Article

The number of opioid-related overdose deaths in the United States is estimated to have increased 6-fold over the past 2 decades.¹ In 2017, more than two-thirds of drug overdose deaths involved opioids, yielding a mortality rate of 14.9 per 100,000.² Not only does the opioid epidemic currently pose a significant public health crisis characterized by high morbidity and mortality, but it is also projected to worsen in coming years. According to Chen and colleagues, opioid overdose deaths are estimated to increase by 147% from 2015 to 2025.³ That projects almost 82,000 US deaths annually and > 700,000 deaths in this period—even before accounting for surges in opioid overdoses and opioid-related mortality coinciding with the COVID-19 pandemic.^3,4

As health systems and communities globally struggle with unprecedented losses and stressors introduced by the pandemic, emerging data warrants escalating concerns with regard to increased vulnerability to relapse and overdose among those with mental health and substance use disorders (SUDs). In a recent report, the American Medical Association estimates that opioid-related deaths have increased in more than 40 states with the COVID-19 pandemic.⁴

Veterans are twice as likely to experience a fatal opioid overdose compared with their civilian counterparts.⁵ While several risk mitigation strategies have been employed in recent years to improve opioid prescribing and safety within the US Department of Veterans Affairs (VA), veterans continue to overdose on opioids, both prescribed and obtained illicitly.⁶ Variables shown to be strongly associated with opioid overdose risk include presence of mental health disorders, SUDs, medical conditions involving impaired drug metabolism or excretion, respiratory disorders, higher doses of opioids, concomitant use of sedative medications, and history of overdose.^6-8 Many veterans struggle with chronic pain and those prescribed high doses of opioids were more likely to have comorbid pain diagnoses, mental health disorders, and SUDs.⁹ Dashboards and predictive models, such as the Stratification Tool for Opioid Risk Mitigation (STORM) and the Risk Index for Overdose or Serious Opioid-induced Respiratory Depression (RIOSORD), incorporate such factors to stratify overdose risk among veterans, in an effort to prioritize high-risk individuals for review and provision of care.^6,10,11 Despite recent recognition that overdose prevention likely requires a holistic approach that addresses the biopsychosocial factors contributing to opioid-related morbidity and mortality, it is unclear whether veterans are receiving adequate and appropriate treatment for contributing conditions.

There are currently no existing studies that describe health service utilization (HSU), medication interventions, and rates of opioid-related adverse events (ORAEs) among veterans after survival of a nonfatal opioid overdose (NFO). Clinical characteristics of veterans treated for opioid overdose at a VA emergency department (ED) have previously been described by Clement and Stock.¹² Despite improvements that have been made in VA opioid prescribing and safety, knowledge gaps remain with regard to best practices for opioid overdose prevention. The aim of this study was to characterize HSU and medication interventions in veterans following NFO, as well as the frequency of ORAEs after overdose. The findings of this study may aid in the identification of areas for targeted improvement in the prevention and reduction of opioid overdoses and adverse opioid-related sequelae.

Methods

This retrospective descriptive study was conducted at VA San Diego Healthcare System (VASDHCS) in California. Subjects included were veterans administered naloxone in the ED for suspected opioid overdose between July 1, 2013 and April 1, 2017. The study population was identified through data retrieved from automated drug dispensing systems, which was then confirmed through manual chart review of notes associated with the index ED visit. Inclusion criteria included documented increased respiration or responsiveness following naloxone administration. Subjects were excluded if they demonstrated lack of response to naloxone, overdosed secondary to inpatient administration of opioids, received palliative or hospice care during the study period, or were lost to follow-up.

Data were collected via retrospective chart review and included date of index ED visit, demographics, active prescriptions, urine drug screen (UDS) results, benzodiazepine (BZD) use corroborated by positive UDS or mention of BZD in index visit chart notes, whether overdose was determined to be a suicide attempt, and naloxone kit dispensing. Patient data was collected for 2 years following overdose, including: ORAEs; ED visits; hospitalizations; repeat overdoses; fatal overdose; whether subjects were still alive; follow-up visits for pain management, mental health, and addiction treatment services; and visits to the psychiatric emergency clinic. Clinical characteristics, such as mental health disorder diagnoses, SUDs, and relevant medical conditions also were collected. Statistical analysis was performed using Microsoft Excel and included only descriptive statistics.

Results

Ninety-three patients received naloxone in the VASDHCS ED. Thirty-five met inclusion criteria and were included in the primary analysis. All subjects received IV naloxone with a mean 0.8 mg IV boluses (range, 0.1-4.4 mg).

Most patients were male with a mean age of 59.8 years (Table 1). Almost all overdoses were nonintentional except for 3 suicide attempts that were reviewed by the Suicide Prevention Committee. Three patients had previously been treated for opioid overdose at the VA with a documented positive clinical response to naloxone administration.

Discussions

This retrospective study is representative of many veterans receiving VA care, despite the small sample size. Clinical characteristics observed in the study population were generally consistent with those published by Clement and Stock, including high rates of medical and psychiatric comorbidities.¹² Subjects in both studies were prescribed comparable dosages of opioids; among those prescribed opioids but not BZDs through the VA, the mean MEDD was 117 mg in our study compared with 126 mg in the Clement and Stock study. Since implementation of the Opioid Safety Initiative (OSI) in 2013, opioid prescribing practices have improved nationwide across VA facilities, including successful reduction in the numbers of patients prescribed high-dose opioids and concurrent BZDs.¹³

Despite the tools and resources available to clinicians, discontinuing opioid therapy remains a difficult process. Concerns related to mental health and/or substance-use related decompensations often exist in the setting of rapid dose reductions or abrupt discontinuation of opioids.⁶ Although less than half of patients in the present study with an active opioid prescription at time of index overdose had their opioids discontinued within 2 years, it is reassuring to note the much higher rate of those with subsequent decreases in their prescribed doses, as well as the 50% reduction in BZD coprescribing. Ultimately, these findings remain consistent with the VA goals of mitigating harm, improving opioid prescribing, and ensuring the safe use of opioid medications when clinically appropriate.

Moreover, recent evidence suggests that interventions focused solely on opioid prescribing practices are becoming increasingly limited in their impact on reducing opioid-related deaths and will likely be insufficient for addressing the opioid epidemic as it continues to evolve. According to Chen and colleagues, opioid overdose deaths are projected to increase over the next several years, while further reduction in the incidence of prescription opioid misuse is estimated to decrease overdose deaths by only 3% to 5.3%. In the context of recent surges in synthetic opioid use, it is projected that 80% of overdose deaths between 2016 and 2025 will be attributable to illicit opioids.³ Such predictions underscore the urgent need to adopt alternative approaches to risk-reducing measures and policy change.

The increased risk of mortality associated with opioid misuse and overdose is well established in the current literature. However, less is known regarding the rate of ORAEs after survival of an NFO. Olfson and colleagues sought to address this knowledge gap by characterizing mortality risks in 76,325 US adults within 1 year following NFO.¹⁴ Among their studied population, all-cause mortality occurred at a rate of 778.3 per 10,000 person-years, which was 24 times greater than that of the general population. This emphasizes the need for the optimization of mental health services, addiction treatment, and medical care for these individuals at higher risk.

Limitations

Certain factors and limitations should be considered when interpreting the results of this study. Given that the study included only veterans, factors such as the demographic and clinical characteristics more commonly observed among these patients should be taken into account and may in turn limit the generalizability of these findings to nonveteran populations. Another major limitation is the small sample size; the study period and by extension, the number of patients able to be included in the present study were restricted by the availability of retrievable data from automated drug dispensing systems. Patients without documented response to naloxone were excluded from the study due to low clinical suspicion for opioid overdose, although the possibility that the dose administered was too low to produce a robust clinical response cannot be definitively ruled out. The lack of reliable methods to capture events and overdoses treated outside of the VA may have resulted in underestimations of the true occurrence of ORAEs following NFO. Information regarding naloxone administration outside VA facilities, such as in transport to the hospital, self-reported, or bystander administration, was similarly limited by lack of reliable methods for retrieving such data and absence of documentation in VA records. Although all interventions and outcomes reported in the present study occurred within 2 years following NFO, further conclusions pertaining to the relative timing of specific interventions and ORAEs cannot be made. Lastly, this study did not investigate the direct impact of opioid risk mitigation initiatives implemented by the VA in the years coinciding with the study period.

Future Directions

Despite these limitations, an important strength of this study is its ability to identify potential areas for targeted improvement and to guide further efforts relating to the prevention of opioid overdose and opioid-related mortality among veterans. Identification of individuals at high risk for opioid overdose and misuse is an imperative first step that allows for the implementation of downstream risk-mitigating interventions. Within the VA, several tools have been developed in recent years to provide clinicians with additional resources and support in this regard.^6,15

No more than half of those diagnosed with mental health disorders and SUDs in the present study received outpatient follow-up care for these conditions within 6 months following NFO, which may suggest high rates of inadequate treatment. Given the strong association between mental health disorders, SUDs, and increased risk of overdose, increasing engagement with mental health and addiction treatment services may be paramount to preventing subsequent ORAEs, including repeat overdose.^6-9,11

Naloxone kit dispensing represents another area for targeted improvement. Interventions may include clinician education and systematic changes, such as implementing protocols that boost the likelihood of high-risk individuals being provided with naloxone at the earliest opportunity. Bystander-administered naloxone programs can also be considered for increasing naloxone access and reducing opioid-related mortality.¹⁶

Finally, despite evidence supporting the benefit of MAT in OUD treatment and reducing all-cause and opioid-related mortality after NFO, the low rates of MAT observed in this study are consistent with previous reports that these medications remain underutilized.¹⁷ Screening for OUD, in conjunction with increasing access to and utilization of OUD treatment modalities, is an established and integral component of overdose prevention efforts. For VA clinicians, the Psychotropic Drug Safety Initiative (PDSI) dashboard can be used to identify patients diagnosed with OUD who are not yet on MAT.¹⁸ Initiatives to expand MAT access through the ED have the potential to provide life-saving interventions and bridge care in the interim until patients are able to become established with a long-term health care practitioner.¹⁹

Conclusions

This is the first study to describe HSU, medication interventions, and ORAEs among veterans who survive NFO. Studies have shown that veterans with a history of NFO are at increased risk of subsequent AEs and premature death.^6,7,10,14 As such, NFOs represent crucial opportunities to identify high-risk individuals and ensure provision of adequate care. Recent data supports the development of a holistic, multimodal approach focused on adequate treatment of conditions that contribute to opioid-related risks, including mental health disorders, SUDs, pain diagnoses, and medical comorbidities.^3,14 Interventions designed to improve access, engagement, and retention in such care therefore play a pivotal role in overdose prevention and reducing mortality.

Although existing risk mitigation initiatives have improved opioid prescribing and safety within the VA, the findings of this study suggest that there remains room for improvement, and the need for well-coordinated efforts to reduce risks associated with both prescribed and illicit opioid use cannot be overstated. Rates of overdose deaths not only remain high but are projected to continue increasing in coming years, despite advances in clinical practice aimed at reducing harms associated with opioid use. The present findings aim to help identify processes with the potential to reduce rates of overdose, death, and adverse sequelae in high-risk populations. However, future studies are warranted to expand on these findings and contribute to ongoing efforts in reducing opioid-related harms and overdose deaths. This study may provide critical insight to inform further investigations to guide such interventions and highlight tools that health care facilities even outside the VA can consider implementing.

Acknowledgments

The authors would like to thank Jonathan Lacro, PharmD, BCPP, for his guidance with this important clinical topic and navigating IRB submissions.

The number of opioid-related overdose deaths in the United States is estimated to have increased 6-fold over the past 2 decades.¹ In 2017, more than two-thirds of drug overdose deaths involved opioids, yielding a mortality rate of 14.9 per 100,000.² Not only does the opioid epidemic currently pose a significant public health crisis characterized by high morbidity and mortality, but it is also projected to worsen in coming years. According to Chen and colleagues, opioid overdose deaths are estimated to increase by 147% from 2015 to 2025.³ That projects almost 82,000 US deaths annually and > 700,000 deaths in this period—even before accounting for surges in opioid overdoses and opioid-related mortality coinciding with the COVID-19 pandemic.^3,4

As health systems and communities globally struggle with unprecedented losses and stressors introduced by the pandemic, emerging data warrants escalating concerns with regard to increased vulnerability to relapse and overdose among those with mental health and substance use disorders (SUDs). In a recent report, the American Medical Association estimates that opioid-related deaths have increased in more than 40 states with the COVID-19 pandemic.⁴

Veterans are twice as likely to experience a fatal opioid overdose compared with their civilian counterparts.⁵ While several risk mitigation strategies have been employed in recent years to improve opioid prescribing and safety within the US Department of Veterans Affairs (VA), veterans continue to overdose on opioids, both prescribed and obtained illicitly.⁶ Variables shown to be strongly associated with opioid overdose risk include presence of mental health disorders, SUDs, medical conditions involving impaired drug metabolism or excretion, respiratory disorders, higher doses of opioids, concomitant use of sedative medications, and history of overdose.^6-8 Many veterans struggle with chronic pain and those prescribed high doses of opioids were more likely to have comorbid pain diagnoses, mental health disorders, and SUDs.⁹ Dashboards and predictive models, such as the Stratification Tool for Opioid Risk Mitigation (STORM) and the Risk Index for Overdose or Serious Opioid-induced Respiratory Depression (RIOSORD), incorporate such factors to stratify overdose risk among veterans, in an effort to prioritize high-risk individuals for review and provision of care.^6,10,11 Despite recent recognition that overdose prevention likely requires a holistic approach that addresses the biopsychosocial factors contributing to opioid-related morbidity and mortality, it is unclear whether veterans are receiving adequate and appropriate treatment for contributing conditions.

There are currently no existing studies that describe health service utilization (HSU), medication interventions, and rates of opioid-related adverse events (ORAEs) among veterans after survival of a nonfatal opioid overdose (NFO). Clinical characteristics of veterans treated for opioid overdose at a VA emergency department (ED) have previously been described by Clement and Stock.¹² Despite improvements that have been made in VA opioid prescribing and safety, knowledge gaps remain with regard to best practices for opioid overdose prevention. The aim of this study was to characterize HSU and medication interventions in veterans following NFO, as well as the frequency of ORAEs after overdose. The findings of this study may aid in the identification of areas for targeted improvement in the prevention and reduction of opioid overdoses and adverse opioid-related sequelae.

Methods

This retrospective descriptive study was conducted at VA San Diego Healthcare System (VASDHCS) in California. Subjects included were veterans administered naloxone in the ED for suspected opioid overdose between July 1, 2013 and April 1, 2017. The study population was identified through data retrieved from automated drug dispensing systems, which was then confirmed through manual chart review of notes associated with the index ED visit. Inclusion criteria included documented increased respiration or responsiveness following naloxone administration. Subjects were excluded if they demonstrated lack of response to naloxone, overdosed secondary to inpatient administration of opioids, received palliative or hospice care during the study period, or were lost to follow-up.

Data were collected via retrospective chart review and included date of index ED visit, demographics, active prescriptions, urine drug screen (UDS) results, benzodiazepine (BZD) use corroborated by positive UDS or mention of BZD in index visit chart notes, whether overdose was determined to be a suicide attempt, and naloxone kit dispensing. Patient data was collected for 2 years following overdose, including: ORAEs; ED visits; hospitalizations; repeat overdoses; fatal overdose; whether subjects were still alive; follow-up visits for pain management, mental health, and addiction treatment services; and visits to the psychiatric emergency clinic. Clinical characteristics, such as mental health disorder diagnoses, SUDs, and relevant medical conditions also were collected. Statistical analysis was performed using Microsoft Excel and included only descriptive statistics.

Results

Ninety-three patients received naloxone in the VASDHCS ED. Thirty-five met inclusion criteria and were included in the primary analysis. All subjects received IV naloxone with a mean 0.8 mg IV boluses (range, 0.1-4.4 mg).

Most patients were male with a mean age of 59.8 years (Table 1). Almost all overdoses were nonintentional except for 3 suicide attempts that were reviewed by the Suicide Prevention Committee. Three patients had previously been treated for opioid overdose at the VA with a documented positive clinical response to naloxone administration.

Discussions

This retrospective study is representative of many veterans receiving VA care, despite the small sample size. Clinical characteristics observed in the study population were generally consistent with those published by Clement and Stock, including high rates of medical and psychiatric comorbidities.¹² Subjects in both studies were prescribed comparable dosages of opioids; among those prescribed opioids but not BZDs through the VA, the mean MEDD was 117 mg in our study compared with 126 mg in the Clement and Stock study. Since implementation of the Opioid Safety Initiative (OSI) in 2013, opioid prescribing practices have improved nationwide across VA facilities, including successful reduction in the numbers of patients prescribed high-dose opioids and concurrent BZDs.¹³

Despite the tools and resources available to clinicians, discontinuing opioid therapy remains a difficult process. Concerns related to mental health and/or substance-use related decompensations often exist in the setting of rapid dose reductions or abrupt discontinuation of opioids.⁶ Although less than half of patients in the present study with an active opioid prescription at time of index overdose had their opioids discontinued within 2 years, it is reassuring to note the much higher rate of those with subsequent decreases in their prescribed doses, as well as the 50% reduction in BZD coprescribing. Ultimately, these findings remain consistent with the VA goals of mitigating harm, improving opioid prescribing, and ensuring the safe use of opioid medications when clinically appropriate.

Moreover, recent evidence suggests that interventions focused solely on opioid prescribing practices are becoming increasingly limited in their impact on reducing opioid-related deaths and will likely be insufficient for addressing the opioid epidemic as it continues to evolve. According to Chen and colleagues, opioid overdose deaths are projected to increase over the next several years, while further reduction in the incidence of prescription opioid misuse is estimated to decrease overdose deaths by only 3% to 5.3%. In the context of recent surges in synthetic opioid use, it is projected that 80% of overdose deaths between 2016 and 2025 will be attributable to illicit opioids.³ Such predictions underscore the urgent need to adopt alternative approaches to risk-reducing measures and policy change.

The increased risk of mortality associated with opioid misuse and overdose is well established in the current literature. However, less is known regarding the rate of ORAEs after survival of an NFO. Olfson and colleagues sought to address this knowledge gap by characterizing mortality risks in 76,325 US adults within 1 year following NFO.¹⁴ Among their studied population, all-cause mortality occurred at a rate of 778.3 per 10,000 person-years, which was 24 times greater than that of the general population. This emphasizes the need for the optimization of mental health services, addiction treatment, and medical care for these individuals at higher risk.

Limitations

Certain factors and limitations should be considered when interpreting the results of this study. Given that the study included only veterans, factors such as the demographic and clinical characteristics more commonly observed among these patients should be taken into account and may in turn limit the generalizability of these findings to nonveteran populations. Another major limitation is the small sample size; the study period and by extension, the number of patients able to be included in the present study were restricted by the availability of retrievable data from automated drug dispensing systems. Patients without documented response to naloxone were excluded from the study due to low clinical suspicion for opioid overdose, although the possibility that the dose administered was too low to produce a robust clinical response cannot be definitively ruled out. The lack of reliable methods to capture events and overdoses treated outside of the VA may have resulted in underestimations of the true occurrence of ORAEs following NFO. Information regarding naloxone administration outside VA facilities, such as in transport to the hospital, self-reported, or bystander administration, was similarly limited by lack of reliable methods for retrieving such data and absence of documentation in VA records. Although all interventions and outcomes reported in the present study occurred within 2 years following NFO, further conclusions pertaining to the relative timing of specific interventions and ORAEs cannot be made. Lastly, this study did not investigate the direct impact of opioid risk mitigation initiatives implemented by the VA in the years coinciding with the study period.

Future Directions

Despite these limitations, an important strength of this study is its ability to identify potential areas for targeted improvement and to guide further efforts relating to the prevention of opioid overdose and opioid-related mortality among veterans. Identification of individuals at high risk for opioid overdose and misuse is an imperative first step that allows for the implementation of downstream risk-mitigating interventions. Within the VA, several tools have been developed in recent years to provide clinicians with additional resources and support in this regard.^6,15

No more than half of those diagnosed with mental health disorders and SUDs in the present study received outpatient follow-up care for these conditions within 6 months following NFO, which may suggest high rates of inadequate treatment. Given the strong association between mental health disorders, SUDs, and increased risk of overdose, increasing engagement with mental health and addiction treatment services may be paramount to preventing subsequent ORAEs, including repeat overdose.^6-9,11

Naloxone kit dispensing represents another area for targeted improvement. Interventions may include clinician education and systematic changes, such as implementing protocols that boost the likelihood of high-risk individuals being provided with naloxone at the earliest opportunity. Bystander-administered naloxone programs can also be considered for increasing naloxone access and reducing opioid-related mortality.¹⁶

Finally, despite evidence supporting the benefit of MAT in OUD treatment and reducing all-cause and opioid-related mortality after NFO, the low rates of MAT observed in this study are consistent with previous reports that these medications remain underutilized.¹⁷ Screening for OUD, in conjunction with increasing access to and utilization of OUD treatment modalities, is an established and integral component of overdose prevention efforts. For VA clinicians, the Psychotropic Drug Safety Initiative (PDSI) dashboard can be used to identify patients diagnosed with OUD who are not yet on MAT.¹⁸ Initiatives to expand MAT access through the ED have the potential to provide life-saving interventions and bridge care in the interim until patients are able to become established with a long-term health care practitioner.¹⁹

Conclusions

This is the first study to describe HSU, medication interventions, and ORAEs among veterans who survive NFO. Studies have shown that veterans with a history of NFO are at increased risk of subsequent AEs and premature death.^6,7,10,14 As such, NFOs represent crucial opportunities to identify high-risk individuals and ensure provision of adequate care. Recent data supports the development of a holistic, multimodal approach focused on adequate treatment of conditions that contribute to opioid-related risks, including mental health disorders, SUDs, pain diagnoses, and medical comorbidities.^3,14 Interventions designed to improve access, engagement, and retention in such care therefore play a pivotal role in overdose prevention and reducing mortality.

Although existing risk mitigation initiatives have improved opioid prescribing and safety within the VA, the findings of this study suggest that there remains room for improvement, and the need for well-coordinated efforts to reduce risks associated with both prescribed and illicit opioid use cannot be overstated. Rates of overdose deaths not only remain high but are projected to continue increasing in coming years, despite advances in clinical practice aimed at reducing harms associated with opioid use. The present findings aim to help identify processes with the potential to reduce rates of overdose, death, and adverse sequelae in high-risk populations. However, future studies are warranted to expand on these findings and contribute to ongoing efforts in reducing opioid-related harms and overdose deaths. This study may provide critical insight to inform further investigations to guide such interventions and highlight tools that health care facilities even outside the VA can consider implementing.

Acknowledgments

The authors would like to thank Jonathan Lacro, PharmD, BCPP, for his guidance with this important clinical topic and navigating IRB submissions.

The number of opioid-related overdose deaths in the United States is estimated to have increased 6-fold over the past 2 decades.¹ In 2017, more than two-thirds of drug overdose deaths involved opioids, yielding a mortality rate of 14.9 per 100,000.² Not only does the opioid epidemic currently pose a significant public health crisis characterized by high morbidity and mortality, but it is also projected to worsen in coming years. According to Chen and colleagues, opioid overdose deaths are estimated to increase by 147% from 2015 to 2025.³ That projects almost 82,000 US deaths annually and > 700,000 deaths in this period—even before accounting for surges in opioid overdoses and opioid-related mortality coinciding with the COVID-19 pandemic.^3,4

As health systems and communities globally struggle with unprecedented losses and stressors introduced by the pandemic, emerging data warrants escalating concerns with regard to increased vulnerability to relapse and overdose among those with mental health and substance use disorders (SUDs). In a recent report, the American Medical Association estimates that opioid-related deaths have increased in more than 40 states with the COVID-19 pandemic.⁴

Veterans are twice as likely to experience a fatal opioid overdose compared with their civilian counterparts.⁵ While several risk mitigation strategies have been employed in recent years to improve opioid prescribing and safety within the US Department of Veterans Affairs (VA), veterans continue to overdose on opioids, both prescribed and obtained illicitly.⁶ Variables shown to be strongly associated with opioid overdose risk include presence of mental health disorders, SUDs, medical conditions involving impaired drug metabolism or excretion, respiratory disorders, higher doses of opioids, concomitant use of sedative medications, and history of overdose.^6-8 Many veterans struggle with chronic pain and those prescribed high doses of opioids were more likely to have comorbid pain diagnoses, mental health disorders, and SUDs.⁹ Dashboards and predictive models, such as the Stratification Tool for Opioid Risk Mitigation (STORM) and the Risk Index for Overdose or Serious Opioid-induced Respiratory Depression (RIOSORD), incorporate such factors to stratify overdose risk among veterans, in an effort to prioritize high-risk individuals for review and provision of care.^6,10,11 Despite recent recognition that overdose prevention likely requires a holistic approach that addresses the biopsychosocial factors contributing to opioid-related morbidity and mortality, it is unclear whether veterans are receiving adequate and appropriate treatment for contributing conditions.

There are currently no existing studies that describe health service utilization (HSU), medication interventions, and rates of opioid-related adverse events (ORAEs) among veterans after survival of a nonfatal opioid overdose (NFO). Clinical characteristics of veterans treated for opioid overdose at a VA emergency department (ED) have previously been described by Clement and Stock.¹² Despite improvements that have been made in VA opioid prescribing and safety, knowledge gaps remain with regard to best practices for opioid overdose prevention. The aim of this study was to characterize HSU and medication interventions in veterans following NFO, as well as the frequency of ORAEs after overdose. The findings of this study may aid in the identification of areas for targeted improvement in the prevention and reduction of opioid overdoses and adverse opioid-related sequelae.

Methods

This retrospective descriptive study was conducted at VA San Diego Healthcare System (VASDHCS) in California. Subjects included were veterans administered naloxone in the ED for suspected opioid overdose between July 1, 2013 and April 1, 2017. The study population was identified through data retrieved from automated drug dispensing systems, which was then confirmed through manual chart review of notes associated with the index ED visit. Inclusion criteria included documented increased respiration or responsiveness following naloxone administration. Subjects were excluded if they demonstrated lack of response to naloxone, overdosed secondary to inpatient administration of opioids, received palliative or hospice care during the study period, or were lost to follow-up.

Data were collected via retrospective chart review and included date of index ED visit, demographics, active prescriptions, urine drug screen (UDS) results, benzodiazepine (BZD) use corroborated by positive UDS or mention of BZD in index visit chart notes, whether overdose was determined to be a suicide attempt, and naloxone kit dispensing. Patient data was collected for 2 years following overdose, including: ORAEs; ED visits; hospitalizations; repeat overdoses; fatal overdose; whether subjects were still alive; follow-up visits for pain management, mental health, and addiction treatment services; and visits to the psychiatric emergency clinic. Clinical characteristics, such as mental health disorder diagnoses, SUDs, and relevant medical conditions also were collected. Statistical analysis was performed using Microsoft Excel and included only descriptive statistics.

Results

Ninety-three patients received naloxone in the VASDHCS ED. Thirty-five met inclusion criteria and were included in the primary analysis. All subjects received IV naloxone with a mean 0.8 mg IV boluses (range, 0.1-4.4 mg).

Most patients were male with a mean age of 59.8 years (Table 1). Almost all overdoses were nonintentional except for 3 suicide attempts that were reviewed by the Suicide Prevention Committee. Three patients had previously been treated for opioid overdose at the VA with a documented positive clinical response to naloxone administration.

Discussions

This retrospective study is representative of many veterans receiving VA care, despite the small sample size. Clinical characteristics observed in the study population were generally consistent with those published by Clement and Stock, including high rates of medical and psychiatric comorbidities.¹² Subjects in both studies were prescribed comparable dosages of opioids; among those prescribed opioids but not BZDs through the VA, the mean MEDD was 117 mg in our study compared with 126 mg in the Clement and Stock study. Since implementation of the Opioid Safety Initiative (OSI) in 2013, opioid prescribing practices have improved nationwide across VA facilities, including successful reduction in the numbers of patients prescribed high-dose opioids and concurrent BZDs.¹³

Despite the tools and resources available to clinicians, discontinuing opioid therapy remains a difficult process. Concerns related to mental health and/or substance-use related decompensations often exist in the setting of rapid dose reductions or abrupt discontinuation of opioids.⁶ Although less than half of patients in the present study with an active opioid prescription at time of index overdose had their opioids discontinued within 2 years, it is reassuring to note the much higher rate of those with subsequent decreases in their prescribed doses, as well as the 50% reduction in BZD coprescribing. Ultimately, these findings remain consistent with the VA goals of mitigating harm, improving opioid prescribing, and ensuring the safe use of opioid medications when clinically appropriate.

Moreover, recent evidence suggests that interventions focused solely on opioid prescribing practices are becoming increasingly limited in their impact on reducing opioid-related deaths and will likely be insufficient for addressing the opioid epidemic as it continues to evolve. According to Chen and colleagues, opioid overdose deaths are projected to increase over the next several years, while further reduction in the incidence of prescription opioid misuse is estimated to decrease overdose deaths by only 3% to 5.3%. In the context of recent surges in synthetic opioid use, it is projected that 80% of overdose deaths between 2016 and 2025 will be attributable to illicit opioids.³ Such predictions underscore the urgent need to adopt alternative approaches to risk-reducing measures and policy change.

The increased risk of mortality associated with opioid misuse and overdose is well established in the current literature. However, less is known regarding the rate of ORAEs after survival of an NFO. Olfson and colleagues sought to address this knowledge gap by characterizing mortality risks in 76,325 US adults within 1 year following NFO.¹⁴ Among their studied population, all-cause mortality occurred at a rate of 778.3 per 10,000 person-years, which was 24 times greater than that of the general population. This emphasizes the need for the optimization of mental health services, addiction treatment, and medical care for these individuals at higher risk.

Limitations

Certain factors and limitations should be considered when interpreting the results of this study. Given that the study included only veterans, factors such as the demographic and clinical characteristics more commonly observed among these patients should be taken into account and may in turn limit the generalizability of these findings to nonveteran populations. Another major limitation is the small sample size; the study period and by extension, the number of patients able to be included in the present study were restricted by the availability of retrievable data from automated drug dispensing systems. Patients without documented response to naloxone were excluded from the study due to low clinical suspicion for opioid overdose, although the possibility that the dose administered was too low to produce a robust clinical response cannot be definitively ruled out. The lack of reliable methods to capture events and overdoses treated outside of the VA may have resulted in underestimations of the true occurrence of ORAEs following NFO. Information regarding naloxone administration outside VA facilities, such as in transport to the hospital, self-reported, or bystander administration, was similarly limited by lack of reliable methods for retrieving such data and absence of documentation in VA records. Although all interventions and outcomes reported in the present study occurred within 2 years following NFO, further conclusions pertaining to the relative timing of specific interventions and ORAEs cannot be made. Lastly, this study did not investigate the direct impact of opioid risk mitigation initiatives implemented by the VA in the years coinciding with the study period.

Future Directions

Despite these limitations, an important strength of this study is its ability to identify potential areas for targeted improvement and to guide further efforts relating to the prevention of opioid overdose and opioid-related mortality among veterans. Identification of individuals at high risk for opioid overdose and misuse is an imperative first step that allows for the implementation of downstream risk-mitigating interventions. Within the VA, several tools have been developed in recent years to provide clinicians with additional resources and support in this regard.^6,15

No more than half of those diagnosed with mental health disorders and SUDs in the present study received outpatient follow-up care for these conditions within 6 months following NFO, which may suggest high rates of inadequate treatment. Given the strong association between mental health disorders, SUDs, and increased risk of overdose, increasing engagement with mental health and addiction treatment services may be paramount to preventing subsequent ORAEs, including repeat overdose.^6-9,11

Naloxone kit dispensing represents another area for targeted improvement. Interventions may include clinician education and systematic changes, such as implementing protocols that boost the likelihood of high-risk individuals being provided with naloxone at the earliest opportunity. Bystander-administered naloxone programs can also be considered for increasing naloxone access and reducing opioid-related mortality.¹⁶

Finally, despite evidence supporting the benefit of MAT in OUD treatment and reducing all-cause and opioid-related mortality after NFO, the low rates of MAT observed in this study are consistent with previous reports that these medications remain underutilized.¹⁷ Screening for OUD, in conjunction with increasing access to and utilization of OUD treatment modalities, is an established and integral component of overdose prevention efforts. For VA clinicians, the Psychotropic Drug Safety Initiative (PDSI) dashboard can be used to identify patients diagnosed with OUD who are not yet on MAT.¹⁸ Initiatives to expand MAT access through the ED have the potential to provide life-saving interventions and bridge care in the interim until patients are able to become established with a long-term health care practitioner.¹⁹

Conclusions

This is the first study to describe HSU, medication interventions, and ORAEs among veterans who survive NFO. Studies have shown that veterans with a history of NFO are at increased risk of subsequent AEs and premature death.^6,7,10,14 As such, NFOs represent crucial opportunities to identify high-risk individuals and ensure provision of adequate care. Recent data supports the development of a holistic, multimodal approach focused on adequate treatment of conditions that contribute to opioid-related risks, including mental health disorders, SUDs, pain diagnoses, and medical comorbidities.^3,14 Interventions designed to improve access, engagement, and retention in such care therefore play a pivotal role in overdose prevention and reducing mortality.

Although existing risk mitigation initiatives have improved opioid prescribing and safety within the VA, the findings of this study suggest that there remains room for improvement, and the need for well-coordinated efforts to reduce risks associated with both prescribed and illicit opioid use cannot be overstated. Rates of overdose deaths not only remain high but are projected to continue increasing in coming years, despite advances in clinical practice aimed at reducing harms associated with opioid use. The present findings aim to help identify processes with the potential to reduce rates of overdose, death, and adverse sequelae in high-risk populations. However, future studies are warranted to expand on these findings and contribute to ongoing efforts in reducing opioid-related harms and overdose deaths. This study may provide critical insight to inform further investigations to guide such interventions and highlight tools that health care facilities even outside the VA can consider implementing.

Acknowledgments

The authors would like to thank Jonathan Lacro, PharmD, BCPP, for his guidance with this important clinical topic and navigating IRB submissions.

Nontuberculous Mycobacterium (NTM) is a ubiquitous organism known to cause a variety of infections in susceptible hosts; however, pulmonary infection is the most common. Mycobacterium avium complex (MAC) is the most prevalent cause of NTM-related pulmonary disease (NTM-PD) and is associated with underlying structural lung disease, such as chronic obstructive pulmonary disease (COPD) and noncystic fibrosis bronchiectasis.^1-3

Diagnosis of NTM-PD requires (1) symptoms or radiographic abnormality; and (2) at least 2 sputum cultures positive with the same organism or at least 1 positive culture result on bronchoscopy (wash, lavage, or biopsy).¹ Notably, the natural history of untreated NTM-PD varies, though even mild disease may progress substantially.^4-6 Progressive disease is more likely to occur in those with a positive smear or more extensive radiographic findings at the initial diagnosis.⁷ A nationwide Medicare-based study showed that patients with NTM-PD had a higher rate of all-cause mortality than did patients without NTM-PD.⁸ In a study of 123 patients from Taiwan with MAC-PD, lack of treatment was an independent predictor of mortality.⁹ Given the risk of progressive morbidity and mortality, recent guidelines recommend initiation of a susceptibility driven, macrolide-based, 3-drug treatment regimen over watchful waiting.¹⁰

MAC-PD is increasingly recognized among US veterans.^11,12 The Jesse Brown Veterans Affairs Medical Center (JBVAMC) in south/west Chicago serves a large, predominantly Black male population of veterans many of whom are socioeconomically underresourced, and half are aged ≥ 65 years. We observed that initiation of guideline-directed therapy in veterans with MAC-PD at JBVAMC varied among health care professionals (HCPs) in the pulmonary clinic. Therefore, the purpose of this retrospective study was to describe and compare the characteristics of veterans without HIV were diagnosed with MAC-PD and managed at JBVAMC.

Methods

The hospital microbiology department identified veterans diagnosed with NTM at JBVAMC between October 2008 and July 2019. Veterans included in the study were considered to have MAC-PD per American Thoracic Society (ATS)/Infectious Diseases Society of America (ISDA) guidelines and those diagnosed with HIV were excluded from analysis. The electronic health record (EHR) was queried for pertinent demographics, smoking history, comorbidities, and symptoms at the time of a positive mycobacterial culture. Computed tomography (CT) and pulmonary function tests (PFTs) performed within 1 year of diagnosis were included. PFTs were assessed in accordance with Global Initiative for Obstructive Lung Disease (GOLD) criteria, with normal forced expiratory volume in 1 second (FEV₁) and forced vital capacity (FVC) values defined as ≥ 80% and a normal FEV₁/FVC ratio defined as ≥ 70. The diffusion capacity of lung for carbon monoxide (DLCO) was assessed per 2017 European Respiratory Society (ERS) technical standards and was considered reduced if below the lower limit of normal.¹³ Information regarding treatment decisions, initiation, and cessation were collected. All-cause mortality was recorded if available in the EHR at the time of data collection.

Statistical analysis was performed using Mann-Whitney U and Fisher exact tests where appropriate. P < .05 was considered statistically significant. The study was approved by the JBVAMC Institutional Review Board.

Results

We identified 43 veterans who had a positive culture for MAC; however, only 19 veterans met the diagnostic criteria for MAC-PD and were included in the study (Table). The cohort included predominantly Black and male veterans with a median age of 74 years at time of diagnosis (range, 45-92). Sixteen veterans had underlying lung disease (84.2%), and 16 (84.2%) were current or former smokers. Common comorbidities included COPD, obstructive sleep apnea, gastroesophageal reflux disease, and lung cancer. Respiratory symptoms were reported in 17 veterans (89.5%), 15 (78.9%) had a chronic cough, and 10 (52.6%) had dyspnea. Fifteen veterans had a chest CT scan within 1 year of diagnosis: A nodular and tree-in-bud pattern was most commonly found in 13 (86.7%) of veterans. Thirteen veterans had PFTs within 1 year of MAC-PD diagnosis, of whom 6 had a restrictive pattern with percent predicted FVC < 80%, and 9 had evidence of obstruction with FEV₁/FVC < 70. DLCO was below the lower limit of normal in 18 veterans. Finally, 6 veterans were deceased at the time of the study.

Of the 19 veterans, guideline-directed, combination antimycobacterial therapy for MAC-PD was initiated in only 10 (52.6%) patients due to presence of symptoms and/or imaging abnormalities. Treatment was deferred due to improved symptoms, concern for adverse events (AEs), or lost to follow-up. Five veterans stopped treatment prematurely due to AEs, lost to follow-up, or all-cause mortality. Assessment of differences between treated and untreated groups revealed no significant difference in race, sex, age, body mass index (BMI), symptom presence, or chest CT abnormalities. There was no statistically significant difference in all-cause mortality (40% and 22.2% in treated and untreated group, respectively).

To further understand the differences of this cohort, the 13 veterans alive at time of the study were compared with the 6 who had since died of all-cause mortality. No statistically significant differences were found.

Discussion

Consistent with previous reports in the literature, veterans in our cohort were predominantly current or former smoking males with underlying COPD and bronchiectasis.^1-3,11,12 Chest CT findings varied: Most veterans presented not only with nodules and tree-in-bud opacities, but also a high frequency of fibrosis and emphysema. PFTs revealed a variety of obstruction and restrictive patterns, and most veterans had a reduced DLCO, though it is unclear whether this is reflective of underlying emphysema, fibrosis, or an alternative cardiopulmonary disease.^13,14

While underlying structural lung disease may have been a risk factor for MAC-PD in this cohort, the contribution of environmental and domiciliary factors in metropolitan Chicago neighborhoods is unknown. JBVAMC serves an underresourced population who live in the west and south Chicago neighborhoods. Household factors, ambient and indoor air pollution, and potential contamination of the water supply and surface soil may contribute to the prevalence of MAC-PD in this group.^15-19 Further studies are warranted to characterize MAC-PD and its treatment in veterans without HIV who reside in underresourced urban communities in the US.

Recent ATS, European Society of Clinical Microbiology and Infectious Diseases, and IDSA guidelines recommend combination antimycobacterial therapy for patients who meet clinical, radiographic, and microbiologic criteria for the diagnosis of MAC-PD.¹⁰ Patients who meet these diagnostic criteria, particularly patients with smear positivity or fibrocavitary disease, should be treated because of risk of unfavorable outcomes.^15,20-22 However, we found that the initiation of guideline-recommended antimycobacterial therapy in veterans without HIV with MAC-PD were inconsistent among HCPs. The reasons underlying this phenomenon were not apparent beyond cited reasons for treatment initiation or deference. Despite this inconsistency, there was no clear difference in age, BMI, symptom burden, radiographic abnormality, or all-cause mortality between treatment groups. Existing studies support slow but substantial progression of untreated MAC-PD, and while treatment prevents deterioration of the disease, it does not prevent progression of bronchiectasis.⁶ The natural history of MAC-PD in this veteran cohort has yet to be fully elucidated. Furthermore, the 50% treatment dropout rate was higher than previously reported rates (11-33%).⁵ However, the small number of veterans in this study precludes meaningful comparison with similar reports in the literature.

We did note a relatively high all-cause mortality in this cohort (n = 6, 32%); however, this rate is comparable to the all-cause mortality rate of 27% observed in a 2018 meta-analysis of 9035 patients with MAC-PD.²³ Although there was no major difference in those deceased and those alive at the time of data collection in our study, previously described predictors of mortality included male sex, advanced age, presence of fibrocavitary disease, decreased FVC, and presence of comorbidities.^8,23 Larger prospective studies evaluating veterans with MAC-PD are needed to further evaluate contributors to mortality in veterans with MAC-PD.

Limitations

The limitations of this small, single-center, retrospective study prevent a robust, generalizable comparison between groups. Further studies are warranted to characterize MAC-PD and its treatment in veterans without HIV who reside in underresourced urban communities in the US.^24-26

Conclusions

These data suggest that clinical, imaging, and treatment attributes of MAC-PD in veterans without HIV who reside in metropolitan Chicago are heterogeneous and are associated with a relatively high mortality rate. Although there was no difference in the attributes or outcomes of veterans who did and did not initiate treatment despite current recommendations, further studies are needed to better explore these relationships.

Nontuberculous Mycobacterium (NTM) is a ubiquitous organism known to cause a variety of infections in susceptible hosts; however, pulmonary infection is the most common. Mycobacterium avium complex (MAC) is the most prevalent cause of NTM-related pulmonary disease (NTM-PD) and is associated with underlying structural lung disease, such as chronic obstructive pulmonary disease (COPD) and noncystic fibrosis bronchiectasis.^1-3

Diagnosis of NTM-PD requires (1) symptoms or radiographic abnormality; and (2) at least 2 sputum cultures positive with the same organism or at least 1 positive culture result on bronchoscopy (wash, lavage, or biopsy).¹ Notably, the natural history of untreated NTM-PD varies, though even mild disease may progress substantially.^4-6 Progressive disease is more likely to occur in those with a positive smear or more extensive radiographic findings at the initial diagnosis.⁷ A nationwide Medicare-based study showed that patients with NTM-PD had a higher rate of all-cause mortality than did patients without NTM-PD.⁸ In a study of 123 patients from Taiwan with MAC-PD, lack of treatment was an independent predictor of mortality.⁹ Given the risk of progressive morbidity and mortality, recent guidelines recommend initiation of a susceptibility driven, macrolide-based, 3-drug treatment regimen over watchful waiting.¹⁰

MAC-PD is increasingly recognized among US veterans.^11,12 The Jesse Brown Veterans Affairs Medical Center (JBVAMC) in south/west Chicago serves a large, predominantly Black male population of veterans many of whom are socioeconomically underresourced, and half are aged ≥ 65 years. We observed that initiation of guideline-directed therapy in veterans with MAC-PD at JBVAMC varied among health care professionals (HCPs) in the pulmonary clinic. Therefore, the purpose of this retrospective study was to describe and compare the characteristics of veterans without HIV were diagnosed with MAC-PD and managed at JBVAMC.

Methods

The hospital microbiology department identified veterans diagnosed with NTM at JBVAMC between October 2008 and July 2019. Veterans included in the study were considered to have MAC-PD per American Thoracic Society (ATS)/Infectious Diseases Society of America (ISDA) guidelines and those diagnosed with HIV were excluded from analysis. The electronic health record (EHR) was queried for pertinent demographics, smoking history, comorbidities, and symptoms at the time of a positive mycobacterial culture. Computed tomography (CT) and pulmonary function tests (PFTs) performed within 1 year of diagnosis were included. PFTs were assessed in accordance with Global Initiative for Obstructive Lung Disease (GOLD) criteria, with normal forced expiratory volume in 1 second (FEV₁) and forced vital capacity (FVC) values defined as ≥ 80% and a normal FEV₁/FVC ratio defined as ≥ 70. The diffusion capacity of lung for carbon monoxide (DLCO) was assessed per 2017 European Respiratory Society (ERS) technical standards and was considered reduced if below the lower limit of normal.¹³ Information regarding treatment decisions, initiation, and cessation were collected. All-cause mortality was recorded if available in the EHR at the time of data collection.

Statistical analysis was performed using Mann-Whitney U and Fisher exact tests where appropriate. P < .05 was considered statistically significant. The study was approved by the JBVAMC Institutional Review Board.

Results

We identified 43 veterans who had a positive culture for MAC; however, only 19 veterans met the diagnostic criteria for MAC-PD and were included in the study (Table). The cohort included predominantly Black and male veterans with a median age of 74 years at time of diagnosis (range, 45-92). Sixteen veterans had underlying lung disease (84.2%), and 16 (84.2%) were current or former smokers. Common comorbidities included COPD, obstructive sleep apnea, gastroesophageal reflux disease, and lung cancer. Respiratory symptoms were reported in 17 veterans (89.5%), 15 (78.9%) had a chronic cough, and 10 (52.6%) had dyspnea. Fifteen veterans had a chest CT scan within 1 year of diagnosis: A nodular and tree-in-bud pattern was most commonly found in 13 (86.7%) of veterans. Thirteen veterans had PFTs within 1 year of MAC-PD diagnosis, of whom 6 had a restrictive pattern with percent predicted FVC < 80%, and 9 had evidence of obstruction with FEV₁/FVC < 70. DLCO was below the lower limit of normal in 18 veterans. Finally, 6 veterans were deceased at the time of the study.

Of the 19 veterans, guideline-directed, combination antimycobacterial therapy for MAC-PD was initiated in only 10 (52.6%) patients due to presence of symptoms and/or imaging abnormalities. Treatment was deferred due to improved symptoms, concern for adverse events (AEs), or lost to follow-up. Five veterans stopped treatment prematurely due to AEs, lost to follow-up, or all-cause mortality. Assessment of differences between treated and untreated groups revealed no significant difference in race, sex, age, body mass index (BMI), symptom presence, or chest CT abnormalities. There was no statistically significant difference in all-cause mortality (40% and 22.2% in treated and untreated group, respectively).

To further understand the differences of this cohort, the 13 veterans alive at time of the study were compared with the 6 who had since died of all-cause mortality. No statistically significant differences were found.

Discussion

Consistent with previous reports in the literature, veterans in our cohort were predominantly current or former smoking males with underlying COPD and bronchiectasis.^1-3,11,12 Chest CT findings varied: Most veterans presented not only with nodules and tree-in-bud opacities, but also a high frequency of fibrosis and emphysema. PFTs revealed a variety of obstruction and restrictive patterns, and most veterans had a reduced DLCO, though it is unclear whether this is reflective of underlying emphysema, fibrosis, or an alternative cardiopulmonary disease.^13,14

While underlying structural lung disease may have been a risk factor for MAC-PD in this cohort, the contribution of environmental and domiciliary factors in metropolitan Chicago neighborhoods is unknown. JBVAMC serves an underresourced population who live in the west and south Chicago neighborhoods. Household factors, ambient and indoor air pollution, and potential contamination of the water supply and surface soil may contribute to the prevalence of MAC-PD in this group.^15-19 Further studies are warranted to characterize MAC-PD and its treatment in veterans without HIV who reside in underresourced urban communities in the US.

Recent ATS, European Society of Clinical Microbiology and Infectious Diseases, and IDSA guidelines recommend combination antimycobacterial therapy for patients who meet clinical, radiographic, and microbiologic criteria for the diagnosis of MAC-PD.¹⁰ Patients who meet these diagnostic criteria, particularly patients with smear positivity or fibrocavitary disease, should be treated because of risk of unfavorable outcomes.^15,20-22 However, we found that the initiation of guideline-recommended antimycobacterial therapy in veterans without HIV with MAC-PD were inconsistent among HCPs. The reasons underlying this phenomenon were not apparent beyond cited reasons for treatment initiation or deference. Despite this inconsistency, there was no clear difference in age, BMI, symptom burden, radiographic abnormality, or all-cause mortality between treatment groups. Existing studies support slow but substantial progression of untreated MAC-PD, and while treatment prevents deterioration of the disease, it does not prevent progression of bronchiectasis.⁶ The natural history of MAC-PD in this veteran cohort has yet to be fully elucidated. Furthermore, the 50% treatment dropout rate was higher than previously reported rates (11-33%).⁵ However, the small number of veterans in this study precludes meaningful comparison with similar reports in the literature.

We did note a relatively high all-cause mortality in this cohort (n = 6, 32%); however, this rate is comparable to the all-cause mortality rate of 27% observed in a 2018 meta-analysis of 9035 patients with MAC-PD.²³ Although there was no major difference in those deceased and those alive at the time of data collection in our study, previously described predictors of mortality included male sex, advanced age, presence of fibrocavitary disease, decreased FVC, and presence of comorbidities.^8,23 Larger prospective studies evaluating veterans with MAC-PD are needed to further evaluate contributors to mortality in veterans with MAC-PD.

Limitations

The limitations of this small, single-center, retrospective study prevent a robust, generalizable comparison between groups. Further studies are warranted to characterize MAC-PD and its treatment in veterans without HIV who reside in underresourced urban communities in the US.^24-26

Conclusions

These data suggest that clinical, imaging, and treatment attributes of MAC-PD in veterans without HIV who reside in metropolitan Chicago are heterogeneous and are associated with a relatively high mortality rate. Although there was no difference in the attributes or outcomes of veterans who did and did not initiate treatment despite current recommendations, further studies are needed to better explore these relationships.

Nontuberculous Mycobacterium (NTM) is a ubiquitous organism known to cause a variety of infections in susceptible hosts; however, pulmonary infection is the most common. Mycobacterium avium complex (MAC) is the most prevalent cause of NTM-related pulmonary disease (NTM-PD) and is associated with underlying structural lung disease, such as chronic obstructive pulmonary disease (COPD) and noncystic fibrosis bronchiectasis.^1-3

Diagnosis of NTM-PD requires (1) symptoms or radiographic abnormality; and (2) at least 2 sputum cultures positive with the same organism or at least 1 positive culture result on bronchoscopy (wash, lavage, or biopsy).¹ Notably, the natural history of untreated NTM-PD varies, though even mild disease may progress substantially.^4-6 Progressive disease is more likely to occur in those with a positive smear or more extensive radiographic findings at the initial diagnosis.⁷ A nationwide Medicare-based study showed that patients with NTM-PD had a higher rate of all-cause mortality than did patients without NTM-PD.⁸ In a study of 123 patients from Taiwan with MAC-PD, lack of treatment was an independent predictor of mortality.⁹ Given the risk of progressive morbidity and mortality, recent guidelines recommend initiation of a susceptibility driven, macrolide-based, 3-drug treatment regimen over watchful waiting.¹⁰

MAC-PD is increasingly recognized among US veterans.^11,12 The Jesse Brown Veterans Affairs Medical Center (JBVAMC) in south/west Chicago serves a large, predominantly Black male population of veterans many of whom are socioeconomically underresourced, and half are aged ≥ 65 years. We observed that initiation of guideline-directed therapy in veterans with MAC-PD at JBVAMC varied among health care professionals (HCPs) in the pulmonary clinic. Therefore, the purpose of this retrospective study was to describe and compare the characteristics of veterans without HIV were diagnosed with MAC-PD and managed at JBVAMC.

Methods

The hospital microbiology department identified veterans diagnosed with NTM at JBVAMC between October 2008 and July 2019. Veterans included in the study were considered to have MAC-PD per American Thoracic Society (ATS)/Infectious Diseases Society of America (ISDA) guidelines and those diagnosed with HIV were excluded from analysis. The electronic health record (EHR) was queried for pertinent demographics, smoking history, comorbidities, and symptoms at the time of a positive mycobacterial culture. Computed tomography (CT) and pulmonary function tests (PFTs) performed within 1 year of diagnosis were included. PFTs were assessed in accordance with Global Initiative for Obstructive Lung Disease (GOLD) criteria, with normal forced expiratory volume in 1 second (FEV₁) and forced vital capacity (FVC) values defined as ≥ 80% and a normal FEV₁/FVC ratio defined as ≥ 70. The diffusion capacity of lung for carbon monoxide (DLCO) was assessed per 2017 European Respiratory Society (ERS) technical standards and was considered reduced if below the lower limit of normal.¹³ Information regarding treatment decisions, initiation, and cessation were collected. All-cause mortality was recorded if available in the EHR at the time of data collection.

Statistical analysis was performed using Mann-Whitney U and Fisher exact tests where appropriate. P < .05 was considered statistically significant. The study was approved by the JBVAMC Institutional Review Board.

Results

We identified 43 veterans who had a positive culture for MAC; however, only 19 veterans met the diagnostic criteria for MAC-PD and were included in the study (Table). The cohort included predominantly Black and male veterans with a median age of 74 years at time of diagnosis (range, 45-92). Sixteen veterans had underlying lung disease (84.2%), and 16 (84.2%) were current or former smokers. Common comorbidities included COPD, obstructive sleep apnea, gastroesophageal reflux disease, and lung cancer. Respiratory symptoms were reported in 17 veterans (89.5%), 15 (78.9%) had a chronic cough, and 10 (52.6%) had dyspnea. Fifteen veterans had a chest CT scan within 1 year of diagnosis: A nodular and tree-in-bud pattern was most commonly found in 13 (86.7%) of veterans. Thirteen veterans had PFTs within 1 year of MAC-PD diagnosis, of whom 6 had a restrictive pattern with percent predicted FVC < 80%, and 9 had evidence of obstruction with FEV₁/FVC < 70. DLCO was below the lower limit of normal in 18 veterans. Finally, 6 veterans were deceased at the time of the study.

Of the 19 veterans, guideline-directed, combination antimycobacterial therapy for MAC-PD was initiated in only 10 (52.6%) patients due to presence of symptoms and/or imaging abnormalities. Treatment was deferred due to improved symptoms, concern for adverse events (AEs), or lost to follow-up. Five veterans stopped treatment prematurely due to AEs, lost to follow-up, or all-cause mortality. Assessment of differences between treated and untreated groups revealed no significant difference in race, sex, age, body mass index (BMI), symptom presence, or chest CT abnormalities. There was no statistically significant difference in all-cause mortality (40% and 22.2% in treated and untreated group, respectively).

To further understand the differences of this cohort, the 13 veterans alive at time of the study were compared with the 6 who had since died of all-cause mortality. No statistically significant differences were found.

Discussion

Consistent with previous reports in the literature, veterans in our cohort were predominantly current or former smoking males with underlying COPD and bronchiectasis.^1-3,11,12 Chest CT findings varied: Most veterans presented not only with nodules and tree-in-bud opacities, but also a high frequency of fibrosis and emphysema. PFTs revealed a variety of obstruction and restrictive patterns, and most veterans had a reduced DLCO, though it is unclear whether this is reflective of underlying emphysema, fibrosis, or an alternative cardiopulmonary disease.^13,14

While underlying structural lung disease may have been a risk factor for MAC-PD in this cohort, the contribution of environmental and domiciliary factors in metropolitan Chicago neighborhoods is unknown. JBVAMC serves an underresourced population who live in the west and south Chicago neighborhoods. Household factors, ambient and indoor air pollution, and potential contamination of the water supply and surface soil may contribute to the prevalence of MAC-PD in this group.^15-19 Further studies are warranted to characterize MAC-PD and its treatment in veterans without HIV who reside in underresourced urban communities in the US.

Recent ATS, European Society of Clinical Microbiology and Infectious Diseases, and IDSA guidelines recommend combination antimycobacterial therapy for patients who meet clinical, radiographic, and microbiologic criteria for the diagnosis of MAC-PD.¹⁰ Patients who meet these diagnostic criteria, particularly patients with smear positivity or fibrocavitary disease, should be treated because of risk of unfavorable outcomes.^15,20-22 However, we found that the initiation of guideline-recommended antimycobacterial therapy in veterans without HIV with MAC-PD were inconsistent among HCPs. The reasons underlying this phenomenon were not apparent beyond cited reasons for treatment initiation or deference. Despite this inconsistency, there was no clear difference in age, BMI, symptom burden, radiographic abnormality, or all-cause mortality between treatment groups. Existing studies support slow but substantial progression of untreated MAC-PD, and while treatment prevents deterioration of the disease, it does not prevent progression of bronchiectasis.⁶ The natural history of MAC-PD in this veteran cohort has yet to be fully elucidated. Furthermore, the 50% treatment dropout rate was higher than previously reported rates (11-33%).⁵ However, the small number of veterans in this study precludes meaningful comparison with similar reports in the literature.

We did note a relatively high all-cause mortality in this cohort (n = 6, 32%); however, this rate is comparable to the all-cause mortality rate of 27% observed in a 2018 meta-analysis of 9035 patients with MAC-PD.²³ Although there was no major difference in those deceased and those alive at the time of data collection in our study, previously described predictors of mortality included male sex, advanced age, presence of fibrocavitary disease, decreased FVC, and presence of comorbidities.^8,23 Larger prospective studies evaluating veterans with MAC-PD are needed to further evaluate contributors to mortality in veterans with MAC-PD.

Limitations

The limitations of this small, single-center, retrospective study prevent a robust, generalizable comparison between groups. Further studies are warranted to characterize MAC-PD and its treatment in veterans without HIV who reside in underresourced urban communities in the US.^24-26

Conclusions

These data suggest that clinical, imaging, and treatment attributes of MAC-PD in veterans without HIV who reside in metropolitan Chicago are heterogeneous and are associated with a relatively high mortality rate. Although there was no difference in the attributes or outcomes of veterans who did and did not initiate treatment despite current recommendations, further studies are needed to better explore these relationships.

As one of 4 statutory missions, the US Department of Veterans Affairs (VA) educates and trains health professionals to enhance the quality of and timely access to care provided to veterans within the Veterans Health Administration (VHA). To achieve its mission to educate, the VA has conducted health professions education and training in partnership with affiliated US academic institutions for the past 76 years in accordance with the landmark 1946 Policy Memorandum No. 2.^1,2

Despite its long-term success affiliating with medical schools, VA has continued to be challenged by physician staff shortages with wide variability in the number and specialty of available health care professionals across facilities.^3,4 A 2020 VA Office of Inspector General report on VHA occupational staffing shortages concluded that numerous physician specialties were difficult to recruit due to a lack of qualified applicants, noncompetitive salary, and less desirable geographic locations.³

Federal health professions scholarship programs and loan repayment programs have long been used to address physician shortages.⁴ Focusing on physician shortages in underserved areas in the US, the Emergency Health Personnel Act of 1970 and its subsequent amendments paved the way for various federal medical school scholarship and loan repayment programs.⁵ Similarly, physician shortages in the armed forces were mitigated through the Uniformed Services Health Professions Revitalization Act of 1972 (USHPRA).^6,7

In 2018, Congress passed the VA MISSION (Maintaining Internal Systems and Strengthening Integrated Outside Networks) Act, which included sections designed to alleviate physician shortages in the VHA.⁸ These sections authorized scholarships similar to those offered by the US Department of Defense (DoD) and loan repayment programs. Section 301 created the Health Professions Scholarship Program (HPSP), which offers scholarships for physicians and dentists. Section 302 increased the maximum debt reduction through the Education Debt Reduction Program (EDRP). Section 303 authorizes the Specialty Education Loan Repayment Program (SELRP), which provides for repayment of educational loans for physicians in specialties deemed necessary for VA. Finally, Section 304 created the Veterans Healing Veterans (VHV), a pilot scholarship specifically for veteran medical students.
 

Program Characteristics

Health Professions Scholarship

The VA HPSP is a program for physicians and dentists that extends from 2020 to 2033. The HPSP funds the costs of tuition, fees, and provides a stipend with a service obligation of 18 months for each year of support. The program is authorized for 10 years and must provide a minimum of 50 scholarships annually for physicians or dentists based on VHA needs. Applications are screened based on criteria that include a commitment to rural or underserved populations, veteran status, grade point average, essays, and letters of recommendation. Although the minimum required number of scholarships annually is 50, VA anticipates providing 1000 scholarships over 10 years with an aim to significantly increase the number physicians at VHA facilities (Table 1).

Veterans Healing Veterans

Implemented in 2020, the VHV was a 1-year pilot program. It offered scholarships to 2 veterans attending medical school at each of the 5 Teague-Cranston and the 4 Historically Black College and University (HBCU) medical schools (Table 2). The intent of the program was to determine the feasibility of increasing the pool of veteran physicians at VHA. Eligible applicants were notified of the scholarship opportunity through the American Medical College Application Service or through the medical school. Applicants must have separated from military service within the preceding 10 years of being admitted to medical school. In exchange for full tuition, fees, a monthly stipend, and rotation travel costs, the recipients accepted a 4-year clinical service obligation at VA facilities after completing their residency training.

Specialty Education Loan Repayment

The SELRP is a loan repayment program available to recently graduated physicians. Applicants must have graduated from an accredited medical or osteopathic school, matched to an accredited residency program and be ≥ 2 years from completion of residency. The specialties qualifying for SELRP are determined through an analysis of succession planning by the VA Office of Workforce Management and Consulting and change based on VA physician workforce needs. The SELRP provides loan repayment in the amount of $40,000 per year for up to 4 years, with a service obligation of 1 year for each $40,000 of support. In April 2021, VA began accepting applications from the eligible specialties of family medicine, internal medicine, gastroenterology, psychiatry, emergency medicine, and geriatrics.

Education Debt Reduction

The EDRP offers debt relief to clinicians in the most difficult to recruit professions, including physicians (generalists and specialists), registered nurses, licensed practical nurses, social workers, and psychologists. The list of difficult to recruit positions is developed annually by VA facilities. Annual reimbursements through the program may be used for tuition and expenses, such as fees, books, supplies, equipment, and other materials. In 2018, through the MISSION Act Section 302, the annual loan repayment was increased from $24,000 to $40,000, and the maximum level of support was increased from $120,000 to $200,000 over 5 years. Recipients receive reimbursement for loan repayment at the end of each year or service period and recipients are not required to remain in VA for 5 years.

Program Results

Health Professions Scholarship

For academic years 2020/2021 and 2021/2022, 126 HPSP applications from both allopathic and osteopathic schools were submitted and 51 scholarships were awarded (Table 3). Assuming an average residency length of 4 years, VHA estimates that these awards will yield 204 service-year equivalents by 2029.

Veterans Healing Veterans

In the VHV program, scholarship recipients came from 5 Teague-Cranston schools; 2 at University of South Carolina, 2 at East Tennessee State University, 2 at Wright State University, 1 at Texas A&M College of Medicine, 1 at Marshall University; and 3 HBCUs; 2 at Howard University, 1 at Morehouse School of Medicine and 1 at Meharry Medical College. The Charles R. Drew University of Medicine and Science did not nominate any students for the scholarship. Assuming all recipients complete postgraduate training, the VHV scholarship program will provide an additional 12 veteran physicians to serve at VA for at least 4 years each (48 service years).

Specialty Education Loan Repayment

Fourteen applicants have been approved, including 5 in psychiatry, 4 in family medicine, 3 in internal medicine, 1 in emergency medicine, and 1 in geriatrics. The mean loan repayment is anticipated to be $110,000 and equating to 38.5 VA service years or a mean of 2.3 years of service obligation per individual for the first cohort. The program has no termination date, and with continued funding, VA anticipates granting 100 loan repayments annually.

Education Debt Reduction

Since 2018, 1,546 VA physicians have received EDRP awards. Due to the increased reimbursement provided through the MISSION Act, average physician award amounts have increased from $96,090 in 2018 to $142,557 in 2019 and $148,302 in 2020.

Conclusions

The VA physician scholarship and loan repayment programs outlined in the MISSION Act build on the success of existing federal scholarship programs by providing opportunities for physician trainees to alleviate educational debt and explore a VA health professions career.

Looking ahead, VA must focus on measuring the success of the MISSION scholarship and loan repayment programs by tracking rates of acceptance and student graduation, residency and fellowship completion, and placement in VA medical facilities—both for the service obligation and future employment. Ultimately, the total impact on VA staffing, especially at rural and underresourced sites, will determine the success of the MISSION programs.

As one of 4 statutory missions, the US Department of Veterans Affairs (VA) educates and trains health professionals to enhance the quality of and timely access to care provided to veterans within the Veterans Health Administration (VHA). To achieve its mission to educate, the VA has conducted health professions education and training in partnership with affiliated US academic institutions for the past 76 years in accordance with the landmark 1946 Policy Memorandum No. 2.^1,2

Despite its long-term success affiliating with medical schools, VA has continued to be challenged by physician staff shortages with wide variability in the number and specialty of available health care professionals across facilities.^3,4 A 2020 VA Office of Inspector General report on VHA occupational staffing shortages concluded that numerous physician specialties were difficult to recruit due to a lack of qualified applicants, noncompetitive salary, and less desirable geographic locations.³

Federal health professions scholarship programs and loan repayment programs have long been used to address physician shortages.⁴ Focusing on physician shortages in underserved areas in the US, the Emergency Health Personnel Act of 1970 and its subsequent amendments paved the way for various federal medical school scholarship and loan repayment programs.⁵ Similarly, physician shortages in the armed forces were mitigated through the Uniformed Services Health Professions Revitalization Act of 1972 (USHPRA).^6,7

In 2018, Congress passed the VA MISSION (Maintaining Internal Systems and Strengthening Integrated Outside Networks) Act, which included sections designed to alleviate physician shortages in the VHA.⁸ These sections authorized scholarships similar to those offered by the US Department of Defense (DoD) and loan repayment programs. Section 301 created the Health Professions Scholarship Program (HPSP), which offers scholarships for physicians and dentists. Section 302 increased the maximum debt reduction through the Education Debt Reduction Program (EDRP). Section 303 authorizes the Specialty Education Loan Repayment Program (SELRP), which provides for repayment of educational loans for physicians in specialties deemed necessary for VA. Finally, Section 304 created the Veterans Healing Veterans (VHV), a pilot scholarship specifically for veteran medical students.
 

Program Characteristics

Health Professions Scholarship

The VA HPSP is a program for physicians and dentists that extends from 2020 to 2033. The HPSP funds the costs of tuition, fees, and provides a stipend with a service obligation of 18 months for each year of support. The program is authorized for 10 years and must provide a minimum of 50 scholarships annually for physicians or dentists based on VHA needs. Applications are screened based on criteria that include a commitment to rural or underserved populations, veteran status, grade point average, essays, and letters of recommendation. Although the minimum required number of scholarships annually is 50, VA anticipates providing 1000 scholarships over 10 years with an aim to significantly increase the number physicians at VHA facilities (Table 1).

Veterans Healing Veterans

Implemented in 2020, the VHV was a 1-year pilot program. It offered scholarships to 2 veterans attending medical school at each of the 5 Teague-Cranston and the 4 Historically Black College and University (HBCU) medical schools (Table 2). The intent of the program was to determine the feasibility of increasing the pool of veteran physicians at VHA. Eligible applicants were notified of the scholarship opportunity through the American Medical College Application Service or through the medical school. Applicants must have separated from military service within the preceding 10 years of being admitted to medical school. In exchange for full tuition, fees, a monthly stipend, and rotation travel costs, the recipients accepted a 4-year clinical service obligation at VA facilities after completing their residency training.

Specialty Education Loan Repayment

The SELRP is a loan repayment program available to recently graduated physicians. Applicants must have graduated from an accredited medical or osteopathic school, matched to an accredited residency program and be ≥ 2 years from completion of residency. The specialties qualifying for SELRP are determined through an analysis of succession planning by the VA Office of Workforce Management and Consulting and change based on VA physician workforce needs. The SELRP provides loan repayment in the amount of $40,000 per year for up to 4 years, with a service obligation of 1 year for each $40,000 of support. In April 2021, VA began accepting applications from the eligible specialties of family medicine, internal medicine, gastroenterology, psychiatry, emergency medicine, and geriatrics.

Education Debt Reduction

The EDRP offers debt relief to clinicians in the most difficult to recruit professions, including physicians (generalists and specialists), registered nurses, licensed practical nurses, social workers, and psychologists. The list of difficult to recruit positions is developed annually by VA facilities. Annual reimbursements through the program may be used for tuition and expenses, such as fees, books, supplies, equipment, and other materials. In 2018, through the MISSION Act Section 302, the annual loan repayment was increased from $24,000 to $40,000, and the maximum level of support was increased from $120,000 to $200,000 over 5 years. Recipients receive reimbursement for loan repayment at the end of each year or service period and recipients are not required to remain in VA for 5 years.

Program Results

Health Professions Scholarship

For academic years 2020/2021 and 2021/2022, 126 HPSP applications from both allopathic and osteopathic schools were submitted and 51 scholarships were awarded (Table 3). Assuming an average residency length of 4 years, VHA estimates that these awards will yield 204 service-year equivalents by 2029.

Veterans Healing Veterans

In the VHV program, scholarship recipients came from 5 Teague-Cranston schools; 2 at University of South Carolina, 2 at East Tennessee State University, 2 at Wright State University, 1 at Texas A&M College of Medicine, 1 at Marshall University; and 3 HBCUs; 2 at Howard University, 1 at Morehouse School of Medicine and 1 at Meharry Medical College. The Charles R. Drew University of Medicine and Science did not nominate any students for the scholarship. Assuming all recipients complete postgraduate training, the VHV scholarship program will provide an additional 12 veteran physicians to serve at VA for at least 4 years each (48 service years).

Specialty Education Loan Repayment

Fourteen applicants have been approved, including 5 in psychiatry, 4 in family medicine, 3 in internal medicine, 1 in emergency medicine, and 1 in geriatrics. The mean loan repayment is anticipated to be $110,000 and equating to 38.5 VA service years or a mean of 2.3 years of service obligation per individual for the first cohort. The program has no termination date, and with continued funding, VA anticipates granting 100 loan repayments annually.

Education Debt Reduction

Since 2018, 1,546 VA physicians have received EDRP awards. Due to the increased reimbursement provided through the MISSION Act, average physician award amounts have increased from $96,090 in 2018 to $142,557 in 2019 and $148,302 in 2020.

Conclusions

The VA physician scholarship and loan repayment programs outlined in the MISSION Act build on the success of existing federal scholarship programs by providing opportunities for physician trainees to alleviate educational debt and explore a VA health professions career.

Looking ahead, VA must focus on measuring the success of the MISSION scholarship and loan repayment programs by tracking rates of acceptance and student graduation, residency and fellowship completion, and placement in VA medical facilities—both for the service obligation and future employment. Ultimately, the total impact on VA staffing, especially at rural and underresourced sites, will determine the success of the MISSION programs.

As one of 4 statutory missions, the US Department of Veterans Affairs (VA) educates and trains health professionals to enhance the quality of and timely access to care provided to veterans within the Veterans Health Administration (VHA). To achieve its mission to educate, the VA has conducted health professions education and training in partnership with affiliated US academic institutions for the past 76 years in accordance with the landmark 1946 Policy Memorandum No. 2.^1,2

Despite its long-term success affiliating with medical schools, VA has continued to be challenged by physician staff shortages with wide variability in the number and specialty of available health care professionals across facilities.^3,4 A 2020 VA Office of Inspector General report on VHA occupational staffing shortages concluded that numerous physician specialties were difficult to recruit due to a lack of qualified applicants, noncompetitive salary, and less desirable geographic locations.³

Federal health professions scholarship programs and loan repayment programs have long been used to address physician shortages.⁴ Focusing on physician shortages in underserved areas in the US, the Emergency Health Personnel Act of 1970 and its subsequent amendments paved the way for various federal medical school scholarship and loan repayment programs.⁵ Similarly, physician shortages in the armed forces were mitigated through the Uniformed Services Health Professions Revitalization Act of 1972 (USHPRA).^6,7

In 2018, Congress passed the VA MISSION (Maintaining Internal Systems and Strengthening Integrated Outside Networks) Act, which included sections designed to alleviate physician shortages in the VHA.⁸ These sections authorized scholarships similar to those offered by the US Department of Defense (DoD) and loan repayment programs. Section 301 created the Health Professions Scholarship Program (HPSP), which offers scholarships for physicians and dentists. Section 302 increased the maximum debt reduction through the Education Debt Reduction Program (EDRP). Section 303 authorizes the Specialty Education Loan Repayment Program (SELRP), which provides for repayment of educational loans for physicians in specialties deemed necessary for VA. Finally, Section 304 created the Veterans Healing Veterans (VHV), a pilot scholarship specifically for veteran medical students.
 

Program Characteristics

Health Professions Scholarship

The VA HPSP is a program for physicians and dentists that extends from 2020 to 2033. The HPSP funds the costs of tuition, fees, and provides a stipend with a service obligation of 18 months for each year of support. The program is authorized for 10 years and must provide a minimum of 50 scholarships annually for physicians or dentists based on VHA needs. Applications are screened based on criteria that include a commitment to rural or underserved populations, veteran status, grade point average, essays, and letters of recommendation. Although the minimum required number of scholarships annually is 50, VA anticipates providing 1000 scholarships over 10 years with an aim to significantly increase the number physicians at VHA facilities (Table 1).

Veterans Healing Veterans

Implemented in 2020, the VHV was a 1-year pilot program. It offered scholarships to 2 veterans attending medical school at each of the 5 Teague-Cranston and the 4 Historically Black College and University (HBCU) medical schools (Table 2). The intent of the program was to determine the feasibility of increasing the pool of veteran physicians at VHA. Eligible applicants were notified of the scholarship opportunity through the American Medical College Application Service or through the medical school. Applicants must have separated from military service within the preceding 10 years of being admitted to medical school. In exchange for full tuition, fees, a monthly stipend, and rotation travel costs, the recipients accepted a 4-year clinical service obligation at VA facilities after completing their residency training.

Specialty Education Loan Repayment

The SELRP is a loan repayment program available to recently graduated physicians. Applicants must have graduated from an accredited medical or osteopathic school, matched to an accredited residency program and be ≥ 2 years from completion of residency. The specialties qualifying for SELRP are determined through an analysis of succession planning by the VA Office of Workforce Management and Consulting and change based on VA physician workforce needs. The SELRP provides loan repayment in the amount of $40,000 per year for up to 4 years, with a service obligation of 1 year for each $40,000 of support. In April 2021, VA began accepting applications from the eligible specialties of family medicine, internal medicine, gastroenterology, psychiatry, emergency medicine, and geriatrics.

Education Debt Reduction

The EDRP offers debt relief to clinicians in the most difficult to recruit professions, including physicians (generalists and specialists), registered nurses, licensed practical nurses, social workers, and psychologists. The list of difficult to recruit positions is developed annually by VA facilities. Annual reimbursements through the program may be used for tuition and expenses, such as fees, books, supplies, equipment, and other materials. In 2018, through the MISSION Act Section 302, the annual loan repayment was increased from $24,000 to $40,000, and the maximum level of support was increased from $120,000 to $200,000 over 5 years. Recipients receive reimbursement for loan repayment at the end of each year or service period and recipients are not required to remain in VA for 5 years.

Program Results

Health Professions Scholarship

For academic years 2020/2021 and 2021/2022, 126 HPSP applications from both allopathic and osteopathic schools were submitted and 51 scholarships were awarded (Table 3). Assuming an average residency length of 4 years, VHA estimates that these awards will yield 204 service-year equivalents by 2029.

Veterans Healing Veterans

In the VHV program, scholarship recipients came from 5 Teague-Cranston schools; 2 at University of South Carolina, 2 at East Tennessee State University, 2 at Wright State University, 1 at Texas A&M College of Medicine, 1 at Marshall University; and 3 HBCUs; 2 at Howard University, 1 at Morehouse School of Medicine and 1 at Meharry Medical College. The Charles R. Drew University of Medicine and Science did not nominate any students for the scholarship. Assuming all recipients complete postgraduate training, the VHV scholarship program will provide an additional 12 veteran physicians to serve at VA for at least 4 years each (48 service years).

Specialty Education Loan Repayment

Fourteen applicants have been approved, including 5 in psychiatry, 4 in family medicine, 3 in internal medicine, 1 in emergency medicine, and 1 in geriatrics. The mean loan repayment is anticipated to be $110,000 and equating to 38.5 VA service years or a mean of 2.3 years of service obligation per individual for the first cohort. The program has no termination date, and with continued funding, VA anticipates granting 100 loan repayments annually.

Education Debt Reduction

Since 2018, 1,546 VA physicians have received EDRP awards. Due to the increased reimbursement provided through the MISSION Act, average physician award amounts have increased from $96,090 in 2018 to $142,557 in 2019 and $148,302 in 2020.

Conclusions

The VA physician scholarship and loan repayment programs outlined in the MISSION Act build on the success of existing federal scholarship programs by providing opportunities for physician trainees to alleviate educational debt and explore a VA health professions career.

Looking ahead, VA must focus on measuring the success of the MISSION scholarship and loan repayment programs by tracking rates of acceptance and student graduation, residency and fellowship completion, and placement in VA medical facilities—both for the service obligation and future employment. Ultimately, the total impact on VA staffing, especially at rural and underresourced sites, will determine the success of the MISSION programs.

There are several malignant and nonmalignant conditions that can lead to central airway obstruction (CAO) resulting in lobar collapse. The clinical consequences range from significant dyspnea to respiratory failure. Airway stenting has been used to maintain patency of obstructed airways and relieve symptoms. Before lung cancer screening became more common, approximately 10% of lung cancers at presentation had evidence of CAO.¹

On occasion, an endobronchial malignancy involves the right mainstem (RMS) bronchus near the orifice of the right upper lobe (RUL).² Such strategically located lesions pose a challenge to relieve the RMS obstruction through stenting, securing airway patency into the bronchus intermedius (BI) while avoiding obstruction of the RUL bronchus. The use of endobronchial silicone stents, hybrid covered stents, as well as self-expanding metal stents (SEMS) is an established mode of relieving CAO due to malignant disease.³ We reviewed the literature for approaches that were available before and after the date of the index case reported here.

Case Presentation

A 65-year-old veteran with a history of smoking presented to a US Department of Veterans Affairs Medical Center (VAMC) in 2011, with hemoptysis of 2-week duration. Computed tomography (CT) of the chest revealed a 5.3 × 4.2 × 6.5 cm right mediastinal mass and a 3.0 × 2.8 × 3 cm right hilar mass. Flexible bronchoscopy revealed > 80% occlusion of the RMS and BI due to a medially located mass sparing the RUL orifice, which was patent (Figure 1). Airways distal to the BI were free of disease. Endobronchial biopsies revealed poorly differentiated non-small cell carcinoma of the lung. The patient was referred to the interventional pulmonary service for further airway management.

Under general anesthesia and through a size-9 endotracheal tube, piecemeal debulking of the mass using a cryoprobe was performed. Argon photocoagulation (APC) was used to control bleeding. Balloon bronchoplasty was performed next with pulmonary Boston Scientific CRE balloon at the BI and the RMS bronchus. Under fluoroscopic guidance, a 12 × 30 mm self-expanding hybrid Merit Medical AERO stent was placed distally into the BI. Next, a 14 × 30 mm AERO stent was placed proximally in the RMS bronchus with its distal end telescoped into the smaller distal stent for a distance of 3 to 4 mm at a slanted angle. The overlap was deliberately performed at the level of RUL takeoff. Forcing the distal end of the proximal larger stent into a smaller stent created mechanical stress. The angled alignment channeled this mechanical stress so that the distal end of the proximal stent flared open laterally into the RUL orifice to allow for ventilation (Figure 2). On follow-up 6 months later, all 3 airways remained patent with stents in place (Figure 3).

Literature Review

A literature review revealed multiple approaches to preserving a 3-way patent airway at the takeoff of the RUL (Table). One approach to alleviating such an obstruction favors placing a straight silicone stent from the RMS into the BI, closing off the orifice of the RUL (Figure 4A).⁴ However, this entails sacrificing ventilation of the RUL. An alternative suggested by Peled and colleagues was carried out successfully in 3 patients. After placing a stent to relieve the obstruction, a Nd:YAG laser is used to create a window in the stent in proximity to the RUL orifice, which allows preservation or ventilations to the RUL (Figure 4B).⁵

A third effective approach utilizes silicone Y stents, which are usually employed for relief of obstruction at the level of the main carina.^6,7 Instead of deploying them at the main carina, they would be deployed at the secondary carina, which the RUL makes with the BI, often with customized cutting for adjustment of the stent limbs to the appropriate size of the RUL and BI (Figure 4C). This approach has been successfully used to maintain RUL ventilation.²

A fourth technique involves using an Oki stent, a dedicated bifurcated silicone stent, which was first described in 2013. It is designed for the RMS bronchus around the RUL and BI bifurcation, enabling the stent to maintain airway patency in the right lung without affecting the trachea and carina (Figure 4D). The arm located in the RUL prevents migration.⁸ A fifth technique involves deploying a precisely selected Oki stent specially modified based on a printed 3-dimensional (3D) model of the airways after computer-aided simulation.⁹A sixth technique employs de novo custom printing stents based on 3D models of the tracheobronchial tree constructed based on CT imaging. This approach creates more accurately fitting stents.¹

Discussion

The RUL contributes roughly 5 to 10% of the total oxygenation capacity of the lung.¹⁰ In patients with lung cancer and limited pulmonary reserve, preserving ventilation to the RUL can be clinically important. The chosen method to relieve endobronchial obstruction depends on several variables, including expertise, ability of the patient to undergo general anesthesia for rigid or flexible bronchoscopy, stent availability, and airway anatomy.

This case illustrates a new method to deal with lesions close to the RUL orifice. This maneuver may not be possible with all types of stents. AERO stents are fully covered (Figure 4E). In contrast, stents that are uncovered at both distal ends, such as a Boston Scientific Ultraflex stent, may not be adequate for such a maneuver. Intercalating uncovered ends of SEMS may allow for tumor in-growth through the uncovered metal mesh near the RUL orifice and may paradoxically compromise both the RUL and BI. The diameter of AERO stents is slightly larger at its ends.¹¹ This helps prevent migration, which in this case maintained the crucial overlap of the stents. On the other hand, use of AERO stents may be associated with a higher risk of infection.¹² Precise measurements of the airway diameter are essential given the difference in internal and external stent diameter with silicone stents.

Silicone stents migrate more readily than SEMS and may not be well suited for the procedure we performed. In our case, we wished to maintain ventilation for the RUL; hence, we elected not to bypass it with a silicone stent. We did not have access to a YAG. Moreover, laser carries more energy than APC. Nd:YAG laser has been reported to cause airway fire when used with silicone stents.¹³ Several authors have reported the use of silicone Y stents at the primary or secondary carina to preserve luminal patency.^6,7 Airway anatomy and the angle of the Y may require modification of these stents prior to their use. Cutting stents may compromise their integrity. The bifurcating limb prevents migration which can be a significant concern with the tubular silicone stents. An important consideration for patients in advanced stages of malignancy is that placement of such stent requires undergoing general anesthesia and rigid bronchoscopy, unlike with AERO and metal stents that can be deployed with fiberoptic bronchoscopy under moderate sedation. As such, we did not elect to use a silicone Y stent. Accumulation of secretions or formation of granulation tissue at the orifices can result in recurrence of obstruction.¹⁴

Advances in 3D printing seem to be the future of customized airway stenting. This could help clinicians overcome the challenges of improperly sized stents and distorted airway anatomy. Cases have reported successful use of 3D-printed patient-specific airway prostheses.^15,16 However, their use is not common practice, as there is a limited amount of materials that are flexible, biocompatible, and approved by the US Food and Drug Administration (FDA) for medical use. Infection control is another layer of consideration in such stents. Standardization of materials and regulation of personalized devices and their cleansing protocols is neccesary.¹⁷ At the time of this case, Oki stents and 3D printing were not available in the market. This report provides a viable alternative to use AERO stents for this maneuver.

Conclusions

Patients presenting with malignant CAO near the RUL require a personalized approach to treatment, considering their overall health, functional status, nature and location of CAO, and degree of symptoms. Once a decision is made to stent the airway, careful assessment of airway anatomy, delineation of obstruction, available expertise, and types of stents available needs to be made to preserve ventilation to the nondiseased RUL. Airway stents are expensive and need to be used wisely for palliation and allowing for a quality life while the patient receives more definitive targeted therapy.

Acknowledgments

The authors would like to gratefully acknowledge Dr Jenny Kim, who referred the patient to the interventional service and helped obtain consent for publishing the case.

There are several malignant and nonmalignant conditions that can lead to central airway obstruction (CAO) resulting in lobar collapse. The clinical consequences range from significant dyspnea to respiratory failure. Airway stenting has been used to maintain patency of obstructed airways and relieve symptoms. Before lung cancer screening became more common, approximately 10% of lung cancers at presentation had evidence of CAO.¹

On occasion, an endobronchial malignancy involves the right mainstem (RMS) bronchus near the orifice of the right upper lobe (RUL).² Such strategically located lesions pose a challenge to relieve the RMS obstruction through stenting, securing airway patency into the bronchus intermedius (BI) while avoiding obstruction of the RUL bronchus. The use of endobronchial silicone stents, hybrid covered stents, as well as self-expanding metal stents (SEMS) is an established mode of relieving CAO due to malignant disease.³ We reviewed the literature for approaches that were available before and after the date of the index case reported here.

Case Presentation

A 65-year-old veteran with a history of smoking presented to a US Department of Veterans Affairs Medical Center (VAMC) in 2011, with hemoptysis of 2-week duration. Computed tomography (CT) of the chest revealed a 5.3 × 4.2 × 6.5 cm right mediastinal mass and a 3.0 × 2.8 × 3 cm right hilar mass. Flexible bronchoscopy revealed > 80% occlusion of the RMS and BI due to a medially located mass sparing the RUL orifice, which was patent (Figure 1). Airways distal to the BI were free of disease. Endobronchial biopsies revealed poorly differentiated non-small cell carcinoma of the lung. The patient was referred to the interventional pulmonary service for further airway management.

Under general anesthesia and through a size-9 endotracheal tube, piecemeal debulking of the mass using a cryoprobe was performed. Argon photocoagulation (APC) was used to control bleeding. Balloon bronchoplasty was performed next with pulmonary Boston Scientific CRE balloon at the BI and the RMS bronchus. Under fluoroscopic guidance, a 12 × 30 mm self-expanding hybrid Merit Medical AERO stent was placed distally into the BI. Next, a 14 × 30 mm AERO stent was placed proximally in the RMS bronchus with its distal end telescoped into the smaller distal stent for a distance of 3 to 4 mm at a slanted angle. The overlap was deliberately performed at the level of RUL takeoff. Forcing the distal end of the proximal larger stent into a smaller stent created mechanical stress. The angled alignment channeled this mechanical stress so that the distal end of the proximal stent flared open laterally into the RUL orifice to allow for ventilation (Figure 2). On follow-up 6 months later, all 3 airways remained patent with stents in place (Figure 3).

Literature Review

A literature review revealed multiple approaches to preserving a 3-way patent airway at the takeoff of the RUL (Table). One approach to alleviating such an obstruction favors placing a straight silicone stent from the RMS into the BI, closing off the orifice of the RUL (Figure 4A).⁴ However, this entails sacrificing ventilation of the RUL. An alternative suggested by Peled and colleagues was carried out successfully in 3 patients. After placing a stent to relieve the obstruction, a Nd:YAG laser is used to create a window in the stent in proximity to the RUL orifice, which allows preservation or ventilations to the RUL (Figure 4B).⁵

A third effective approach utilizes silicone Y stents, which are usually employed for relief of obstruction at the level of the main carina.^6,7 Instead of deploying them at the main carina, they would be deployed at the secondary carina, which the RUL makes with the BI, often with customized cutting for adjustment of the stent limbs to the appropriate size of the RUL and BI (Figure 4C). This approach has been successfully used to maintain RUL ventilation.²

A fourth technique involves using an Oki stent, a dedicated bifurcated silicone stent, which was first described in 2013. It is designed for the RMS bronchus around the RUL and BI bifurcation, enabling the stent to maintain airway patency in the right lung without affecting the trachea and carina (Figure 4D). The arm located in the RUL prevents migration.⁸ A fifth technique involves deploying a precisely selected Oki stent specially modified based on a printed 3-dimensional (3D) model of the airways after computer-aided simulation.⁹A sixth technique employs de novo custom printing stents based on 3D models of the tracheobronchial tree constructed based on CT imaging. This approach creates more accurately fitting stents.¹

Discussion

The RUL contributes roughly 5 to 10% of the total oxygenation capacity of the lung.¹⁰ In patients with lung cancer and limited pulmonary reserve, preserving ventilation to the RUL can be clinically important. The chosen method to relieve endobronchial obstruction depends on several variables, including expertise, ability of the patient to undergo general anesthesia for rigid or flexible bronchoscopy, stent availability, and airway anatomy.

This case illustrates a new method to deal with lesions close to the RUL orifice. This maneuver may not be possible with all types of stents. AERO stents are fully covered (Figure 4E). In contrast, stents that are uncovered at both distal ends, such as a Boston Scientific Ultraflex stent, may not be adequate for such a maneuver. Intercalating uncovered ends of SEMS may allow for tumor in-growth through the uncovered metal mesh near the RUL orifice and may paradoxically compromise both the RUL and BI. The diameter of AERO stents is slightly larger at its ends.¹¹ This helps prevent migration, which in this case maintained the crucial overlap of the stents. On the other hand, use of AERO stents may be associated with a higher risk of infection.¹² Precise measurements of the airway diameter are essential given the difference in internal and external stent diameter with silicone stents.

Silicone stents migrate more readily than SEMS and may not be well suited for the procedure we performed. In our case, we wished to maintain ventilation for the RUL; hence, we elected not to bypass it with a silicone stent. We did not have access to a YAG. Moreover, laser carries more energy than APC. Nd:YAG laser has been reported to cause airway fire when used with silicone stents.¹³ Several authors have reported the use of silicone Y stents at the primary or secondary carina to preserve luminal patency.^6,7 Airway anatomy and the angle of the Y may require modification of these stents prior to their use. Cutting stents may compromise their integrity. The bifurcating limb prevents migration which can be a significant concern with the tubular silicone stents. An important consideration for patients in advanced stages of malignancy is that placement of such stent requires undergoing general anesthesia and rigid bronchoscopy, unlike with AERO and metal stents that can be deployed with fiberoptic bronchoscopy under moderate sedation. As such, we did not elect to use a silicone Y stent. Accumulation of secretions or formation of granulation tissue at the orifices can result in recurrence of obstruction.¹⁴

Advances in 3D printing seem to be the future of customized airway stenting. This could help clinicians overcome the challenges of improperly sized stents and distorted airway anatomy. Cases have reported successful use of 3D-printed patient-specific airway prostheses.^15,16 However, their use is not common practice, as there is a limited amount of materials that are flexible, biocompatible, and approved by the US Food and Drug Administration (FDA) for medical use. Infection control is another layer of consideration in such stents. Standardization of materials and regulation of personalized devices and their cleansing protocols is neccesary.¹⁷ At the time of this case, Oki stents and 3D printing were not available in the market. This report provides a viable alternative to use AERO stents for this maneuver.

Conclusions

Patients presenting with malignant CAO near the RUL require a personalized approach to treatment, considering their overall health, functional status, nature and location of CAO, and degree of symptoms. Once a decision is made to stent the airway, careful assessment of airway anatomy, delineation of obstruction, available expertise, and types of stents available needs to be made to preserve ventilation to the nondiseased RUL. Airway stents are expensive and need to be used wisely for palliation and allowing for a quality life while the patient receives more definitive targeted therapy.

Acknowledgments

The authors would like to gratefully acknowledge Dr Jenny Kim, who referred the patient to the interventional service and helped obtain consent for publishing the case.

There are several malignant and nonmalignant conditions that can lead to central airway obstruction (CAO) resulting in lobar collapse. The clinical consequences range from significant dyspnea to respiratory failure. Airway stenting has been used to maintain patency of obstructed airways and relieve symptoms. Before lung cancer screening became more common, approximately 10% of lung cancers at presentation had evidence of CAO.¹

On occasion, an endobronchial malignancy involves the right mainstem (RMS) bronchus near the orifice of the right upper lobe (RUL).² Such strategically located lesions pose a challenge to relieve the RMS obstruction through stenting, securing airway patency into the bronchus intermedius (BI) while avoiding obstruction of the RUL bronchus. The use of endobronchial silicone stents, hybrid covered stents, as well as self-expanding metal stents (SEMS) is an established mode of relieving CAO due to malignant disease.³ We reviewed the literature for approaches that were available before and after the date of the index case reported here.

Case Presentation

A 65-year-old veteran with a history of smoking presented to a US Department of Veterans Affairs Medical Center (VAMC) in 2011, with hemoptysis of 2-week duration. Computed tomography (CT) of the chest revealed a 5.3 × 4.2 × 6.5 cm right mediastinal mass and a 3.0 × 2.8 × 3 cm right hilar mass. Flexible bronchoscopy revealed > 80% occlusion of the RMS and BI due to a medially located mass sparing the RUL orifice, which was patent (Figure 1). Airways distal to the BI were free of disease. Endobronchial biopsies revealed poorly differentiated non-small cell carcinoma of the lung. The patient was referred to the interventional pulmonary service for further airway management.

Under general anesthesia and through a size-9 endotracheal tube, piecemeal debulking of the mass using a cryoprobe was performed. Argon photocoagulation (APC) was used to control bleeding. Balloon bronchoplasty was performed next with pulmonary Boston Scientific CRE balloon at the BI and the RMS bronchus. Under fluoroscopic guidance, a 12 × 30 mm self-expanding hybrid Merit Medical AERO stent was placed distally into the BI. Next, a 14 × 30 mm AERO stent was placed proximally in the RMS bronchus with its distal end telescoped into the smaller distal stent for a distance of 3 to 4 mm at a slanted angle. The overlap was deliberately performed at the level of RUL takeoff. Forcing the distal end of the proximal larger stent into a smaller stent created mechanical stress. The angled alignment channeled this mechanical stress so that the distal end of the proximal stent flared open laterally into the RUL orifice to allow for ventilation (Figure 2). On follow-up 6 months later, all 3 airways remained patent with stents in place (Figure 3).

Literature Review

A literature review revealed multiple approaches to preserving a 3-way patent airway at the takeoff of the RUL (Table). One approach to alleviating such an obstruction favors placing a straight silicone stent from the RMS into the BI, closing off the orifice of the RUL (Figure 4A).⁴ However, this entails sacrificing ventilation of the RUL. An alternative suggested by Peled and colleagues was carried out successfully in 3 patients. After placing a stent to relieve the obstruction, a Nd:YAG laser is used to create a window in the stent in proximity to the RUL orifice, which allows preservation or ventilations to the RUL (Figure 4B).⁵

A third effective approach utilizes silicone Y stents, which are usually employed for relief of obstruction at the level of the main carina.^6,7 Instead of deploying them at the main carina, they would be deployed at the secondary carina, which the RUL makes with the BI, often with customized cutting for adjustment of the stent limbs to the appropriate size of the RUL and BI (Figure 4C). This approach has been successfully used to maintain RUL ventilation.²

A fourth technique involves using an Oki stent, a dedicated bifurcated silicone stent, which was first described in 2013. It is designed for the RMS bronchus around the RUL and BI bifurcation, enabling the stent to maintain airway patency in the right lung without affecting the trachea and carina (Figure 4D). The arm located in the RUL prevents migration.⁸ A fifth technique involves deploying a precisely selected Oki stent specially modified based on a printed 3-dimensional (3D) model of the airways after computer-aided simulation.⁹A sixth technique employs de novo custom printing stents based on 3D models of the tracheobronchial tree constructed based on CT imaging. This approach creates more accurately fitting stents.¹

Discussion

The RUL contributes roughly 5 to 10% of the total oxygenation capacity of the lung.¹⁰ In patients with lung cancer and limited pulmonary reserve, preserving ventilation to the RUL can be clinically important. The chosen method to relieve endobronchial obstruction depends on several variables, including expertise, ability of the patient to undergo general anesthesia for rigid or flexible bronchoscopy, stent availability, and airway anatomy.

This case illustrates a new method to deal with lesions close to the RUL orifice. This maneuver may not be possible with all types of stents. AERO stents are fully covered (Figure 4E). In contrast, stents that are uncovered at both distal ends, such as a Boston Scientific Ultraflex stent, may not be adequate for such a maneuver. Intercalating uncovered ends of SEMS may allow for tumor in-growth through the uncovered metal mesh near the RUL orifice and may paradoxically compromise both the RUL and BI. The diameter of AERO stents is slightly larger at its ends.¹¹ This helps prevent migration, which in this case maintained the crucial overlap of the stents. On the other hand, use of AERO stents may be associated with a higher risk of infection.¹² Precise measurements of the airway diameter are essential given the difference in internal and external stent diameter with silicone stents.

Silicone stents migrate more readily than SEMS and may not be well suited for the procedure we performed. In our case, we wished to maintain ventilation for the RUL; hence, we elected not to bypass it with a silicone stent. We did not have access to a YAG. Moreover, laser carries more energy than APC. Nd:YAG laser has been reported to cause airway fire when used with silicone stents.¹³ Several authors have reported the use of silicone Y stents at the primary or secondary carina to preserve luminal patency.^6,7 Airway anatomy and the angle of the Y may require modification of these stents prior to their use. Cutting stents may compromise their integrity. The bifurcating limb prevents migration which can be a significant concern with the tubular silicone stents. An important consideration for patients in advanced stages of malignancy is that placement of such stent requires undergoing general anesthesia and rigid bronchoscopy, unlike with AERO and metal stents that can be deployed with fiberoptic bronchoscopy under moderate sedation. As such, we did not elect to use a silicone Y stent. Accumulation of secretions or formation of granulation tissue at the orifices can result in recurrence of obstruction.¹⁴

Advances in 3D printing seem to be the future of customized airway stenting. This could help clinicians overcome the challenges of improperly sized stents and distorted airway anatomy. Cases have reported successful use of 3D-printed patient-specific airway prostheses.^15,16 However, their use is not common practice, as there is a limited amount of materials that are flexible, biocompatible, and approved by the US Food and Drug Administration (FDA) for medical use. Infection control is another layer of consideration in such stents. Standardization of materials and regulation of personalized devices and their cleansing protocols is neccesary.¹⁷ At the time of this case, Oki stents and 3D printing were not available in the market. This report provides a viable alternative to use AERO stents for this maneuver.

Conclusions

Patients presenting with malignant CAO near the RUL require a personalized approach to treatment, considering their overall health, functional status, nature and location of CAO, and degree of symptoms. Once a decision is made to stent the airway, careful assessment of airway anatomy, delineation of obstruction, available expertise, and types of stents available needs to be made to preserve ventilation to the nondiseased RUL. Airway stents are expensive and need to be used wisely for palliation and allowing for a quality life while the patient receives more definitive targeted therapy.

Acknowledgments

The authors would like to gratefully acknowledge Dr Jenny Kim, who referred the patient to the interventional service and helped obtain consent for publishing the case.

Suicide is the tenth leading cause of death in the United States claiming nearly 48,000 individuals in 2019 and is the second leading cause of death among individuals aged 10 to 34 years.¹ From 1999 to 2019, the suicide rate increased by 33%.¹ In a retrospective study evaluating suicide risk in > 29,000 men, veterans had a greater risk for suicide in all age groups except for the oldest when compared with nonveterans.² Another study of > 800,000 veterans found that younger veterans were most at risk for suicide.³ Veterans with completed suicides have a high incidence of affective disorders comorbid with substance use disorders, and therefore it is imperative to optimally treat these conditions to address suicidality.⁴ Additionally, a retrospective case-control study of veterans who died by suicide matched to controls identified that the cases had significantly higher rates of mental health conditions and suicidal ideation. Given that the veteran population is at higher risk of suicide, research of treatments to address suicidal ideation in veterans is needed.⁵

Lithium and Antisuicidal Properties

Lithium is the oldest treatment for bipolar disorder and is a long-standing first-line option due to its well-established efficacy as a mood stabilizer.⁶ Lithium’s antisuicidal properties separate it from the other pharmacologic options for bipolar disorder. A possible explanation for lithium’s unique antisuicidal properties is that these effects are mediated by its impact on aggression and impulsivity, which are both linked to an increased suicide risk.^7,8 A meta-analysis by Baldessarini and colleagues demonstrated that patients with mood disorders who were prescribed lithium had a 5 times lower risk of suicide and attempts than did those not treated with lithium.⁹ Lithium’s current place in therapy is in the treatment of bipolar disorder and major depressive disorder augmentation.^10-12Smith and colleagues found that in a cohort study of 21,194 veterans diagnosed with mental health conditions and initiated on lithium or valproate, there were no significant differences in associations with suicide observed between these agents over 365 days; however, there was a significant increased risk of suicide among patients discontinuing or modifying lithium within the first 180 days of treatment.¹³

Currently, lithium is thought to be underutilized at the US Department of Veterans Affairs (VA) Michael E. DeBakey VA Medical Center (MEDVAMC) in Houston, Texas, based on the number of prescriptions of lithium in the large population of veterans seen by mental health clinicians. MEDVAMC is a 538-bed academic teaching hospital serving approximately 130,000 veterans in southeast Texas. The Mental Health Care Line has 73 inpatient beds and an outpatient clinic serving > 12,000 patients annually. By retrospectively evaluating changes in suicidality in a sample of veterans prescribed lithium, we may be able to better understand the role that lithium plays in a population that has a higher suicide rate than does the general population. The primary objective of this study was to evaluate the change in number of suicide attempts from 3 months prior to lithium initiation to 3 months following a 6-month duration of lithium use. The secondary objective was to determine the change in suicidal ideation from the period prior to lithium use to the period following 6 months of lithium use.

Methods

This was a single-site, retrospective chart review conducted between October 2017 and April 2018. Prior to data collection, the MEDVAMC Research and Development committee approved the study as quality assurance research. Patients with an active lithium prescription were identified using the VA Lithium Lab Monitoring Dashboard, which includes all patients on lithium, their lithium level, and other data such as upcoming appointments.

Inclusion criteria consisted of adults who were aged ≥ 18 years, had an active lithium prescription on the date of data extraction, and had an active lithium prescription for at least 6 months. Patients were excluded if they had < 3 months of data before and/or after lithium was used for 6 months, and if they were initiated on lithium outside MEDVAMC. Cumulatively, patients had to have at least 12 months of data: 3 months prior to lithium use, at least 6 months of lithium use, and 9 months after lithium initiation.

Suicide Attempt and Suicidal Ideation Identification

When determining the number of suicide attempts, we recorded 4 data points: Veterans Crisis Line notes documenting suicide attempts, hospital admissions for suicide attempts, suicide behavior reports within the indicated time frame, and mental health progress notes documenting suicide attempts. Suicidal ideation was measured in 4 ways. First, we looked at the percentage of outpatient mental health progress notes documenting suicidal ideation. Second, using the Patient Health Questionnaire-9 (PHQ-9) depression assessments, we looked at the percentage of patients that indicated several days, more than half the days, or nearly every day to the question, “Thoughts that you would be better off dead or of hurting yourself in some way.”¹⁴ Third, we recorded the percentage of suicide risk assessments that patients responded yes to both questions on current preoccupation with suicidal thoughts and serious intent and plan to commit suicide with access to guns, stashed pills, or other means. Finally, we noted the percentage of suicide risk assessments where the assessment of risk was moderate or high.

A retrospective electronic health record (EHR) review was performed and the following information was obtained: patient demographics, lithium refill history, concomitant psychotropic medications and psychotherapy, lithium levels, comorbidities at lithium initiation, presence of a high-risk suicide flag in the EHR, suicide risk assessments, suicide behavior reports, Veteran Crisis Line notes, PHQ-9 assessments, and hospital admission and mental health outpatient notes. The lithium therapeutic range of 0.6-1.2 mmol/L is indicated for bipolar disorder and not other indications where the dose is typically titrated to effect rather than level. Medication possession ratio (MPR) was also calculated for lithium (sum of days’ supply for all fills in period ÷ number of days in period). A high-risk suicide flag alerts clinicians and staff that a mental health professional considers the veteran at risk for suicide.¹⁵ Statistical analysis was performed using the paired t test for means to assess proportional differences between variables for the primary and secondary outcomes. Descriptive statistics were used to describe the baseline characteristics.

Results

A total of 214 patients with an active prescription for lithium were identified on the Lithium Lab Monitoring Dashboard on October 31, 2017. After exclusion criteria were applied, 98 patients were included in the study (Figure 1). The 2 most common reasons for exclusion were due to patients not being on lithium for at least 6 months and being initiated on lithium at an outside facility. One patient was enrolled in a lithium research study (the medication ordered was lithium/placebo) and another patient refused all psychotropic medications according to the progress notes.

Most of the 98 patients (82.7%) were male with average age 50.5 years (Table 1). Almost half the patients (n = 47) were concomitantly participating in psychotherapy, and 50 (51.0%) patients received at least 1 antipsychotic medication. Twenty-nine patients had an active prescription for an additional mood stabilizer, and only 4 (4.1%) patients received lithium as monotherapy. Only 75 (76.5%) patients had a lithium level drawn during the 6 months of therapy, with 28 (37.3%) patients having a therapeutic lithium level (0.6 - 1.2 mmol/L). Seventy-one patients (72.4% ) were adherent to lithium therapy with a MPR > 0.8.¹⁶ Participants had 13 different psychiatric diagnoses at the time of lithium initiation; the most common were bipolar spectrum disorder (n = 38; 38.8%), depressive disorder (n = 27; 27.6%), and posttraumatic stress disorder (PTSD) (n = 26; 26.5%). Of note, 5 patients had a diagnosis of only PTSD without a concomitant mood disorder.

Discussion

The results of this study suggest lithium may have a role in reducing suicidality in a veteran population. There was a statistically significant reduction in hospitalizations for suicide attempt and suicidal ideation after at least 6 months of lithium use. These results are comparable with a previously published study that observed significant decreases in suicidal behavior and/or hospitalization risks among veterans taking lithium compared with those not taking lithium.¹⁷ Our study was similar in respect to the reduced hospitalizations among a veteran population; however, the previous study did not report a difference in suicide attempts and lithium use. This could be related to the longer follow-up time in the previous study (3 years) vs our study (9 months).

Our study identified a significant reduction in Veteran Crisis Line calls after at least 6 months of lithium use. While a reduction in suicidal ideations could be implicated in the decrease in crisis line calls, there may be a confounding variable. It is possible that after lithium initiation, veterans had more frequent contact with health care practitioners due to laboratory test monitoring and follow-up visits and thus had concerns/crises addressed during these interactions ultimately leading to a decreased utilization of the crisis line. Interestingly, there was a reduction in mental health outpatient notes from the prelithium period to the 3-month period that followed 6 months of lithium therapy. However, our study did not report on the number of mental health outpatient notes or visits during the 6-month lithium duration. Additionally, time of year/season could have an impact on suicidality, but this relationship was not evaluated in this study.

The presence of high-risk suicide flags also decreased from the prelithium period to the period 3 months following 6 months of lithium use. High-risk flags are reviewed by the suicide prevention coordinators and mental health professionals every 90 days; therefore, the patients with flags had multiple opportunities for review and thus renewal or discontinuation during the study period. A similar rationale can be applied to the high-risk flag as with the Veteran Crisis Line reduction, although this change could also be representative of a decrease in suicidality. Our study is different from other lithium studies because it included patients with a multitude of psychiatric diagnoses rather than just mood disorders. Five of the patients had a diagnosis of only PTSD and no documented mood disorder at the time of lithium initiation. Additional research is needed on the impact of lithium on suicidality in veterans with PTSD and psychiatric conditions other than mood disorders.

Underutilization of Lithium

Despite widespread knowledge of lithium’s antisuicidal effects, it is underutilized as a mood stabilizer in the US.¹⁸ There are various modifiable barriers that impact the prescribing as well as use of lithium. Clinicians may not be fully aware of lithium’s antisuicidal properties and may also have a low level of confidence in patients’ likelihood of adherence to laboratory monitoring.^18,19 Due to the narrow therapeutic index of lithium, the consequences of nonadherence to monitoring can be dangerous, which may deter mental health professionals from prescribing this antisuicidal agent. At MEDVAMC, only 72.4% of patients with a lithium prescription had a lithium level drawn within a 6-month period. This could be attributed to patient nonadherence (eg, the test was ordered but the patient did not go) or clinician nonadherence (eg, test was not ordered).

With increased clinician education as well as clinics dedicated to lithium management that allow for closer follow-up, facilities may see an increased level of comfort with lithium use. Lithium management clinics that provide close follow-up may also help address patient-related concerns about adverse effects and allow for close monitoring. To facilitate lithium monitoring at MEDVAMC, mental health practitioners and pharmacists developed a lithium test monitoring menu that serves as a “one-stop shop” for lithium baseline and ongoing test results.

In the future, we may study the impact of this test monitoring menu on lithium prescribing. One may also consider whether lithium levels need to be monitored at different frequencies (eg, less frequently for depression than bipolar disorder) depending on the diagnoses. A better understanding of the necessity for therapeutic monitoring may potentially reduce barriers to prescribing for patients who do not have indications that have a recommended therapeutic range (eg, bipolar disorder).

Lithium Adherence

A primary patient-related concern for low lithium utilization is poor adherence. In this sample, 71 patients (72.4%) were considered fully adherent. This was higher than the rate of 54.1% reported by Sajatovic and colleagues in a study evaluating adherence to lithium and other anticonvulsants in veterans with bipolar disorder.²⁰ Patients’ beliefs about medications and overall health as well as knowledge of the illness and treatment may impact adherence.²¹ The literature indicates that strategies such as cognitive behavioral therapy (CBT) and didactic lectures positively impact patients’ attitudes about lithium, which ultimately influences adherence.^21-23 Involving a family member or significant other in psychotherapy may also improve lithium adherence.²¹ Specifically in the VA, to address knowledge deficits and improve overall adherence, the Lithium Lab Monitoring Dashboard could be used to identify and invite new lithium starts to educational groups about lithium. These groups could also serve as lithium management clinics.

Limitations

There were several limitations to this study. This was a single-site, retrospective chart review with a small sample size. We studied a cross-section of veterans with only active prescriptions, which limited the sample size. The results should be interpreted cautiously because < 40% of patients who had a level drawn were in the therapeutic range. Patients whose lithium levels were outside of the therapeutic range may have not been fully adherent to the medication. Further analysis based on reason for lithium prescription (eg, bipolar disorder vs depression vs aggression/impulsivity in PTSD) may be helpful in better understanding the results.

Additionally, while we collected data on concomitant mood stabilizers and antipsychotics, we did not collect data on concurrent antidepressant therapy and only 4% of patients were on lithium monotherapy. Data regarding veterans undergoing concurrent CBT during their lithium trial were not assessed in this study and could be considered a confounding factor for future studies. We included any Veteran Crisis Line call in our results regardless of the reason for the call, which could have led to overreporting of this suicidality marker.

Given its small sample size, this study should be considered as hypothesis-generating. Further studies are needed to address lithium’s antisuicidal effects in specific diagnoses (eg, PTSD, anxiety, schizoaffective disorder) to better understand its place in therapy. Studies evaluating the relationship between dosing and suicidality may help provide insight into whether the antisuicidal effect of lithium is dose-dependent and whether a specific dose range rather than a therapeutic level should be targeted for antisuicidal purposes.

Conclusions

People treated for an affective disorder have a 30-times greater risk of suicide than do those in the general population; however, as lithium can reduce the risk of suicide and self-harm, it should continue to have an important role in clinical practice.²⁴ At MEDVAMC, we observed a statistically significant reduction in hospitalization for suicide attempts and suicidal ideation in veterans prescribed lithium following nonfatal suicide behavior and suicidal ideation. Prospective randomized placebo-controlled studies are needed to better understand lithium’s antisuicidal effects.

Suicide is the tenth leading cause of death in the United States claiming nearly 48,000 individuals in 2019 and is the second leading cause of death among individuals aged 10 to 34 years.¹ From 1999 to 2019, the suicide rate increased by 33%.¹ In a retrospective study evaluating suicide risk in > 29,000 men, veterans had a greater risk for suicide in all age groups except for the oldest when compared with nonveterans.² Another study of > 800,000 veterans found that younger veterans were most at risk for suicide.³ Veterans with completed suicides have a high incidence of affective disorders comorbid with substance use disorders, and therefore it is imperative to optimally treat these conditions to address suicidality.⁴ Additionally, a retrospective case-control study of veterans who died by suicide matched to controls identified that the cases had significantly higher rates of mental health conditions and suicidal ideation. Given that the veteran population is at higher risk of suicide, research of treatments to address suicidal ideation in veterans is needed.⁵

Lithium and Antisuicidal Properties

Lithium is the oldest treatment for bipolar disorder and is a long-standing first-line option due to its well-established efficacy as a mood stabilizer.⁶ Lithium’s antisuicidal properties separate it from the other pharmacologic options for bipolar disorder. A possible explanation for lithium’s unique antisuicidal properties is that these effects are mediated by its impact on aggression and impulsivity, which are both linked to an increased suicide risk.^7,8 A meta-analysis by Baldessarini and colleagues demonstrated that patients with mood disorders who were prescribed lithium had a 5 times lower risk of suicide and attempts than did those not treated with lithium.⁹ Lithium’s current place in therapy is in the treatment of bipolar disorder and major depressive disorder augmentation.^10-12Smith and colleagues found that in a cohort study of 21,194 veterans diagnosed with mental health conditions and initiated on lithium or valproate, there were no significant differences in associations with suicide observed between these agents over 365 days; however, there was a significant increased risk of suicide among patients discontinuing or modifying lithium within the first 180 days of treatment.¹³

Currently, lithium is thought to be underutilized at the US Department of Veterans Affairs (VA) Michael E. DeBakey VA Medical Center (MEDVAMC) in Houston, Texas, based on the number of prescriptions of lithium in the large population of veterans seen by mental health clinicians. MEDVAMC is a 538-bed academic teaching hospital serving approximately 130,000 veterans in southeast Texas. The Mental Health Care Line has 73 inpatient beds and an outpatient clinic serving > 12,000 patients annually. By retrospectively evaluating changes in suicidality in a sample of veterans prescribed lithium, we may be able to better understand the role that lithium plays in a population that has a higher suicide rate than does the general population. The primary objective of this study was to evaluate the change in number of suicide attempts from 3 months prior to lithium initiation to 3 months following a 6-month duration of lithium use. The secondary objective was to determine the change in suicidal ideation from the period prior to lithium use to the period following 6 months of lithium use.

Methods

This was a single-site, retrospective chart review conducted between October 2017 and April 2018. Prior to data collection, the MEDVAMC Research and Development committee approved the study as quality assurance research. Patients with an active lithium prescription were identified using the VA Lithium Lab Monitoring Dashboard, which includes all patients on lithium, their lithium level, and other data such as upcoming appointments.

Inclusion criteria consisted of adults who were aged ≥ 18 years, had an active lithium prescription on the date of data extraction, and had an active lithium prescription for at least 6 months. Patients were excluded if they had < 3 months of data before and/or after lithium was used for 6 months, and if they were initiated on lithium outside MEDVAMC. Cumulatively, patients had to have at least 12 months of data: 3 months prior to lithium use, at least 6 months of lithium use, and 9 months after lithium initiation.

Suicide Attempt and Suicidal Ideation Identification

When determining the number of suicide attempts, we recorded 4 data points: Veterans Crisis Line notes documenting suicide attempts, hospital admissions for suicide attempts, suicide behavior reports within the indicated time frame, and mental health progress notes documenting suicide attempts. Suicidal ideation was measured in 4 ways. First, we looked at the percentage of outpatient mental health progress notes documenting suicidal ideation. Second, using the Patient Health Questionnaire-9 (PHQ-9) depression assessments, we looked at the percentage of patients that indicated several days, more than half the days, or nearly every day to the question, “Thoughts that you would be better off dead or of hurting yourself in some way.”¹⁴ Third, we recorded the percentage of suicide risk assessments that patients responded yes to both questions on current preoccupation with suicidal thoughts and serious intent and plan to commit suicide with access to guns, stashed pills, or other means. Finally, we noted the percentage of suicide risk assessments where the assessment of risk was moderate or high.

A retrospective electronic health record (EHR) review was performed and the following information was obtained: patient demographics, lithium refill history, concomitant psychotropic medications and psychotherapy, lithium levels, comorbidities at lithium initiation, presence of a high-risk suicide flag in the EHR, suicide risk assessments, suicide behavior reports, Veteran Crisis Line notes, PHQ-9 assessments, and hospital admission and mental health outpatient notes. The lithium therapeutic range of 0.6-1.2 mmol/L is indicated for bipolar disorder and not other indications where the dose is typically titrated to effect rather than level. Medication possession ratio (MPR) was also calculated for lithium (sum of days’ supply for all fills in period ÷ number of days in period). A high-risk suicide flag alerts clinicians and staff that a mental health professional considers the veteran at risk for suicide.¹⁵ Statistical analysis was performed using the paired t test for means to assess proportional differences between variables for the primary and secondary outcomes. Descriptive statistics were used to describe the baseline characteristics.

Results

A total of 214 patients with an active prescription for lithium were identified on the Lithium Lab Monitoring Dashboard on October 31, 2017. After exclusion criteria were applied, 98 patients were included in the study (Figure 1). The 2 most common reasons for exclusion were due to patients not being on lithium for at least 6 months and being initiated on lithium at an outside facility. One patient was enrolled in a lithium research study (the medication ordered was lithium/placebo) and another patient refused all psychotropic medications according to the progress notes.

Most of the 98 patients (82.7%) were male with average age 50.5 years (Table 1). Almost half the patients (n = 47) were concomitantly participating in psychotherapy, and 50 (51.0%) patients received at least 1 antipsychotic medication. Twenty-nine patients had an active prescription for an additional mood stabilizer, and only 4 (4.1%) patients received lithium as monotherapy. Only 75 (76.5%) patients had a lithium level drawn during the 6 months of therapy, with 28 (37.3%) patients having a therapeutic lithium level (0.6 - 1.2 mmol/L). Seventy-one patients (72.4% ) were adherent to lithium therapy with a MPR > 0.8.¹⁶ Participants had 13 different psychiatric diagnoses at the time of lithium initiation; the most common were bipolar spectrum disorder (n = 38; 38.8%), depressive disorder (n = 27; 27.6%), and posttraumatic stress disorder (PTSD) (n = 26; 26.5%). Of note, 5 patients had a diagnosis of only PTSD without a concomitant mood disorder.

Discussion

The results of this study suggest lithium may have a role in reducing suicidality in a veteran population. There was a statistically significant reduction in hospitalizations for suicide attempt and suicidal ideation after at least 6 months of lithium use. These results are comparable with a previously published study that observed significant decreases in suicidal behavior and/or hospitalization risks among veterans taking lithium compared with those not taking lithium.¹⁷ Our study was similar in respect to the reduced hospitalizations among a veteran population; however, the previous study did not report a difference in suicide attempts and lithium use. This could be related to the longer follow-up time in the previous study (3 years) vs our study (9 months).

Our study identified a significant reduction in Veteran Crisis Line calls after at least 6 months of lithium use. While a reduction in suicidal ideations could be implicated in the decrease in crisis line calls, there may be a confounding variable. It is possible that after lithium initiation, veterans had more frequent contact with health care practitioners due to laboratory test monitoring and follow-up visits and thus had concerns/crises addressed during these interactions ultimately leading to a decreased utilization of the crisis line. Interestingly, there was a reduction in mental health outpatient notes from the prelithium period to the 3-month period that followed 6 months of lithium therapy. However, our study did not report on the number of mental health outpatient notes or visits during the 6-month lithium duration. Additionally, time of year/season could have an impact on suicidality, but this relationship was not evaluated in this study.

The presence of high-risk suicide flags also decreased from the prelithium period to the period 3 months following 6 months of lithium use. High-risk flags are reviewed by the suicide prevention coordinators and mental health professionals every 90 days; therefore, the patients with flags had multiple opportunities for review and thus renewal or discontinuation during the study period. A similar rationale can be applied to the high-risk flag as with the Veteran Crisis Line reduction, although this change could also be representative of a decrease in suicidality. Our study is different from other lithium studies because it included patients with a multitude of psychiatric diagnoses rather than just mood disorders. Five of the patients had a diagnosis of only PTSD and no documented mood disorder at the time of lithium initiation. Additional research is needed on the impact of lithium on suicidality in veterans with PTSD and psychiatric conditions other than mood disorders.

Underutilization of Lithium

Despite widespread knowledge of lithium’s antisuicidal effects, it is underutilized as a mood stabilizer in the US.¹⁸ There are various modifiable barriers that impact the prescribing as well as use of lithium. Clinicians may not be fully aware of lithium’s antisuicidal properties and may also have a low level of confidence in patients’ likelihood of adherence to laboratory monitoring.^18,19 Due to the narrow therapeutic index of lithium, the consequences of nonadherence to monitoring can be dangerous, which may deter mental health professionals from prescribing this antisuicidal agent. At MEDVAMC, only 72.4% of patients with a lithium prescription had a lithium level drawn within a 6-month period. This could be attributed to patient nonadherence (eg, the test was ordered but the patient did not go) or clinician nonadherence (eg, test was not ordered).

With increased clinician education as well as clinics dedicated to lithium management that allow for closer follow-up, facilities may see an increased level of comfort with lithium use. Lithium management clinics that provide close follow-up may also help address patient-related concerns about adverse effects and allow for close monitoring. To facilitate lithium monitoring at MEDVAMC, mental health practitioners and pharmacists developed a lithium test monitoring menu that serves as a “one-stop shop” for lithium baseline and ongoing test results.

In the future, we may study the impact of this test monitoring menu on lithium prescribing. One may also consider whether lithium levels need to be monitored at different frequencies (eg, less frequently for depression than bipolar disorder) depending on the diagnoses. A better understanding of the necessity for therapeutic monitoring may potentially reduce barriers to prescribing for patients who do not have indications that have a recommended therapeutic range (eg, bipolar disorder).

Lithium Adherence

A primary patient-related concern for low lithium utilization is poor adherence. In this sample, 71 patients (72.4%) were considered fully adherent. This was higher than the rate of 54.1% reported by Sajatovic and colleagues in a study evaluating adherence to lithium and other anticonvulsants in veterans with bipolar disorder.²⁰ Patients’ beliefs about medications and overall health as well as knowledge of the illness and treatment may impact adherence.²¹ The literature indicates that strategies such as cognitive behavioral therapy (CBT) and didactic lectures positively impact patients’ attitudes about lithium, which ultimately influences adherence.^21-23 Involving a family member or significant other in psychotherapy may also improve lithium adherence.²¹ Specifically in the VA, to address knowledge deficits and improve overall adherence, the Lithium Lab Monitoring Dashboard could be used to identify and invite new lithium starts to educational groups about lithium. These groups could also serve as lithium management clinics.

Limitations

There were several limitations to this study. This was a single-site, retrospective chart review with a small sample size. We studied a cross-section of veterans with only active prescriptions, which limited the sample size. The results should be interpreted cautiously because < 40% of patients who had a level drawn were in the therapeutic range. Patients whose lithium levels were outside of the therapeutic range may have not been fully adherent to the medication. Further analysis based on reason for lithium prescription (eg, bipolar disorder vs depression vs aggression/impulsivity in PTSD) may be helpful in better understanding the results.

Additionally, while we collected data on concomitant mood stabilizers and antipsychotics, we did not collect data on concurrent antidepressant therapy and only 4% of patients were on lithium monotherapy. Data regarding veterans undergoing concurrent CBT during their lithium trial were not assessed in this study and could be considered a confounding factor for future studies. We included any Veteran Crisis Line call in our results regardless of the reason for the call, which could have led to overreporting of this suicidality marker.

Given its small sample size, this study should be considered as hypothesis-generating. Further studies are needed to address lithium’s antisuicidal effects in specific diagnoses (eg, PTSD, anxiety, schizoaffective disorder) to better understand its place in therapy. Studies evaluating the relationship between dosing and suicidality may help provide insight into whether the antisuicidal effect of lithium is dose-dependent and whether a specific dose range rather than a therapeutic level should be targeted for antisuicidal purposes.

Conclusions

People treated for an affective disorder have a 30-times greater risk of suicide than do those in the general population; however, as lithium can reduce the risk of suicide and self-harm, it should continue to have an important role in clinical practice.²⁴ At MEDVAMC, we observed a statistically significant reduction in hospitalization for suicide attempts and suicidal ideation in veterans prescribed lithium following nonfatal suicide behavior and suicidal ideation. Prospective randomized placebo-controlled studies are needed to better understand lithium’s antisuicidal effects.

Suicide is the tenth leading cause of death in the United States claiming nearly 48,000 individuals in 2019 and is the second leading cause of death among individuals aged 10 to 34 years.¹ From 1999 to 2019, the suicide rate increased by 33%.¹ In a retrospective study evaluating suicide risk in > 29,000 men, veterans had a greater risk for suicide in all age groups except for the oldest when compared with nonveterans.² Another study of > 800,000 veterans found that younger veterans were most at risk for suicide.³ Veterans with completed suicides have a high incidence of affective disorders comorbid with substance use disorders, and therefore it is imperative to optimally treat these conditions to address suicidality.⁴ Additionally, a retrospective case-control study of veterans who died by suicide matched to controls identified that the cases had significantly higher rates of mental health conditions and suicidal ideation. Given that the veteran population is at higher risk of suicide, research of treatments to address suicidal ideation in veterans is needed.⁵

Lithium and Antisuicidal Properties

Lithium is the oldest treatment for bipolar disorder and is a long-standing first-line option due to its well-established efficacy as a mood stabilizer.⁶ Lithium’s antisuicidal properties separate it from the other pharmacologic options for bipolar disorder. A possible explanation for lithium’s unique antisuicidal properties is that these effects are mediated by its impact on aggression and impulsivity, which are both linked to an increased suicide risk.^7,8 A meta-analysis by Baldessarini and colleagues demonstrated that patients with mood disorders who were prescribed lithium had a 5 times lower risk of suicide and attempts than did those not treated with lithium.⁹ Lithium’s current place in therapy is in the treatment of bipolar disorder and major depressive disorder augmentation.^10-12Smith and colleagues found that in a cohort study of 21,194 veterans diagnosed with mental health conditions and initiated on lithium or valproate, there were no significant differences in associations with suicide observed between these agents over 365 days; however, there was a significant increased risk of suicide among patients discontinuing or modifying lithium within the first 180 days of treatment.¹³

Currently, lithium is thought to be underutilized at the US Department of Veterans Affairs (VA) Michael E. DeBakey VA Medical Center (MEDVAMC) in Houston, Texas, based on the number of prescriptions of lithium in the large population of veterans seen by mental health clinicians. MEDVAMC is a 538-bed academic teaching hospital serving approximately 130,000 veterans in southeast Texas. The Mental Health Care Line has 73 inpatient beds and an outpatient clinic serving > 12,000 patients annually. By retrospectively evaluating changes in suicidality in a sample of veterans prescribed lithium, we may be able to better understand the role that lithium plays in a population that has a higher suicide rate than does the general population. The primary objective of this study was to evaluate the change in number of suicide attempts from 3 months prior to lithium initiation to 3 months following a 6-month duration of lithium use. The secondary objective was to determine the change in suicidal ideation from the period prior to lithium use to the period following 6 months of lithium use.

Methods

This was a single-site, retrospective chart review conducted between October 2017 and April 2018. Prior to data collection, the MEDVAMC Research and Development committee approved the study as quality assurance research. Patients with an active lithium prescription were identified using the VA Lithium Lab Monitoring Dashboard, which includes all patients on lithium, their lithium level, and other data such as upcoming appointments.

Inclusion criteria consisted of adults who were aged ≥ 18 years, had an active lithium prescription on the date of data extraction, and had an active lithium prescription for at least 6 months. Patients were excluded if they had < 3 months of data before and/or after lithium was used for 6 months, and if they were initiated on lithium outside MEDVAMC. Cumulatively, patients had to have at least 12 months of data: 3 months prior to lithium use, at least 6 months of lithium use, and 9 months after lithium initiation.

Suicide Attempt and Suicidal Ideation Identification

When determining the number of suicide attempts, we recorded 4 data points: Veterans Crisis Line notes documenting suicide attempts, hospital admissions for suicide attempts, suicide behavior reports within the indicated time frame, and mental health progress notes documenting suicide attempts. Suicidal ideation was measured in 4 ways. First, we looked at the percentage of outpatient mental health progress notes documenting suicidal ideation. Second, using the Patient Health Questionnaire-9 (PHQ-9) depression assessments, we looked at the percentage of patients that indicated several days, more than half the days, or nearly every day to the question, “Thoughts that you would be better off dead or of hurting yourself in some way.”¹⁴ Third, we recorded the percentage of suicide risk assessments that patients responded yes to both questions on current preoccupation with suicidal thoughts and serious intent and plan to commit suicide with access to guns, stashed pills, or other means. Finally, we noted the percentage of suicide risk assessments where the assessment of risk was moderate or high.

A retrospective electronic health record (EHR) review was performed and the following information was obtained: patient demographics, lithium refill history, concomitant psychotropic medications and psychotherapy, lithium levels, comorbidities at lithium initiation, presence of a high-risk suicide flag in the EHR, suicide risk assessments, suicide behavior reports, Veteran Crisis Line notes, PHQ-9 assessments, and hospital admission and mental health outpatient notes. The lithium therapeutic range of 0.6-1.2 mmol/L is indicated for bipolar disorder and not other indications where the dose is typically titrated to effect rather than level. Medication possession ratio (MPR) was also calculated for lithium (sum of days’ supply for all fills in period ÷ number of days in period). A high-risk suicide flag alerts clinicians and staff that a mental health professional considers the veteran at risk for suicide.¹⁵ Statistical analysis was performed using the paired t test for means to assess proportional differences between variables for the primary and secondary outcomes. Descriptive statistics were used to describe the baseline characteristics.

Results

A total of 214 patients with an active prescription for lithium were identified on the Lithium Lab Monitoring Dashboard on October 31, 2017. After exclusion criteria were applied, 98 patients were included in the study (Figure 1). The 2 most common reasons for exclusion were due to patients not being on lithium for at least 6 months and being initiated on lithium at an outside facility. One patient was enrolled in a lithium research study (the medication ordered was lithium/placebo) and another patient refused all psychotropic medications according to the progress notes.

Most of the 98 patients (82.7%) were male with average age 50.5 years (Table 1). Almost half the patients (n = 47) were concomitantly participating in psychotherapy, and 50 (51.0%) patients received at least 1 antipsychotic medication. Twenty-nine patients had an active prescription for an additional mood stabilizer, and only 4 (4.1%) patients received lithium as monotherapy. Only 75 (76.5%) patients had a lithium level drawn during the 6 months of therapy, with 28 (37.3%) patients having a therapeutic lithium level (0.6 - 1.2 mmol/L). Seventy-one patients (72.4% ) were adherent to lithium therapy with a MPR > 0.8.¹⁶ Participants had 13 different psychiatric diagnoses at the time of lithium initiation; the most common were bipolar spectrum disorder (n = 38; 38.8%), depressive disorder (n = 27; 27.6%), and posttraumatic stress disorder (PTSD) (n = 26; 26.5%). Of note, 5 patients had a diagnosis of only PTSD without a concomitant mood disorder.

Discussion

The results of this study suggest lithium may have a role in reducing suicidality in a veteran population. There was a statistically significant reduction in hospitalizations for suicide attempt and suicidal ideation after at least 6 months of lithium use. These results are comparable with a previously published study that observed significant decreases in suicidal behavior and/or hospitalization risks among veterans taking lithium compared with those not taking lithium.¹⁷ Our study was similar in respect to the reduced hospitalizations among a veteran population; however, the previous study did not report a difference in suicide attempts and lithium use. This could be related to the longer follow-up time in the previous study (3 years) vs our study (9 months).

Our study identified a significant reduction in Veteran Crisis Line calls after at least 6 months of lithium use. While a reduction in suicidal ideations could be implicated in the decrease in crisis line calls, there may be a confounding variable. It is possible that after lithium initiation, veterans had more frequent contact with health care practitioners due to laboratory test monitoring and follow-up visits and thus had concerns/crises addressed during these interactions ultimately leading to a decreased utilization of the crisis line. Interestingly, there was a reduction in mental health outpatient notes from the prelithium period to the 3-month period that followed 6 months of lithium therapy. However, our study did not report on the number of mental health outpatient notes or visits during the 6-month lithium duration. Additionally, time of year/season could have an impact on suicidality, but this relationship was not evaluated in this study.

The presence of high-risk suicide flags also decreased from the prelithium period to the period 3 months following 6 months of lithium use. High-risk flags are reviewed by the suicide prevention coordinators and mental health professionals every 90 days; therefore, the patients with flags had multiple opportunities for review and thus renewal or discontinuation during the study period. A similar rationale can be applied to the high-risk flag as with the Veteran Crisis Line reduction, although this change could also be representative of a decrease in suicidality. Our study is different from other lithium studies because it included patients with a multitude of psychiatric diagnoses rather than just mood disorders. Five of the patients had a diagnosis of only PTSD and no documented mood disorder at the time of lithium initiation. Additional research is needed on the impact of lithium on suicidality in veterans with PTSD and psychiatric conditions other than mood disorders.

Underutilization of Lithium

Despite widespread knowledge of lithium’s antisuicidal effects, it is underutilized as a mood stabilizer in the US.¹⁸ There are various modifiable barriers that impact the prescribing as well as use of lithium. Clinicians may not be fully aware of lithium’s antisuicidal properties and may also have a low level of confidence in patients’ likelihood of adherence to laboratory monitoring.^18,19 Due to the narrow therapeutic index of lithium, the consequences of nonadherence to monitoring can be dangerous, which may deter mental health professionals from prescribing this antisuicidal agent. At MEDVAMC, only 72.4% of patients with a lithium prescription had a lithium level drawn within a 6-month period. This could be attributed to patient nonadherence (eg, the test was ordered but the patient did not go) or clinician nonadherence (eg, test was not ordered).

With increased clinician education as well as clinics dedicated to lithium management that allow for closer follow-up, facilities may see an increased level of comfort with lithium use. Lithium management clinics that provide close follow-up may also help address patient-related concerns about adverse effects and allow for close monitoring. To facilitate lithium monitoring at MEDVAMC, mental health practitioners and pharmacists developed a lithium test monitoring menu that serves as a “one-stop shop” for lithium baseline and ongoing test results.

In the future, we may study the impact of this test monitoring menu on lithium prescribing. One may also consider whether lithium levels need to be monitored at different frequencies (eg, less frequently for depression than bipolar disorder) depending on the diagnoses. A better understanding of the necessity for therapeutic monitoring may potentially reduce barriers to prescribing for patients who do not have indications that have a recommended therapeutic range (eg, bipolar disorder).

Lithium Adherence

A primary patient-related concern for low lithium utilization is poor adherence. In this sample, 71 patients (72.4%) were considered fully adherent. This was higher than the rate of 54.1% reported by Sajatovic and colleagues in a study evaluating adherence to lithium and other anticonvulsants in veterans with bipolar disorder.²⁰ Patients’ beliefs about medications and overall health as well as knowledge of the illness and treatment may impact adherence.²¹ The literature indicates that strategies such as cognitive behavioral therapy (CBT) and didactic lectures positively impact patients’ attitudes about lithium, which ultimately influences adherence.^21-23 Involving a family member or significant other in psychotherapy may also improve lithium adherence.²¹ Specifically in the VA, to address knowledge deficits and improve overall adherence, the Lithium Lab Monitoring Dashboard could be used to identify and invite new lithium starts to educational groups about lithium. These groups could also serve as lithium management clinics.

Limitations

There were several limitations to this study. This was a single-site, retrospective chart review with a small sample size. We studied a cross-section of veterans with only active prescriptions, which limited the sample size. The results should be interpreted cautiously because < 40% of patients who had a level drawn were in the therapeutic range. Patients whose lithium levels were outside of the therapeutic range may have not been fully adherent to the medication. Further analysis based on reason for lithium prescription (eg, bipolar disorder vs depression vs aggression/impulsivity in PTSD) may be helpful in better understanding the results.

Additionally, while we collected data on concomitant mood stabilizers and antipsychotics, we did not collect data on concurrent antidepressant therapy and only 4% of patients were on lithium monotherapy. Data regarding veterans undergoing concurrent CBT during their lithium trial were not assessed in this study and could be considered a confounding factor for future studies. We included any Veteran Crisis Line call in our results regardless of the reason for the call, which could have led to overreporting of this suicidality marker.

Given its small sample size, this study should be considered as hypothesis-generating. Further studies are needed to address lithium’s antisuicidal effects in specific diagnoses (eg, PTSD, anxiety, schizoaffective disorder) to better understand its place in therapy. Studies evaluating the relationship between dosing and suicidality may help provide insight into whether the antisuicidal effect of lithium is dose-dependent and whether a specific dose range rather than a therapeutic level should be targeted for antisuicidal purposes.

Conclusions

People treated for an affective disorder have a 30-times greater risk of suicide than do those in the general population; however, as lithium can reduce the risk of suicide and self-harm, it should continue to have an important role in clinical practice.²⁴ At MEDVAMC, we observed a statistically significant reduction in hospitalization for suicide attempts and suicidal ideation in veterans prescribed lithium following nonfatal suicide behavior and suicidal ideation. Prospective randomized placebo-controlled studies are needed to better understand lithium’s antisuicidal effects.

March is Women’s History Month. Many women have served in all branches of government health care over centuries and are worthy of celebrating. These nurses, physicians, pharmacists, and other allied health professionals devoted their time and talents, compassion, and competence to deliver and improve the care of wounded service members, disabled veterans, and the underresourced in our communities. To honor the collective contribution of women to federal practice in the Indian Health Service, Public Health Service Core, US Department of Veterans Affairs (VA) and the US Department of Defense, this column examines one pioneer in women’s federal practice—Margaret D. Craighill, MD—who epitomizes the spirit of the selfless dedication that generations of women have given to public service. Craighill is an ideal choice to represent this noble cadre of women as her career spanned active military duty, public health, and the Veterans Health Administration.

Craighill was a graduate of several of the finest institutions of medical training in the United States. Born in Southport, North Carolina, in 1898, she earned her undergraduate degree Phi Beta Kappa and master’s degree from the University of Wisconsin.² She set her sights on becoming a physician at a period in American history when many prominent medical schools accepted few women. A marked exception—due to the fund raising and lobbying of influential women—was the prestigious Johns Hopkins University School of Medicine.³ She graduated in 1924 and held a postgraduate position at Yale Medical School. She then worked as a physiologist at a military arsenal, a pathologist, a general surgeon, and completed a residency in obstetrics and gynecology. This broad training gave her the diverse expertise she would need for her future work.⁴

Craighill came from a military family: Her father was a colonel in the engineering corps, and her grandfather rose to become chief engineer of the Army.⁵ Along with many of America’s best and brightest, Craighill left her successful medical career as dean of the Women’s Medical College of Pennsylvania to join the war effort. Author Alan G. Knight points out, more than in civilian medicine, gender stereotypes kept women from entering the military: Women were expected and accepted as nurses, not doctors.⁵ But in 1943 Congress passed and President Roosevelt signed the Sparkman-Johnson Bill, enabling women to enter the then all-male Army and Navy Medical Corps. Craighill took advantage of this opportunity and accepted an appointment to the Women’s Army Corps (WAC) as a major in 1943 at age 45 years, becoming the first woman physician to be commissioned an officer in the Army.

Major Craighill’s initial assignment was to the Office of the Surgeon General in the Preventive Medicine Division as the consultant for health and welfare of women. Here, she served as liaison to another innovation in women’s history in military medicine—the WAC. Journeying 56,000 miles to war zones in multiple countries, she assessed the health of 160,000 Army nurses and other staff whose focus was public health and infectious disease and hygiene. The history of women in medicine in and out of federal service is marked by overcoming innumerable biases and barriers. Craighill faced the prevailing presumption that women were unfit for military duty. In an early example of evidence-based medicine, she disproved this theory, showing that women were faring well doing hard jobs in tough environments.⁴

Their fortitude is more remarkable considering induction examinations for women during World War II were cursory and not tailored to address women’s health care needs. Based on her visits to WACs in theater and at home, Craighill observed recruits suffering from previously undiagnosed gynecologic and psychiatric conditions that adversely affected their health and function. She advocated for comprehensive standardized examinations that would detect many of these disorders.⁵

Craighill promoted other prejudices of her era. WAC command wanted to win public approval of women in the service and was concerned that lesbian relationships and “heterosexual promiscuity” would damage their public relations aims. They pressured Craighill to develop induction examinations that would screen lesbians and women with behavioral problems. She urged tolerance of homosexual behavior until it was proven.

Though clearly discriminatory and personally offensive to gay persons in federal service, we must recognize that only last year did the Pentagon move to overturn the prior administration’s prohibition against transgender persons serving in uniform.⁶ In this light Craighill, as the first female physician-leader in a 1940s military, adopted a relatively progressive stance.

Craighill rose to the rank of lieutenant colonel and received the Legion of Merit award for her exemplary wartime service. In 1945, she earned another first when she was appointed to be a consultant on the medical care of women veterans. For women veterans, gaining access to newly earned benefits and receiving appropriate care were serious problems that Craighill worked to solve. For many women veterans, those challenges remain, and Craighill’s legacy summons us to take up the charge to empower women in federal health professions to enhance the quality of care women veterans receive in all sectors of US medicine.

Critics and advocates agree that the VA still has a long way to go to achieve equity and excellence in our care for women veterans.^7,8 Craighill’s position stands as a landmark in this effort. During her VA tenure, Craighill entered a residency in the first class of the Menninger School of Psychiatry in Topeka, Kansas, and completed psychoanalytic training. Her wartime experiences had convinced her of the need to provide high-quality mental health care to women veterans. She put her new psychosomatic knowledge and skills to use, serving as the chief of a women’s health clinic at the VA Hospital in Topeka and published several important scholarly papers.^5,9Craighill went on to have a distinguished career in academic medicine, underscoring the long and valuable relationship of US medicine and the scholarly medical community. Once her psychiatric training was finished, she returned to private practice, ending her career as chief psychiatrist at Connecticut College for Women.

Craighill made a significant contribution to the role of women in federal practice. She was a visionary in her conviction that women, whether physicians, nurses, or other health care professionals, had the gifts and the grit to serve with distinction and valor and that their military service entitled them in war and peace to gender-sensitive health care. As the epigraph for this editorial shows, Craighill knew the path for women in federal practice or service while not easy is well worth treading. Her pioneering career can inspire all those women who today and in the future choose to follow in her footsteps.

March is Women’s History Month. Many women have served in all branches of government health care over centuries and are worthy of celebrating. These nurses, physicians, pharmacists, and other allied health professionals devoted their time and talents, compassion, and competence to deliver and improve the care of wounded service members, disabled veterans, and the underresourced in our communities. To honor the collective contribution of women to federal practice in the Indian Health Service, Public Health Service Core, US Department of Veterans Affairs (VA) and the US Department of Defense, this column examines one pioneer in women’s federal practice—Margaret D. Craighill, MD—who epitomizes the spirit of the selfless dedication that generations of women have given to public service. Craighill is an ideal choice to represent this noble cadre of women as her career spanned active military duty, public health, and the Veterans Health Administration.

Craighill was a graduate of several of the finest institutions of medical training in the United States. Born in Southport, North Carolina, in 1898, she earned her undergraduate degree Phi Beta Kappa and master’s degree from the University of Wisconsin.² She set her sights on becoming a physician at a period in American history when many prominent medical schools accepted few women. A marked exception—due to the fund raising and lobbying of influential women—was the prestigious Johns Hopkins University School of Medicine.³ She graduated in 1924 and held a postgraduate position at Yale Medical School. She then worked as a physiologist at a military arsenal, a pathologist, a general surgeon, and completed a residency in obstetrics and gynecology. This broad training gave her the diverse expertise she would need for her future work.⁴

Craighill came from a military family: Her father was a colonel in the engineering corps, and her grandfather rose to become chief engineer of the Army.⁵ Along with many of America’s best and brightest, Craighill left her successful medical career as dean of the Women’s Medical College of Pennsylvania to join the war effort. Author Alan G. Knight points out, more than in civilian medicine, gender stereotypes kept women from entering the military: Women were expected and accepted as nurses, not doctors.⁵ But in 1943 Congress passed and President Roosevelt signed the Sparkman-Johnson Bill, enabling women to enter the then all-male Army and Navy Medical Corps. Craighill took advantage of this opportunity and accepted an appointment to the Women’s Army Corps (WAC) as a major in 1943 at age 45 years, becoming the first woman physician to be commissioned an officer in the Army.

Major Craighill’s initial assignment was to the Office of the Surgeon General in the Preventive Medicine Division as the consultant for health and welfare of women. Here, she served as liaison to another innovation in women’s history in military medicine—the WAC. Journeying 56,000 miles to war zones in multiple countries, she assessed the health of 160,000 Army nurses and other staff whose focus was public health and infectious disease and hygiene. The history of women in medicine in and out of federal service is marked by overcoming innumerable biases and barriers. Craighill faced the prevailing presumption that women were unfit for military duty. In an early example of evidence-based medicine, she disproved this theory, showing that women were faring well doing hard jobs in tough environments.⁴

Their fortitude is more remarkable considering induction examinations for women during World War II were cursory and not tailored to address women’s health care needs. Based on her visits to WACs in theater and at home, Craighill observed recruits suffering from previously undiagnosed gynecologic and psychiatric conditions that adversely affected their health and function. She advocated for comprehensive standardized examinations that would detect many of these disorders.⁵

Craighill promoted other prejudices of her era. WAC command wanted to win public approval of women in the service and was concerned that lesbian relationships and “heterosexual promiscuity” would damage their public relations aims. They pressured Craighill to develop induction examinations that would screen lesbians and women with behavioral problems. She urged tolerance of homosexual behavior until it was proven.

Though clearly discriminatory and personally offensive to gay persons in federal service, we must recognize that only last year did the Pentagon move to overturn the prior administration’s prohibition against transgender persons serving in uniform.⁶ In this light Craighill, as the first female physician-leader in a 1940s military, adopted a relatively progressive stance.

Craighill rose to the rank of lieutenant colonel and received the Legion of Merit award for her exemplary wartime service. In 1945, she earned another first when she was appointed to be a consultant on the medical care of women veterans. For women veterans, gaining access to newly earned benefits and receiving appropriate care were serious problems that Craighill worked to solve. For many women veterans, those challenges remain, and Craighill’s legacy summons us to take up the charge to empower women in federal health professions to enhance the quality of care women veterans receive in all sectors of US medicine.

Critics and advocates agree that the VA still has a long way to go to achieve equity and excellence in our care for women veterans.^7,8 Craighill’s position stands as a landmark in this effort. During her VA tenure, Craighill entered a residency in the first class of the Menninger School of Psychiatry in Topeka, Kansas, and completed psychoanalytic training. Her wartime experiences had convinced her of the need to provide high-quality mental health care to women veterans. She put her new psychosomatic knowledge and skills to use, serving as the chief of a women’s health clinic at the VA Hospital in Topeka and published several important scholarly papers.^5,9Craighill went on to have a distinguished career in academic medicine, underscoring the long and valuable relationship of US medicine and the scholarly medical community. Once her psychiatric training was finished, she returned to private practice, ending her career as chief psychiatrist at Connecticut College for Women.

Craighill made a significant contribution to the role of women in federal practice. She was a visionary in her conviction that women, whether physicians, nurses, or other health care professionals, had the gifts and the grit to serve with distinction and valor and that their military service entitled them in war and peace to gender-sensitive health care. As the epigraph for this editorial shows, Craighill knew the path for women in federal practice or service while not easy is well worth treading. Her pioneering career can inspire all those women who today and in the future choose to follow in her footsteps.

March is Women’s History Month. Many women have served in all branches of government health care over centuries and are worthy of celebrating. These nurses, physicians, pharmacists, and other allied health professionals devoted their time and talents, compassion, and competence to deliver and improve the care of wounded service members, disabled veterans, and the underresourced in our communities. To honor the collective contribution of women to federal practice in the Indian Health Service, Public Health Service Core, US Department of Veterans Affairs (VA) and the US Department of Defense, this column examines one pioneer in women’s federal practice—Margaret D. Craighill, MD—who epitomizes the spirit of the selfless dedication that generations of women have given to public service. Craighill is an ideal choice to represent this noble cadre of women as her career spanned active military duty, public health, and the Veterans Health Administration.

Craighill was a graduate of several of the finest institutions of medical training in the United States. Born in Southport, North Carolina, in 1898, she earned her undergraduate degree Phi Beta Kappa and master’s degree from the University of Wisconsin.² She set her sights on becoming a physician at a period in American history when many prominent medical schools accepted few women. A marked exception—due to the fund raising and lobbying of influential women—was the prestigious Johns Hopkins University School of Medicine.³ She graduated in 1924 and held a postgraduate position at Yale Medical School. She then worked as a physiologist at a military arsenal, a pathologist, a general surgeon, and completed a residency in obstetrics and gynecology. This broad training gave her the diverse expertise she would need for her future work.⁴

Craighill came from a military family: Her father was a colonel in the engineering corps, and her grandfather rose to become chief engineer of the Army.⁵ Along with many of America’s best and brightest, Craighill left her successful medical career as dean of the Women’s Medical College of Pennsylvania to join the war effort. Author Alan G. Knight points out, more than in civilian medicine, gender stereotypes kept women from entering the military: Women were expected and accepted as nurses, not doctors.⁵ But in 1943 Congress passed and President Roosevelt signed the Sparkman-Johnson Bill, enabling women to enter the then all-male Army and Navy Medical Corps. Craighill took advantage of this opportunity and accepted an appointment to the Women’s Army Corps (WAC) as a major in 1943 at age 45 years, becoming the first woman physician to be commissioned an officer in the Army.

Major Craighill’s initial assignment was to the Office of the Surgeon General in the Preventive Medicine Division as the consultant for health and welfare of women. Here, she served as liaison to another innovation in women’s history in military medicine—the WAC. Journeying 56,000 miles to war zones in multiple countries, she assessed the health of 160,000 Army nurses and other staff whose focus was public health and infectious disease and hygiene. The history of women in medicine in and out of federal service is marked by overcoming innumerable biases and barriers. Craighill faced the prevailing presumption that women were unfit for military duty. In an early example of evidence-based medicine, she disproved this theory, showing that women were faring well doing hard jobs in tough environments.⁴

Their fortitude is more remarkable considering induction examinations for women during World War II were cursory and not tailored to address women’s health care needs. Based on her visits to WACs in theater and at home, Craighill observed recruits suffering from previously undiagnosed gynecologic and psychiatric conditions that adversely affected their health and function. She advocated for comprehensive standardized examinations that would detect many of these disorders.⁵

Craighill promoted other prejudices of her era. WAC command wanted to win public approval of women in the service and was concerned that lesbian relationships and “heterosexual promiscuity” would damage their public relations aims. They pressured Craighill to develop induction examinations that would screen lesbians and women with behavioral problems. She urged tolerance of homosexual behavior until it was proven.

Though clearly discriminatory and personally offensive to gay persons in federal service, we must recognize that only last year did the Pentagon move to overturn the prior administration’s prohibition against transgender persons serving in uniform.⁶ In this light Craighill, as the first female physician-leader in a 1940s military, adopted a relatively progressive stance.

Craighill rose to the rank of lieutenant colonel and received the Legion of Merit award for her exemplary wartime service. In 1945, she earned another first when she was appointed to be a consultant on the medical care of women veterans. For women veterans, gaining access to newly earned benefits and receiving appropriate care were serious problems that Craighill worked to solve. For many women veterans, those challenges remain, and Craighill’s legacy summons us to take up the charge to empower women in federal health professions to enhance the quality of care women veterans receive in all sectors of US medicine.

Critics and advocates agree that the VA still has a long way to go to achieve equity and excellence in our care for women veterans.^7,8 Craighill’s position stands as a landmark in this effort. During her VA tenure, Craighill entered a residency in the first class of the Menninger School of Psychiatry in Topeka, Kansas, and completed psychoanalytic training. Her wartime experiences had convinced her of the need to provide high-quality mental health care to women veterans. She put her new psychosomatic knowledge and skills to use, serving as the chief of a women’s health clinic at the VA Hospital in Topeka and published several important scholarly papers.^5,9Craighill went on to have a distinguished career in academic medicine, underscoring the long and valuable relationship of US medicine and the scholarly medical community. Once her psychiatric training was finished, she returned to private practice, ending her career as chief psychiatrist at Connecticut College for Women.

Craighill made a significant contribution to the role of women in federal practice. She was a visionary in her conviction that women, whether physicians, nurses, or other health care professionals, had the gifts and the grit to serve with distinction and valor and that their military service entitled them in war and peace to gender-sensitive health care. As the epigraph for this editorial shows, Craighill knew the path for women in federal practice or service while not easy is well worth treading. Her pioneering career can inspire all those women who today and in the future choose to follow in her footsteps.

In patients with persistent atopic dermatitis (AD) who are taking dupilumab, is there benefit of patch testing to determine if allergic contact dermatitis (ACD) also is contributing to their disease? Results of patch testing are likely be influenced by the immunomodulatory effects of dupilumab. Similar to the recommendation for patients to refrain from using topical or systemic corticosteroids for 1 week or more prior to patch testing to eliminate false negatives, we reviewed the literature to create practice guidelines for dermatologists regarding patch testing while a patient is taking dupilumab.

Pathophysiology and Pathomechanism

Dupilumab functions through the blockade of T helper 2 (T_H2) cells; ACD is propagated through the T helper 1 (T_H1) cellular pathway. However, patients with ACD that is unresponsive to allergen avoidance and traditional therapies, such as topical and oral corticosteroids, have responded to dupilumab. The more common reports of this responsiveness are with fragrances; multiple case series described patients with ACD to fragrance mix I¹ and balsam of Peru^1,2 who improved on dupilumab when other treatments failed. There also are reports of response when ACD was secondary to nickel,^2,3p-phenylenediamine,¹ Compositae,⁴ and non–formaldehyde-releasing preservatives (non-FRPs).⁵ Therefore, not all ACD is propagated through the T_H1 cellular pathway.

As noted in these cases, ACD can be a response to an allergen whose pathogenesis involves the T_H2 pathway or when patient characteristics favor a T_H2 response. It has been suggested that AD patients are more susceptible to T_H2-mediated contact sensitization to less-potent allergens, such as fragrances.⁶

Patch Test Results

Positive patch test results for allergens have been reported while patients are on dupilumab therapy, including a few studies in which results prior to starting dupilumab were compared with those while patients were on dupilumab therapy. In a retrospective chart review of 48 patients on dupilumab for AD with persistent disease, 23 patients were patch tested before and during dupilumab therapy. In these patients, the majority of contact allergies were persistent and only 10% (13/125) of patch test–positive results resolved on dupilumab therapy.⁷ Contact allergies that resolved included those to emulsifiers (propylene glycol, Amerchol L101 [lanolin-containing products found in cosmetics and other goods], dimethylaminopropylamine), fragrances (fragrance mix I, balsam of Peru), sunscreens (sulisobenzone, phenylbenzimidazole-5-sulfonic acid), and metals (vanadium chloride, phenylmercuric acetate).⁷ The following results observed in individual cases demonstrated conflicting findings: persistence of allergy to non-FRPs (methylisothiazolinone [MI]) but resolution of allergy to formaldehyde⁸; persistence of allergy to corticosteroids (budesonide and alclometasone)⁹; persistence of allergy to an antibiotic (neomycin sulfate) but resolution of allergies to a different antibiotic (bacitracin), glues (ethyl acrylate), bleach, and glutaraldehyde⁹; persistence of nickel allergy but resolution of allergies to fragrances (cinnamic aldehyde, balsam of Peru) and non-FRPs (methylchloroisothiazolinone or MI)¹⁰; and persistence of allergies to non-FRPs (MI) and FRPs (bronopol) but resolution of allergies to nickel, fragrances (hydroperoxides of linalool), and Compositae.¹¹ Additional case reports of positive patch test results while on dupilumab but with no pretreatment results for comparison include allergies to rubber additives,^12-14 nickel,¹⁴ textile dyes,¹⁴ cosmetic and hair care additives,^12,14,15 corticosteroids,¹⁵ FRPs,¹⁵ fragrances,^15,16 emulsifiers,¹⁶ and non-FRPs.¹⁷

An evident theme in the dupilumab patch-testing literature has been that results are variable and case specific: a given patient with ACD to an allergen will respond to dupilumab treatment and have subsequent negative patch testing, while another patient will not respond to dupilumab treatment and have persistent positive patch testing. This is likely because, in certain individuals, the allergen-immune system combination shifts ACD pathogenesis from a purely T_H1 response to at least a partial T_H2 response, thus allowing for benefit from dupilumab therapy. T helper 1 cell–mediated ACD should not be affected by dupilumab; therefore, reliable results can be elucidated from patch testing despite the drug.

Final Thoughts

We propose that AD patients with residual disease after taking dupilumab undergo patch testing. Positive results indicate allergens that are not inhibited by the drug. Patients will need to follow strict allergen avoidance to resolve this component of their disease; failure to improve might suggest the result was a nonrelevant positive.

If patch testing is negative, an alternative cause for residual disease must be sought. We do not recommend stopping dupilumab prior to patch testing to avoid a disease flare from AD or possible T_H2-mediated ACD.

In patients with persistent atopic dermatitis (AD) who are taking dupilumab, is there benefit of patch testing to determine if allergic contact dermatitis (ACD) also is contributing to their disease? Results of patch testing are likely be influenced by the immunomodulatory effects of dupilumab. Similar to the recommendation for patients to refrain from using topical or systemic corticosteroids for 1 week or more prior to patch testing to eliminate false negatives, we reviewed the literature to create practice guidelines for dermatologists regarding patch testing while a patient is taking dupilumab.

Pathophysiology and Pathomechanism

Dupilumab functions through the blockade of T helper 2 (T_H2) cells; ACD is propagated through the T helper 1 (T_H1) cellular pathway. However, patients with ACD that is unresponsive to allergen avoidance and traditional therapies, such as topical and oral corticosteroids, have responded to dupilumab. The more common reports of this responsiveness are with fragrances; multiple case series described patients with ACD to fragrance mix I¹ and balsam of Peru^1,2 who improved on dupilumab when other treatments failed. There also are reports of response when ACD was secondary to nickel,^2,3p-phenylenediamine,¹ Compositae,⁴ and non–formaldehyde-releasing preservatives (non-FRPs).⁵ Therefore, not all ACD is propagated through the T_H1 cellular pathway.

As noted in these cases, ACD can be a response to an allergen whose pathogenesis involves the T_H2 pathway or when patient characteristics favor a T_H2 response. It has been suggested that AD patients are more susceptible to T_H2-mediated contact sensitization to less-potent allergens, such as fragrances.⁶

Patch Test Results

Positive patch test results for allergens have been reported while patients are on dupilumab therapy, including a few studies in which results prior to starting dupilumab were compared with those while patients were on dupilumab therapy. In a retrospective chart review of 48 patients on dupilumab for AD with persistent disease, 23 patients were patch tested before and during dupilumab therapy. In these patients, the majority of contact allergies were persistent and only 10% (13/125) of patch test–positive results resolved on dupilumab therapy.⁷ Contact allergies that resolved included those to emulsifiers (propylene glycol, Amerchol L101 [lanolin-containing products found in cosmetics and other goods], dimethylaminopropylamine), fragrances (fragrance mix I, balsam of Peru), sunscreens (sulisobenzone, phenylbenzimidazole-5-sulfonic acid), and metals (vanadium chloride, phenylmercuric acetate).⁷ The following results observed in individual cases demonstrated conflicting findings: persistence of allergy to non-FRPs (methylisothiazolinone [MI]) but resolution of allergy to formaldehyde⁸; persistence of allergy to corticosteroids (budesonide and alclometasone)⁹; persistence of allergy to an antibiotic (neomycin sulfate) but resolution of allergies to a different antibiotic (bacitracin), glues (ethyl acrylate), bleach, and glutaraldehyde⁹; persistence of nickel allergy but resolution of allergies to fragrances (cinnamic aldehyde, balsam of Peru) and non-FRPs (methylchloroisothiazolinone or MI)¹⁰; and persistence of allergies to non-FRPs (MI) and FRPs (bronopol) but resolution of allergies to nickel, fragrances (hydroperoxides of linalool), and Compositae.¹¹ Additional case reports of positive patch test results while on dupilumab but with no pretreatment results for comparison include allergies to rubber additives,^12-14 nickel,¹⁴ textile dyes,¹⁴ cosmetic and hair care additives,^12,14,15 corticosteroids,¹⁵ FRPs,¹⁵ fragrances,^15,16 emulsifiers,¹⁶ and non-FRPs.¹⁷

An evident theme in the dupilumab patch-testing literature has been that results are variable and case specific: a given patient with ACD to an allergen will respond to dupilumab treatment and have subsequent negative patch testing, while another patient will not respond to dupilumab treatment and have persistent positive patch testing. This is likely because, in certain individuals, the allergen-immune system combination shifts ACD pathogenesis from a purely T_H1 response to at least a partial T_H2 response, thus allowing for benefit from dupilumab therapy. T helper 1 cell–mediated ACD should not be affected by dupilumab; therefore, reliable results can be elucidated from patch testing despite the drug.

Final Thoughts

We propose that AD patients with residual disease after taking dupilumab undergo patch testing. Positive results indicate allergens that are not inhibited by the drug. Patients will need to follow strict allergen avoidance to resolve this component of their disease; failure to improve might suggest the result was a nonrelevant positive.

If patch testing is negative, an alternative cause for residual disease must be sought. We do not recommend stopping dupilumab prior to patch testing to avoid a disease flare from AD or possible T_H2-mediated ACD.

In patients with persistent atopic dermatitis (AD) who are taking dupilumab, is there benefit of patch testing to determine if allergic contact dermatitis (ACD) also is contributing to their disease? Results of patch testing are likely be influenced by the immunomodulatory effects of dupilumab. Similar to the recommendation for patients to refrain from using topical or systemic corticosteroids for 1 week or more prior to patch testing to eliminate false negatives, we reviewed the literature to create practice guidelines for dermatologists regarding patch testing while a patient is taking dupilumab.

Pathophysiology and Pathomechanism

Dupilumab functions through the blockade of T helper 2 (T_H2) cells; ACD is propagated through the T helper 1 (T_H1) cellular pathway. However, patients with ACD that is unresponsive to allergen avoidance and traditional therapies, such as topical and oral corticosteroids, have responded to dupilumab. The more common reports of this responsiveness are with fragrances; multiple case series described patients with ACD to fragrance mix I¹ and balsam of Peru^1,2 who improved on dupilumab when other treatments failed. There also are reports of response when ACD was secondary to nickel,^2,3p-phenylenediamine,¹ Compositae,⁴ and non–formaldehyde-releasing preservatives (non-FRPs).⁵ Therefore, not all ACD is propagated through the T_H1 cellular pathway.

As noted in these cases, ACD can be a response to an allergen whose pathogenesis involves the T_H2 pathway or when patient characteristics favor a T_H2 response. It has been suggested that AD patients are more susceptible to T_H2-mediated contact sensitization to less-potent allergens, such as fragrances.⁶

Patch Test Results

Positive patch test results for allergens have been reported while patients are on dupilumab therapy, including a few studies in which results prior to starting dupilumab were compared with those while patients were on dupilumab therapy. In a retrospective chart review of 48 patients on dupilumab for AD with persistent disease, 23 patients were patch tested before and during dupilumab therapy. In these patients, the majority of contact allergies were persistent and only 10% (13/125) of patch test–positive results resolved on dupilumab therapy.⁷ Contact allergies that resolved included those to emulsifiers (propylene glycol, Amerchol L101 [lanolin-containing products found in cosmetics and other goods], dimethylaminopropylamine), fragrances (fragrance mix I, balsam of Peru), sunscreens (sulisobenzone, phenylbenzimidazole-5-sulfonic acid), and metals (vanadium chloride, phenylmercuric acetate).⁷ The following results observed in individual cases demonstrated conflicting findings: persistence of allergy to non-FRPs (methylisothiazolinone [MI]) but resolution of allergy to formaldehyde⁸; persistence of allergy to corticosteroids (budesonide and alclometasone)⁹; persistence of allergy to an antibiotic (neomycin sulfate) but resolution of allergies to a different antibiotic (bacitracin), glues (ethyl acrylate), bleach, and glutaraldehyde⁹; persistence of nickel allergy but resolution of allergies to fragrances (cinnamic aldehyde, balsam of Peru) and non-FRPs (methylchloroisothiazolinone or MI)¹⁰; and persistence of allergies to non-FRPs (MI) and FRPs (bronopol) but resolution of allergies to nickel, fragrances (hydroperoxides of linalool), and Compositae.¹¹ Additional case reports of positive patch test results while on dupilumab but with no pretreatment results for comparison include allergies to rubber additives,^12-14 nickel,¹⁴ textile dyes,¹⁴ cosmetic and hair care additives,^12,14,15 corticosteroids,¹⁵ FRPs,¹⁵ fragrances,^15,16 emulsifiers,¹⁶ and non-FRPs.¹⁷

An evident theme in the dupilumab patch-testing literature has been that results are variable and case specific: a given patient with ACD to an allergen will respond to dupilumab treatment and have subsequent negative patch testing, while another patient will not respond to dupilumab treatment and have persistent positive patch testing. This is likely because, in certain individuals, the allergen-immune system combination shifts ACD pathogenesis from a purely T_H1 response to at least a partial T_H2 response, thus allowing for benefit from dupilumab therapy. T helper 1 cell–mediated ACD should not be affected by dupilumab; therefore, reliable results can be elucidated from patch testing despite the drug.

Final Thoughts

We propose that AD patients with residual disease after taking dupilumab undergo patch testing. Positive results indicate allergens that are not inhibited by the drug. Patients will need to follow strict allergen avoidance to resolve this component of their disease; failure to improve might suggest the result was a nonrelevant positive.

If patch testing is negative, an alternative cause for residual disease must be sought. We do not recommend stopping dupilumab prior to patch testing to avoid a disease flare from AD or possible T_H2-mediated ACD.

Practice Gap

The topical calcineurin inhibitors (TCIs) tacrolimus and pimecrolimus are US Food and Drug Administration approved for the treatment of atopic dermatitis.¹ In addition, these 2 drugs are utilized off label for many other dermatologic conditions, including vitiligo, psoriasis, and periorificial dermatitis. They can be used safely for prolonged periods and on sensitive areas, including the face.

Treatment with a TCI provides advantages over topical steroids, which can cause atrophy, telangiectasia, dyspigmentation, ocular hypertension, cataracts, and tachyphylaxis after prolonged use. Adverse events resulting from use of a TCI most commonly include transient burning, warmth, and erythema in areas of application. Patients typically acclimate to these effects after a few consecutive days of use.

Localized flushing after alcohol ingestion is a known potential side effect of TCIs¹; however, this association may be underappreciated and underreported to patients.

Counseling Patients Taking TCIs

Topical calcineurin inhibitors cause alcohol-induced flushing on areas of application (Figures 1 and 2) in approximately 3.4% to 6.9% of patients.¹ The reaction has been reported with both topical TCIs but more often is noted with tacrolimus.^2,3 Typically, flushing begins 2 to 4 weeks after treatment is initiated and within 5 to 20 minutes after alcohol intake.⁴ The phenomenon is self-limited; erythema typically resolves in 20 to 60 minutes.

Topical calcineurin inhibitors are hypothesized to cause alcohol-induced flushing by locally inhibiting acetaldehyde dehydrogenase, an enzyme necessary for alcohol metabolism. This leads to accumulation of acetaldehyde, a by-product of alcohol metabolism, which indirectly causes concentrated vasodilation by means of prostaglandins, histamines, and other vasodilatory mediators. The combination of ethanol and a TCI also might induce release of neuropeptides, which could cause vasodilation.⁴

Alcohol-related flushing commonly is seen among individuals who are aldehyde dehydrogenase 2 (ALDH2) deficient; it is sometimes accompanied by nausea, headache, and tachycardia. The same pathway is implicated in disulfiram reactions, to a more intense and systemic degree, to discourage alcohol intake.

Oral calcineurin inhibitors are not reported to cause generalized flushing, perhaps because of differences in the relative dose. For example, topical tacrolimus 0.1% is 1 mg/g that is applied to a relatively small body surface area; oral calcineurin inhibitors are dosed at a range of 1 to 15 mg for an entire person.

Techniques to Avoid Flushing

Discontinuing a TCI altogether leads to resolution of associated adverse effects, including flushing, typically within weeks to 1 month.⁵ Alternatively, oral aspirin (81 mg) might eliminate or attenuate symptoms, as documented in a double-blind, controlled trial in which relief of TCI-induced flushing after consuming wine was investigated.⁶

Another approach (albeit nontraditional) is for patients who experience this phenomenon to “wet their whistles” with an alcoholic drink before a social engagement. After flushing resolves in 20 to 60 minutes, subsequent drinks do not appear to elicit symptoms again in most patients. That said, we stop short of calling this tip “doctor’s orders.”

Practical Implication

Counseling patients who will be using a TCI—tacrolimus or pimecrolimus—about the potential for these drugs to produce localized flushing after alcohol ingestion as well as techniques for lessening or eliminating this adverse effect are important facets of their dermatologic care.

Practice Gap

The topical calcineurin inhibitors (TCIs) tacrolimus and pimecrolimus are US Food and Drug Administration approved for the treatment of atopic dermatitis.¹ In addition, these 2 drugs are utilized off label for many other dermatologic conditions, including vitiligo, psoriasis, and periorificial dermatitis. They can be used safely for prolonged periods and on sensitive areas, including the face.

Treatment with a TCI provides advantages over topical steroids, which can cause atrophy, telangiectasia, dyspigmentation, ocular hypertension, cataracts, and tachyphylaxis after prolonged use. Adverse events resulting from use of a TCI most commonly include transient burning, warmth, and erythema in areas of application. Patients typically acclimate to these effects after a few consecutive days of use.

Localized flushing after alcohol ingestion is a known potential side effect of TCIs¹; however, this association may be underappreciated and underreported to patients.

Counseling Patients Taking TCIs

Topical calcineurin inhibitors cause alcohol-induced flushing on areas of application (Figures 1 and 2) in approximately 3.4% to 6.9% of patients.¹ The reaction has been reported with both topical TCIs but more often is noted with tacrolimus.^2,3 Typically, flushing begins 2 to 4 weeks after treatment is initiated and within 5 to 20 minutes after alcohol intake.⁴ The phenomenon is self-limited; erythema typically resolves in 20 to 60 minutes.

Topical calcineurin inhibitors are hypothesized to cause alcohol-induced flushing by locally inhibiting acetaldehyde dehydrogenase, an enzyme necessary for alcohol metabolism. This leads to accumulation of acetaldehyde, a by-product of alcohol metabolism, which indirectly causes concentrated vasodilation by means of prostaglandins, histamines, and other vasodilatory mediators. The combination of ethanol and a TCI also might induce release of neuropeptides, which could cause vasodilation.⁴

Alcohol-related flushing commonly is seen among individuals who are aldehyde dehydrogenase 2 (ALDH2) deficient; it is sometimes accompanied by nausea, headache, and tachycardia. The same pathway is implicated in disulfiram reactions, to a more intense and systemic degree, to discourage alcohol intake.

Oral calcineurin inhibitors are not reported to cause generalized flushing, perhaps because of differences in the relative dose. For example, topical tacrolimus 0.1% is 1 mg/g that is applied to a relatively small body surface area; oral calcineurin inhibitors are dosed at a range of 1 to 15 mg for an entire person.

Techniques to Avoid Flushing

Discontinuing a TCI altogether leads to resolution of associated adverse effects, including flushing, typically within weeks to 1 month.⁵ Alternatively, oral aspirin (81 mg) might eliminate or attenuate symptoms, as documented in a double-blind, controlled trial in which relief of TCI-induced flushing after consuming wine was investigated.⁶

Another approach (albeit nontraditional) is for patients who experience this phenomenon to “wet their whistles” with an alcoholic drink before a social engagement. After flushing resolves in 20 to 60 minutes, subsequent drinks do not appear to elicit symptoms again in most patients. That said, we stop short of calling this tip “doctor’s orders.”

Practical Implication

Counseling patients who will be using a TCI—tacrolimus or pimecrolimus—about the potential for these drugs to produce localized flushing after alcohol ingestion as well as techniques for lessening or eliminating this adverse effect are important facets of their dermatologic care.

Practice Gap

The topical calcineurin inhibitors (TCIs) tacrolimus and pimecrolimus are US Food and Drug Administration approved for the treatment of atopic dermatitis.¹ In addition, these 2 drugs are utilized off label for many other dermatologic conditions, including vitiligo, psoriasis, and periorificial dermatitis. They can be used safely for prolonged periods and on sensitive areas, including the face.

Treatment with a TCI provides advantages over topical steroids, which can cause atrophy, telangiectasia, dyspigmentation, ocular hypertension, cataracts, and tachyphylaxis after prolonged use. Adverse events resulting from use of a TCI most commonly include transient burning, warmth, and erythema in areas of application. Patients typically acclimate to these effects after a few consecutive days of use.

Localized flushing after alcohol ingestion is a known potential side effect of TCIs¹; however, this association may be underappreciated and underreported to patients.

Counseling Patients Taking TCIs

Topical calcineurin inhibitors cause alcohol-induced flushing on areas of application (Figures 1 and 2) in approximately 3.4% to 6.9% of patients.¹ The reaction has been reported with both topical TCIs but more often is noted with tacrolimus.^2,3 Typically, flushing begins 2 to 4 weeks after treatment is initiated and within 5 to 20 minutes after alcohol intake.⁴ The phenomenon is self-limited; erythema typically resolves in 20 to 60 minutes.

Topical calcineurin inhibitors are hypothesized to cause alcohol-induced flushing by locally inhibiting acetaldehyde dehydrogenase, an enzyme necessary for alcohol metabolism. This leads to accumulation of acetaldehyde, a by-product of alcohol metabolism, which indirectly causes concentrated vasodilation by means of prostaglandins, histamines, and other vasodilatory mediators. The combination of ethanol and a TCI also might induce release of neuropeptides, which could cause vasodilation.⁴

Alcohol-related flushing commonly is seen among individuals who are aldehyde dehydrogenase 2 (ALDH2) deficient; it is sometimes accompanied by nausea, headache, and tachycardia. The same pathway is implicated in disulfiram reactions, to a more intense and systemic degree, to discourage alcohol intake.

Oral calcineurin inhibitors are not reported to cause generalized flushing, perhaps because of differences in the relative dose. For example, topical tacrolimus 0.1% is 1 mg/g that is applied to a relatively small body surface area; oral calcineurin inhibitors are dosed at a range of 1 to 15 mg for an entire person.

Techniques to Avoid Flushing

Discontinuing a TCI altogether leads to resolution of associated adverse effects, including flushing, typically within weeks to 1 month.⁵ Alternatively, oral aspirin (81 mg) might eliminate or attenuate symptoms, as documented in a double-blind, controlled trial in which relief of TCI-induced flushing after consuming wine was investigated.⁶

Another approach (albeit nontraditional) is for patients who experience this phenomenon to “wet their whistles” with an alcoholic drink before a social engagement. After flushing resolves in 20 to 60 minutes, subsequent drinks do not appear to elicit symptoms again in most patients. That said, we stop short of calling this tip “doctor’s orders.”

Practical Implication

Counseling patients who will be using a TCI—tacrolimus or pimecrolimus—about the potential for these drugs to produce localized flushing after alcohol ingestion as well as techniques for lessening or eliminating this adverse effect are important facets of their dermatologic care.

Lichen planus is a chronic inflammatory mucocutaneous disease that usually affects the skin and/or the genital and oral mucosae.^1,2 This disease classically presents with clinical relapses or outbreaks that alternate with periods of remission or latency. Oral lichen planus (OLP) can present with or without extraoral manifestation. It sometimes is difficult to differentiate OLP from oral lichenoid reactions, which can be related to dental materials, some drugs, and systemic conditions or can be idiopathic.^1,2

Oral lichen planus is one of the most common noninfectious diseases of the oral cavity, with a reported prevalence of 1% worldwide and marked geographical differences. In Europe, the prevalence of OLP ranges from 1% to 2%.^3,4 It is more frequent in women (1.5:1 to 2:1) and usually appears in the fourth and fifth decades of life.^1-4

The causes of OLP have not been entirely elucidated, but it is broadly accepted that there is a deregulation on different T lymphocytes that in turn causes effects on CD8 lymphocytes in response to an external noxa. This unknown “trigger” or starting factor also produces an impact on basal keratinocytes. Therefore, the pathogenesis of lichen planus is influenced by a series of cellular events mediated by different cytokines.^2,5,6 Among these, tumor necrosis factor α and IL-1 are known to have important roles in the disease. More recently, other cytokines, such as IL-4, secreted by type 2 helper T cells, also have been related to the development and progression of the oral lesions.^5,6 In addition to the factors that generate the onset of the disease, there are others that may precipitate clinical outbreaks. Different factors have been related to the progression of the disease, influencing the initiation, perpetuation, and/or worsening of OLP lesions.^1,2 Exactly how these factors affect disease progression is another challenging question. The list of possible or potential factors related to disease progression is long; nonetheless, in the vast majority, a clear explanation at a molecular level has not been clearly demonstrated.^2,5

Conventionally, 6 clinical presentations of OLP lesions divided into 2 main groups have been described in the oral cavity: white forms (reticular, papular, and plaquelike) and red forms (erythematous, atrophic-erosive, and bullous).^1,7-9

Oral lichen planus mainly is treated with topically or systemically administered steroids based on the presence of symptoms such as pain and inability to perform daily activities (eg, eating, talking).^5,10 The treatment of choice often is based on the professional’s experience, as there are no broadly accepted national or international clinical practice guidelines on steroid type, administration route, dose, vehicle for administration, or maintenance.¹¹ Despite this lack of unified criteria, different topical and systemic steroid administration protocols allow a reduction in the symptoms or even the disappearance of the red lesions to be achieved in many cases. Unfortunately, there are many patients with lesions refractory to standard treatments for OLP.¹² Several alternatives for these patients have been described in the literature, though on many occasions these alternatives present substantial side effects for the patient.¹³ The search for an effective treatment without side effects is still challenging. One of the treatments tested under this premise has been the application of plasma rich in growth factors (PRGF) by means of infiltration or topical application, in both cases obtaining good results without side effects.¹⁴

We sought to analyze the information from a case series of patients treated at the Eduardo Anitua Clinic (Vitoria-Gasteiz, Spain) and describe the results and follow-up of patients with erosive OLP refractory to standard therapy who have been successfully treated by local infiltration of PRGF as the only treatment.

Material and Methods

Patients—We included data from the database of the clinical center with de-identified information of patients with erosive OLP diagnosed clinically and histopathologically who did not respond to conventional treatment (ie, topical and/or systemic corticosteroids [depending on the case]) as well as patients who presented with extensive erosive OLP with systemic involvement and whose systemic treatment was not effective in resolving oral manifestations.

Therapies Administered and Evaluations—Lesions refractory to conventional corticosteroid protocols had been previously treated for 30 days with 0.5% triamcinolone acetonide mouth rinse followed by a cycle of 1% triamcinolone acetonide mouth rinse. Subsequently, a cycle of oral corticosteroids (prednisone for 30 days: 1 mg/kg/d in a single morning dose with staged reduction after the first week) had been administered. One dayafter the corticosteroid treatment was suspended, the patients were treated by PRGF-Endoret (BTI Biotechnology Institute) infiltration following the protocol described by Anitua et al.^15,16

Before starting the infiltrations with PRGF, the patient had been asked to rate the pain level on a visual analog scale (VAS) of 1 to 10, with 10 being the most intense imaginable pain. Pain score was subsequently rated and registered during every visit. An initial photograph of the lesion also was obtained to establish a starting point for further comparisons of clinical evolution of the lesions.

Prior to each infiltration, the plasma was separated into 2 fractions. The second fraction was the one that corresponded to the highest number of platelets and included the 2 mL of plasma just above the white series (or buffy coat). This fraction of plasma was the one used to infiltrate the lesions.

Plasma rich in growth factors was activated just before infiltration. The activation was done by adding 10% calcium chloride. Once activated, it was infiltrated into the active lesion using a 31-G × 1/6-in hypodermic needle and a 2-mL Luer-lock syringe. Infiltrations were performed without anesthesia. Four punctures were made for each ulcerative lesion, dividing the lesion into 4 points: upper, lower, right, and left. Plasma rich in growth factors was infiltrated until a slight blanching was observed in the surrounding tissue. At that moment, the infiltration was stopped and was carried out in the next infiltration site.

One treatment session was performed per week, with follow-up 1 week after treatment. In the control visit, the state of the lesions was re-evaluated, and it was decided whether new infiltrations were needed. The treatment was finished when complete epithelialization of the lesion was visualized or the associated symptoms disappeared. At each visit, photographs were taken, and the patient assessed the severity of pain on the VAS.

Statistical Analysis—A Shapiro-Wilk test was carried out with the obtained data to check the normal distribution of the sample. The evolution of pain during the study was compared by paired t test. The qualitative variables were described by means of a frequency analysis. Quantitative variables were described by the mean and the SD. The data were analyzed with SPSS V15.0 for Windows (SPSS Inc). P<.05 showed statistical significance.

Results

A total of 15 patients were included in the study, all with atrophic-erosive lichen planus. Two patients were male, and 13 were female. The mean age (SD) of the patients included in the study was 55.27 (14.19) years. The mean number of outbreaks per year (SD) was 3.2 (1.7), with a range of 1 to 8 outbreaks.

Healing of OLP Lesions—The number of treatment sessions to achieve complete healing varied among the patients (Figures 1 and 2). Ten patients (66.7%) required a single session, 2 patients (13.3%) required 2 sessions, and 3 patients (20%) required 3 sessions. The mean time (SD) without lesions for the patients who required a single session was 10.9 (5.2) months (range, 6–24 months).

Pain Assessment—The mean (SD) score obtained on the VAS before treatment with PRGF was 8.27 (1.16); this score dropped to 1.27 (1.53) after the first treatment session and was a statistically significant difference (P=.006).

For those patients requiring more than 1 session, the mean (SD) pain scores decreased by 0.75 (0.97) points and 0 points after the first and second sessions of treatment, respectively. The mean (SD) amount of PRGF infiltrated in each patient in the first session was 2.60 (0.63) mL. In the second session, the mean (SD) amount was 1.2 (0.33) mL; these differences were statistically significant (P=.008). In the last session, the mean (SD) amount was 1.1 (0.22) mL.

Follow-up and Adverse Effects—The mean (SD) follow-up time was 47.16 (15.78) months. The patients were free of symptoms, and there were no adverse effects derived from the treatment during follow-up.

Comment

The primary goal of OLP treatment is to stop the outbreaks.^1,9,13 The lack of potency of corticosteroids in some patients with OLP could be due in part to the inadequate selection of the vehicle (ointment/oral rinse) for the extension and characteristics of the lesion or because of an inappropriate prescription dose, time, and/or frequency, as described by González-Moles.¹⁷ However, even when using an appropriate protocol, some lesions are resistant to topical treatment and require other therapeutic modalities.^1,9,13 Previously proposed topical treatments include different immunosuppressants, such as the mammalian target of rapamycin, tacrolimus ointment 0.1%, pimecrolimus cream 1%, or cyclosporine A (50–100 mg/mL) formulations.¹⁸ Nevertheless, these drugs seem to have a greater number of side effects than topical steroids, and tacrolimus has been associated with cases of oral malignancy after continuing treatment.¹⁵

Severe and/or recalcitrant lesions and extraoral involvement have been successfully treated with systemic prednisone (40–80 mg/d).^1,9,13 Nevertheless, systemic corticosteroid toxicity requires that these treatments should be used only when necessary at the lowest possible dose and for the shortest possible duration.¹⁹ Other nonpharmacologic options for treatment are photodynamic, UV, and low-level laser therapy.^20,21 They have been accepted as supplementary modalities in different inflammatory skin conditions but present important technical requirements. Their effectiveness in corticosteroid-resistant cases have not been definitively assessed. Interestingly, promising results recently have been reported by Bennardo et al²² when comparing the efficacy of autologous platelet concentrates with triamcinolone injection.

In our study, the use of PRGF stopped the lesions’ evolution since the first treatment session, reducing them by 6.5-fold. The positive effects observed may have been promoted by the activity of different proteins present in PRGF (eg, platelet-derived growth factor, vascular endothelial growth factor, transforming growth factor, epidermal growth factor, fibroblast growth factor, fibronectin). These molecules contribute to collagen synthesis; angiogenesis; endothelial cell migration and proliferation; or keratinocyte cell migration, proliferation, differentiation, growth, and migration—phenomena that are essential for healing and re-epithelialization.^23-25

Different studies also have supported an anti-inflammatory effect of PRGF mediated by an inhibition of the transcription of nuclear factor–κB and the expression of cyclooxygenase-2 and chemokine receptor type 4 produced by its high content of hepatocyte growth factor or the reduction of inflammatory marker expression, such as intercellular adhesion molecule 1. The development of an efficient 3-dimensional fibrin scaffold formation that occurs after PRGF administration also could facilitate healing, helping some cell populations to guide their position and function.^23-25

Limitations of our study include the small number of patients and the absence of a control group. The higher number of female patients in the study did not seem to affect the results, as differences related to gender have not been reported when treating patients with OLP with autologous platelet concentrates or other modalities of treatment.

Conclusion

Results from our study indicate that the use of PRGF could be a new treatment option for OLP cases refractory to conventional therapy. No complications were observed during the treatment procedure or during the complete follow-up period. Nonetheless, new prospective studies with a greater number of patients and longer follow-up periods are needed to confirm these preliminary results.

Lichen planus is a chronic inflammatory mucocutaneous disease that usually affects the skin and/or the genital and oral mucosae.^1,2 This disease classically presents with clinical relapses or outbreaks that alternate with periods of remission or latency. Oral lichen planus (OLP) can present with or without extraoral manifestation. It sometimes is difficult to differentiate OLP from oral lichenoid reactions, which can be related to dental materials, some drugs, and systemic conditions or can be idiopathic.^1,2

Oral lichen planus is one of the most common noninfectious diseases of the oral cavity, with a reported prevalence of 1% worldwide and marked geographical differences. In Europe, the prevalence of OLP ranges from 1% to 2%.^3,4 It is more frequent in women (1.5:1 to 2:1) and usually appears in the fourth and fifth decades of life.^1-4

The causes of OLP have not been entirely elucidated, but it is broadly accepted that there is a deregulation on different T lymphocytes that in turn causes effects on CD8 lymphocytes in response to an external noxa. This unknown “trigger” or starting factor also produces an impact on basal keratinocytes. Therefore, the pathogenesis of lichen planus is influenced by a series of cellular events mediated by different cytokines.^2,5,6 Among these, tumor necrosis factor α and IL-1 are known to have important roles in the disease. More recently, other cytokines, such as IL-4, secreted by type 2 helper T cells, also have been related to the development and progression of the oral lesions.^5,6 In addition to the factors that generate the onset of the disease, there are others that may precipitate clinical outbreaks. Different factors have been related to the progression of the disease, influencing the initiation, perpetuation, and/or worsening of OLP lesions.^1,2 Exactly how these factors affect disease progression is another challenging question. The list of possible or potential factors related to disease progression is long; nonetheless, in the vast majority, a clear explanation at a molecular level has not been clearly demonstrated.^2,5

Conventionally, 6 clinical presentations of OLP lesions divided into 2 main groups have been described in the oral cavity: white forms (reticular, papular, and plaquelike) and red forms (erythematous, atrophic-erosive, and bullous).^1,7-9

Oral lichen planus mainly is treated with topically or systemically administered steroids based on the presence of symptoms such as pain and inability to perform daily activities (eg, eating, talking).^5,10 The treatment of choice often is based on the professional’s experience, as there are no broadly accepted national or international clinical practice guidelines on steroid type, administration route, dose, vehicle for administration, or maintenance.¹¹ Despite this lack of unified criteria, different topical and systemic steroid administration protocols allow a reduction in the symptoms or even the disappearance of the red lesions to be achieved in many cases. Unfortunately, there are many patients with lesions refractory to standard treatments for OLP.¹² Several alternatives for these patients have been described in the literature, though on many occasions these alternatives present substantial side effects for the patient.¹³ The search for an effective treatment without side effects is still challenging. One of the treatments tested under this premise has been the application of plasma rich in growth factors (PRGF) by means of infiltration or topical application, in both cases obtaining good results without side effects.¹⁴

We sought to analyze the information from a case series of patients treated at the Eduardo Anitua Clinic (Vitoria-Gasteiz, Spain) and describe the results and follow-up of patients with erosive OLP refractory to standard therapy who have been successfully treated by local infiltration of PRGF as the only treatment.

Material and Methods

Patients—We included data from the database of the clinical center with de-identified information of patients with erosive OLP diagnosed clinically and histopathologically who did not respond to conventional treatment (ie, topical and/or systemic corticosteroids [depending on the case]) as well as patients who presented with extensive erosive OLP with systemic involvement and whose systemic treatment was not effective in resolving oral manifestations.

Therapies Administered and Evaluations—Lesions refractory to conventional corticosteroid protocols had been previously treated for 30 days with 0.5% triamcinolone acetonide mouth rinse followed by a cycle of 1% triamcinolone acetonide mouth rinse. Subsequently, a cycle of oral corticosteroids (prednisone for 30 days: 1 mg/kg/d in a single morning dose with staged reduction after the first week) had been administered. One dayafter the corticosteroid treatment was suspended, the patients were treated by PRGF-Endoret (BTI Biotechnology Institute) infiltration following the protocol described by Anitua et al.^15,16

Before starting the infiltrations with PRGF, the patient had been asked to rate the pain level on a visual analog scale (VAS) of 1 to 10, with 10 being the most intense imaginable pain. Pain score was subsequently rated and registered during every visit. An initial photograph of the lesion also was obtained to establish a starting point for further comparisons of clinical evolution of the lesions.

Prior to each infiltration, the plasma was separated into 2 fractions. The second fraction was the one that corresponded to the highest number of platelets and included the 2 mL of plasma just above the white series (or buffy coat). This fraction of plasma was the one used to infiltrate the lesions.

Plasma rich in growth factors was activated just before infiltration. The activation was done by adding 10% calcium chloride. Once activated, it was infiltrated into the active lesion using a 31-G × 1/6-in hypodermic needle and a 2-mL Luer-lock syringe. Infiltrations were performed without anesthesia. Four punctures were made for each ulcerative lesion, dividing the lesion into 4 points: upper, lower, right, and left. Plasma rich in growth factors was infiltrated until a slight blanching was observed in the surrounding tissue. At that moment, the infiltration was stopped and was carried out in the next infiltration site.

One treatment session was performed per week, with follow-up 1 week after treatment. In the control visit, the state of the lesions was re-evaluated, and it was decided whether new infiltrations were needed. The treatment was finished when complete epithelialization of the lesion was visualized or the associated symptoms disappeared. At each visit, photographs were taken, and the patient assessed the severity of pain on the VAS.

Statistical Analysis—A Shapiro-Wilk test was carried out with the obtained data to check the normal distribution of the sample. The evolution of pain during the study was compared by paired t test. The qualitative variables were described by means of a frequency analysis. Quantitative variables were described by the mean and the SD. The data were analyzed with SPSS V15.0 for Windows (SPSS Inc). P<.05 showed statistical significance.

Results

A total of 15 patients were included in the study, all with atrophic-erosive lichen planus. Two patients were male, and 13 were female. The mean age (SD) of the patients included in the study was 55.27 (14.19) years. The mean number of outbreaks per year (SD) was 3.2 (1.7), with a range of 1 to 8 outbreaks.

Healing of OLP Lesions—The number of treatment sessions to achieve complete healing varied among the patients (Figures 1 and 2). Ten patients (66.7%) required a single session, 2 patients (13.3%) required 2 sessions, and 3 patients (20%) required 3 sessions. The mean time (SD) without lesions for the patients who required a single session was 10.9 (5.2) months (range, 6–24 months).

Pain Assessment—The mean (SD) score obtained on the VAS before treatment with PRGF was 8.27 (1.16); this score dropped to 1.27 (1.53) after the first treatment session and was a statistically significant difference (P=.006).

For those patients requiring more than 1 session, the mean (SD) pain scores decreased by 0.75 (0.97) points and 0 points after the first and second sessions of treatment, respectively. The mean (SD) amount of PRGF infiltrated in each patient in the first session was 2.60 (0.63) mL. In the second session, the mean (SD) amount was 1.2 (0.33) mL; these differences were statistically significant (P=.008). In the last session, the mean (SD) amount was 1.1 (0.22) mL.

Follow-up and Adverse Effects—The mean (SD) follow-up time was 47.16 (15.78) months. The patients were free of symptoms, and there were no adverse effects derived from the treatment during follow-up.

Comment

The primary goal of OLP treatment is to stop the outbreaks.^1,9,13 The lack of potency of corticosteroids in some patients with OLP could be due in part to the inadequate selection of the vehicle (ointment/oral rinse) for the extension and characteristics of the lesion or because of an inappropriate prescription dose, time, and/or frequency, as described by González-Moles.¹⁷ However, even when using an appropriate protocol, some lesions are resistant to topical treatment and require other therapeutic modalities.^1,9,13 Previously proposed topical treatments include different immunosuppressants, such as the mammalian target of rapamycin, tacrolimus ointment 0.1%, pimecrolimus cream 1%, or cyclosporine A (50–100 mg/mL) formulations.¹⁸ Nevertheless, these drugs seem to have a greater number of side effects than topical steroids, and tacrolimus has been associated with cases of oral malignancy after continuing treatment.¹⁵

Severe and/or recalcitrant lesions and extraoral involvement have been successfully treated with systemic prednisone (40–80 mg/d).^1,9,13 Nevertheless, systemic corticosteroid toxicity requires that these treatments should be used only when necessary at the lowest possible dose and for the shortest possible duration.¹⁹ Other nonpharmacologic options for treatment are photodynamic, UV, and low-level laser therapy.^20,21 They have been accepted as supplementary modalities in different inflammatory skin conditions but present important technical requirements. Their effectiveness in corticosteroid-resistant cases have not been definitively assessed. Interestingly, promising results recently have been reported by Bennardo et al²² when comparing the efficacy of autologous platelet concentrates with triamcinolone injection.

In our study, the use of PRGF stopped the lesions’ evolution since the first treatment session, reducing them by 6.5-fold. The positive effects observed may have been promoted by the activity of different proteins present in PRGF (eg, platelet-derived growth factor, vascular endothelial growth factor, transforming growth factor, epidermal growth factor, fibroblast growth factor, fibronectin). These molecules contribute to collagen synthesis; angiogenesis; endothelial cell migration and proliferation; or keratinocyte cell migration, proliferation, differentiation, growth, and migration—phenomena that are essential for healing and re-epithelialization.^23-25

Different studies also have supported an anti-inflammatory effect of PRGF mediated by an inhibition of the transcription of nuclear factor–κB and the expression of cyclooxygenase-2 and chemokine receptor type 4 produced by its high content of hepatocyte growth factor or the reduction of inflammatory marker expression, such as intercellular adhesion molecule 1. The development of an efficient 3-dimensional fibrin scaffold formation that occurs after PRGF administration also could facilitate healing, helping some cell populations to guide their position and function.^23-25

Limitations of our study include the small number of patients and the absence of a control group. The higher number of female patients in the study did not seem to affect the results, as differences related to gender have not been reported when treating patients with OLP with autologous platelet concentrates or other modalities of treatment.

Conclusion

Results from our study indicate that the use of PRGF could be a new treatment option for OLP cases refractory to conventional therapy. No complications were observed during the treatment procedure or during the complete follow-up period. Nonetheless, new prospective studies with a greater number of patients and longer follow-up periods are needed to confirm these preliminary results.

Lichen planus is a chronic inflammatory mucocutaneous disease that usually affects the skin and/or the genital and oral mucosae.^1,2 This disease classically presents with clinical relapses or outbreaks that alternate with periods of remission or latency. Oral lichen planus (OLP) can present with or without extraoral manifestation. It sometimes is difficult to differentiate OLP from oral lichenoid reactions, which can be related to dental materials, some drugs, and systemic conditions or can be idiopathic.^1,2

Oral lichen planus is one of the most common noninfectious diseases of the oral cavity, with a reported prevalence of 1% worldwide and marked geographical differences. In Europe, the prevalence of OLP ranges from 1% to 2%.^3,4 It is more frequent in women (1.5:1 to 2:1) and usually appears in the fourth and fifth decades of life.^1-4

The causes of OLP have not been entirely elucidated, but it is broadly accepted that there is a deregulation on different T lymphocytes that in turn causes effects on CD8 lymphocytes in response to an external noxa. This unknown “trigger” or starting factor also produces an impact on basal keratinocytes. Therefore, the pathogenesis of lichen planus is influenced by a series of cellular events mediated by different cytokines.^2,5,6 Among these, tumor necrosis factor α and IL-1 are known to have important roles in the disease. More recently, other cytokines, such as IL-4, secreted by type 2 helper T cells, also have been related to the development and progression of the oral lesions.^5,6 In addition to the factors that generate the onset of the disease, there are others that may precipitate clinical outbreaks. Different factors have been related to the progression of the disease, influencing the initiation, perpetuation, and/or worsening of OLP lesions.^1,2 Exactly how these factors affect disease progression is another challenging question. The list of possible or potential factors related to disease progression is long; nonetheless, in the vast majority, a clear explanation at a molecular level has not been clearly demonstrated.^2,5

Conventionally, 6 clinical presentations of OLP lesions divided into 2 main groups have been described in the oral cavity: white forms (reticular, papular, and plaquelike) and red forms (erythematous, atrophic-erosive, and bullous).^1,7-9

Oral lichen planus mainly is treated with topically or systemically administered steroids based on the presence of symptoms such as pain and inability to perform daily activities (eg, eating, talking).^5,10 The treatment of choice often is based on the professional’s experience, as there are no broadly accepted national or international clinical practice guidelines on steroid type, administration route, dose, vehicle for administration, or maintenance.¹¹ Despite this lack of unified criteria, different topical and systemic steroid administration protocols allow a reduction in the symptoms or even the disappearance of the red lesions to be achieved in many cases. Unfortunately, there are many patients with lesions refractory to standard treatments for OLP.¹² Several alternatives for these patients have been described in the literature, though on many occasions these alternatives present substantial side effects for the patient.¹³ The search for an effective treatment without side effects is still challenging. One of the treatments tested under this premise has been the application of plasma rich in growth factors (PRGF) by means of infiltration or topical application, in both cases obtaining good results without side effects.¹⁴

We sought to analyze the information from a case series of patients treated at the Eduardo Anitua Clinic (Vitoria-Gasteiz, Spain) and describe the results and follow-up of patients with erosive OLP refractory to standard therapy who have been successfully treated by local infiltration of PRGF as the only treatment.

Material and Methods

Patients—We included data from the database of the clinical center with de-identified information of patients with erosive OLP diagnosed clinically and histopathologically who did not respond to conventional treatment (ie, topical and/or systemic corticosteroids [depending on the case]) as well as patients who presented with extensive erosive OLP with systemic involvement and whose systemic treatment was not effective in resolving oral manifestations.

Therapies Administered and Evaluations—Lesions refractory to conventional corticosteroid protocols had been previously treated for 30 days with 0.5% triamcinolone acetonide mouth rinse followed by a cycle of 1% triamcinolone acetonide mouth rinse. Subsequently, a cycle of oral corticosteroids (prednisone for 30 days: 1 mg/kg/d in a single morning dose with staged reduction after the first week) had been administered. One dayafter the corticosteroid treatment was suspended, the patients were treated by PRGF-Endoret (BTI Biotechnology Institute) infiltration following the protocol described by Anitua et al.^15,16

Before starting the infiltrations with PRGF, the patient had been asked to rate the pain level on a visual analog scale (VAS) of 1 to 10, with 10 being the most intense imaginable pain. Pain score was subsequently rated and registered during every visit. An initial photograph of the lesion also was obtained to establish a starting point for further comparisons of clinical evolution of the lesions.

Prior to each infiltration, the plasma was separated into 2 fractions. The second fraction was the one that corresponded to the highest number of platelets and included the 2 mL of plasma just above the white series (or buffy coat). This fraction of plasma was the one used to infiltrate the lesions.

Plasma rich in growth factors was activated just before infiltration. The activation was done by adding 10% calcium chloride. Once activated, it was infiltrated into the active lesion using a 31-G × 1/6-in hypodermic needle and a 2-mL Luer-lock syringe. Infiltrations were performed without anesthesia. Four punctures were made for each ulcerative lesion, dividing the lesion into 4 points: upper, lower, right, and left. Plasma rich in growth factors was infiltrated until a slight blanching was observed in the surrounding tissue. At that moment, the infiltration was stopped and was carried out in the next infiltration site.

One treatment session was performed per week, with follow-up 1 week after treatment. In the control visit, the state of the lesions was re-evaluated, and it was decided whether new infiltrations were needed. The treatment was finished when complete epithelialization of the lesion was visualized or the associated symptoms disappeared. At each visit, photographs were taken, and the patient assessed the severity of pain on the VAS.

Statistical Analysis—A Shapiro-Wilk test was carried out with the obtained data to check the normal distribution of the sample. The evolution of pain during the study was compared by paired t test. The qualitative variables were described by means of a frequency analysis. Quantitative variables were described by the mean and the SD. The data were analyzed with SPSS V15.0 for Windows (SPSS Inc). P<.05 showed statistical significance.

Results

A total of 15 patients were included in the study, all with atrophic-erosive lichen planus. Two patients were male, and 13 were female. The mean age (SD) of the patients included in the study was 55.27 (14.19) years. The mean number of outbreaks per year (SD) was 3.2 (1.7), with a range of 1 to 8 outbreaks.

Healing of OLP Lesions—The number of treatment sessions to achieve complete healing varied among the patients (Figures 1 and 2). Ten patients (66.7%) required a single session, 2 patients (13.3%) required 2 sessions, and 3 patients (20%) required 3 sessions. The mean time (SD) without lesions for the patients who required a single session was 10.9 (5.2) months (range, 6–24 months).

Pain Assessment—The mean (SD) score obtained on the VAS before treatment with PRGF was 8.27 (1.16); this score dropped to 1.27 (1.53) after the first treatment session and was a statistically significant difference (P=.006).

For those patients requiring more than 1 session, the mean (SD) pain scores decreased by 0.75 (0.97) points and 0 points after the first and second sessions of treatment, respectively. The mean (SD) amount of PRGF infiltrated in each patient in the first session was 2.60 (0.63) mL. In the second session, the mean (SD) amount was 1.2 (0.33) mL; these differences were statistically significant (P=.008). In the last session, the mean (SD) amount was 1.1 (0.22) mL.

Follow-up and Adverse Effects—The mean (SD) follow-up time was 47.16 (15.78) months. The patients were free of symptoms, and there were no adverse effects derived from the treatment during follow-up.

Comment

The primary goal of OLP treatment is to stop the outbreaks.^1,9,13 The lack of potency of corticosteroids in some patients with OLP could be due in part to the inadequate selection of the vehicle (ointment/oral rinse) for the extension and characteristics of the lesion or because of an inappropriate prescription dose, time, and/or frequency, as described by González-Moles.¹⁷ However, even when using an appropriate protocol, some lesions are resistant to topical treatment and require other therapeutic modalities.^1,9,13 Previously proposed topical treatments include different immunosuppressants, such as the mammalian target of rapamycin, tacrolimus ointment 0.1%, pimecrolimus cream 1%, or cyclosporine A (50–100 mg/mL) formulations.¹⁸ Nevertheless, these drugs seem to have a greater number of side effects than topical steroids, and tacrolimus has been associated with cases of oral malignancy after continuing treatment.¹⁵

Severe and/or recalcitrant lesions and extraoral involvement have been successfully treated with systemic prednisone (40–80 mg/d).^1,9,13 Nevertheless, systemic corticosteroid toxicity requires that these treatments should be used only when necessary at the lowest possible dose and for the shortest possible duration.¹⁹ Other nonpharmacologic options for treatment are photodynamic, UV, and low-level laser therapy.^20,21 They have been accepted as supplementary modalities in different inflammatory skin conditions but present important technical requirements. Their effectiveness in corticosteroid-resistant cases have not been definitively assessed. Interestingly, promising results recently have been reported by Bennardo et al²² when comparing the efficacy of autologous platelet concentrates with triamcinolone injection.

In our study, the use of PRGF stopped the lesions’ evolution since the first treatment session, reducing them by 6.5-fold. The positive effects observed may have been promoted by the activity of different proteins present in PRGF (eg, platelet-derived growth factor, vascular endothelial growth factor, transforming growth factor, epidermal growth factor, fibroblast growth factor, fibronectin). These molecules contribute to collagen synthesis; angiogenesis; endothelial cell migration and proliferation; or keratinocyte cell migration, proliferation, differentiation, growth, and migration—phenomena that are essential for healing and re-epithelialization.^23-25

Different studies also have supported an anti-inflammatory effect of PRGF mediated by an inhibition of the transcription of nuclear factor–κB and the expression of cyclooxygenase-2 and chemokine receptor type 4 produced by its high content of hepatocyte growth factor or the reduction of inflammatory marker expression, such as intercellular adhesion molecule 1. The development of an efficient 3-dimensional fibrin scaffold formation that occurs after PRGF administration also could facilitate healing, helping some cell populations to guide their position and function.^23-25

Limitations of our study include the small number of patients and the absence of a control group. The higher number of female patients in the study did not seem to affect the results, as differences related to gender have not been reported when treating patients with OLP with autologous platelet concentrates or other modalities of treatment.

Conclusion

Results from our study indicate that the use of PRGF could be a new treatment option for OLP cases refractory to conventional therapy. No complications were observed during the treatment procedure or during the complete follow-up period. Nonetheless, new prospective studies with a greater number of patients and longer follow-up periods are needed to confirm these preliminary results.

The number of monoclonal antibodies developed for therapeutic use has rapidly expanded over the last decade due to their generally favorable adverse effect (AE) profiles and efficacy.¹ Tumor necrosis factor α inhibitors and general integrin antagonists are well-known examples of such monoclonal antibodies. Common conditions utilizing immunotherapy include inflammatory bowel diseases (IBDs), such as Crohn disease and ulcerative colitis (UC).²

The monoclonal antibody vedolizumab, approved in 2014 for moderate to severe UC and Crohn disease, selectively antagonizes α4β7 integrin to target a specific population of gastrointestinal T lymphocytes, preventing their mobilization to areas of inflammation.³ Adverse effects in patients treated with vedolizumab occur at a rate comparable to placebo and largely are considered nonserious^4,5; the most commonly reported AE is disease exacerbation (13%–17% of patients).^5,6 Published reports of cutaneous AEs at administration of vedolizumab include urticaria during infusion, appearance of cutaneous manifestations characteristic of IBD, psoriasis, Henoch-Schönlein purpura, and Sweet syndrome.^7-10

We present the case of a 61-year-old woman with UC who developed reactive granulomatous dermatitis (RGD), interstitial granulomatous dermatitis (IGD) type secondary to vedolizumab. This adverse reaction has not, to our knowledge, been previously reported.

Case Report

A 61-year-old woman with a medical history of UC treated with vedolizumab and myelodysplastic syndrome treated with intravenous immunoglobulin (due to hypogammaglobulinemia following allogeneic stem cell transplantation 14 months prior) presented with a concern of a rash. The patient had been in a baseline state of health until 1 week after receiving her second dose of vedolizumab, at which time she developed a mildly pruritic maculopapular rash on the back and chest. Triamcinolone ointment and hydroxyzine were recommended during an initial telehealth consultation with an oncologist with minimal improvement. The rash continued to spread distally with worsening pruritus.

The patient returned to her oncologist for a routine follow-up appointment 5 days after initial teleconsultation. She reported poor oral intake due to oropharyngeal pain and a worsening rash; her husband added a report of recent onset of somnolence. She was admitted to the hospital, and intravenous fluids were administered.

At admission, the patient was hypotensive; vital signs were otherwise normal. Physical examination revealed the oropharynx was erythematous. Pink lichenoid papules coalescing into plaques were present diffusely across the trunk, arms, and legs; the hands, feet, and face were spared (Figure 1).

Photograph courtesy of William Beuerlein, DO (Jacksonville, Florida).

A complete blood cell count and comprehensive metabolic panel were unremarkable. A lumbar puncture, chest radiograph, blood cultures, urinalysis, and urine cultures did not identify a clear infectious cause for the rash, though the workup for infection did raise concern about active cytomegalovirus (CMV) infection with colitis and pneumonitis. Computed tomography of the head showed no acute hemorrhage.

Dermatology was consulted and determined that the appearance of the rash was most consistent with a lichenoid drug eruption, likely secondary to vedolizumab that was administered 1 week before the rash onset. Analysis of a skin biopsy revealed a dense dermal histiocytic and lymphocytic infiltrate in close approximation to blood vessels, confirmed by immunohistochemical staining for CD45, CD43, CD68, CD34, c-KIT, and myeloperoxidase (Figures 2A and 2B). Colloidal iron staining of the specimen revealed no mucin (Figure 2C).

Photographs courtesy of William Beuerlein, DO (Jacksonville, Florida) and Angela Niehaus, MD (WinstonSalem, North Carolina).

Taken together, the clinical presentation and histopathologic findings were determined to be most consistent with RGD, IGD type, with secondary vasculitis due to vedolizumab. The patient was treated with triamcinolone ointment and low-dose prednisone. Vedolizumab was discontinued. The rash resolved several weeks after cessation of vedolizumab.

Comment

This case describes the development of RGD, IGD type, as an AE of vedolizumab for the treatment of IBD. Reactive granulomatous dermatitis encompasses a spectrum of cutaneous reactions that includes the diagnosis formerly distinctly identified as IGD.¹¹ This variety of RGD is characterized by histologic findings of heavy histiocytic inflammation in the reticular layer of the dermis with interstitial and perivascular neutrophils, lymphocytes, and histiocytes, as well as the absence of mucin. Interstitial granulomatous dermatitis–type reactions commonly are associated with autoimmune conditions and medications, with accumulating examples occurring in the setting of other biologic therapies, including the IL-6 receptor inhibitor tocilizumab; the programmed death receptor-1 inhibitor nivolumab; and the tumor necrosis factor α inhibitors infliximab, etanercept, and adalimumab.^12-15

Although our patient represents CMV infection while being treated with vedolizumab, the relationship between the two is unclear. Development of CMV infection while receiving vedolizumab has been reported in the literature in a patient who was concurrently immunosuppressed with azathioprine.¹⁶ In contrast, vedolizumab administration has been utilized as a treatment of CMV infection in IBD patients, either alone or in combination with antiviral agents, with successful resolution of infection.^17,18 Additional observations of the interaction between CMV infection and vedolizumab would be required to determine if the onset of CMV infection in this patient represents an additional risk of the medication.

Identifying a relationship between a monoclonal antibody therapy, such as vedolizumab, and RGD, IGD type, might be difficult in clinical practice, particularly if this type of reaction has not been previously associated with the culprit medication. In our patient, onset of cutaneous findings in relation to dosing of vedolizumab and exclusion of other possible causes of the rash supported the decision to stop vedolizumab. However, this decision often is challenging in patients with multiple concurrent medical conditions and those whose therapeutic options are limited.

Conclusion

Ulcerative colitis is not an uncommon condition; utilization of targeted monoclonal antibodies as a treatment strategy is expanding.^2,19 As implementation of vedolizumab as a targeted biologic therapy for this disease increases, additional cases of IGD might emerge with greater frequency. Because IBD and autoimmune conditions have a tendency to coincide, awareness of the reaction presented here might be particularly important for dermatologists managing cutaneous manifestations of autoimmune conditions, as patients might present with a clinical picture complicated by preexisting skin findings.²⁰ Furthermore, as reports of RGD, IGD type, in response to several monoclonal antibodies accumulate, it is prudent for all physicians to be aware of this potential complication of this class of medication so that they can make educated decisions about continuing monoclonal antibody therapy.

The number of monoclonal antibodies developed for therapeutic use has rapidly expanded over the last decade due to their generally favorable adverse effect (AE) profiles and efficacy.¹ Tumor necrosis factor α inhibitors and general integrin antagonists are well-known examples of such monoclonal antibodies. Common conditions utilizing immunotherapy include inflammatory bowel diseases (IBDs), such as Crohn disease and ulcerative colitis (UC).²

The monoclonal antibody vedolizumab, approved in 2014 for moderate to severe UC and Crohn disease, selectively antagonizes α4β7 integrin to target a specific population of gastrointestinal T lymphocytes, preventing their mobilization to areas of inflammation.³ Adverse effects in patients treated with vedolizumab occur at a rate comparable to placebo and largely are considered nonserious^4,5; the most commonly reported AE is disease exacerbation (13%–17% of patients).^5,6 Published reports of cutaneous AEs at administration of vedolizumab include urticaria during infusion, appearance of cutaneous manifestations characteristic of IBD, psoriasis, Henoch-Schönlein purpura, and Sweet syndrome.^7-10

We present the case of a 61-year-old woman with UC who developed reactive granulomatous dermatitis (RGD), interstitial granulomatous dermatitis (IGD) type secondary to vedolizumab. This adverse reaction has not, to our knowledge, been previously reported.

Case Report

A 61-year-old woman with a medical history of UC treated with vedolizumab and myelodysplastic syndrome treated with intravenous immunoglobulin (due to hypogammaglobulinemia following allogeneic stem cell transplantation 14 months prior) presented with a concern of a rash. The patient had been in a baseline state of health until 1 week after receiving her second dose of vedolizumab, at which time she developed a mildly pruritic maculopapular rash on the back and chest. Triamcinolone ointment and hydroxyzine were recommended during an initial telehealth consultation with an oncologist with minimal improvement. The rash continued to spread distally with worsening pruritus.

The patient returned to her oncologist for a routine follow-up appointment 5 days after initial teleconsultation. She reported poor oral intake due to oropharyngeal pain and a worsening rash; her husband added a report of recent onset of somnolence. She was admitted to the hospital, and intravenous fluids were administered.

At admission, the patient was hypotensive; vital signs were otherwise normal. Physical examination revealed the oropharynx was erythematous. Pink lichenoid papules coalescing into plaques were present diffusely across the trunk, arms, and legs; the hands, feet, and face were spared (Figure 1).

Photograph courtesy of William Beuerlein, DO (Jacksonville, Florida).

A complete blood cell count and comprehensive metabolic panel were unremarkable. A lumbar puncture, chest radiograph, blood cultures, urinalysis, and urine cultures did not identify a clear infectious cause for the rash, though the workup for infection did raise concern about active cytomegalovirus (CMV) infection with colitis and pneumonitis. Computed tomography of the head showed no acute hemorrhage.

Dermatology was consulted and determined that the appearance of the rash was most consistent with a lichenoid drug eruption, likely secondary to vedolizumab that was administered 1 week before the rash onset. Analysis of a skin biopsy revealed a dense dermal histiocytic and lymphocytic infiltrate in close approximation to blood vessels, confirmed by immunohistochemical staining for CD45, CD43, CD68, CD34, c-KIT, and myeloperoxidase (Figures 2A and 2B). Colloidal iron staining of the specimen revealed no mucin (Figure 2C).

Photographs courtesy of William Beuerlein, DO (Jacksonville, Florida) and Angela Niehaus, MD (WinstonSalem, North Carolina).

Taken together, the clinical presentation and histopathologic findings were determined to be most consistent with RGD, IGD type, with secondary vasculitis due to vedolizumab. The patient was treated with triamcinolone ointment and low-dose prednisone. Vedolizumab was discontinued. The rash resolved several weeks after cessation of vedolizumab.

Comment

This case describes the development of RGD, IGD type, as an AE of vedolizumab for the treatment of IBD. Reactive granulomatous dermatitis encompasses a spectrum of cutaneous reactions that includes the diagnosis formerly distinctly identified as IGD.¹¹ This variety of RGD is characterized by histologic findings of heavy histiocytic inflammation in the reticular layer of the dermis with interstitial and perivascular neutrophils, lymphocytes, and histiocytes, as well as the absence of mucin. Interstitial granulomatous dermatitis–type reactions commonly are associated with autoimmune conditions and medications, with accumulating examples occurring in the setting of other biologic therapies, including the IL-6 receptor inhibitor tocilizumab; the programmed death receptor-1 inhibitor nivolumab; and the tumor necrosis factor α inhibitors infliximab, etanercept, and adalimumab.^12-15

Although our patient represents CMV infection while being treated with vedolizumab, the relationship between the two is unclear. Development of CMV infection while receiving vedolizumab has been reported in the literature in a patient who was concurrently immunosuppressed with azathioprine.¹⁶ In contrast, vedolizumab administration has been utilized as a treatment of CMV infection in IBD patients, either alone or in combination with antiviral agents, with successful resolution of infection.^17,18 Additional observations of the interaction between CMV infection and vedolizumab would be required to determine if the onset of CMV infection in this patient represents an additional risk of the medication.

Identifying a relationship between a monoclonal antibody therapy, such as vedolizumab, and RGD, IGD type, might be difficult in clinical practice, particularly if this type of reaction has not been previously associated with the culprit medication. In our patient, onset of cutaneous findings in relation to dosing of vedolizumab and exclusion of other possible causes of the rash supported the decision to stop vedolizumab. However, this decision often is challenging in patients with multiple concurrent medical conditions and those whose therapeutic options are limited.

Conclusion

Ulcerative colitis is not an uncommon condition; utilization of targeted monoclonal antibodies as a treatment strategy is expanding.^2,19 As implementation of vedolizumab as a targeted biologic therapy for this disease increases, additional cases of IGD might emerge with greater frequency. Because IBD and autoimmune conditions have a tendency to coincide, awareness of the reaction presented here might be particularly important for dermatologists managing cutaneous manifestations of autoimmune conditions, as patients might present with a clinical picture complicated by preexisting skin findings.²⁰ Furthermore, as reports of RGD, IGD type, in response to several monoclonal antibodies accumulate, it is prudent for all physicians to be aware of this potential complication of this class of medication so that they can make educated decisions about continuing monoclonal antibody therapy.

The number of monoclonal antibodies developed for therapeutic use has rapidly expanded over the last decade due to their generally favorable adverse effect (AE) profiles and efficacy.¹ Tumor necrosis factor α inhibitors and general integrin antagonists are well-known examples of such monoclonal antibodies. Common conditions utilizing immunotherapy include inflammatory bowel diseases (IBDs), such as Crohn disease and ulcerative colitis (UC).²

The monoclonal antibody vedolizumab, approved in 2014 for moderate to severe UC and Crohn disease, selectively antagonizes α4β7 integrin to target a specific population of gastrointestinal T lymphocytes, preventing their mobilization to areas of inflammation.³ Adverse effects in patients treated with vedolizumab occur at a rate comparable to placebo and largely are considered nonserious^4,5; the most commonly reported AE is disease exacerbation (13%–17% of patients).^5,6 Published reports of cutaneous AEs at administration of vedolizumab include urticaria during infusion, appearance of cutaneous manifestations characteristic of IBD, psoriasis, Henoch-Schönlein purpura, and Sweet syndrome.^7-10

We present the case of a 61-year-old woman with UC who developed reactive granulomatous dermatitis (RGD), interstitial granulomatous dermatitis (IGD) type secondary to vedolizumab. This adverse reaction has not, to our knowledge, been previously reported.

Case Report

A 61-year-old woman with a medical history of UC treated with vedolizumab and myelodysplastic syndrome treated with intravenous immunoglobulin (due to hypogammaglobulinemia following allogeneic stem cell transplantation 14 months prior) presented with a concern of a rash. The patient had been in a baseline state of health until 1 week after receiving her second dose of vedolizumab, at which time she developed a mildly pruritic maculopapular rash on the back and chest. Triamcinolone ointment and hydroxyzine were recommended during an initial telehealth consultation with an oncologist with minimal improvement. The rash continued to spread distally with worsening pruritus.

The patient returned to her oncologist for a routine follow-up appointment 5 days after initial teleconsultation. She reported poor oral intake due to oropharyngeal pain and a worsening rash; her husband added a report of recent onset of somnolence. She was admitted to the hospital, and intravenous fluids were administered.

At admission, the patient was hypotensive; vital signs were otherwise normal. Physical examination revealed the oropharynx was erythematous. Pink lichenoid papules coalescing into plaques were present diffusely across the trunk, arms, and legs; the hands, feet, and face were spared (Figure 1).

Photograph courtesy of William Beuerlein, DO (Jacksonville, Florida).

A complete blood cell count and comprehensive metabolic panel were unremarkable. A lumbar puncture, chest radiograph, blood cultures, urinalysis, and urine cultures did not identify a clear infectious cause for the rash, though the workup for infection did raise concern about active cytomegalovirus (CMV) infection with colitis and pneumonitis. Computed tomography of the head showed no acute hemorrhage.

Dermatology was consulted and determined that the appearance of the rash was most consistent with a lichenoid drug eruption, likely secondary to vedolizumab that was administered 1 week before the rash onset. Analysis of a skin biopsy revealed a dense dermal histiocytic and lymphocytic infiltrate in close approximation to blood vessels, confirmed by immunohistochemical staining for CD45, CD43, CD68, CD34, c-KIT, and myeloperoxidase (Figures 2A and 2B). Colloidal iron staining of the specimen revealed no mucin (Figure 2C).

Photographs courtesy of William Beuerlein, DO (Jacksonville, Florida) and Angela Niehaus, MD (WinstonSalem, North Carolina).

Taken together, the clinical presentation and histopathologic findings were determined to be most consistent with RGD, IGD type, with secondary vasculitis due to vedolizumab. The patient was treated with triamcinolone ointment and low-dose prednisone. Vedolizumab was discontinued. The rash resolved several weeks after cessation of vedolizumab.

Comment

This case describes the development of RGD, IGD type, as an AE of vedolizumab for the treatment of IBD. Reactive granulomatous dermatitis encompasses a spectrum of cutaneous reactions that includes the diagnosis formerly distinctly identified as IGD.¹¹ This variety of RGD is characterized by histologic findings of heavy histiocytic inflammation in the reticular layer of the dermis with interstitial and perivascular neutrophils, lymphocytes, and histiocytes, as well as the absence of mucin. Interstitial granulomatous dermatitis–type reactions commonly are associated with autoimmune conditions and medications, with accumulating examples occurring in the setting of other biologic therapies, including the IL-6 receptor inhibitor tocilizumab; the programmed death receptor-1 inhibitor nivolumab; and the tumor necrosis factor α inhibitors infliximab, etanercept, and adalimumab.^12-15

Although our patient represents CMV infection while being treated with vedolizumab, the relationship between the two is unclear. Development of CMV infection while receiving vedolizumab has been reported in the literature in a patient who was concurrently immunosuppressed with azathioprine.¹⁶ In contrast, vedolizumab administration has been utilized as a treatment of CMV infection in IBD patients, either alone or in combination with antiviral agents, with successful resolution of infection.^17,18 Additional observations of the interaction between CMV infection and vedolizumab would be required to determine if the onset of CMV infection in this patient represents an additional risk of the medication.

Identifying a relationship between a monoclonal antibody therapy, such as vedolizumab, and RGD, IGD type, might be difficult in clinical practice, particularly if this type of reaction has not been previously associated with the culprit medication. In our patient, onset of cutaneous findings in relation to dosing of vedolizumab and exclusion of other possible causes of the rash supported the decision to stop vedolizumab. However, this decision often is challenging in patients with multiple concurrent medical conditions and those whose therapeutic options are limited.

Conclusion

Ulcerative colitis is not an uncommon condition; utilization of targeted monoclonal antibodies as a treatment strategy is expanding.^2,19 As implementation of vedolizumab as a targeted biologic therapy for this disease increases, additional cases of IGD might emerge with greater frequency. Because IBD and autoimmune conditions have a tendency to coincide, awareness of the reaction presented here might be particularly important for dermatologists managing cutaneous manifestations of autoimmune conditions, as patients might present with a clinical picture complicated by preexisting skin findings.²⁰ Furthermore, as reports of RGD, IGD type, in response to several monoclonal antibodies accumulate, it is prudent for all physicians to be aware of this potential complication of this class of medication so that they can make educated decisions about continuing monoclonal antibody therapy.