the start of CGM use and at the time of data collection. The most recent CGM order prescribed was also collected.

Veterans were included if they were aged ≥ 18 years, had an active order for a CGM, T2DM diagnosis, an insulin prescription, and previously used test strips for glucose monitoring. Patients with T1DM, those who accessed CGMs or care in the community, and patients without HbA_1c values pre-CGM, were excluded.

Statistical Analysis

The primary endpoint of change in HbA_1c level before and after CGM use was compared using a paired t test. A 0.5% change in HbA_1c was considered clinically significant, as suggested in other studies.^8,9 P < .05 was considered statistically significant. Analysis for continuous baseline characteristics, including age and total daily insulin, were reported as mean values. Nominal characteristics including sex, race, diabetes medications, and prescriber type are reported as percentages.

Results

A total of 402 veterans were identified with an active CGM at the time of initial data collection in January 2024 and 175 met inclusion criteria. Sixty patients were excluded due to diabetes managed through a community HCP, 38 had T1DM, and 129 lacked HbA_1c within all specified time periods. The 175 veterans were randomized, and 150 were selected to perform a chart review for data collection. The mean age was 70 years, most were male and identified as White (Table 1). The majority of patients were managed by endocrinology (53.3%), followed by primary care (24.0%), and pharmacy (22.7%) (Table 2). The mean baseline HbA_1c was 8.6%.

The difference in HbA_1c before and after use of CGM was -0.97% (P = .0001). Prior to use of a CGM the change in HbA_1c was minimal, with an increase of 0.003% with the use of selfmonitoring glucose. After use of a CGM, HbA_1c decreased by 0.971%. This reduction in HbA_1c would also be considered clinically significant as the change was > 0.5%. The mean pre-, at start, and post-CGM HbA_1c levels were 8.6%, 8.6%, and 7.6%, respectively (Figure 2). Pharmacy prescribers had a 0.7% reduction in HbA_1c post-CGM, the least of all prescribers. While most age groups saw a reduction in HbA_1c, those aged ≥ 80 years had an increase of 0.18% (Table 3). There was an overall mean reduction in insulin of 22 units, which was similar between all prescribers.

Discussion

The primary endpoint of difference in change of HbA_1c before and after CGM use was found to be statistically and clinically significant, with a nearly 1% reduction in HbA_1c, which was similar to the reduction found by Vigersky and colleagues. ⁵ Across all prescribers, post-CGM HbA_1c levels were similar; however, patients with CGM prescribed by pharmacists had the smallest change in HbA_1c. VA pharmacists primarily assess veterans taking insulin who have HbA_1c levels that are below the goal with the aim of decreasing insulin to reduce the risk of hypoglycemia, which could result in increased HbA1c levels. This may also explain the observed increase in post-CGM HbA_1c levels in patients aged ≥ 80 years. Patients under the care of pharmacists also had baseline mean HbA1c levels that were lower than primary care and endocrinology prescribers and were closer to their HbA_1c goal at baseline, which likely was reflected in the smaller reduction in post-CGM HbA_1c level.

While there was a decrease in HbA_1c levels with CGM use, there were also changes to medications during this timeframe that also may have impacted HbA_1c levels. The most common diabetes medications started during CGM use were GLP-1 agonists and SGLT2-inhibitors. Additionally, there was a reduction in the total daily dose of insulin in the study population. These results demonstrate the potential benefits of CGMs for prescribers who take advantage of the CGM glucose data available to assist with medication adjustments. Another consideration for differences in changes of HbA_1c among prescriber types is the opportunity for more frequent follow- up visits with pharmacy or endocrinology compared with primary care. If veterans are followed more closely, it may be associated with improved HbA_1c control. Further research investigating changes in HbA_1c levels based on followup frequency may be useful.

Strengths and Limitations

The crossover design was a strength of this study. This design reduced confounding variables by having veterans serve as their own controls. In addition, the collection of multiple secondary outcomes adds to the knowledge base for future studies. This study focused on a unique population of veterans with T2DM who were taking insulin, an area that previously had very little data available to determine the benefits of CGM use.

Although the use of a CGM showed statistical significance in lowering HbA_1c, many veterans were started on new diabetes medication during the period of CGM use, which also likely contributed to the reduction in HbA_1c and may have confounded the results. The study was limited by its small population size due to time constraints of chart reviews and the limited generalizability of results outside of the VA system. The majority of patients were from a single site, male and identified as White, which may not be reflective of other VA and community health care systems. It was also noted that the time from the initiation of CGM use to the most recent HbA_1c level varied from 6 months to several years. Additionally, veterans managed by community-based HCPs with complex diabetes cases were excluded.

Conclusions

This study demonstrated a clinically and statistically significant reduction in HbA_1c with the use of a CGM compared to fingerstick monitoring in veterans with T2DM who were being treated with insulin. The change in post-CGM HbA_1c levels across prescribers was similar. In the subgroup analysis of change in HbA_1c among age groups, there was a lower HbA_1c reduction in individuals aged ≥ 80 years. The results from this study support the idea that CGM use may be beneficial for patients who require a reduction in HbA_1c by allowing more precise adjustments to medications and optimization of therapy, as well as the potential to reduce insulin requirements, which is especially valuable in the older adult veteran population.

Background

Cancer clinical trials (CCTs) are central to improving cancer care. However, generalizability of findings from CCTs is difficult due to the lack of diversity in most United States CCTs. Clinical trial accrual of underrepresented groups, is low throughout the United States and is approximately 4-5% in most CCTs. Reasons for low accrual in this population are multifactorial. Despite numerous factors related to accruing racial and ethnic underrepresented groups, many institutions have sought to address these barriers. We conducted a scoping review to identify evidence-based approaches to increase participation in cancer treatment clinical trials.

Methods

We reviewed the Salisbury VA Medical Center Oncology clinical trial database from October 2019 to June 2024. The participants in these clinical trials required consent. These clinical trials included treatment interventional as well as non-treatment interventional. Fifteen studies were included and over 260 Veterans participated.

Results

Key themes emerged that included a focus on patient education, cultural competency, and building capacity in the clinics to care for the Veteran population at three separate sites in the Salisbury VA system. The Black Veteran accrual rate of 29% was achieved. This accrual rate is representative of our VA catchment population of 33% for Black Veterans, and is five times the national average.

Conclusions

The research team’s success in enrolling Black Veterans in clinical trials is attributed to several factors. The demographic composition of Veterans served by the Salisbury, Charlotte, and Kernersville VA provided a diverse population that included a 33% Black group. The type of clinical trials focused on patients who were most impacted by the disease. The VA did afford less barriers to access to health care.

Background

Cancer clinical trials (CCTs) are central to improving cancer care. However, generalizability of findings from CCTs is difficult due to the lack of diversity in most United States CCTs. Clinical trial accrual of underrepresented groups, is low throughout the United States and is approximately 4-5% in most CCTs. Reasons for low accrual in this population are multifactorial. Despite numerous factors related to accruing racial and ethnic underrepresented groups, many institutions have sought to address these barriers. We conducted a scoping review to identify evidence-based approaches to increase participation in cancer treatment clinical trials.

Methods

We reviewed the Salisbury VA Medical Center Oncology clinical trial database from October 2019 to June 2024. The participants in these clinical trials required consent. These clinical trials included treatment interventional as well as non-treatment interventional. Fifteen studies were included and over 260 Veterans participated.

Results

Key themes emerged that included a focus on patient education, cultural competency, and building capacity in the clinics to care for the Veteran population at three separate sites in the Salisbury VA system. The Black Veteran accrual rate of 29% was achieved. This accrual rate is representative of our VA catchment population of 33% for Black Veterans, and is five times the national average.

Conclusions

The research team’s success in enrolling Black Veterans in clinical trials is attributed to several factors. The demographic composition of Veterans served by the Salisbury, Charlotte, and Kernersville VA provided a diverse population that included a 33% Black group. The type of clinical trials focused on patients who were most impacted by the disease. The VA did afford less barriers to access to health care.

Background

Cancer clinical trials (CCTs) are central to improving cancer care. However, generalizability of findings from CCTs is difficult due to the lack of diversity in most United States CCTs. Clinical trial accrual of underrepresented groups, is low throughout the United States and is approximately 4-5% in most CCTs. Reasons for low accrual in this population are multifactorial. Despite numerous factors related to accruing racial and ethnic underrepresented groups, many institutions have sought to address these barriers. We conducted a scoping review to identify evidence-based approaches to increase participation in cancer treatment clinical trials.

Methods

We reviewed the Salisbury VA Medical Center Oncology clinical trial database from October 2019 to June 2024. The participants in these clinical trials required consent. These clinical trials included treatment interventional as well as non-treatment interventional. Fifteen studies were included and over 260 Veterans participated.

Results

Key themes emerged that included a focus on patient education, cultural competency, and building capacity in the clinics to care for the Veteran population at three separate sites in the Salisbury VA system. The Black Veteran accrual rate of 29% was achieved. This accrual rate is representative of our VA catchment population of 33% for Black Veterans, and is five times the national average.

Conclusions

The research team’s success in enrolling Black Veterans in clinical trials is attributed to several factors. The demographic composition of Veterans served by the Salisbury, Charlotte, and Kernersville VA provided a diverse population that included a 33% Black group. The type of clinical trials focused on patients who were most impacted by the disease. The VA did afford less barriers to access to health care.