Federal Practitioner

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.

Clinical awareness of cancers associated with Camp Lejeune water contamination exposure remains limited despite legal and policy advances. Gaps persist in early symptom recognition and timely diagnostic evaluation before a definitive cancer diagnosis among exposed personnel. This may represent missed opportunities for earlier identification of volatile organic compounds (VOCs)-related cancers and for less invasive treatment options for veterans in this high-risk population.

Federal health care practitioners (HCPs), especially those in primary care and internal medicine, are uniquely positioned to bridge this gap. By improving the recognition of symptoms, pertinent physical examination findings, and implementing a diagnostic screening panel, HCPs can support accurate diagnoses and facilitate earlier treatment to improve health and quality of life for this population.

From 1953 to 1985, as many as 1 million military personnel, civilian workers, and their families stationed at US Marine Corps Base Camp Lejeune were unknowingly exposed to toxic and carcinogenic chemicals in drinking and bathing water.¹ Three of the 8 main water sources on base were contaminated with VOCs, which are associated with multiple cancers.^1-3

The US Department of Veterans Affairs (VA) recognizes 15 conditions associated with Camp Lejeune contaminated water exposure for VA benefits, including 10 cancers: adult leukemia; aplastic anemia and other myelodysplastic syndromes (MDS); bladder, esophageal, kidney, liver, breast (male and female), and lung cancers; multiple myeloma; and non-Hodgkin lymphoma (NHL).⁴

BACKGROUND

Established in 1942, Camp Lejeune is an important Marine Corps training installation. Between 1953 and 1985, multiple on-base water systems were contaminated with VOCs, including trichloroethylene (TCE), perchloroethylene (PCE), benzene, and vinyl chloride, due to improper waste disposal and industrial runoff from on- and off-base sources.⁵ Tarawa Terrace water treatment plant (WTP) was contaminated primarily with PCE from November 1957 to February 1987. Hadnot Point WTP was contaminated with TCE from August 1953 to December 1984, along with PCE, and benzene, toluene, ethylbenzene, and xylene (BTEX). Holcomb Boulevard WTP, established in 1972, was contaminated with TCE from June 1972 to February 1985.² These contaminants entered the drinking and bathing water supply over decades, and exposure often occurred concurrently across = 1 VOC, compounding health risks.^2,3 This prolonged 32-year VOC exposure window underlies current concerns regarding long-term cancer risk among affected service members, civilian employees, and family members. Epidemiologic research has found statistically significant associations between VOC exposure and multiple cancers, neurologic conditions, and reproductive issues.⁶ Specifically, TCE is associated with higher risks of hematologic cancers, multiple myeloma, NHL, and kidney cancer.³ PCE is linked with kidney cancer, benzene with multiple myeloma and NHL, and vinyl chloride with hepatobiliary cancers.³ A cohort mortality study compared Camp Lejeune personnel with a control group at Camp Pendleton from 1972 to 1985 and found a 3-fold higher incidence or mortality rate for kidney, esophageal, and female breast cancers, leukemia, and lymphoma among exposed Camp Lejeune personnel.⁶ Notably, personnel assigned to Camp Lejeune for as little as 6 months faced up to a 6-fold increase in cancer risk; the average military assignment between 1975 and 1985 was 18 months.^3,6

Honoring America's Veterans and Caring for Camp Lejeune Families Act of 2012, the Sergeant First Class Heath Robinson Honoring Our Promise to Address Comprehensive Toxics (PACT) Act of 2022, the Camp Lejeune Justice Act of 2022, and the pending Ensuring Justice for Camp Lejeune Victims Act of 2025 provide health care and legal resources for personnel and families affected by Camp Lejeune’s contaminated water.^6-8 These laws acknowledge associations between exposure and specific health conditions and expanded health care, benefits, and legal recourse for affected veterans, survivors, and their families.^8,9

CANCERS LINKED TO CAMP LEJEUNE

Camp Lejeune VOC-contaminated water exposure is associated with solid tumor and hematologic cancers. Symptoms, physical examination findings, and diagnostic considerations vary by cancer type (Table 1).

Bladder Cancer

The US incidence rate of bladder cancer for both males and females is 18 per 100,000 individuals per year, with a death rate of 4.1 per 100,000 individuals per year, and a 2.1% lifetime diagnosis risk.¹⁰ Personnel exposed to VOCs at Camp Lejeune had a 9% higher risk of developing bladder cancer and a 2% increased mortality compared with an unexposed control group at Camp Pendleton.^1,7 Other bladder cancer subtypes at increased risk are papillary transitional cell carcinoma, nonpapillary transition cell carcinoma, and urothelial carcinoma.⁷ This is consistent with prior research that found PCE exposure is associated with an increased risk for bladder cancer.^3,7,11 Smoking and tobacco use remain significant risk factors for bladder cancer.¹²

Symptomatology. The most common symptom associated with bladder cancer is painless hematuria (gross or microscopic). Other often delayed symptoms include urinary frequency, urgency, or nocturia.^13,14

Diagnostics. Screening tests include urinalysis for hematuria, urine cytology, and cystoscopy with biopsy as the gold standard for diagnosis and staging.^15,16

Kidney Cancer

The US incidence rate of kidney cancer and renal pelvis cancer for both males and females is 17.5 per 100,000 individuals per year, with a death rate of 3.4 per 100,000, and a 1.8% lifetime diagnosis risk.¹⁷ Camp Lejeune personnel exposed to VOCs had a 6% increased risk of developing kidney cancer and renal pelvis cancer and a 21% higher mortality risk compared with Camp Pendleton controls.^1,7 Subtypes at risk include renal cell carcinoma and papillary carcinoma.⁷ This is consistent with prior research that found exposures to TCE and PCE are associated with a 3-fold increased risk of kidney cancer.^3,7

Symptomatology. Hematuria, flank pain, and a palpable abdominal mass are common symptoms associated with kidney cancer. In advanced stages, other symptoms may include left-sided varicocele, anemia, weight loss, fatigue, fever, and night sweats.¹⁸

Diagnostics. Screening tests include urinalysis to assess the presence of blood, complete blood count (CBC) to assess anemia, calcium (elevated), and lactate dehydrogenase (LDH), which may be elevated. Imaging strategies include abdominal computed tomography (CT), magnetic resonance imaging (MRI), or ultrasound.¹⁹

Esophageal Cancer

The US incidence rate of esophageal cancer for both males and females is 4.2 per 100,000 individuals per year, the death rate is 3.7 per 100,000 individuals per year, and a 0.5% lifetime diagnosis risk.²⁰ VOC-exposed Camp Lejeune personnel had a 27% increased incidence and 25% increased mortality compared with the control group.^1,7 Esophageal cancer subtypes at elevated risk include squamous cell carcinoma and adenocarcinoma. This is consistent with prior research that found Camp Lejeune water exposure is associated with a 3-fold increased risk for esophageal cancer.⁷ Additional risk factors include history of smoking and alcohol use.²¹

Symptomatology. Esophageal cancer is often asymptomatic with potential symptoms that include dysphagia, hoarseness, and weight loss in advanced disease.²²

Diagnostics. Endoscopy with biopsy is the definitive method for diagnosis.²³

Liver Cancer

The US incidence rate of liver cancer and intrahepatic bile duct cancer for both males and females is 9.4 per 100,000 individuals per year, with a death rate of 6.6 per 100,000 individuals per year, and a 1.1% lifetime diagnosis risk.²⁴ VOC-exposed personnel had a 1% higher mortality than controls.¹

Symptomatology. Liver cancer is often asymptomatic and appears in late stages.²⁵ Common symptoms include right upper quadrant pain, early satiety, nausea, vomiting, loss of appetite, weight loss, ascites, jaundice, and abnormal bleeding or bruising.^25,26

Diagnostics. Diagnostic tests may include an ultrasound, CT, or MRI. Additional laboratory testing may include liver function, a-fetoprotein blood, CBC, renal function, calcium, and hepatitis panel screening for hepatitis B and C.^27,28

Lung Cancer

The US incidence rate of lung cancer for both males and females is 47.8 per 100,000 individuals per year, with a death rate of 31.5 per 100,000 individuals per year, and a 5.4% lifetime diagnosis risk.²⁹ VOC-exposed personnel had a 16% increased risk and 19% higher mortality.^1,7 Subtypes include large cell, small cell, non-small cell, squamous cell, and adenocarcinoma.⁷ Smoking is an additional risk factor.³⁰

Symptomatology. Symptoms of lung cancer include cough, shortness of breath, chest pain worse with deep breathing, unexplained weight loss, fatigue, night sweats, and recurrent fevers. Advanced stages may metastasize or spread to the liver, bones, and brain.³¹

Diagnostics. Low-dose CT and chest X-ray are used for screening.³²

Breast Cancer

The US incidence rate of female breast cancer is 130.8 per 100,000 individuals per year, with a death rate of 19.2 per 100,000 individuals per year, and a 13.0% lifetime risk of diagnosis.³³ For female VOC-exposed personnel, there was an equal risk of developing breast cancer as the control group.¹ However, exposed females at Camp Lejeune had a 23% higher mortality risk compared to the control group.⁷ Breast cancer subtypes among females include ductal carcinoma, lobular carcinoma, and ductal-lobular carcinoma.¹

The US incidence rate of male breast cancer is 1.3 per 100,000 individuals per year, with a death rate of 0.3 per 100,000 individuals per year.^34,35 The lifetime risk for males developing breast cancer is 137.7 per 100,000 and about 70 to 100 times less common in men than women.³⁶

Male personnel exposed at Camp Lejeune had a 4% increased risk for developing breast cancer compared to Camp Pendleton.⁷ However, mortality was lower in the Camp Lejeune group.¹ Although male breast cancer is rare, males at Camp Lejeune had a higher incidence, indicating a link between TCE, PCE, vinyl chloride exposures and male breast cancer.³⁷ Male breast cancer is more often diagnosed in advanced stages than female breast cancer due to the lack of awareness or absence of routine screenings.³⁸ The most common breast cancer type in males is invasive ductal carcinoma, accounting for 85% to 90% of cases; lobular carcinoma is the second most common type.³⁹

Symptomatology. In both females and males, breast cancer symptoms include painless, firm mass or lump in the breast (left breast slightly more common than right), skin changes or dimpling, nipple retraction or turning inward, and nipple discharge. Breast cancer can spread to the lymph nodes and can be appreciated in axilla or clavicular regions.⁴⁰

Diagnostics. The diagnostic evaluation for breast cancer is similar for females and males. It includes a clinical breast examination, diagnostic mammogram, and ultrasound.⁴¹ Mammograms can distinguish between gynecomastia and cancer, especially in males.⁴² A core or fine needle biopsy is needed to confirm diagnosis.⁴¹

Adult Leukemia

The US incidence rate of leukemia for both male and female was 14.4 per 100,000 individuals per year, with a death rate of 5.8 per 100,000 individuals per year, and a 1.5% lifetime diagnosis risk.⁴³

VOC-exposed personnel had a 7% higher risk of developing leukemia and a 13% increased mortality risk compared with the control group.^1,7 Subtypes of leukemia at risk included a 38% increased incidence of acute myeloid/monocytic leukemia (AML) and a 2% increased incidence of chronic lymphocytic leukemia (CLL).¹ Benzene and TCE exposures are known risk factors for AML and other leukemias.⁷ Personnel at Camp Lejeune had 3 times the incidence or mortality for leukemia, specifically AML mortality at 20%.⁷ Smoking is an additional risk factor for certain leukemias, especially AML.³⁰

Symptomatology. Symptoms associated with leukemia are often nonspecific and may include fatigue, pallor, easy bruising or bleeding (skin or gums), recurrent infections secondary to neutropenia, fever, night sweats, pain or feeling full after a small meal due to enlarged spleen or liver, and weight loss.^44,45

Diagnostics. An initial screening includes a CBC with differential, a peripheral smear to detect the presence of blast cells, as well as Auer rods in myeloid blast cells in AML or smudge cells in CLL. Confirmatory tests may include bone marrow biopsy or flow cytometry. A referral to a hematologist is recommended for any suspected leukemia.^46,47

Myelodysplastic Syndromes

Aplastic anemia and MDS are considered rare disorders.⁴⁸ Aplastic anemia is a nonmalignant bone marrow failure disorder with pancytopenia and hypocellular bone marrow due to the loss of hematopoietic stem cells.⁴⁸ MDS is a type of hematopoietic cancer where the bone marrow produces abnormal blood cells or does not make enough healthy cells.⁴⁹ This can lead to an increased risk for infection, cytopenias, neutropenia, refractory anemia, and thrombocytopenia, and progression to AML in some patients.⁴⁹

The reported US incidence of MDS from 1975 to 2013 was 6.7 per 100,000 for males and 3.7 per 100,000 for females.⁵⁰ Benzene exposure is linked to MDS and a known cause of AML.¹ VOC-exposed personnel had a 68% increased risk of developing MDS and a 2.3-fold increased mortality risk compared to controls.^1,7

Symptomatology. Some patients are asymptomatic at diagnosis.⁵¹ Symptoms related to cytopenia include fatigue, pallor, purpura, petechiae, bleeding of skin, gum, or nose, recurrent infections, fever, bone pain, loss of appetite, and weight loss.^50,51

Diagnostics. Initial workup includes a CBC with differential to assess for anemia, white blood cell and absolute neutrophil counts (low), and thrombocytopenia.⁵² A peripheral blood smear may show myeloid blast cells. A bone marrow aspiration and biopsy, flow cytometry, and cytogenetic or molecular testing may be performed. If MDS is suspected, a referral to a hematologist should be considered.⁵²

Multiple Myeloma

The US incidence rate of multiple myeloma for both males and females is 7.3 per 100,000 individuals per year, with a mortality rate of 2.9 per 100,000 individuals per year, and a 0.8% lifetime diagnosis risk.⁵³ VOC-exposed personnel had a 13% increased risk of developing multiple myeloma and an 8% increased mortality risk compared to unexposed personnel.^1,7

Symptomatology. Multiple myeloma may be asymptomatic in early stages. The most common presenting symptom is bone pain, especially in the back, hips, and long bones, due to hypercalcemia from increased reabsorption, plasma cell tumor overgrowth in the bone marrow, and lytic lesions.⁵⁴ Additional symptoms include fatigue and pallor related to anemia, leukopenia, thrombocytopenia, recurrent infections, extreme thirst, frequent urination, dehydration, confusion associated with hypercalcemia, peripheral neuropathy, loss of appetite, weight loss, and renal impairment or failure.⁵⁴

Diagnostics. Testing considerations include a CBC with a peripheral blood smear to evaluate anemia and rouleaux formation of red blood cells (seen in > 50% of patients with multiple myeloma), comprehensive metabolic panel (CMP) to assess kidney function, calcium levels (elevated), serum and urine protein electrophoresis with immunofixation to detect monoclonal protein (detected in > 80% of patients with multiple myeloma) and Bence-Jones proteins, serum free light chain assay, and a bone marrow biopsy for diagnosis.^55,56

MRI of the spine and pelvis is the most sensitive to detecting bone marrow involvement and focal lesions before lytic lesion progression occurs and for assessing spinal cord compression.⁵⁷ PET/CT is more sensitive at detecting extramedullary disease, outside of the spine, and for patients that cannot undergo MRI.⁵⁷ A whole-body low-dose CT, either alone or with PET, is more sensitive than an X-ray at detecting lytic lesions, fractures, or osteoporosis associated with multiple myeloma.⁵⁷

Non-Hodgkin Lymphoma

The US incidence rate of NHL for both males and females are 18.7 per 100,000 individuals per year, the death rate is 4.9 per 100,000 individuals per year, and a 2% lifetime diagnosis risk.⁵⁸ VOC-exposed personnel had a 1% higher risk of developing NHL and a decreased mortality risk compared to the control group.^1,7 Specific NHL subtypes with increased risk in the exposed cohort are mantle cell (26%), follicular (7%), Burkitt (53%), and marginal zone B-cell (45%).⁷

Symptomatology. NHL often presents with painless lymphadenopathy or enlarged lymph nodes involving the cervical, axillary, inguinal regions.^59,60 Other symptoms include frequent infections, unexplained bruising, weight loss, and “B symptoms,” such as fever and night sweats.^59,60 Some patients develop a mediastinal mass in the thorax, which if large may lead to cough or shortness of breath.⁵⁹

Diagnostics. The initial diagnostic workup includes CBC with differential and LDH, which may be elevated.^60,61 Imaging may begin with a chest X-ray to assess for a mediastinal mass; however, CTs of the chest, abdomen, and pelvis provide more detail to better assess for NHL. Whole body PET/CT is considered the gold standard for assessing and staging systemic involvement. If enlarged lymph nodes are present, a biopsy can confirm the subtype of NHL.^60,61

PHYSICAL EXAMINATION

A focused physical examination may aid HCPs in early detection of the cancers associated with Camp Lejeune (Table 2). The physical examination can guide diagnostic testing and imaging for further assessment and workup for VOC-related cancers.

Proposed Diagnostic Screening Panel

Primary care and internal medicine HCPs have the opportunity to improve patient health outcomes by implementing a targeted diagnostic screening panel for identified veterans previously stationed at Camp Lejeune. Early identification of cancers associated with VOCs exposure can facilitate earlier treatment interventions and improve health and quality of life outcomes. The following diagnostic screening panel outlines a potential cost-effective strategy for evaluating and detecting the 10 cancers associated with VOC exposure in Camp Lejeune water.

Baseline Screening

Implementing a diagnostic screening panel in this high-risk cohort can lead to earlier diagnosis, reduce mortality, and improve patient outcomes through early intervention, which in turn may result in less invasive treatment. This approach may also reduce health care costs by avoiding costs associated with delayed diagnosis and advanced-stage cancer care (Tables 3 and 4).

A baseline panel of tests for exposed veterans could include:

• A CBC with differential and peripheral smear to assess for anemia, leukemia, thrombocytopenia, and blast cells associated with leukemias, MDS, multiple myeloma, and NHL.^{19,46,47,52,55,56,60,61}
• CMP evaluates calcium, total protein, renal and liver renal function. Elevated test results may indicate kidney or liver cancer or multiple myeloma.^{19,27,28,55,56}
• LDH testing may reveal levels that are elevated from tissue damage or high cell turnover in kidney cancer, multiple myeloma, and NHL.^{19,55,56,60,61}
• Urinalysis with microscopy may detect hematuria, proteinuria and cellular casts in bladder and kidney cancers.^13,24,19
• Low-dose CTs of the chest, abdomen, and pelvis are recommended for early identification of any masses or lymphadenopathy in lung, kidney, liver cancers, and NHL.^{19,27,28,32,60,61}

COST EFFICIENCY

Screening Panel Cost

According to the Medicare Clinical Laboratory Fee Schedule payment cap for 2018, the mean cost for the proposed blood workup was $35 (CBC, $10; CMP, $13; LDH, $8; urinalysis, $4).⁶² Medicare procedure price schedule for 2025 includes $351 for a CT of the abdomen and pelvis with and without contrast (Current Procedural Terminology [CPT] code 74177) and $187 for a CT of the chest with and without contrast (CPT code 71270).^63,64 The total proposed diagnostic screening panel payment cost about $572.

Cancer Care Cost

The average cost for initial cancer care across all cancer sites from 2007 to 2013 was $43,516 per patient; Camp Lejeune-associated cancers ranged from $26,443 for bladder cancer to $89,947 for esophageal cancer care.⁶⁴ Further, the last year of life cost across all cancer sites averaged $109,727, and Camp Lejeune-associated cancer types ranged from $76,101 for breast cancer to $169,588 for leukemia.⁶⁵

CONCLUSIONS

From 1953 to 1985, up to 1 million military personnel, civilian workers, and their families stationed at Camp Lejeune were unknowingly exposed to toxic and carcinogenic VOCs, which are associated with = 10 cancers, including bladder, kidney, esophageal, liver, lung, breast, and hematologic malignancies.^1-4 Some veterans may be asymptomatic, whereas others present with subtle or specific symptoms that can vary by individual and the type and stage of cancer. HCPs have an opportunity to improve patient outcomes through awareness in identifying symptoms associated with Camp Lejeune water exposure and performing a thorough baseline physical examination, especially noting lymphadenopathy, unexplained weight loss, or masses, which can guide further diagnostic evaluation. Timely screening can identify cancers earlier, reducing delays in care, mitigating the cost burden associated with advanced-stage cancer treatment, improving survival outcomes, and enhancing quality of life. Primary care and internal medicine HCPs specifically play a crucial role in early recognition, physical assessment, and appropriate screening tools. A proposed panel includes CBC with differential and peripheral smear, CMP, LDH, urinalysis, and low-dose CTs of the chest, abdomen and pelvis. Implementation should be guided by clinical judgment and patient-specific risk factors. The proposed diagnostic screening panel is a small price to pay for those who served in any capacity at Camp Lejeune.

Clinical awareness of cancers associated with Camp Lejeune water contamination exposure remains limited despite legal and policy advances. Gaps persist in early symptom recognition and timely diagnostic evaluation before a definitive cancer diagnosis among exposed personnel. This may represent missed opportunities for earlier identification of volatile organic compounds (VOCs)-related cancers and for less invasive treatment options for veterans in this high-risk population.

Federal health care practitioners (HCPs), especially those in primary care and internal medicine, are uniquely positioned to bridge this gap. By improving the recognition of symptoms, pertinent physical examination findings, and implementing a diagnostic screening panel, HCPs can support accurate diagnoses and facilitate earlier treatment to improve health and quality of life for this population.

From 1953 to 1985, as many as 1 million military personnel, civilian workers, and their families stationed at US Marine Corps Base Camp Lejeune were unknowingly exposed to toxic and carcinogenic chemicals in drinking and bathing water.¹ Three of the 8 main water sources on base were contaminated with VOCs, which are associated with multiple cancers.^1-3

The US Department of Veterans Affairs (VA) recognizes 15 conditions associated with Camp Lejeune contaminated water exposure for VA benefits, including 10 cancers: adult leukemia; aplastic anemia and other myelodysplastic syndromes (MDS); bladder, esophageal, kidney, liver, breast (male and female), and lung cancers; multiple myeloma; and non-Hodgkin lymphoma (NHL).⁴

BACKGROUND

Established in 1942, Camp Lejeune is an important Marine Corps training installation. Between 1953 and 1985, multiple on-base water systems were contaminated with VOCs, including trichloroethylene (TCE), perchloroethylene (PCE), benzene, and vinyl chloride, due to improper waste disposal and industrial runoff from on- and off-base sources.⁵ Tarawa Terrace water treatment plant (WTP) was contaminated primarily with PCE from November 1957 to February 1987. Hadnot Point WTP was contaminated with TCE from August 1953 to December 1984, along with PCE, and benzene, toluene, ethylbenzene, and xylene (BTEX). Holcomb Boulevard WTP, established in 1972, was contaminated with TCE from June 1972 to February 1985.² These contaminants entered the drinking and bathing water supply over decades, and exposure often occurred concurrently across = 1 VOC, compounding health risks.^2,3 This prolonged 32-year VOC exposure window underlies current concerns regarding long-term cancer risk among affected service members, civilian employees, and family members. Epidemiologic research has found statistically significant associations between VOC exposure and multiple cancers, neurologic conditions, and reproductive issues.⁶ Specifically, TCE is associated with higher risks of hematologic cancers, multiple myeloma, NHL, and kidney cancer.³ PCE is linked with kidney cancer, benzene with multiple myeloma and NHL, and vinyl chloride with hepatobiliary cancers.³ A cohort mortality study compared Camp Lejeune personnel with a control group at Camp Pendleton from 1972 to 1985 and found a 3-fold higher incidence or mortality rate for kidney, esophageal, and female breast cancers, leukemia, and lymphoma among exposed Camp Lejeune personnel.⁶ Notably, personnel assigned to Camp Lejeune for as little as 6 months faced up to a 6-fold increase in cancer risk; the average military assignment between 1975 and 1985 was 18 months.^3,6

Honoring America's Veterans and Caring for Camp Lejeune Families Act of 2012, the Sergeant First Class Heath Robinson Honoring Our Promise to Address Comprehensive Toxics (PACT) Act of 2022, the Camp Lejeune Justice Act of 2022, and the pending Ensuring Justice for Camp Lejeune Victims Act of 2025 provide health care and legal resources for personnel and families affected by Camp Lejeune’s contaminated water.^6-8 These laws acknowledge associations between exposure and specific health conditions and expanded health care, benefits, and legal recourse for affected veterans, survivors, and their families.^8,9

CANCERS LINKED TO CAMP LEJEUNE

Camp Lejeune VOC-contaminated water exposure is associated with solid tumor and hematologic cancers. Symptoms, physical examination findings, and diagnostic considerations vary by cancer type (Table 1).

Bladder Cancer

The US incidence rate of bladder cancer for both males and females is 18 per 100,000 individuals per year, with a death rate of 4.1 per 100,000 individuals per year, and a 2.1% lifetime diagnosis risk.¹⁰ Personnel exposed to VOCs at Camp Lejeune had a 9% higher risk of developing bladder cancer and a 2% increased mortality compared with an unexposed control group at Camp Pendleton.^1,7 Other bladder cancer subtypes at increased risk are papillary transitional cell carcinoma, nonpapillary transition cell carcinoma, and urothelial carcinoma.⁷ This is consistent with prior research that found PCE exposure is associated with an increased risk for bladder cancer.^3,7,11 Smoking and tobacco use remain significant risk factors for bladder cancer.¹²

Symptomatology. The most common symptom associated with bladder cancer is painless hematuria (gross or microscopic). Other often delayed symptoms include urinary frequency, urgency, or nocturia.^13,14

Diagnostics. Screening tests include urinalysis for hematuria, urine cytology, and cystoscopy with biopsy as the gold standard for diagnosis and staging.^15,16

Kidney Cancer

The US incidence rate of kidney cancer and renal pelvis cancer for both males and females is 17.5 per 100,000 individuals per year, with a death rate of 3.4 per 100,000, and a 1.8% lifetime diagnosis risk.¹⁷ Camp Lejeune personnel exposed to VOCs had a 6% increased risk of developing kidney cancer and renal pelvis cancer and a 21% higher mortality risk compared with Camp Pendleton controls.^1,7 Subtypes at risk include renal cell carcinoma and papillary carcinoma.⁷ This is consistent with prior research that found exposures to TCE and PCE are associated with a 3-fold increased risk of kidney cancer.^3,7

Symptomatology. Hematuria, flank pain, and a palpable abdominal mass are common symptoms associated with kidney cancer. In advanced stages, other symptoms may include left-sided varicocele, anemia, weight loss, fatigue, fever, and night sweats.¹⁸

Diagnostics. Screening tests include urinalysis to assess the presence of blood, complete blood count (CBC) to assess anemia, calcium (elevated), and lactate dehydrogenase (LDH), which may be elevated. Imaging strategies include abdominal computed tomography (CT), magnetic resonance imaging (MRI), or ultrasound.¹⁹

Esophageal Cancer

The US incidence rate of esophageal cancer for both males and females is 4.2 per 100,000 individuals per year, the death rate is 3.7 per 100,000 individuals per year, and a 0.5% lifetime diagnosis risk.²⁰ VOC-exposed Camp Lejeune personnel had a 27% increased incidence and 25% increased mortality compared with the control group.^1,7 Esophageal cancer subtypes at elevated risk include squamous cell carcinoma and adenocarcinoma. This is consistent with prior research that found Camp Lejeune water exposure is associated with a 3-fold increased risk for esophageal cancer.⁷ Additional risk factors include history of smoking and alcohol use.²¹

Symptomatology. Esophageal cancer is often asymptomatic with potential symptoms that include dysphagia, hoarseness, and weight loss in advanced disease.²²

Diagnostics. Endoscopy with biopsy is the definitive method for diagnosis.²³

Liver Cancer

The US incidence rate of liver cancer and intrahepatic bile duct cancer for both males and females is 9.4 per 100,000 individuals per year, with a death rate of 6.6 per 100,000 individuals per year, and a 1.1% lifetime diagnosis risk.²⁴ VOC-exposed personnel had a 1% higher mortality than controls.¹

Symptomatology. Liver cancer is often asymptomatic and appears in late stages.²⁵ Common symptoms include right upper quadrant pain, early satiety, nausea, vomiting, loss of appetite, weight loss, ascites, jaundice, and abnormal bleeding or bruising.^25,26

Diagnostics. Diagnostic tests may include an ultrasound, CT, or MRI. Additional laboratory testing may include liver function, a-fetoprotein blood, CBC, renal function, calcium, and hepatitis panel screening for hepatitis B and C.^27,28

Lung Cancer

The US incidence rate of lung cancer for both males and females is 47.8 per 100,000 individuals per year, with a death rate of 31.5 per 100,000 individuals per year, and a 5.4% lifetime diagnosis risk.²⁹ VOC-exposed personnel had a 16% increased risk and 19% higher mortality.^1,7 Subtypes include large cell, small cell, non-small cell, squamous cell, and adenocarcinoma.⁷ Smoking is an additional risk factor.³⁰

Symptomatology. Symptoms of lung cancer include cough, shortness of breath, chest pain worse with deep breathing, unexplained weight loss, fatigue, night sweats, and recurrent fevers. Advanced stages may metastasize or spread to the liver, bones, and brain.³¹

Diagnostics. Low-dose CT and chest X-ray are used for screening.³²

Breast Cancer

The US incidence rate of female breast cancer is 130.8 per 100,000 individuals per year, with a death rate of 19.2 per 100,000 individuals per year, and a 13.0% lifetime risk of diagnosis.³³ For female VOC-exposed personnel, there was an equal risk of developing breast cancer as the control group.¹ However, exposed females at Camp Lejeune had a 23% higher mortality risk compared to the control group.⁷ Breast cancer subtypes among females include ductal carcinoma, lobular carcinoma, and ductal-lobular carcinoma.¹

The US incidence rate of male breast cancer is 1.3 per 100,000 individuals per year, with a death rate of 0.3 per 100,000 individuals per year.^34,35 The lifetime risk for males developing breast cancer is 137.7 per 100,000 and about 70 to 100 times less common in men than women.³⁶

Male personnel exposed at Camp Lejeune had a 4% increased risk for developing breast cancer compared to Camp Pendleton.⁷ However, mortality was lower in the Camp Lejeune group.¹ Although male breast cancer is rare, males at Camp Lejeune had a higher incidence, indicating a link between TCE, PCE, vinyl chloride exposures and male breast cancer.³⁷ Male breast cancer is more often diagnosed in advanced stages than female breast cancer due to the lack of awareness or absence of routine screenings.³⁸ The most common breast cancer type in males is invasive ductal carcinoma, accounting for 85% to 90% of cases; lobular carcinoma is the second most common type.³⁹

Symptomatology. In both females and males, breast cancer symptoms include painless, firm mass or lump in the breast (left breast slightly more common than right), skin changes or dimpling, nipple retraction or turning inward, and nipple discharge. Breast cancer can spread to the lymph nodes and can be appreciated in axilla or clavicular regions.⁴⁰

Diagnostics. The diagnostic evaluation for breast cancer is similar for females and males. It includes a clinical breast examination, diagnostic mammogram, and ultrasound.⁴¹ Mammograms can distinguish between gynecomastia and cancer, especially in males.⁴² A core or fine needle biopsy is needed to confirm diagnosis.⁴¹

Adult Leukemia

The US incidence rate of leukemia for both male and female was 14.4 per 100,000 individuals per year, with a death rate of 5.8 per 100,000 individuals per year, and a 1.5% lifetime diagnosis risk.⁴³

VOC-exposed personnel had a 7% higher risk of developing leukemia and a 13% increased mortality risk compared with the control group.^1,7 Subtypes of leukemia at risk included a 38% increased incidence of acute myeloid/monocytic leukemia (AML) and a 2% increased incidence of chronic lymphocytic leukemia (CLL).¹ Benzene and TCE exposures are known risk factors for AML and other leukemias.⁷ Personnel at Camp Lejeune had 3 times the incidence or mortality for leukemia, specifically AML mortality at 20%.⁷ Smoking is an additional risk factor for certain leukemias, especially AML.³⁰

Symptomatology. Symptoms associated with leukemia are often nonspecific and may include fatigue, pallor, easy bruising or bleeding (skin or gums), recurrent infections secondary to neutropenia, fever, night sweats, pain or feeling full after a small meal due to enlarged spleen or liver, and weight loss.^44,45

Diagnostics. An initial screening includes a CBC with differential, a peripheral smear to detect the presence of blast cells, as well as Auer rods in myeloid blast cells in AML or smudge cells in CLL. Confirmatory tests may include bone marrow biopsy or flow cytometry. A referral to a hematologist is recommended for any suspected leukemia.^46,47

Myelodysplastic Syndromes

Aplastic anemia and MDS are considered rare disorders.⁴⁸ Aplastic anemia is a nonmalignant bone marrow failure disorder with pancytopenia and hypocellular bone marrow due to the loss of hematopoietic stem cells.⁴⁸ MDS is a type of hematopoietic cancer where the bone marrow produces abnormal blood cells or does not make enough healthy cells.⁴⁹ This can lead to an increased risk for infection, cytopenias, neutropenia, refractory anemia, and thrombocytopenia, and progression to AML in some patients.⁴⁹

The reported US incidence of MDS from 1975 to 2013 was 6.7 per 100,000 for males and 3.7 per 100,000 for females.⁵⁰ Benzene exposure is linked to MDS and a known cause of AML.¹ VOC-exposed personnel had a 68% increased risk of developing MDS and a 2.3-fold increased mortality risk compared to controls.^1,7

Symptomatology. Some patients are asymptomatic at diagnosis.⁵¹ Symptoms related to cytopenia include fatigue, pallor, purpura, petechiae, bleeding of skin, gum, or nose, recurrent infections, fever, bone pain, loss of appetite, and weight loss.^50,51

Diagnostics. Initial workup includes a CBC with differential to assess for anemia, white blood cell and absolute neutrophil counts (low), and thrombocytopenia.⁵² A peripheral blood smear may show myeloid blast cells. A bone marrow aspiration and biopsy, flow cytometry, and cytogenetic or molecular testing may be performed. If MDS is suspected, a referral to a hematologist should be considered.⁵²

Multiple Myeloma

The US incidence rate of multiple myeloma for both males and females is 7.3 per 100,000 individuals per year, with a mortality rate of 2.9 per 100,000 individuals per year, and a 0.8% lifetime diagnosis risk.⁵³ VOC-exposed personnel had a 13% increased risk of developing multiple myeloma and an 8% increased mortality risk compared to unexposed personnel.^1,7

Symptomatology. Multiple myeloma may be asymptomatic in early stages. The most common presenting symptom is bone pain, especially in the back, hips, and long bones, due to hypercalcemia from increased reabsorption, plasma cell tumor overgrowth in the bone marrow, and lytic lesions.⁵⁴ Additional symptoms include fatigue and pallor related to anemia, leukopenia, thrombocytopenia, recurrent infections, extreme thirst, frequent urination, dehydration, confusion associated with hypercalcemia, peripheral neuropathy, loss of appetite, weight loss, and renal impairment or failure.⁵⁴

Diagnostics. Testing considerations include a CBC with a peripheral blood smear to evaluate anemia and rouleaux formation of red blood cells (seen in > 50% of patients with multiple myeloma), comprehensive metabolic panel (CMP) to assess kidney function, calcium levels (elevated), serum and urine protein electrophoresis with immunofixation to detect monoclonal protein (detected in > 80% of patients with multiple myeloma) and Bence-Jones proteins, serum free light chain assay, and a bone marrow biopsy for diagnosis.^55,56

MRI of the spine and pelvis is the most sensitive to detecting bone marrow involvement and focal lesions before lytic lesion progression occurs and for assessing spinal cord compression.⁵⁷ PET/CT is more sensitive at detecting extramedullary disease, outside of the spine, and for patients that cannot undergo MRI.⁵⁷ A whole-body low-dose CT, either alone or with PET, is more sensitive than an X-ray at detecting lytic lesions, fractures, or osteoporosis associated with multiple myeloma.⁵⁷

Non-Hodgkin Lymphoma

The US incidence rate of NHL for both males and females are 18.7 per 100,000 individuals per year, the death rate is 4.9 per 100,000 individuals per year, and a 2% lifetime diagnosis risk.⁵⁸ VOC-exposed personnel had a 1% higher risk of developing NHL and a decreased mortality risk compared to the control group.^1,7 Specific NHL subtypes with increased risk in the exposed cohort are mantle cell (26%), follicular (7%), Burkitt (53%), and marginal zone B-cell (45%).⁷

Symptomatology. NHL often presents with painless lymphadenopathy or enlarged lymph nodes involving the cervical, axillary, inguinal regions.^59,60 Other symptoms include frequent infections, unexplained bruising, weight loss, and “B symptoms,” such as fever and night sweats.^59,60 Some patients develop a mediastinal mass in the thorax, which if large may lead to cough or shortness of breath.⁵⁹

Diagnostics. The initial diagnostic workup includes CBC with differential and LDH, which may be elevated.^60,61 Imaging may begin with a chest X-ray to assess for a mediastinal mass; however, CTs of the chest, abdomen, and pelvis provide more detail to better assess for NHL. Whole body PET/CT is considered the gold standard for assessing and staging systemic involvement. If enlarged lymph nodes are present, a biopsy can confirm the subtype of NHL.^60,61

PHYSICAL EXAMINATION

A focused physical examination may aid HCPs in early detection of the cancers associated with Camp Lejeune (Table 2). The physical examination can guide diagnostic testing and imaging for further assessment and workup for VOC-related cancers.

Proposed Diagnostic Screening Panel

Primary care and internal medicine HCPs have the opportunity to improve patient health outcomes by implementing a targeted diagnostic screening panel for identified veterans previously stationed at Camp Lejeune. Early identification of cancers associated with VOCs exposure can facilitate earlier treatment interventions and improve health and quality of life outcomes. The following diagnostic screening panel outlines a potential cost-effective strategy for evaluating and detecting the 10 cancers associated with VOC exposure in Camp Lejeune water.

Baseline Screening

Implementing a diagnostic screening panel in this high-risk cohort can lead to earlier diagnosis, reduce mortality, and improve patient outcomes through early intervention, which in turn may result in less invasive treatment. This approach may also reduce health care costs by avoiding costs associated with delayed diagnosis and advanced-stage cancer care (Tables 3 and 4).

A baseline panel of tests for exposed veterans could include:

• A CBC with differential and peripheral smear to assess for anemia, leukemia, thrombocytopenia, and blast cells associated with leukemias, MDS, multiple myeloma, and NHL.^{19,46,47,52,55,56,60,61}
• CMP evaluates calcium, total protein, renal and liver renal function. Elevated test results may indicate kidney or liver cancer or multiple myeloma.^{19,27,28,55,56}
• LDH testing may reveal levels that are elevated from tissue damage or high cell turnover in kidney cancer, multiple myeloma, and NHL.^{19,55,56,60,61}
• Urinalysis with microscopy may detect hematuria, proteinuria and cellular casts in bladder and kidney cancers.^13,24,19
• Low-dose CTs of the chest, abdomen, and pelvis are recommended for early identification of any masses or lymphadenopathy in lung, kidney, liver cancers, and NHL.^{19,27,28,32,60,61}

COST EFFICIENCY

Screening Panel Cost

According to the Medicare Clinical Laboratory Fee Schedule payment cap for 2018, the mean cost for the proposed blood workup was $35 (CBC, $10; CMP, $13; LDH, $8; urinalysis, $4).⁶² Medicare procedure price schedule for 2025 includes $351 for a CT of the abdomen and pelvis with and without contrast (Current Procedural Terminology [CPT] code 74177) and $187 for a CT of the chest with and without contrast (CPT code 71270).^63,64 The total proposed diagnostic screening panel payment cost about $572.

Cancer Care Cost

The average cost for initial cancer care across all cancer sites from 2007 to 2013 was $43,516 per patient; Camp Lejeune-associated cancers ranged from $26,443 for bladder cancer to $89,947 for esophageal cancer care.⁶⁴ Further, the last year of life cost across all cancer sites averaged $109,727, and Camp Lejeune-associated cancer types ranged from $76,101 for breast cancer to $169,588 for leukemia.⁶⁵

CONCLUSIONS

From 1953 to 1985, up to 1 million military personnel, civilian workers, and their families stationed at Camp Lejeune were unknowingly exposed to toxic and carcinogenic VOCs, which are associated with = 10 cancers, including bladder, kidney, esophageal, liver, lung, breast, and hematologic malignancies.^1-4 Some veterans may be asymptomatic, whereas others present with subtle or specific symptoms that can vary by individual and the type and stage of cancer. HCPs have an opportunity to improve patient outcomes through awareness in identifying symptoms associated with Camp Lejeune water exposure and performing a thorough baseline physical examination, especially noting lymphadenopathy, unexplained weight loss, or masses, which can guide further diagnostic evaluation. Timely screening can identify cancers earlier, reducing delays in care, mitigating the cost burden associated with advanced-stage cancer treatment, improving survival outcomes, and enhancing quality of life. Primary care and internal medicine HCPs specifically play a crucial role in early recognition, physical assessment, and appropriate screening tools. A proposed panel includes CBC with differential and peripheral smear, CMP, LDH, urinalysis, and low-dose CTs of the chest, abdomen and pelvis. Implementation should be guided by clinical judgment and patient-specific risk factors. The proposed diagnostic screening panel is a small price to pay for those who served in any capacity at Camp Lejeune.

Clinical awareness of cancers associated with Camp Lejeune water contamination exposure remains limited despite legal and policy advances. Gaps persist in early symptom recognition and timely diagnostic evaluation before a definitive cancer diagnosis among exposed personnel. This may represent missed opportunities for earlier identification of volatile organic compounds (VOCs)-related cancers and for less invasive treatment options for veterans in this high-risk population.

Federal health care practitioners (HCPs), especially those in primary care and internal medicine, are uniquely positioned to bridge this gap. By improving the recognition of symptoms, pertinent physical examination findings, and implementing a diagnostic screening panel, HCPs can support accurate diagnoses and facilitate earlier treatment to improve health and quality of life for this population.

From 1953 to 1985, as many as 1 million military personnel, civilian workers, and their families stationed at US Marine Corps Base Camp Lejeune were unknowingly exposed to toxic and carcinogenic chemicals in drinking and bathing water.¹ Three of the 8 main water sources on base were contaminated with VOCs, which are associated with multiple cancers.^1-3

The US Department of Veterans Affairs (VA) recognizes 15 conditions associated with Camp Lejeune contaminated water exposure for VA benefits, including 10 cancers: adult leukemia; aplastic anemia and other myelodysplastic syndromes (MDS); bladder, esophageal, kidney, liver, breast (male and female), and lung cancers; multiple myeloma; and non-Hodgkin lymphoma (NHL).⁴

BACKGROUND

Established in 1942, Camp Lejeune is an important Marine Corps training installation. Between 1953 and 1985, multiple on-base water systems were contaminated with VOCs, including trichloroethylene (TCE), perchloroethylene (PCE), benzene, and vinyl chloride, due to improper waste disposal and industrial runoff from on- and off-base sources.⁵ Tarawa Terrace water treatment plant (WTP) was contaminated primarily with PCE from November 1957 to February 1987. Hadnot Point WTP was contaminated with TCE from August 1953 to December 1984, along with PCE, and benzene, toluene, ethylbenzene, and xylene (BTEX). Holcomb Boulevard WTP, established in 1972, was contaminated with TCE from June 1972 to February 1985.² These contaminants entered the drinking and bathing water supply over decades, and exposure often occurred concurrently across = 1 VOC, compounding health risks.^2,3 This prolonged 32-year VOC exposure window underlies current concerns regarding long-term cancer risk among affected service members, civilian employees, and family members. Epidemiologic research has found statistically significant associations between VOC exposure and multiple cancers, neurologic conditions, and reproductive issues.⁶ Specifically, TCE is associated with higher risks of hematologic cancers, multiple myeloma, NHL, and kidney cancer.³ PCE is linked with kidney cancer, benzene with multiple myeloma and NHL, and vinyl chloride with hepatobiliary cancers.³ A cohort mortality study compared Camp Lejeune personnel with a control group at Camp Pendleton from 1972 to 1985 and found a 3-fold higher incidence or mortality rate for kidney, esophageal, and female breast cancers, leukemia, and lymphoma among exposed Camp Lejeune personnel.⁶ Notably, personnel assigned to Camp Lejeune for as little as 6 months faced up to a 6-fold increase in cancer risk; the average military assignment between 1975 and 1985 was 18 months.^3,6

Honoring America's Veterans and Caring for Camp Lejeune Families Act of 2012, the Sergeant First Class Heath Robinson Honoring Our Promise to Address Comprehensive Toxics (PACT) Act of 2022, the Camp Lejeune Justice Act of 2022, and the pending Ensuring Justice for Camp Lejeune Victims Act of 2025 provide health care and legal resources for personnel and families affected by Camp Lejeune’s contaminated water.^6-8 These laws acknowledge associations between exposure and specific health conditions and expanded health care, benefits, and legal recourse for affected veterans, survivors, and their families.^8,9

CANCERS LINKED TO CAMP LEJEUNE

Camp Lejeune VOC-contaminated water exposure is associated with solid tumor and hematologic cancers. Symptoms, physical examination findings, and diagnostic considerations vary by cancer type (Table 1).

Bladder Cancer

The US incidence rate of bladder cancer for both males and females is 18 per 100,000 individuals per year, with a death rate of 4.1 per 100,000 individuals per year, and a 2.1% lifetime diagnosis risk.¹⁰ Personnel exposed to VOCs at Camp Lejeune had a 9% higher risk of developing bladder cancer and a 2% increased mortality compared with an unexposed control group at Camp Pendleton.^1,7 Other bladder cancer subtypes at increased risk are papillary transitional cell carcinoma, nonpapillary transition cell carcinoma, and urothelial carcinoma.⁷ This is consistent with prior research that found PCE exposure is associated with an increased risk for bladder cancer.^3,7,11 Smoking and tobacco use remain significant risk factors for bladder cancer.¹²

Symptomatology. The most common symptom associated with bladder cancer is painless hematuria (gross or microscopic). Other often delayed symptoms include urinary frequency, urgency, or nocturia.^13,14

Diagnostics. Screening tests include urinalysis for hematuria, urine cytology, and cystoscopy with biopsy as the gold standard for diagnosis and staging.^15,16

Kidney Cancer

The US incidence rate of kidney cancer and renal pelvis cancer for both males and females is 17.5 per 100,000 individuals per year, with a death rate of 3.4 per 100,000, and a 1.8% lifetime diagnosis risk.¹⁷ Camp Lejeune personnel exposed to VOCs had a 6% increased risk of developing kidney cancer and renal pelvis cancer and a 21% higher mortality risk compared with Camp Pendleton controls.^1,7 Subtypes at risk include renal cell carcinoma and papillary carcinoma.⁷ This is consistent with prior research that found exposures to TCE and PCE are associated with a 3-fold increased risk of kidney cancer.^3,7

Symptomatology. Hematuria, flank pain, and a palpable abdominal mass are common symptoms associated with kidney cancer. In advanced stages, other symptoms may include left-sided varicocele, anemia, weight loss, fatigue, fever, and night sweats.¹⁸

Diagnostics. Screening tests include urinalysis to assess the presence of blood, complete blood count (CBC) to assess anemia, calcium (elevated), and lactate dehydrogenase (LDH), which may be elevated. Imaging strategies include abdominal computed tomography (CT), magnetic resonance imaging (MRI), or ultrasound.¹⁹

Esophageal Cancer

The US incidence rate of esophageal cancer for both males and females is 4.2 per 100,000 individuals per year, the death rate is 3.7 per 100,000 individuals per year, and a 0.5% lifetime diagnosis risk.²⁰ VOC-exposed Camp Lejeune personnel had a 27% increased incidence and 25% increased mortality compared with the control group.^1,7 Esophageal cancer subtypes at elevated risk include squamous cell carcinoma and adenocarcinoma. This is consistent with prior research that found Camp Lejeune water exposure is associated with a 3-fold increased risk for esophageal cancer.⁷ Additional risk factors include history of smoking and alcohol use.²¹

Symptomatology. Esophageal cancer is often asymptomatic with potential symptoms that include dysphagia, hoarseness, and weight loss in advanced disease.²²

Diagnostics. Endoscopy with biopsy is the definitive method for diagnosis.²³

Liver Cancer

The US incidence rate of liver cancer and intrahepatic bile duct cancer for both males and females is 9.4 per 100,000 individuals per year, with a death rate of 6.6 per 100,000 individuals per year, and a 1.1% lifetime diagnosis risk.²⁴ VOC-exposed personnel had a 1% higher mortality than controls.¹

Symptomatology. Liver cancer is often asymptomatic and appears in late stages.²⁵ Common symptoms include right upper quadrant pain, early satiety, nausea, vomiting, loss of appetite, weight loss, ascites, jaundice, and abnormal bleeding or bruising.^25,26

Diagnostics. Diagnostic tests may include an ultrasound, CT, or MRI. Additional laboratory testing may include liver function, a-fetoprotein blood, CBC, renal function, calcium, and hepatitis panel screening for hepatitis B and C.^27,28

Lung Cancer

The US incidence rate of lung cancer for both males and females is 47.8 per 100,000 individuals per year, with a death rate of 31.5 per 100,000 individuals per year, and a 5.4% lifetime diagnosis risk.²⁹ VOC-exposed personnel had a 16% increased risk and 19% higher mortality.^1,7 Subtypes include large cell, small cell, non-small cell, squamous cell, and adenocarcinoma.⁷ Smoking is an additional risk factor.³⁰

Symptomatology. Symptoms of lung cancer include cough, shortness of breath, chest pain worse with deep breathing, unexplained weight loss, fatigue, night sweats, and recurrent fevers. Advanced stages may metastasize or spread to the liver, bones, and brain.³¹

Diagnostics. Low-dose CT and chest X-ray are used for screening.³²

Breast Cancer

The US incidence rate of female breast cancer is 130.8 per 100,000 individuals per year, with a death rate of 19.2 per 100,000 individuals per year, and a 13.0% lifetime risk of diagnosis.³³ For female VOC-exposed personnel, there was an equal risk of developing breast cancer as the control group.¹ However, exposed females at Camp Lejeune had a 23% higher mortality risk compared to the control group.⁷ Breast cancer subtypes among females include ductal carcinoma, lobular carcinoma, and ductal-lobular carcinoma.¹

The US incidence rate of male breast cancer is 1.3 per 100,000 individuals per year, with a death rate of 0.3 per 100,000 individuals per year.^34,35 The lifetime risk for males developing breast cancer is 137.7 per 100,000 and about 70 to 100 times less common in men than women.³⁶

Male personnel exposed at Camp Lejeune had a 4% increased risk for developing breast cancer compared to Camp Pendleton.⁷ However, mortality was lower in the Camp Lejeune group.¹ Although male breast cancer is rare, males at Camp Lejeune had a higher incidence, indicating a link between TCE, PCE, vinyl chloride exposures and male breast cancer.³⁷ Male breast cancer is more often diagnosed in advanced stages than female breast cancer due to the lack of awareness or absence of routine screenings.³⁸ The most common breast cancer type in males is invasive ductal carcinoma, accounting for 85% to 90% of cases; lobular carcinoma is the second most common type.³⁹

Symptomatology. In both females and males, breast cancer symptoms include painless, firm mass or lump in the breast (left breast slightly more common than right), skin changes or dimpling, nipple retraction or turning inward, and nipple discharge. Breast cancer can spread to the lymph nodes and can be appreciated in axilla or clavicular regions.⁴⁰

Diagnostics. The diagnostic evaluation for breast cancer is similar for females and males. It includes a clinical breast examination, diagnostic mammogram, and ultrasound.⁴¹ Mammograms can distinguish between gynecomastia and cancer, especially in males.⁴² A core or fine needle biopsy is needed to confirm diagnosis.⁴¹

Adult Leukemia

The US incidence rate of leukemia for both male and female was 14.4 per 100,000 individuals per year, with a death rate of 5.8 per 100,000 individuals per year, and a 1.5% lifetime diagnosis risk.⁴³

VOC-exposed personnel had a 7% higher risk of developing leukemia and a 13% increased mortality risk compared with the control group.^1,7 Subtypes of leukemia at risk included a 38% increased incidence of acute myeloid/monocytic leukemia (AML) and a 2% increased incidence of chronic lymphocytic leukemia (CLL).¹ Benzene and TCE exposures are known risk factors for AML and other leukemias.⁷ Personnel at Camp Lejeune had 3 times the incidence or mortality for leukemia, specifically AML mortality at 20%.⁷ Smoking is an additional risk factor for certain leukemias, especially AML.³⁰

Symptomatology. Symptoms associated with leukemia are often nonspecific and may include fatigue, pallor, easy bruising or bleeding (skin or gums), recurrent infections secondary to neutropenia, fever, night sweats, pain or feeling full after a small meal due to enlarged spleen or liver, and weight loss.^44,45

Diagnostics. An initial screening includes a CBC with differential, a peripheral smear to detect the presence of blast cells, as well as Auer rods in myeloid blast cells in AML or smudge cells in CLL. Confirmatory tests may include bone marrow biopsy or flow cytometry. A referral to a hematologist is recommended for any suspected leukemia.^46,47

Myelodysplastic Syndromes

Aplastic anemia and MDS are considered rare disorders.⁴⁸ Aplastic anemia is a nonmalignant bone marrow failure disorder with pancytopenia and hypocellular bone marrow due to the loss of hematopoietic stem cells.⁴⁸ MDS is a type of hematopoietic cancer where the bone marrow produces abnormal blood cells or does not make enough healthy cells.⁴⁹ This can lead to an increased risk for infection, cytopenias, neutropenia, refractory anemia, and thrombocytopenia, and progression to AML in some patients.⁴⁹

The reported US incidence of MDS from 1975 to 2013 was 6.7 per 100,000 for males and 3.7 per 100,000 for females.⁵⁰ Benzene exposure is linked to MDS and a known cause of AML.¹ VOC-exposed personnel had a 68% increased risk of developing MDS and a 2.3-fold increased mortality risk compared to controls.^1,7

Symptomatology. Some patients are asymptomatic at diagnosis.⁵¹ Symptoms related to cytopenia include fatigue, pallor, purpura, petechiae, bleeding of skin, gum, or nose, recurrent infections, fever, bone pain, loss of appetite, and weight loss.^50,51

Diagnostics. Initial workup includes a CBC with differential to assess for anemia, white blood cell and absolute neutrophil counts (low), and thrombocytopenia.⁵² A peripheral blood smear may show myeloid blast cells. A bone marrow aspiration and biopsy, flow cytometry, and cytogenetic or molecular testing may be performed. If MDS is suspected, a referral to a hematologist should be considered.⁵²

Multiple Myeloma

The US incidence rate of multiple myeloma for both males and females is 7.3 per 100,000 individuals per year, with a mortality rate of 2.9 per 100,000 individuals per year, and a 0.8% lifetime diagnosis risk.⁵³ VOC-exposed personnel had a 13% increased risk of developing multiple myeloma and an 8% increased mortality risk compared to unexposed personnel.^1,7

Symptomatology. Multiple myeloma may be asymptomatic in early stages. The most common presenting symptom is bone pain, especially in the back, hips, and long bones, due to hypercalcemia from increased reabsorption, plasma cell tumor overgrowth in the bone marrow, and lytic lesions.⁵⁴ Additional symptoms include fatigue and pallor related to anemia, leukopenia, thrombocytopenia, recurrent infections, extreme thirst, frequent urination, dehydration, confusion associated with hypercalcemia, peripheral neuropathy, loss of appetite, weight loss, and renal impairment or failure.⁵⁴

Diagnostics. Testing considerations include a CBC with a peripheral blood smear to evaluate anemia and rouleaux formation of red blood cells (seen in > 50% of patients with multiple myeloma), comprehensive metabolic panel (CMP) to assess kidney function, calcium levels (elevated), serum and urine protein electrophoresis with immunofixation to detect monoclonal protein (detected in > 80% of patients with multiple myeloma) and Bence-Jones proteins, serum free light chain assay, and a bone marrow biopsy for diagnosis.^55,56

MRI of the spine and pelvis is the most sensitive to detecting bone marrow involvement and focal lesions before lytic lesion progression occurs and for assessing spinal cord compression.⁵⁷ PET/CT is more sensitive at detecting extramedullary disease, outside of the spine, and for patients that cannot undergo MRI.⁵⁷ A whole-body low-dose CT, either alone or with PET, is more sensitive than an X-ray at detecting lytic lesions, fractures, or osteoporosis associated with multiple myeloma.⁵⁷

Non-Hodgkin Lymphoma

The US incidence rate of NHL for both males and females are 18.7 per 100,000 individuals per year, the death rate is 4.9 per 100,000 individuals per year, and a 2% lifetime diagnosis risk.⁵⁸ VOC-exposed personnel had a 1% higher risk of developing NHL and a decreased mortality risk compared to the control group.^1,7 Specific NHL subtypes with increased risk in the exposed cohort are mantle cell (26%), follicular (7%), Burkitt (53%), and marginal zone B-cell (45%).⁷

Symptomatology. NHL often presents with painless lymphadenopathy or enlarged lymph nodes involving the cervical, axillary, inguinal regions.^59,60 Other symptoms include frequent infections, unexplained bruising, weight loss, and “B symptoms,” such as fever and night sweats.^59,60 Some patients develop a mediastinal mass in the thorax, which if large may lead to cough or shortness of breath.⁵⁹

Diagnostics. The initial diagnostic workup includes CBC with differential and LDH, which may be elevated.^60,61 Imaging may begin with a chest X-ray to assess for a mediastinal mass; however, CTs of the chest, abdomen, and pelvis provide more detail to better assess for NHL. Whole body PET/CT is considered the gold standard for assessing and staging systemic involvement. If enlarged lymph nodes are present, a biopsy can confirm the subtype of NHL.^60,61

PHYSICAL EXAMINATION

A focused physical examination may aid HCPs in early detection of the cancers associated with Camp Lejeune (Table 2). The physical examination can guide diagnostic testing and imaging for further assessment and workup for VOC-related cancers.

Proposed Diagnostic Screening Panel

Primary care and internal medicine HCPs have the opportunity to improve patient health outcomes by implementing a targeted diagnostic screening panel for identified veterans previously stationed at Camp Lejeune. Early identification of cancers associated with VOCs exposure can facilitate earlier treatment interventions and improve health and quality of life outcomes. The following diagnostic screening panel outlines a potential cost-effective strategy for evaluating and detecting the 10 cancers associated with VOC exposure in Camp Lejeune water.

Baseline Screening

Implementing a diagnostic screening panel in this high-risk cohort can lead to earlier diagnosis, reduce mortality, and improve patient outcomes through early intervention, which in turn may result in less invasive treatment. This approach may also reduce health care costs by avoiding costs associated with delayed diagnosis and advanced-stage cancer care (Tables 3 and 4).

A baseline panel of tests for exposed veterans could include:

• A CBC with differential and peripheral smear to assess for anemia, leukemia, thrombocytopenia, and blast cells associated with leukemias, MDS, multiple myeloma, and NHL.^{19,46,47,52,55,56,60,61}
• CMP evaluates calcium, total protein, renal and liver renal function. Elevated test results may indicate kidney or liver cancer or multiple myeloma.^{19,27,28,55,56}
• LDH testing may reveal levels that are elevated from tissue damage or high cell turnover in kidney cancer, multiple myeloma, and NHL.^{19,55,56,60,61}
• Urinalysis with microscopy may detect hematuria, proteinuria and cellular casts in bladder and kidney cancers.^13,24,19
• Low-dose CTs of the chest, abdomen, and pelvis are recommended for early identification of any masses or lymphadenopathy in lung, kidney, liver cancers, and NHL.^{19,27,28,32,60,61}

COST EFFICIENCY

Screening Panel Cost

According to the Medicare Clinical Laboratory Fee Schedule payment cap for 2018, the mean cost for the proposed blood workup was $35 (CBC, $10; CMP, $13; LDH, $8; urinalysis, $4).⁶² Medicare procedure price schedule for 2025 includes $351 for a CT of the abdomen and pelvis with and without contrast (Current Procedural Terminology [CPT] code 74177) and $187 for a CT of the chest with and without contrast (CPT code 71270).^63,64 The total proposed diagnostic screening panel payment cost about $572.

Cancer Care Cost

The average cost for initial cancer care across all cancer sites from 2007 to 2013 was $43,516 per patient; Camp Lejeune-associated cancers ranged from $26,443 for bladder cancer to $89,947 for esophageal cancer care.⁶⁴ Further, the last year of life cost across all cancer sites averaged $109,727, and Camp Lejeune-associated cancer types ranged from $76,101 for breast cancer to $169,588 for leukemia.⁶⁵

CONCLUSIONS

From 1953 to 1985, up to 1 million military personnel, civilian workers, and their families stationed at Camp Lejeune were unknowingly exposed to toxic and carcinogenic VOCs, which are associated with = 10 cancers, including bladder, kidney, esophageal, liver, lung, breast, and hematologic malignancies.^1-4 Some veterans may be asymptomatic, whereas others present with subtle or specific symptoms that can vary by individual and the type and stage of cancer. HCPs have an opportunity to improve patient outcomes through awareness in identifying symptoms associated with Camp Lejeune water exposure and performing a thorough baseline physical examination, especially noting lymphadenopathy, unexplained weight loss, or masses, which can guide further diagnostic evaluation. Timely screening can identify cancers earlier, reducing delays in care, mitigating the cost burden associated with advanced-stage cancer treatment, improving survival outcomes, and enhancing quality of life. Primary care and internal medicine HCPs specifically play a crucial role in early recognition, physical assessment, and appropriate screening tools. A proposed panel includes CBC with differential and peripheral smear, CMP, LDH, urinalysis, and low-dose CTs of the chest, abdomen and pelvis. Implementation should be guided by clinical judgment and patient-specific risk factors. The proposed diagnostic screening panel is a small price to pay for those who served in any capacity at Camp Lejeune.

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.

Kenneth W. Kizer, MD, MPH, knows a thing or two about transition at the US Department of Veterans Affairs (VA). He served as VA Under Secretary of Health from 1994 to 1999, stepping in during an era of crisis with a mandate for transformation.

Kizer, a Distinguished Professor Emeritus at the University of California, Davis School of Medicine, is among the top thinkers about the VA and its future. He recently spoke with Federal Practitioner about community care, the electronic health record transition, and other challenges facing the Veterans Health Administration (VHA).

At stake, Kizer explained, is an invaluable service for veterans—and much more. “VA is the largest provider of training for... multiple types of health professionals that people use every day,” he said. “There’s also the research, the direct care provided to veterans, and the contingency support the VA provides, which was very well demonstrated during the COVID pandemic. These are things that benefit all Americans, not just veterans.”

When you look at the VA, what do you see?

I see an organization in turmoil, a great health care system struggling with multiple major challenges simultaneously. The VHA is becoming a very large health insurance program without the necessary infrastructure, and costs are rising rapidly. And it is trying to roll out a new EHR and implement new third-party administrator contracts while suffering from significant staffing reductions and very depressed morale.

There are a host of other high-visibility and high-impact issues, including a major reorganization. There’s been a paucity of details about exactly what is going to change, who is going to be doing what, and how the changes will affect staffing and workflow.

How will the loss of 35,000 health care positions affect veterans' care?

If you don’t have enough people, then you’re not going to be able to provide the care that is needed. Years ago, I led a project assessing the Roseburg VA Medical Center in Central Oregon. Among other things, there were a lot of problems with cardiology care. The biggest complaint the cardiologists had, and why the hospital couldn’t keep cardiologists on staff, was that there weren’t enough support staff to do the electrocardiograms. The cardiologists had to do the electrocardiograms themselves, which meant they weren’t doing other things they should be doing. You can amplify that example in a hundred different ways in VA today. If physicians don’t have adequate support, they get frustrated and disenfranchised. And they leave.

One of the fallacies I’ve heard mentioned in some congressional hearings is that it’s mainly a matter of lower pay in the VA. Pay may be an issue somewhere on the list of recruiting challenges, but more important issues higher up are things like the support clinicians receive, the work environment, whether they feel valued, and how easy it is for them to do their work. Case in point: If you put in a new EHR that doesn’t work as well as the existing one, then some doctors are going to leave.

Is VA being pushed toward privatization?

At some point it becomes a self-fulfilling prophecy. If you don’t have the staff to provide the services, then you refer more veterans to the community, and you get in a downward spiral. Patients are going to the community, you lose more staff, you continue to be unable to provide services, and more care goes to the community.

A part of this equation that hasn’t been given adequate attention is VA’s teaching mission. If care is increasingly going to the community, those patients won’t be available for the trainees in teaching programs. That in turn impacts the pipeline of clinicians who will be available to serve the population at large. The negative effects will be seen far beyond the VA.

Why have you expressed concern over VA care fragmentation?¹

Greater than 80% of VA ICU [intensive care unit] care is now being provided in community hospitals. When patients are discharged from those hospitals, they often continue getting follow-up care in the community because VA doesn’t have good mechanisms to reconnect those patients back to VA care.²

[Other researchers] found that the majority of emergency department care for enrolled veterans in New York State was being paid for by entities other than the VA, most commonly Medicare but also Medicaid and private insurance. Where follow-up care occurred often depended on who paid for the emergency department visit, not necessarily what was best for the patient.³

The core problem is that the VA has very little insight into what’s happening when its enrollees get care that is paid for by another payer. VA doesn’t know when their patients are in a private hospital emergency department, so they can’t reach out in real time, and they can’t reconnect with them afterward.

That is very different than for commercial health plans. They know when one of their enrollees is admitted to an out-of-network hospital, and when they are discharged, and they follow up immediately. VA doesn’t have the infrastructure in place to do that.

Why did the VA spend $44 billion on Medicare Advantage double-payments from 2018 to 2021?⁴

That number is much larger now—$87 billion from 2019 to 2023. Here’s the problem: When VA enrollees are also enrolled in a Medicare Advantage plan, the Medicare plan gets paid to provide the care for those veterans. But when those enrollees come to the VA, the VA provides and pays for the care but cannot bill Medicare for the costs. So the federal government ends up paying twice for care of the same person.

In a paper I coauthored last December we showed that in 2023 alone VA spent $23 billion for care of veterans enrolled in Medicare Advantage plan. Those duplicative payments accounted for almost 20% of VA's entire medical care budget.⁵

How can fragmented care be reduced?

Two things really stand out. First, real-time health insurance data sharing across payers is foundational. VA has to know when its patients get care by non-VA providers if it is going to coordinate and provide follow-up care. As a first step, VA and the Centers for Medicare & Medicaid Services need to create a data sharing platform for veterans dually enrolled in VA and Medicare or Medicaid.

This is not a new idea. I tried to do it when I was Under Secretary for Health in the late 1990s, but it never happened for various political reasons. Others have tried since. Maybe now, given how much money is at stake, it will finally get done.

Second, the VA needs to implement rigorous case management for high utilizers. The costs are not evenly distributed across enrollees. Approximately 10% of community care users account for almost 90% of community care expenditures. Common sense says you should intensely manage the care of those high-need patients who account for so much of the costs, try to avoid out-of-network ICU and emergency department care as much as possible, and build relationships with other providers so there are clear mechanisms to reconnect those patients back to VA care after an acute episode is treated outside the VA health system.

Is community care itself the problem?

No. Community care is a good thing for many veterans. It has increased access and made it easier for enrolled veterans to get care in some situations. The problem is that the VA hasn’t built in the mechanisms and processes to share information, manage complex patients, provide follow-up care, or oversee quality in community care.

Historically, VA has been an integrated delivery system that provided the overwhelming majority of care within its own facilities. However, over the last decade it has become a hybrid purchaser-provider system. It has become a very large purchaser of non-VA care, going from about $7 billion to $50 billion in community care spending over the past decade. But the VA hasn’t built the infrastructure—information exchange, case management, utilization review, quality oversight—that a hybrid purchaser-provider system needs to be a prudent purchaser.

What is your perspective on VHA's EHR transition?

The many problems with the rollout of the Oracle/Cerner EHR have been well-documented by the Inspector General, frontline clinicians, and others. The problems have been so bad that implementation has been halted a couple times. They’re now moving forward again, but it remains to be seen whether the problems truly have been fixed.

Still unaddressed is the more fundamental question of whether VistA could have been upgraded and modernized at far less cost and disruption of care. No thorough, deliberative analysis of that was ever done. And some of the ostensible problems with upgrading VistA in years past are no longer an issue.

Given the challenges VA faces, are you optimistic about its future?

While there definitely are problems, they are all solvable. Every challenge the VA is facing can be addressed. The question is when and how, and whether the VA is going to be given a fair chance to work through its challenges.

As for those who look to the private sector and think that’s the solution: They haven’t looked closely enough. The private sector is also struggling with staffing and financing issues, many of the same issues VA is dealing with, just in a somewhat different way. The problems in the private sector will be an increasing challenge for community care going forward.

Overall, my life experience is that dark times are always followed by daylight, so I am confident there are brighter days ahead for VA.

Kenneth W. Kizer, MD, MPH, knows a thing or two about transition at the US Department of Veterans Affairs (VA). He served as VA Under Secretary of Health from 1994 to 1999, stepping in during an era of crisis with a mandate for transformation.

Kizer, a Distinguished Professor Emeritus at the University of California, Davis School of Medicine, is among the top thinkers about the VA and its future. He recently spoke with Federal Practitioner about community care, the electronic health record transition, and other challenges facing the Veterans Health Administration (VHA).

At stake, Kizer explained, is an invaluable service for veterans—and much more. “VA is the largest provider of training for... multiple types of health professionals that people use every day,” he said. “There’s also the research, the direct care provided to veterans, and the contingency support the VA provides, which was very well demonstrated during the COVID pandemic. These are things that benefit all Americans, not just veterans.”

When you look at the VA, what do you see?

I see an organization in turmoil, a great health care system struggling with multiple major challenges simultaneously. The VHA is becoming a very large health insurance program without the necessary infrastructure, and costs are rising rapidly. And it is trying to roll out a new EHR and implement new third-party administrator contracts while suffering from significant staffing reductions and very depressed morale.

There are a host of other high-visibility and high-impact issues, including a major reorganization. There’s been a paucity of details about exactly what is going to change, who is going to be doing what, and how the changes will affect staffing and workflow.

How will the loss of 35,000 health care positions affect veterans' care?

If you don’t have enough people, then you’re not going to be able to provide the care that is needed. Years ago, I led a project assessing the Roseburg VA Medical Center in Central Oregon. Among other things, there were a lot of problems with cardiology care. The biggest complaint the cardiologists had, and why the hospital couldn’t keep cardiologists on staff, was that there weren’t enough support staff to do the electrocardiograms. The cardiologists had to do the electrocardiograms themselves, which meant they weren’t doing other things they should be doing. You can amplify that example in a hundred different ways in VA today. If physicians don’t have adequate support, they get frustrated and disenfranchised. And they leave.

One of the fallacies I’ve heard mentioned in some congressional hearings is that it’s mainly a matter of lower pay in the VA. Pay may be an issue somewhere on the list of recruiting challenges, but more important issues higher up are things like the support clinicians receive, the work environment, whether they feel valued, and how easy it is for them to do their work. Case in point: If you put in a new EHR that doesn’t work as well as the existing one, then some doctors are going to leave.

Is VA being pushed toward privatization?

At some point it becomes a self-fulfilling prophecy. If you don’t have the staff to provide the services, then you refer more veterans to the community, and you get in a downward spiral. Patients are going to the community, you lose more staff, you continue to be unable to provide services, and more care goes to the community.

A part of this equation that hasn’t been given adequate attention is VA’s teaching mission. If care is increasingly going to the community, those patients won’t be available for the trainees in teaching programs. That in turn impacts the pipeline of clinicians who will be available to serve the population at large. The negative effects will be seen far beyond the VA.

Why have you expressed concern over VA care fragmentation?¹

Greater than 80% of VA ICU [intensive care unit] care is now being provided in community hospitals. When patients are discharged from those hospitals, they often continue getting follow-up care in the community because VA doesn’t have good mechanisms to reconnect those patients back to VA care.²

[Other researchers] found that the majority of emergency department care for enrolled veterans in New York State was being paid for by entities other than the VA, most commonly Medicare but also Medicaid and private insurance. Where follow-up care occurred often depended on who paid for the emergency department visit, not necessarily what was best for the patient.³

The core problem is that the VA has very little insight into what’s happening when its enrollees get care that is paid for by another payer. VA doesn’t know when their patients are in a private hospital emergency department, so they can’t reach out in real time, and they can’t reconnect with them afterward.

That is very different than for commercial health plans. They know when one of their enrollees is admitted to an out-of-network hospital, and when they are discharged, and they follow up immediately. VA doesn’t have the infrastructure in place to do that.

Why did the VA spend $44 billion on Medicare Advantage double-payments from 2018 to 2021?⁴

That number is much larger now—$87 billion from 2019 to 2023. Here’s the problem: When VA enrollees are also enrolled in a Medicare Advantage plan, the Medicare plan gets paid to provide the care for those veterans. But when those enrollees come to the VA, the VA provides and pays for the care but cannot bill Medicare for the costs. So the federal government ends up paying twice for care of the same person.

In a paper I coauthored last December we showed that in 2023 alone VA spent $23 billion for care of veterans enrolled in Medicare Advantage plan. Those duplicative payments accounted for almost 20% of VA's entire medical care budget.⁵

How can fragmented care be reduced?

Two things really stand out. First, real-time health insurance data sharing across payers is foundational. VA has to know when its patients get care by non-VA providers if it is going to coordinate and provide follow-up care. As a first step, VA and the Centers for Medicare & Medicaid Services need to create a data sharing platform for veterans dually enrolled in VA and Medicare or Medicaid.

This is not a new idea. I tried to do it when I was Under Secretary for Health in the late 1990s, but it never happened for various political reasons. Others have tried since. Maybe now, given how much money is at stake, it will finally get done.

Second, the VA needs to implement rigorous case management for high utilizers. The costs are not evenly distributed across enrollees. Approximately 10% of community care users account for almost 90% of community care expenditures. Common sense says you should intensely manage the care of those high-need patients who account for so much of the costs, try to avoid out-of-network ICU and emergency department care as much as possible, and build relationships with other providers so there are clear mechanisms to reconnect those patients back to VA care after an acute episode is treated outside the VA health system.

Is community care itself the problem?

No. Community care is a good thing for many veterans. It has increased access and made it easier for enrolled veterans to get care in some situations. The problem is that the VA hasn’t built in the mechanisms and processes to share information, manage complex patients, provide follow-up care, or oversee quality in community care.

Historically, VA has been an integrated delivery system that provided the overwhelming majority of care within its own facilities. However, over the last decade it has become a hybrid purchaser-provider system. It has become a very large purchaser of non-VA care, going from about $7 billion to $50 billion in community care spending over the past decade. But the VA hasn’t built the infrastructure—information exchange, case management, utilization review, quality oversight—that a hybrid purchaser-provider system needs to be a prudent purchaser.

What is your perspective on VHA's EHR transition?

The many problems with the rollout of the Oracle/Cerner EHR have been well-documented by the Inspector General, frontline clinicians, and others. The problems have been so bad that implementation has been halted a couple times. They’re now moving forward again, but it remains to be seen whether the problems truly have been fixed.

Still unaddressed is the more fundamental question of whether VistA could have been upgraded and modernized at far less cost and disruption of care. No thorough, deliberative analysis of that was ever done. And some of the ostensible problems with upgrading VistA in years past are no longer an issue.

Given the challenges VA faces, are you optimistic about its future?

While there definitely are problems, they are all solvable. Every challenge the VA is facing can be addressed. The question is when and how, and whether the VA is going to be given a fair chance to work through its challenges.

As for those who look to the private sector and think that’s the solution: They haven’t looked closely enough. The private sector is also struggling with staffing and financing issues, many of the same issues VA is dealing with, just in a somewhat different way. The problems in the private sector will be an increasing challenge for community care going forward.

Overall, my life experience is that dark times are always followed by daylight, so I am confident there are brighter days ahead for VA.

Kenneth W. Kizer, MD, MPH, knows a thing or two about transition at the US Department of Veterans Affairs (VA). He served as VA Under Secretary of Health from 1994 to 1999, stepping in during an era of crisis with a mandate for transformation.

Kizer, a Distinguished Professor Emeritus at the University of California, Davis School of Medicine, is among the top thinkers about the VA and its future. He recently spoke with Federal Practitioner about community care, the electronic health record transition, and other challenges facing the Veterans Health Administration (VHA).

At stake, Kizer explained, is an invaluable service for veterans—and much more. “VA is the largest provider of training for... multiple types of health professionals that people use every day,” he said. “There’s also the research, the direct care provided to veterans, and the contingency support the VA provides, which was very well demonstrated during the COVID pandemic. These are things that benefit all Americans, not just veterans.”

When you look at the VA, what do you see?

I see an organization in turmoil, a great health care system struggling with multiple major challenges simultaneously. The VHA is becoming a very large health insurance program without the necessary infrastructure, and costs are rising rapidly. And it is trying to roll out a new EHR and implement new third-party administrator contracts while suffering from significant staffing reductions and very depressed morale.

There are a host of other high-visibility and high-impact issues, including a major reorganization. There’s been a paucity of details about exactly what is going to change, who is going to be doing what, and how the changes will affect staffing and workflow.

How will the loss of 35,000 health care positions affect veterans' care?

If you don’t have enough people, then you’re not going to be able to provide the care that is needed. Years ago, I led a project assessing the Roseburg VA Medical Center in Central Oregon. Among other things, there were a lot of problems with cardiology care. The biggest complaint the cardiologists had, and why the hospital couldn’t keep cardiologists on staff, was that there weren’t enough support staff to do the electrocardiograms. The cardiologists had to do the electrocardiograms themselves, which meant they weren’t doing other things they should be doing. You can amplify that example in a hundred different ways in VA today. If physicians don’t have adequate support, they get frustrated and disenfranchised. And they leave.

One of the fallacies I’ve heard mentioned in some congressional hearings is that it’s mainly a matter of lower pay in the VA. Pay may be an issue somewhere on the list of recruiting challenges, but more important issues higher up are things like the support clinicians receive, the work environment, whether they feel valued, and how easy it is for them to do their work. Case in point: If you put in a new EHR that doesn’t work as well as the existing one, then some doctors are going to leave.

Is VA being pushed toward privatization?

At some point it becomes a self-fulfilling prophecy. If you don’t have the staff to provide the services, then you refer more veterans to the community, and you get in a downward spiral. Patients are going to the community, you lose more staff, you continue to be unable to provide services, and more care goes to the community.

A part of this equation that hasn’t been given adequate attention is VA’s teaching mission. If care is increasingly going to the community, those patients won’t be available for the trainees in teaching programs. That in turn impacts the pipeline of clinicians who will be available to serve the population at large. The negative effects will be seen far beyond the VA.

Why have you expressed concern over VA care fragmentation?¹

Greater than 80% of VA ICU [intensive care unit] care is now being provided in community hospitals. When patients are discharged from those hospitals, they often continue getting follow-up care in the community because VA doesn’t have good mechanisms to reconnect those patients back to VA care.²

[Other researchers] found that the majority of emergency department care for enrolled veterans in New York State was being paid for by entities other than the VA, most commonly Medicare but also Medicaid and private insurance. Where follow-up care occurred often depended on who paid for the emergency department visit, not necessarily what was best for the patient.³

The core problem is that the VA has very little insight into what’s happening when its enrollees get care that is paid for by another payer. VA doesn’t know when their patients are in a private hospital emergency department, so they can’t reach out in real time, and they can’t reconnect with them afterward.

That is very different than for commercial health plans. They know when one of their enrollees is admitted to an out-of-network hospital, and when they are discharged, and they follow up immediately. VA doesn’t have the infrastructure in place to do that.

Why did the VA spend $44 billion on Medicare Advantage double-payments from 2018 to 2021?⁴

That number is much larger now—$87 billion from 2019 to 2023. Here’s the problem: When VA enrollees are also enrolled in a Medicare Advantage plan, the Medicare plan gets paid to provide the care for those veterans. But when those enrollees come to the VA, the VA provides and pays for the care but cannot bill Medicare for the costs. So the federal government ends up paying twice for care of the same person.

In a paper I coauthored last December we showed that in 2023 alone VA spent $23 billion for care of veterans enrolled in Medicare Advantage plan. Those duplicative payments accounted for almost 20% of VA's entire medical care budget.⁵

How can fragmented care be reduced?

Two things really stand out. First, real-time health insurance data sharing across payers is foundational. VA has to know when its patients get care by non-VA providers if it is going to coordinate and provide follow-up care. As a first step, VA and the Centers for Medicare & Medicaid Services need to create a data sharing platform for veterans dually enrolled in VA and Medicare or Medicaid.

This is not a new idea. I tried to do it when I was Under Secretary for Health in the late 1990s, but it never happened for various political reasons. Others have tried since. Maybe now, given how much money is at stake, it will finally get done.

Second, the VA needs to implement rigorous case management for high utilizers. The costs are not evenly distributed across enrollees. Approximately 10% of community care users account for almost 90% of community care expenditures. Common sense says you should intensely manage the care of those high-need patients who account for so much of the costs, try to avoid out-of-network ICU and emergency department care as much as possible, and build relationships with other providers so there are clear mechanisms to reconnect those patients back to VA care after an acute episode is treated outside the VA health system.

Is community care itself the problem?

No. Community care is a good thing for many veterans. It has increased access and made it easier for enrolled veterans to get care in some situations. The problem is that the VA hasn’t built in the mechanisms and processes to share information, manage complex patients, provide follow-up care, or oversee quality in community care.

Historically, VA has been an integrated delivery system that provided the overwhelming majority of care within its own facilities. However, over the last decade it has become a hybrid purchaser-provider system. It has become a very large purchaser of non-VA care, going from about $7 billion to $50 billion in community care spending over the past decade. But the VA hasn’t built the infrastructure—information exchange, case management, utilization review, quality oversight—that a hybrid purchaser-provider system needs to be a prudent purchaser.

What is your perspective on VHA's EHR transition?

The many problems with the rollout of the Oracle/Cerner EHR have been well-documented by the Inspector General, frontline clinicians, and others. The problems have been so bad that implementation has been halted a couple times. They’re now moving forward again, but it remains to be seen whether the problems truly have been fixed.

Still unaddressed is the more fundamental question of whether VistA could have been upgraded and modernized at far less cost and disruption of care. No thorough, deliberative analysis of that was ever done. And some of the ostensible problems with upgrading VistA in years past are no longer an issue.

Given the challenges VA faces, are you optimistic about its future?

While there definitely are problems, they are all solvable. Every challenge the VA is facing can be addressed. The question is when and how, and whether the VA is going to be given a fair chance to work through its challenges.

As for those who look to the private sector and think that’s the solution: They haven’t looked closely enough. The private sector is also struggling with staffing and financing issues, many of the same issues VA is dealing with, just in a somewhat different way. The problems in the private sector will be an increasing challenge for community care going forward.

Overall, my life experience is that dark times are always followed by daylight, so I am confident there are brighter days ahead for VA.

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