A new Senate bill would require Medicare to test two tools routinely used by credit card companies to prevent fraud: Artificial intelligence (AI)-trained algorithms to detect suspicious activity and a system to quickly alert Medicare patients on whose behalf payment is being sought.

Senator Mike Braun (R-IN) recently introduced the Medicare Transaction Fraud Prevention Act, which calls for a 2-year test of this approach.

The experiment, targeted to start in 2025, would focus on durable medical equipment and clinical diagnostic laboratory tests and cover Medicare beneficiaries who receive electronic notices about claims.

The legislation would direct the Center for Medicare and Medicaid Services (CMS) to test the use of predictive risk-scoring algorithms in finding fraud. The program would be modeled on the systems that credit card companies already use. Transactions could be scored from 1 (least risky) to 99 (most risky).

CMS would then check directly by email or phone call with selected Medicare enrollees about transactions considered to present a high risk for fraud.

Many consumers have benefited from this approach when used to check for fraud on their credit cards, Braun noted during a November hearing of the Senate Special Committee on Aging. Credit card companies often can intervene before a fraudulent transaction is cleared.

“There’s no reason we wouldn’t want to minimally at least mimic that,” Braun said at the hearing.

Asking Medicare enrollees to verify certain purchases could give CMS increased access to vital predictive data, test proof of concept, and save hundreds of millions of dollars, Braun said.

Concerns Raised

So far, Braun has only one cosponsor for the bill, Senator Bill Cassidy, MD (R-LA), and the bill has drawn some criticism.

Brett Meeks, executive director of the Health Innovation Alliance, a trade group representing technology companies, insurers, and consumer organizations, objected to requiring Medicare enrollees to verify flagged orders. CMS should internally root out fraud through technology, not burden seniors, Meeks told this news organization.

Meeks said he has been following the discussion about the use of AI in addressing Medicare fraud. Had a bill broadly targeted Medicare fraud through AI, his alliance might have backed it, he said. But the current proposed legislation has a narrower focus.

Focusing on durable medical equipment, for example, could have unintended consequences like denying power wheelchairs to people with debilitating conditions like multiple sclerosis, Meeks said.

But Braun’s bill won a quick nod of approval from a researcher who studies the use of AI to detect Medicare fraud. Taghi M. Khoshgoftaar, PhD, director of the Data Mining and Machine Learning Lab at Florida Atlantic University, Boca Raton, Florida, said he sees an advantage to Braun’s approach of involving Medicare enrollees in the protection of their benefits.

The bill does not authorize funding for the pilot project, and it’s unclear what it would cost.

Detecting Medicare Fraud

The federal government has stepped up Medicare fraud investigations in recent years, and more doctors are getting caught.

A study published in 2018 examined cases of physicians excluded from Medicare using data from the US Office of Inspector General (OIG) at the Department of Health and Human Services.

The OIG has the right to exclude clinicians from Medicare for fraud or other reasons. Chen and coauthors looked at Medicare physician exclusions from 2007 to 2017. They found that exclusions due to fraud increased an estimated 14% per year on average from a base level of 139 exclusions in 2007.

In 2019, CMS sought feedback on new ways to use AI to detect fraud. In a public request for information, the agency said Medicare scrutinizes fewer claims for payment than commercial insurers do.

About 99.7% of Medicare fee-for-service claims are processed and paid within 17 days without any medical review, CMS said at the time.

A version of this article appeared on Medscape.com .

A new Senate bill would require Medicare to test two tools routinely used by credit card companies to prevent fraud: Artificial intelligence (AI)-trained algorithms to detect suspicious activity and a system to quickly alert Medicare patients on whose behalf payment is being sought.

Senator Mike Braun (R-IN) recently introduced the Medicare Transaction Fraud Prevention Act, which calls for a 2-year test of this approach.

The experiment, targeted to start in 2025, would focus on durable medical equipment and clinical diagnostic laboratory tests and cover Medicare beneficiaries who receive electronic notices about claims.

The legislation would direct the Center for Medicare and Medicaid Services (CMS) to test the use of predictive risk-scoring algorithms in finding fraud. The program would be modeled on the systems that credit card companies already use. Transactions could be scored from 1 (least risky) to 99 (most risky).

CMS would then check directly by email or phone call with selected Medicare enrollees about transactions considered to present a high risk for fraud.

Many consumers have benefited from this approach when used to check for fraud on their credit cards, Braun noted during a November hearing of the Senate Special Committee on Aging. Credit card companies often can intervene before a fraudulent transaction is cleared.

“There’s no reason we wouldn’t want to minimally at least mimic that,” Braun said at the hearing.

Asking Medicare enrollees to verify certain purchases could give CMS increased access to vital predictive data, test proof of concept, and save hundreds of millions of dollars, Braun said.

Concerns Raised

So far, Braun has only one cosponsor for the bill, Senator Bill Cassidy, MD (R-LA), and the bill has drawn some criticism.

Brett Meeks, executive director of the Health Innovation Alliance, a trade group representing technology companies, insurers, and consumer organizations, objected to requiring Medicare enrollees to verify flagged orders. CMS should internally root out fraud through technology, not burden seniors, Meeks told this news organization.

Meeks said he has been following the discussion about the use of AI in addressing Medicare fraud. Had a bill broadly targeted Medicare fraud through AI, his alliance might have backed it, he said. But the current proposed legislation has a narrower focus.

Focusing on durable medical equipment, for example, could have unintended consequences like denying power wheelchairs to people with debilitating conditions like multiple sclerosis, Meeks said.

But Braun’s bill won a quick nod of approval from a researcher who studies the use of AI to detect Medicare fraud. Taghi M. Khoshgoftaar, PhD, director of the Data Mining and Machine Learning Lab at Florida Atlantic University, Boca Raton, Florida, said he sees an advantage to Braun’s approach of involving Medicare enrollees in the protection of their benefits.

The bill does not authorize funding for the pilot project, and it’s unclear what it would cost.

Detecting Medicare Fraud

The federal government has stepped up Medicare fraud investigations in recent years, and more doctors are getting caught.

A study published in 2018 examined cases of physicians excluded from Medicare using data from the US Office of Inspector General (OIG) at the Department of Health and Human Services.

The OIG has the right to exclude clinicians from Medicare for fraud or other reasons. Chen and coauthors looked at Medicare physician exclusions from 2007 to 2017. They found that exclusions due to fraud increased an estimated 14% per year on average from a base level of 139 exclusions in 2007.

In 2019, CMS sought feedback on new ways to use AI to detect fraud. In a public request for information, the agency said Medicare scrutinizes fewer claims for payment than commercial insurers do.

About 99.7% of Medicare fee-for-service claims are processed and paid within 17 days without any medical review, CMS said at the time.

A version of this article appeared on Medscape.com .

A new Senate bill would require Medicare to test two tools routinely used by credit card companies to prevent fraud: Artificial intelligence (AI)-trained algorithms to detect suspicious activity and a system to quickly alert Medicare patients on whose behalf payment is being sought.

Senator Mike Braun (R-IN) recently introduced the Medicare Transaction Fraud Prevention Act, which calls for a 2-year test of this approach.

The experiment, targeted to start in 2025, would focus on durable medical equipment and clinical diagnostic laboratory tests and cover Medicare beneficiaries who receive electronic notices about claims.

The legislation would direct the Center for Medicare and Medicaid Services (CMS) to test the use of predictive risk-scoring algorithms in finding fraud. The program would be modeled on the systems that credit card companies already use. Transactions could be scored from 1 (least risky) to 99 (most risky).

CMS would then check directly by email or phone call with selected Medicare enrollees about transactions considered to present a high risk for fraud.

Many consumers have benefited from this approach when used to check for fraud on their credit cards, Braun noted during a November hearing of the Senate Special Committee on Aging. Credit card companies often can intervene before a fraudulent transaction is cleared.

“There’s no reason we wouldn’t want to minimally at least mimic that,” Braun said at the hearing.

Asking Medicare enrollees to verify certain purchases could give CMS increased access to vital predictive data, test proof of concept, and save hundreds of millions of dollars, Braun said.

Concerns Raised

So far, Braun has only one cosponsor for the bill, Senator Bill Cassidy, MD (R-LA), and the bill has drawn some criticism.

Brett Meeks, executive director of the Health Innovation Alliance, a trade group representing technology companies, insurers, and consumer organizations, objected to requiring Medicare enrollees to verify flagged orders. CMS should internally root out fraud through technology, not burden seniors, Meeks told this news organization.

Meeks said he has been following the discussion about the use of AI in addressing Medicare fraud. Had a bill broadly targeted Medicare fraud through AI, his alliance might have backed it, he said. But the current proposed legislation has a narrower focus.

Focusing on durable medical equipment, for example, could have unintended consequences like denying power wheelchairs to people with debilitating conditions like multiple sclerosis, Meeks said.

But Braun’s bill won a quick nod of approval from a researcher who studies the use of AI to detect Medicare fraud. Taghi M. Khoshgoftaar, PhD, director of the Data Mining and Machine Learning Lab at Florida Atlantic University, Boca Raton, Florida, said he sees an advantage to Braun’s approach of involving Medicare enrollees in the protection of their benefits.

The bill does not authorize funding for the pilot project, and it’s unclear what it would cost.

Detecting Medicare Fraud

The federal government has stepped up Medicare fraud investigations in recent years, and more doctors are getting caught.

A study published in 2018 examined cases of physicians excluded from Medicare using data from the US Office of Inspector General (OIG) at the Department of Health and Human Services.

The OIG has the right to exclude clinicians from Medicare for fraud or other reasons. Chen and coauthors looked at Medicare physician exclusions from 2007 to 2017. They found that exclusions due to fraud increased an estimated 14% per year on average from a base level of 139 exclusions in 2007.

In 2019, CMS sought feedback on new ways to use AI to detect fraud. In a public request for information, the agency said Medicare scrutinizes fewer claims for payment than commercial insurers do.

About 99.7% of Medicare fee-for-service claims are processed and paid within 17 days without any medical review, CMS said at the time.

A version of this article appeared on Medscape.com .

To the Editor:

Angioimmunoblastic T-cell lymphoma (AITL) is an uncommon peripheral T-cell lymphoma that accounts for 1% to 2% of all forms of non-Hodgkin lymphoma and usually affects middle-aged individuals.¹ It primarily appears on the skin and mimics an inflammatory dermatosis, leading to diagnostic and therapeutic delays.² No gold-standard treatment has been identified for AITL; the prognosis often remains poor, with a 5-year progression-free survival rate of approximately 25%.³ Because of the rarity of AITL and the unmet need of a standard-of-care treatment regimen, relapsing and remitting disease is common and continues to challenge clinicians.

Methotrexate (MTX), a dihydrofolate reductase inhibitor used to treat many autoimmune diseases, is prescribed at a higher dosage (>500 mg/m²) to manage cancers, including refractory AITL.⁴ In blocking dihydrofolate reductase, MTX reduces the folate pool, with the possible adverse effect of bone marrow suppression. Another important toxic effect is acute kidney injury, which may be due to an overdose of MTX or a patient’s predisposition to chronic kidney failure.⁴

A 50-year-old man was admitted to our inpatient clinic for evaluation of acute oral and genital mucositis. He had a 5-year history of AITL. He was previously treated by hematology with 3 lines of chemotherapy for multiple supradiaphragmatic and subdiaphragmatic localizations of lymphoma, without success. Six days prior to the current presentation, the hematologist started high-dose (3.5 g/m²) intravenous MTX therapy. Five days later, the patient developed transfusion-resistant pancytopenia and fever (maximum body temperature, 102.7°F [39.3°C]).

Physical examination at the current presentation revealed massive necrosis of the lower lip (Figure, A) and partial necrosis of the upper lip. Severe purulent balanoposthitis, causing penile edema and phimosis, complicated the clinical condition. Analysis of a specimen from a cutaneous swab of the penis showed infection with Pseudomonas aeruginosa and Enterococcus faecalis. Considering the clinical presentation and time of onset of signs and symptoms, a diagnosis of acute MTX-induced mucositis was made.

Rescue therapy was started immediately, including high-dose intravenous leucovorin (120 mg 4 times daily), oral sulfamethoxazole-trimethoprim (800 mg/160 mg 3 times daily for 3 days per week), and oral levofloxacin (500 mg/d). After 4 days of treatment, the patient was afebrile. Mucositis of the lips had almost resolved (Figure, B), and balanoposthitis also improved after this rescue therapy. Methotrexate was not resumed because rituximab had been started.

To the Editor:

Angioimmunoblastic T-cell lymphoma (AITL) is an uncommon peripheral T-cell lymphoma that accounts for 1% to 2% of all forms of non-Hodgkin lymphoma and usually affects middle-aged individuals.¹ It primarily appears on the skin and mimics an inflammatory dermatosis, leading to diagnostic and therapeutic delays.² No gold-standard treatment has been identified for AITL; the prognosis often remains poor, with a 5-year progression-free survival rate of approximately 25%.³ Because of the rarity of AITL and the unmet need of a standard-of-care treatment regimen, relapsing and remitting disease is common and continues to challenge clinicians.

Methotrexate (MTX), a dihydrofolate reductase inhibitor used to treat many autoimmune diseases, is prescribed at a higher dosage (>500 mg/m²) to manage cancers, including refractory AITL.⁴ In blocking dihydrofolate reductase, MTX reduces the folate pool, with the possible adverse effect of bone marrow suppression. Another important toxic effect is acute kidney injury, which may be due to an overdose of MTX or a patient’s predisposition to chronic kidney failure.⁴

A 50-year-old man was admitted to our inpatient clinic for evaluation of acute oral and genital mucositis. He had a 5-year history of AITL. He was previously treated by hematology with 3 lines of chemotherapy for multiple supradiaphragmatic and subdiaphragmatic localizations of lymphoma, without success. Six days prior to the current presentation, the hematologist started high-dose (3.5 g/m²) intravenous MTX therapy. Five days later, the patient developed transfusion-resistant pancytopenia and fever (maximum body temperature, 102.7°F [39.3°C]).

Physical examination at the current presentation revealed massive necrosis of the lower lip (Figure, A) and partial necrosis of the upper lip. Severe purulent balanoposthitis, causing penile edema and phimosis, complicated the clinical condition. Analysis of a specimen from a cutaneous swab of the penis showed infection with Pseudomonas aeruginosa and Enterococcus faecalis. Considering the clinical presentation and time of onset of signs and symptoms, a diagnosis of acute MTX-induced mucositis was made.

Rescue therapy was started immediately, including high-dose intravenous leucovorin (120 mg 4 times daily), oral sulfamethoxazole-trimethoprim (800 mg/160 mg 3 times daily for 3 days per week), and oral levofloxacin (500 mg/d). After 4 days of treatment, the patient was afebrile. Mucositis of the lips had almost resolved (Figure, B), and balanoposthitis also improved after this rescue therapy. Methotrexate was not resumed because rituximab had been started.

To the Editor:

Angioimmunoblastic T-cell lymphoma (AITL) is an uncommon peripheral T-cell lymphoma that accounts for 1% to 2% of all forms of non-Hodgkin lymphoma and usually affects middle-aged individuals.¹ It primarily appears on the skin and mimics an inflammatory dermatosis, leading to diagnostic and therapeutic delays.² No gold-standard treatment has been identified for AITL; the prognosis often remains poor, with a 5-year progression-free survival rate of approximately 25%.³ Because of the rarity of AITL and the unmet need of a standard-of-care treatment regimen, relapsing and remitting disease is common and continues to challenge clinicians.

Methotrexate (MTX), a dihydrofolate reductase inhibitor used to treat many autoimmune diseases, is prescribed at a higher dosage (>500 mg/m²) to manage cancers, including refractory AITL.⁴ In blocking dihydrofolate reductase, MTX reduces the folate pool, with the possible adverse effect of bone marrow suppression. Another important toxic effect is acute kidney injury, which may be due to an overdose of MTX or a patient’s predisposition to chronic kidney failure.⁴

A 50-year-old man was admitted to our inpatient clinic for evaluation of acute oral and genital mucositis. He had a 5-year history of AITL. He was previously treated by hematology with 3 lines of chemotherapy for multiple supradiaphragmatic and subdiaphragmatic localizations of lymphoma, without success. Six days prior to the current presentation, the hematologist started high-dose (3.5 g/m²) intravenous MTX therapy. Five days later, the patient developed transfusion-resistant pancytopenia and fever (maximum body temperature, 102.7°F [39.3°C]).

Physical examination at the current presentation revealed massive necrosis of the lower lip (Figure, A) and partial necrosis of the upper lip. Severe purulent balanoposthitis, causing penile edema and phimosis, complicated the clinical condition. Analysis of a specimen from a cutaneous swab of the penis showed infection with Pseudomonas aeruginosa and Enterococcus faecalis. Considering the clinical presentation and time of onset of signs and symptoms, a diagnosis of acute MTX-induced mucositis was made.

Rescue therapy was started immediately, including high-dose intravenous leucovorin (120 mg 4 times daily), oral sulfamethoxazole-trimethoprim (800 mg/160 mg 3 times daily for 3 days per week), and oral levofloxacin (500 mg/d). After 4 days of treatment, the patient was afebrile. Mucositis of the lips had almost resolved (Figure, B), and balanoposthitis also improved after this rescue therapy. Methotrexate was not resumed because rituximab had been started.

In patients with stimulant use disorder (SUD), even slight reductions in drug use can lessen depression and reduce cravings, a new analysis showed.

Abstinence has long been the overall goal of SUD treatment, the investigators noted. The findings from this pooled analysis of randomized clinical trials support what investigators noted was a growing recognition that reducing stimulant use can lead to better outcomes.

“This study provides evidence that reducing the overall use of drugs is important and clinically meaningful,” study author Mehdi Farokhnia, MD, MPH, of the National Institute on Drug Abuse, North Bethesda, Maryland, wrote in a press release. “This shift may open opportunities for medication development that can help individuals achieve these improved outcomes, even if complete abstinence is not immediately achievable or wanted.”

The findings were published online on January 10, 2024, in Addiction.
 

Not the Only Indicator of Success

To compare clinical indicators of improvement among those with SUDs who achieved abstinence or reduced their use, investigators pooled data from 13 randomized clinical trials with more than 2000 patients seeking treatment for cocaine or methamphetamine use disorders at centers in the United States from 2001 to 2017.

The trials used similar study protocols, including similar eligibility criteria, recruitment methods, and outcome measures. Participants were 18 or older and met criteria for methamphetamine or cocaine dependence at the beginning of each trial.

Among the participants, 1196 sought treatment for cocaine use disorder and 866 for methamphetamine use disorder. Of those, just 1487 had outcomes available by the end of the trial.

Most participants had no change in the level of use or increased their use through the trial (68%) or transitioned from low (1-4 days a month) to high (5 or more days a month) frequency use.

Nearly one third of participants (32%) stopped or reduced drug use, including 18% who cut down on stimulant use and 14% who abstained altogether.

Participants using methamphetamine were more likely to be in the abstinence vs reduced use category (21.3% vs 13.9%, respectively), whereas participants using cocaine were less likely to be in the abstinence vs reduced use category (9.1% vs 20.9%).

Those who reached abstinence showed better clinical improvement than those who reduced use on most clinical measures (P < .009).

However, there were no significant differences between groups on the Addiction Severity Index (ASI) psychiatric problems subscale and cravings for secondary drugs.

“Our findings suggest that reduced frequency of stimulant use is also associated with improved psychosocial functioning,” the authors wrote. “These findings suggest the need to re-evaluate the traditional approach of exclusively relying on total abstinence as the only indicator of successful treatment, a goal that may not be achievable for all patients, especially after one treatment episode.”

Those who reduced drug intake showed a significant association with nearly all clinical indicators of improvement (P < .010) compared with those who didn’t, except for the ASI psychiatric problems subscale and family/social relationship domains of the Problem Free Functioning scale, and HIV risk behavior.

Study limitations included short follow-up time in most trials and follow-up measures were based only on urine drug screens. There were also a substantial number of missing assessment points.

The study was funded by the National Institute on Drug Abuse and the National Institute of Health. There were no reported disclosures.
 

A version of this article appeared on Medscape.com.

In patients with stimulant use disorder (SUD), even slight reductions in drug use can lessen depression and reduce cravings, a new analysis showed.

Abstinence has long been the overall goal of SUD treatment, the investigators noted. The findings from this pooled analysis of randomized clinical trials support what investigators noted was a growing recognition that reducing stimulant use can lead to better outcomes.

“This study provides evidence that reducing the overall use of drugs is important and clinically meaningful,” study author Mehdi Farokhnia, MD, MPH, of the National Institute on Drug Abuse, North Bethesda, Maryland, wrote in a press release. “This shift may open opportunities for medication development that can help individuals achieve these improved outcomes, even if complete abstinence is not immediately achievable or wanted.”

The findings were published online on January 10, 2024, in Addiction.
 

Not the Only Indicator of Success

To compare clinical indicators of improvement among those with SUDs who achieved abstinence or reduced their use, investigators pooled data from 13 randomized clinical trials with more than 2000 patients seeking treatment for cocaine or methamphetamine use disorders at centers in the United States from 2001 to 2017.

The trials used similar study protocols, including similar eligibility criteria, recruitment methods, and outcome measures. Participants were 18 or older and met criteria for methamphetamine or cocaine dependence at the beginning of each trial.

Among the participants, 1196 sought treatment for cocaine use disorder and 866 for methamphetamine use disorder. Of those, just 1487 had outcomes available by the end of the trial.

Most participants had no change in the level of use or increased their use through the trial (68%) or transitioned from low (1-4 days a month) to high (5 or more days a month) frequency use.

Nearly one third of participants (32%) stopped or reduced drug use, including 18% who cut down on stimulant use and 14% who abstained altogether.

Participants using methamphetamine were more likely to be in the abstinence vs reduced use category (21.3% vs 13.9%, respectively), whereas participants using cocaine were less likely to be in the abstinence vs reduced use category (9.1% vs 20.9%).

Those who reached abstinence showed better clinical improvement than those who reduced use on most clinical measures (P < .009).

However, there were no significant differences between groups on the Addiction Severity Index (ASI) psychiatric problems subscale and cravings for secondary drugs.

“Our findings suggest that reduced frequency of stimulant use is also associated with improved psychosocial functioning,” the authors wrote. “These findings suggest the need to re-evaluate the traditional approach of exclusively relying on total abstinence as the only indicator of successful treatment, a goal that may not be achievable for all patients, especially after one treatment episode.”

Those who reduced drug intake showed a significant association with nearly all clinical indicators of improvement (P < .010) compared with those who didn’t, except for the ASI psychiatric problems subscale and family/social relationship domains of the Problem Free Functioning scale, and HIV risk behavior.

Study limitations included short follow-up time in most trials and follow-up measures were based only on urine drug screens. There were also a substantial number of missing assessment points.

The study was funded by the National Institute on Drug Abuse and the National Institute of Health. There were no reported disclosures.
 

A version of this article appeared on Medscape.com.

In patients with stimulant use disorder (SUD), even slight reductions in drug use can lessen depression and reduce cravings, a new analysis showed.

Abstinence has long been the overall goal of SUD treatment, the investigators noted. The findings from this pooled analysis of randomized clinical trials support what investigators noted was a growing recognition that reducing stimulant use can lead to better outcomes.

“This study provides evidence that reducing the overall use of drugs is important and clinically meaningful,” study author Mehdi Farokhnia, MD, MPH, of the National Institute on Drug Abuse, North Bethesda, Maryland, wrote in a press release. “This shift may open opportunities for medication development that can help individuals achieve these improved outcomes, even if complete abstinence is not immediately achievable or wanted.”

The findings were published online on January 10, 2024, in Addiction.
 

Not the Only Indicator of Success

To compare clinical indicators of improvement among those with SUDs who achieved abstinence or reduced their use, investigators pooled data from 13 randomized clinical trials with more than 2000 patients seeking treatment for cocaine or methamphetamine use disorders at centers in the United States from 2001 to 2017.

The trials used similar study protocols, including similar eligibility criteria, recruitment methods, and outcome measures. Participants were 18 or older and met criteria for methamphetamine or cocaine dependence at the beginning of each trial.

Among the participants, 1196 sought treatment for cocaine use disorder and 866 for methamphetamine use disorder. Of those, just 1487 had outcomes available by the end of the trial.

Most participants had no change in the level of use or increased their use through the trial (68%) or transitioned from low (1-4 days a month) to high (5 or more days a month) frequency use.

Nearly one third of participants (32%) stopped or reduced drug use, including 18% who cut down on stimulant use and 14% who abstained altogether.

Participants using methamphetamine were more likely to be in the abstinence vs reduced use category (21.3% vs 13.9%, respectively), whereas participants using cocaine were less likely to be in the abstinence vs reduced use category (9.1% vs 20.9%).

Those who reached abstinence showed better clinical improvement than those who reduced use on most clinical measures (P < .009).

However, there were no significant differences between groups on the Addiction Severity Index (ASI) psychiatric problems subscale and cravings for secondary drugs.

“Our findings suggest that reduced frequency of stimulant use is also associated with improved psychosocial functioning,” the authors wrote. “These findings suggest the need to re-evaluate the traditional approach of exclusively relying on total abstinence as the only indicator of successful treatment, a goal that may not be achievable for all patients, especially after one treatment episode.”

Those who reduced drug intake showed a significant association with nearly all clinical indicators of improvement (P < .010) compared with those who didn’t, except for the ASI psychiatric problems subscale and family/social relationship domains of the Problem Free Functioning scale, and HIV risk behavior.

Study limitations included short follow-up time in most trials and follow-up measures were based only on urine drug screens. There were also a substantial number of missing assessment points.

The study was funded by the National Institute on Drug Abuse and the National Institute of Health. There were no reported disclosures.
 

A version of this article appeared on Medscape.com.

TOPLINE:

Social frailty, the lack of resources to meet basic social needs, is associated with an increased risk for motoric cognitive risk syndrome (MCR), a predementia syndrome characterized by cognitive complaints and slow gait, results of a large, population-based study suggested.

METHODOLOGY:

• The study used 2011 (Round 1) to 2018 (Round 8) data on a discovery sample of 4657 individuals without MCR or dementia at baseline from the National Health and Aging Trends Study (NHATS), a longitudinal survey of older adult Medicare beneficiaries.
• Researchers also collected data on 3075 newly recruited individuals in Round 5 and followed to Round 8 as an independent validation sample to create a pooled sample of 7732 older adults, mean age 76.06, without MCR at baseline.
• Social frailty, assessed at baseline, included five social items: Going out less, not feeling confident, rarely visiting friends/family, not talking with others, and without live-in partner/spouse (researchers divided participants into normal [zero to one items] and social frailty [two to five items] groups).
• Individuals were considered to have MCR if they had both subjective cognitive complaints and slow gait speed (greater than 1 standard deviation below age-specific level) without dementia or mobility disability.
• Covariates included demographic and lifestyle data, presence of depression and/or anxiety symptoms, and number of chronic diseases.

TAKEAWAY:

• During a median follow-up period of 4 years, 10.35% individuals were diagnosed with MCR.
• After the researchers controlled for confounding factors, those with social frailty had an increased risk for MCR compared with the normal group (pooled sample: hazard ratio [HR], 1.57; 95% CI, 1.34-1.84; P < .001).
• Each additional unfavorable social item was associated with an increased risk for MCR (pooled sample: HR, 1.32; 95% CI, 1.22-1.43; P < .001).
• Results of stratified analyses across subgroups suggested individuals with social frailty had a significantly higher risk for incident MCR than that of those without social frailty, regardless of socioeconomic status, lifestyle factors, chronic diseases, and mental health.

IN PRACTICE:

The findings suggest assessing social frailty using simple questions “is an efficient tool for detecting older individuals with a high risk of MCR,” the authors wrote. They noted that the addition of such a tool in clinical practice may facilitate “timely implementation of prevention strategies.”

SOURCE:

The research was led by Hui Zhang, Human Phenome Institute, Zhangjiang Fudan International Innovation Centre, Fudan University, Shanghai, China. It was published online on January 29, 2024, in Alzheimer’s & Dementia.

LIMITATIONS:

The study was observational, so the association between social frailty and MCR is merely correlational. Due to the lack of genetic information in NHATS data, researchers didn’t evaluate the effect of genetic factors such as apolipoprotein E on the association between social frailty and MCR. Social frailty was assessed at a single time point. In addition, the researchers were unable examine the time sequence between social frailty and MCR and so could not determine the cause of this association.

DISCLOSURES:

The study was supported by the National Natural Science Foundation of China-Youth Science Fund, Shanghai Rising-Star Program, Shanghai Municipal Health Commission and Key Discipline Construction Project of Pudong Health, and Family Planning Commission of Shanghai. The authors reported no relevant conflicts of interest.
 

A version of this article appeared on Medscape.com.

TOPLINE:

Social frailty, the lack of resources to meet basic social needs, is associated with an increased risk for motoric cognitive risk syndrome (MCR), a predementia syndrome characterized by cognitive complaints and slow gait, results of a large, population-based study suggested.

METHODOLOGY:

• The study used 2011 (Round 1) to 2018 (Round 8) data on a discovery sample of 4657 individuals without MCR or dementia at baseline from the National Health and Aging Trends Study (NHATS), a longitudinal survey of older adult Medicare beneficiaries.
• Researchers also collected data on 3075 newly recruited individuals in Round 5 and followed to Round 8 as an independent validation sample to create a pooled sample of 7732 older adults, mean age 76.06, without MCR at baseline.
• Social frailty, assessed at baseline, included five social items: Going out less, not feeling confident, rarely visiting friends/family, not talking with others, and without live-in partner/spouse (researchers divided participants into normal [zero to one items] and social frailty [two to five items] groups).
• Individuals were considered to have MCR if they had both subjective cognitive complaints and slow gait speed (greater than 1 standard deviation below age-specific level) without dementia or mobility disability.
• Covariates included demographic and lifestyle data, presence of depression and/or anxiety symptoms, and number of chronic diseases.

TAKEAWAY:

• During a median follow-up period of 4 years, 10.35% individuals were diagnosed with MCR.
• After the researchers controlled for confounding factors, those with social frailty had an increased risk for MCR compared with the normal group (pooled sample: hazard ratio [HR], 1.57; 95% CI, 1.34-1.84; P < .001).
• Each additional unfavorable social item was associated with an increased risk for MCR (pooled sample: HR, 1.32; 95% CI, 1.22-1.43; P < .001).
• Results of stratified analyses across subgroups suggested individuals with social frailty had a significantly higher risk for incident MCR than that of those without social frailty, regardless of socioeconomic status, lifestyle factors, chronic diseases, and mental health.

IN PRACTICE:

The findings suggest assessing social frailty using simple questions “is an efficient tool for detecting older individuals with a high risk of MCR,” the authors wrote. They noted that the addition of such a tool in clinical practice may facilitate “timely implementation of prevention strategies.”

SOURCE:

The research was led by Hui Zhang, Human Phenome Institute, Zhangjiang Fudan International Innovation Centre, Fudan University, Shanghai, China. It was published online on January 29, 2024, in Alzheimer’s & Dementia.

LIMITATIONS:

The study was observational, so the association between social frailty and MCR is merely correlational. Due to the lack of genetic information in NHATS data, researchers didn’t evaluate the effect of genetic factors such as apolipoprotein E on the association between social frailty and MCR. Social frailty was assessed at a single time point. In addition, the researchers were unable examine the time sequence between social frailty and MCR and so could not determine the cause of this association.

DISCLOSURES:

The study was supported by the National Natural Science Foundation of China-Youth Science Fund, Shanghai Rising-Star Program, Shanghai Municipal Health Commission and Key Discipline Construction Project of Pudong Health, and Family Planning Commission of Shanghai. The authors reported no relevant conflicts of interest.
 

A version of this article appeared on Medscape.com.

TOPLINE:

Social frailty, the lack of resources to meet basic social needs, is associated with an increased risk for motoric cognitive risk syndrome (MCR), a predementia syndrome characterized by cognitive complaints and slow gait, results of a large, population-based study suggested.

METHODOLOGY:

• The study used 2011 (Round 1) to 2018 (Round 8) data on a discovery sample of 4657 individuals without MCR or dementia at baseline from the National Health and Aging Trends Study (NHATS), a longitudinal survey of older adult Medicare beneficiaries.
• Researchers also collected data on 3075 newly recruited individuals in Round 5 and followed to Round 8 as an independent validation sample to create a pooled sample of 7732 older adults, mean age 76.06, without MCR at baseline.
• Social frailty, assessed at baseline, included five social items: Going out less, not feeling confident, rarely visiting friends/family, not talking with others, and without live-in partner/spouse (researchers divided participants into normal [zero to one items] and social frailty [two to five items] groups).
• Individuals were considered to have MCR if they had both subjective cognitive complaints and slow gait speed (greater than 1 standard deviation below age-specific level) without dementia or mobility disability.
• Covariates included demographic and lifestyle data, presence of depression and/or anxiety symptoms, and number of chronic diseases.

TAKEAWAY:

• During a median follow-up period of 4 years, 10.35% individuals were diagnosed with MCR.
• After the researchers controlled for confounding factors, those with social frailty had an increased risk for MCR compared with the normal group (pooled sample: hazard ratio [HR], 1.57; 95% CI, 1.34-1.84; P < .001).
• Each additional unfavorable social item was associated with an increased risk for MCR (pooled sample: HR, 1.32; 95% CI, 1.22-1.43; P < .001).
• Results of stratified analyses across subgroups suggested individuals with social frailty had a significantly higher risk for incident MCR than that of those without social frailty, regardless of socioeconomic status, lifestyle factors, chronic diseases, and mental health.

IN PRACTICE:

The findings suggest assessing social frailty using simple questions “is an efficient tool for detecting older individuals with a high risk of MCR,” the authors wrote. They noted that the addition of such a tool in clinical practice may facilitate “timely implementation of prevention strategies.”

SOURCE:

The research was led by Hui Zhang, Human Phenome Institute, Zhangjiang Fudan International Innovation Centre, Fudan University, Shanghai, China. It was published online on January 29, 2024, in Alzheimer’s & Dementia.

LIMITATIONS:

The study was observational, so the association between social frailty and MCR is merely correlational. Due to the lack of genetic information in NHATS data, researchers didn’t evaluate the effect of genetic factors such as apolipoprotein E on the association between social frailty and MCR. Social frailty was assessed at a single time point. In addition, the researchers were unable examine the time sequence between social frailty and MCR and so could not determine the cause of this association.

DISCLOSURES:

The study was supported by the National Natural Science Foundation of China-Youth Science Fund, Shanghai Rising-Star Program, Shanghai Municipal Health Commission and Key Discipline Construction Project of Pudong Health, and Family Planning Commission of Shanghai. The authors reported no relevant conflicts of interest.
 

A version of this article appeared on Medscape.com.

TOPLINE:

Women with menstrual or perimenopausal symptoms can relieve common physical and psychological issues through cold water swimming, a new study finds.

METHODOLOGY:

• Symptoms of menstrual cycles and perimenopause vary widely but frequently include mood swings, anxiety, depression, fatigue, hot flashes, and sleep disturbances.
• This is the first investigation of whether cold water swimming has an impact on these symptoms.
• Researchers conducted a 42-question online survey of 1114 women who were regularly swam in cold water. More than two-thirds of respondents (68.1%) were between 45 and 59 years of age.
• Some of the data included responses by women who were perimenopausal but still had menstrual symptoms.

TAKEAWAY:

• Researchers found that cold water swimming had multiple beneficial effects on both menstrual and perimenopausal symptoms.
• Women who swam more frequently and for longer reported more beneficial effects than women who swam less often or for less time per swim.
• Reduction of psychological and vasomotor symptoms were most often cited by cold-water swimmers.
• Perimenopausal women who swam regularly in winter and summer saw greater reduction in anxiety and hot flashes than did those who swam in the other seasons.

IN PRACTICE:

“Teaching women to swim safely and encouraging them to swim regularly may have a benefit on the debilitating symptoms associated with the perimenopause,” the authors wrote.

SOURCE:

The study was conducted by researchers from University College, London, and published online in Post Reproductive Health. The corresponding author is Joyce Harper, PhD, professor of reproductive science at EGA Institute for Women’s Health, University College London, England.

LIMITATIONS:

This was an observational study, with no control group. The underlying cause of improved psychological symptoms of perimenopause were not fully evaluated owing to unaccounted multiple variables. Use of an online survey may introduce sampling bias and not align with the population of all menstruating or perimenopausal women. Participants were primarily White and highly educated.

DISCLOSURES:

Dr. Harper disclosed giving paid talks on menopause to businesses and at conferences.

A version of this article appeared on Medscape.com.

TOPLINE:

Women with menstrual or perimenopausal symptoms can relieve common physical and psychological issues through cold water swimming, a new study finds.

METHODOLOGY:

• Symptoms of menstrual cycles and perimenopause vary widely but frequently include mood swings, anxiety, depression, fatigue, hot flashes, and sleep disturbances.
• This is the first investigation of whether cold water swimming has an impact on these symptoms.
• Researchers conducted a 42-question online survey of 1114 women who were regularly swam in cold water. More than two-thirds of respondents (68.1%) were between 45 and 59 years of age.
• Some of the data included responses by women who were perimenopausal but still had menstrual symptoms.

TAKEAWAY:

• Researchers found that cold water swimming had multiple beneficial effects on both menstrual and perimenopausal symptoms.
• Women who swam more frequently and for longer reported more beneficial effects than women who swam less often or for less time per swim.
• Reduction of psychological and vasomotor symptoms were most often cited by cold-water swimmers.
• Perimenopausal women who swam regularly in winter and summer saw greater reduction in anxiety and hot flashes than did those who swam in the other seasons.

IN PRACTICE:

“Teaching women to swim safely and encouraging them to swim regularly may have a benefit on the debilitating symptoms associated with the perimenopause,” the authors wrote.

SOURCE:

The study was conducted by researchers from University College, London, and published online in Post Reproductive Health. The corresponding author is Joyce Harper, PhD, professor of reproductive science at EGA Institute for Women’s Health, University College London, England.

LIMITATIONS:

This was an observational study, with no control group. The underlying cause of improved psychological symptoms of perimenopause were not fully evaluated owing to unaccounted multiple variables. Use of an online survey may introduce sampling bias and not align with the population of all menstruating or perimenopausal women. Participants were primarily White and highly educated.

DISCLOSURES:

Dr. Harper disclosed giving paid talks on menopause to businesses and at conferences.

A version of this article appeared on Medscape.com.

TOPLINE:

Women with menstrual or perimenopausal symptoms can relieve common physical and psychological issues through cold water swimming, a new study finds.

METHODOLOGY:

• Symptoms of menstrual cycles and perimenopause vary widely but frequently include mood swings, anxiety, depression, fatigue, hot flashes, and sleep disturbances.
• This is the first investigation of whether cold water swimming has an impact on these symptoms.
• Researchers conducted a 42-question online survey of 1114 women who were regularly swam in cold water. More than two-thirds of respondents (68.1%) were between 45 and 59 years of age.
• Some of the data included responses by women who were perimenopausal but still had menstrual symptoms.

TAKEAWAY:

• Researchers found that cold water swimming had multiple beneficial effects on both menstrual and perimenopausal symptoms.
• Women who swam more frequently and for longer reported more beneficial effects than women who swam less often or for less time per swim.
• Reduction of psychological and vasomotor symptoms were most often cited by cold-water swimmers.
• Perimenopausal women who swam regularly in winter and summer saw greater reduction in anxiety and hot flashes than did those who swam in the other seasons.

IN PRACTICE:

“Teaching women to swim safely and encouraging them to swim regularly may have a benefit on the debilitating symptoms associated with the perimenopause,” the authors wrote.

SOURCE:

The study was conducted by researchers from University College, London, and published online in Post Reproductive Health. The corresponding author is Joyce Harper, PhD, professor of reproductive science at EGA Institute for Women’s Health, University College London, England.

LIMITATIONS:

This was an observational study, with no control group. The underlying cause of improved psychological symptoms of perimenopause were not fully evaluated owing to unaccounted multiple variables. Use of an online survey may introduce sampling bias and not align with the population of all menstruating or perimenopausal women. Participants were primarily White and highly educated.

DISCLOSURES:

Dr. Harper disclosed giving paid talks on menopause to businesses and at conferences.

A version of this article appeared on Medscape.com.

Clinicians usually measure renal function by using surrogate markers because directly measuring glomerular filtration rate (GFR) is not routinely feasible in a clinical setting.^1,2 Creatinine (Cr) and cystatin C (CysC) are the 2 main surrogate molecules used to estimate GFR.³

Creatine is a molecule nonenzymatically converted into Cr, weighing only 113 Da in skeletal muscles.⁴ It is then filtered at the glomeruli and secreted at the proximal tubules of the kidneys. However, serum Cr (sCr) levels are affected by several factors, including age, biological sex, liver function, diet, and muscle mass.⁵ Historically, sCr levels also are affected by race.⁵ In an early study of factors affecting accurate GFR, researchers reported that self-identified African American patients had a 16% higher GFR than those who did not when using Cr.⁶ Despite this, the inclusion of Cr on a basic metabolic panel has allowed automatic reporting of an estimated GFR using sCr (eGFR_Cr) to be readily available.⁷

In comparison to Cr, CysC is an endogenous protein weighing 13 kDa produced by all nucleated cells.^8,9 CysC is filtered by the kidney at the glomeruli and completely reabsorbed and catabolized by epithelial cells at the proximal tubule.⁹ Since production is not dependent on skeletal muscle, there are fewer physiological impacts on serum concentration of CysC. Levels of CysC may be elevated by factors shown in the Table.

Estimating Glomerular Filtration Rates

Multiple equations were developed to mitigate the impact of extraneous factors on the accuracy of an eGFR_Cr. The first widely used equation that included a variable adjustment for race was the Modification of Diet in Renal Disease study, presented in 2006.¹⁰ The equation increased the accuracy of eGFR_Cr further by adjusting for sex and age. It was followed by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation in 2009, which was more accurate at higher GFR levels.¹¹

CysC was simultaneously studied as an alternative to Cr with multiple equation iterations shown to be viable in various populations as early as 2003.^12-15 However, it was not until 2012 that an equation for the use of CysC was offered for widespread use as an alternative to Cr alongside further refinement of the CKD-EPI equation for Cr.¹⁶ A new formula was presented in 2021 to use both sCr and serum CysC levels to obtain a more accurate estimation of GFR.¹⁷ Research continues its effort to accurately estimate GFR for diagnosing kidney disease and assessing comorbidities relating to decreased kidney function.³

All historical equations attempted to mitigate the potential impact of race on sCr level when calculating eGFR_Crby assigning a separate variable for African American patients. As an unintended adverse effect, these equations may have led to discrimination by having a different equation for African American patients.¹⁸ Moreover, these Cr-based equations remain less accurate in patients with varied muscle mass, such as older patients, bodybuilders, athletes, and individuals with varied extremes of daily protein intake.^1,8,9,19Several medications can also directly affect Cr clearance, reducing its ability to act as a surrogate for kidney function.¹In this case report, we discuss an African American patient with high muscle mass and protein intake who was initially diagnosed with kidney disease based on an elevated Cr and found to be misdiagnosed based on the use of CysC for a more accurate GFR estimation.

A 35-year-old African American man serving in the military and recently diagnosed with HIV was referred to a nephrology clinic for further evaluation of an acute elevation in sCr. Before treatment for HIV, a brief record review showed a baseline Cr of about 1.3 mg/dL, with an eGFR_Cr of 75 mL/min/1.73 m².²⁰ In the same month, the patient was prescribed bictegravir/emtricitabine/tenofovir alafenamide, an HIV drug with nephrotoxic potential.²¹ The patient's total viral load remained low, and CD4 count remained > 500 after initiation of the HIV treatment. He was in his normal state of health and had no known contributory history before his HIV diagnosis. Cr readings peaked at 1.83 mg/dL after starting the HIV treatment and remained elevated to 1.73 mg/dL over the next few months, corresponding to CKD stage 3A. Because bictegravir/emtricitabine/tenofovir alafenamide is cleared by the kidneys and has a nephrotoxic profile, the clinical care team considered dosage adjustment or a medication switch given his observed elevated eGFR_Cr based on the CKD-EPI 2021 equation for Cr alone. It was also noted that the patient had a similar Cr spike to 1.83 mg/dL in 2018 without any identifiable renal insult or symptoms (Figure).

Diagnostic Evaluation

The primary care team ordered a renal ultrasound and referred the patient to the nephrology clinic. The nephrologist ordered the following laboratory studies: urine microalbumin to Cr ratio, basic metabolic panel (BMP), comprehensive metabolic panel (CMP), urinalysis, urine protein, urine Cr, parathyroid hormone level, hemoglobin A_1c, complement component 3/4 panels, antinuclear and antineutrophil cytoplasmic antibodies titers, glomerular basement membrane antibody titer, urine light chains, serum protein electrophoresis, κ/λ ratio, viral hepatitis panel, and rapid plasma reagin testing. Much of this laboratory evaluation served to rule out any secondary causes of kidney disease, including autoimmune disease, monoclonal or polyclonal gammopathies, diabetic nephropathy or glomerulosclerosis, and nephrotic or nephritic syndromes.

All laboratory studies returned within normal limits; no proteinuria was discovered on urinalysis, and no abnormalities were visualized on renal ultrasound. Bictegravir/emtricitabine/tenofovir alafenamide nephrotoxicity was highest among the differential diagnoses due to the timing of Cr elevation coinciding with the initiation of the medications. The patient's CysC level was 0.85 mg/dL with a calculated eGFR_Cys of 125 mL/min/1.73 m². The calculated sCR and serum cystatin C (eGFR_Cr-Cys) using the new 2021 equation and when adjusting for body surface area placed his eGFR at 92 mL/min/1.73 m².²⁰

The patient’s eGFR_Cysreassured the care team that the patient’s renal function was not acutely or chronically impacted by bictegravir/emtricitabine/tenofovir alafenamide, resulting in avoidance of unnecessary dosage adjustment or discontinuation of the HIV treatment. The patient reported a chronic habit of protein and creatine supplementation and bodybuilding, which likely further compounded the discrepancy between eGFR_Cr and eGFR_Cys and explained his previous elevation in Cr in 2018.

Follow-up

The patient underwent serial monitoring that revealed a stable Cr and unremarkable eGFR, ruling out CKD. There has been no evidence of worsening kidney disease to date, and the patient remained on his initial HIV regimen.

This case shows the importance of using CysC as an alternative or confirmatory marker compared with sCr to estimate GFR in patients with high muscle mass and/or high creatine intake, such as many in the US Department of Defense (DoD) and US Department of Veterans Affairs (VA) patient populations. In the presented case, recorded Cr levels climbed from baseline Cr with the initiation of bictegravir/emtricitabine/tenofovir alafenamide. This raised the concern that HIV treatment was leading to the development of kidney damage.²²

Diagnosis of kidney disease as opposed to the normal decline of eGFR with age in individuals without intrinsic CKD requires GFR ≥ 60 mL/min/1.73 m² with kidney damage (proteinuria or radiological abnormalities, etc) or GFR < 135 to 140 mL/min/1.73 m²minus the patient’s age in years.²³ The patient’s Cr peak at 1.83 mg/dL in 2018 led to an inappropriate diagnosis of kidney disease stage 3a based on an eGFR_Cr (2021 equation) of 52 mL/min/1.73 m² when not corrected for body surface area.²⁰ However, using the new 2021 equation using both Cr and CysC, the patient’s eGFR_Cr-Cyswas 92 mL/min/1.73 m² after a correction for body surface area.

The 2009 CKD-EPI recommended the calculation of eGFR based on SCr concentration using age, sex, and race while the 2021 CKD-EPI recommended the exclusion of race.³ Both equations are less accurate in African American patients, individuals taking medications that interfere with Cr secretion and assay, and patients taking creatine supplements, high daily protein intake, or with high muscle mass.⁷ These settings result in a decreased eGFR_Cr without corresponding eGFR_Cys changes. Using SCr and CysC together, the eGFR_Cr-Cys yields improved concordance to measured GFR across race groups compared to GFR estimation based on Cr alone, which can avoid unnecessary expensive diagnostic workup, inappropriate kidney disease diagnosis, incorrect dosing of drugs, and accurately represent the military readiness of patients. Interestingly, in African American patients with recently diagnosed HIV, CKD-EPI using both Cr and CysC without race inclusion led to only a 2.9% overestimation of GFR and was the only equation with no statistically significant bias compared with measured GFR.²⁴

A March 2023 case involving an otherwise healthy 26-year-old male active-duty US Navy member with a history of excessive protein supplement intake and intense exercise < 24 hours before laboratory work was diagnosed with CKD after a measured Cr of 16 mg/dL and an eGFR_Cr of 4 mL/min/1.73 m² without any other evidence of kidney disease. His CysC remained within normal limits, resulting in a normal eGFR_Cys of 121 mL/min/1.73 m², indicating no CKD. His Cr and eGFR recovered 10 days after his clinic visit and cessation of his supplement intake. These findings may not be uncommon given that 65% of active-duty military use protein supplements and 38% use other performance-enhancing supplements, such as creatine, according to a study.²⁵

Unfortunately, the BMP/CMP traditionally used at VA centers use the eGFR_Cr equation, and it is unknown how many primary care practitioners recognize the limitations of these metabolic panels on accurate estimation of kidney function. However, in 2022 an expert panel including VA physicians recommended the immediate use of eGFR_Cr-Cys or eGFR_Cys for confirmatory testing and potentially screening of CKD.²⁶ A small number of VAs have since adopted this recommendation, which should lead to fewer misdiagnoses among US military members as clinicians should now have access to more accurate measurements of GFR.

The VA spends about $18 billion (excluding dialysis) for care for 1.1 to 2.5 million VA patients with CKD.²⁷ The majority of these diagnoses were undoubtedly made using the eGFR_Cr equation, raising the question of how many may be misdiagnosed. Assessment with CysC is currently relatively expensive, but it will likely become more affordable as the use of CysC as a confirmatory test increases.⁵ The cost of a sCr test is about $2.50, while CysC costs about $10.60, with variation from laboratory to laboratory.²⁸ By comparison, a renal ultrasound costs $99 to $140 for uninsured patients.²⁹ Furthermore, the cost of CysC testing is likely to trend downward as more facilities adopt the use of CysC measurements, which can be run on the same analytical equipment currently used for Cr measurements. Currently, most laboratories do not have established assays to use in-house and thus require CysC to be sent out to a laboratory, which increases result time and makes Cr a more attractive option. As more laboratories adopt assays for CysC, the cost of reagents will further decrease.

Given such considerations, confirmation testing of kidney function with CysC in specific patient populations with decreased eGFR_Cr without other features of CKD can offer great medical and financial benefits. A 2023 KDIGO report noted that many individuals may be mistakenly diagnosed with CKD when using eGFR_Cr.³ KDIGO noted that a 2013 meta-analysis of 90,000 individuals found that with a Cr-based eGFR of 45 to 59 mL/min/1.73 m² (42%) had a CysC-based eGFR of ≥ 60 mL/min/1.73 m². An eGFR_Cr of 45 to 59 represents 54% of all patients with CKD, amounting to millions of people (including current and former military personnel).^3,29-31 Correcting a misdiagnosis of CKD would bring significant relief to patients and save millions in health care spending.

Conclusions

In patients who meet CKD criteria using eGFR_Cr but without other features of CKD, we recommend using confirmatory CysC levels and the eGFR_Cr-Cys equation. This will align care with the KDIGO guidelines and could be a cost-effective step toward improving military patient care. Further work in this area should focus on determining the knowledge gaps in primary care practitioners’ understanding of the limits of eGFR_Cr, the potential mitigation of concomitant CysC testing in equivocal CKD cases, and the cost-effectiveness and increased utilization of CysC.

Clinicians usually measure renal function by using surrogate markers because directly measuring glomerular filtration rate (GFR) is not routinely feasible in a clinical setting.^1,2 Creatinine (Cr) and cystatin C (CysC) are the 2 main surrogate molecules used to estimate GFR.³

Creatine is a molecule nonenzymatically converted into Cr, weighing only 113 Da in skeletal muscles.⁴ It is then filtered at the glomeruli and secreted at the proximal tubules of the kidneys. However, serum Cr (sCr) levels are affected by several factors, including age, biological sex, liver function, diet, and muscle mass.⁵ Historically, sCr levels also are affected by race.⁵ In an early study of factors affecting accurate GFR, researchers reported that self-identified African American patients had a 16% higher GFR than those who did not when using Cr.⁶ Despite this, the inclusion of Cr on a basic metabolic panel has allowed automatic reporting of an estimated GFR using sCr (eGFR_Cr) to be readily available.⁷

In comparison to Cr, CysC is an endogenous protein weighing 13 kDa produced by all nucleated cells.^8,9 CysC is filtered by the kidney at the glomeruli and completely reabsorbed and catabolized by epithelial cells at the proximal tubule.⁹ Since production is not dependent on skeletal muscle, there are fewer physiological impacts on serum concentration of CysC. Levels of CysC may be elevated by factors shown in the Table.

Estimating Glomerular Filtration Rates

Multiple equations were developed to mitigate the impact of extraneous factors on the accuracy of an eGFR_Cr. The first widely used equation that included a variable adjustment for race was the Modification of Diet in Renal Disease study, presented in 2006.¹⁰ The equation increased the accuracy of eGFR_Cr further by adjusting for sex and age. It was followed by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation in 2009, which was more accurate at higher GFR levels.¹¹

CysC was simultaneously studied as an alternative to Cr with multiple equation iterations shown to be viable in various populations as early as 2003.^12-15 However, it was not until 2012 that an equation for the use of CysC was offered for widespread use as an alternative to Cr alongside further refinement of the CKD-EPI equation for Cr.¹⁶ A new formula was presented in 2021 to use both sCr and serum CysC levels to obtain a more accurate estimation of GFR.¹⁷ Research continues its effort to accurately estimate GFR for diagnosing kidney disease and assessing comorbidities relating to decreased kidney function.³

All historical equations attempted to mitigate the potential impact of race on sCr level when calculating eGFR_Crby assigning a separate variable for African American patients. As an unintended adverse effect, these equations may have led to discrimination by having a different equation for African American patients.¹⁸ Moreover, these Cr-based equations remain less accurate in patients with varied muscle mass, such as older patients, bodybuilders, athletes, and individuals with varied extremes of daily protein intake.^1,8,9,19Several medications can also directly affect Cr clearance, reducing its ability to act as a surrogate for kidney function.¹In this case report, we discuss an African American patient with high muscle mass and protein intake who was initially diagnosed with kidney disease based on an elevated Cr and found to be misdiagnosed based on the use of CysC for a more accurate GFR estimation.

A 35-year-old African American man serving in the military and recently diagnosed with HIV was referred to a nephrology clinic for further evaluation of an acute elevation in sCr. Before treatment for HIV, a brief record review showed a baseline Cr of about 1.3 mg/dL, with an eGFR_Cr of 75 mL/min/1.73 m².²⁰ In the same month, the patient was prescribed bictegravir/emtricitabine/tenofovir alafenamide, an HIV drug with nephrotoxic potential.²¹ The patient's total viral load remained low, and CD4 count remained > 500 after initiation of the HIV treatment. He was in his normal state of health and had no known contributory history before his HIV diagnosis. Cr readings peaked at 1.83 mg/dL after starting the HIV treatment and remained elevated to 1.73 mg/dL over the next few months, corresponding to CKD stage 3A. Because bictegravir/emtricitabine/tenofovir alafenamide is cleared by the kidneys and has a nephrotoxic profile, the clinical care team considered dosage adjustment or a medication switch given his observed elevated eGFR_Cr based on the CKD-EPI 2021 equation for Cr alone. It was also noted that the patient had a similar Cr spike to 1.83 mg/dL in 2018 without any identifiable renal insult or symptoms (Figure).

Diagnostic Evaluation

The primary care team ordered a renal ultrasound and referred the patient to the nephrology clinic. The nephrologist ordered the following laboratory studies: urine microalbumin to Cr ratio, basic metabolic panel (BMP), comprehensive metabolic panel (CMP), urinalysis, urine protein, urine Cr, parathyroid hormone level, hemoglobin A_1c, complement component 3/4 panels, antinuclear and antineutrophil cytoplasmic antibodies titers, glomerular basement membrane antibody titer, urine light chains, serum protein electrophoresis, κ/λ ratio, viral hepatitis panel, and rapid plasma reagin testing. Much of this laboratory evaluation served to rule out any secondary causes of kidney disease, including autoimmune disease, monoclonal or polyclonal gammopathies, diabetic nephropathy or glomerulosclerosis, and nephrotic or nephritic syndromes.

All laboratory studies returned within normal limits; no proteinuria was discovered on urinalysis, and no abnormalities were visualized on renal ultrasound. Bictegravir/emtricitabine/tenofovir alafenamide nephrotoxicity was highest among the differential diagnoses due to the timing of Cr elevation coinciding with the initiation of the medications. The patient's CysC level was 0.85 mg/dL with a calculated eGFR_Cys of 125 mL/min/1.73 m². The calculated sCR and serum cystatin C (eGFR_Cr-Cys) using the new 2021 equation and when adjusting for body surface area placed his eGFR at 92 mL/min/1.73 m².²⁰

The patient’s eGFR_Cysreassured the care team that the patient’s renal function was not acutely or chronically impacted by bictegravir/emtricitabine/tenofovir alafenamide, resulting in avoidance of unnecessary dosage adjustment or discontinuation of the HIV treatment. The patient reported a chronic habit of protein and creatine supplementation and bodybuilding, which likely further compounded the discrepancy between eGFR_Cr and eGFR_Cys and explained his previous elevation in Cr in 2018.

Follow-up

The patient underwent serial monitoring that revealed a stable Cr and unremarkable eGFR, ruling out CKD. There has been no evidence of worsening kidney disease to date, and the patient remained on his initial HIV regimen.

This case shows the importance of using CysC as an alternative or confirmatory marker compared with sCr to estimate GFR in patients with high muscle mass and/or high creatine intake, such as many in the US Department of Defense (DoD) and US Department of Veterans Affairs (VA) patient populations. In the presented case, recorded Cr levels climbed from baseline Cr with the initiation of bictegravir/emtricitabine/tenofovir alafenamide. This raised the concern that HIV treatment was leading to the development of kidney damage.²²

Diagnosis of kidney disease as opposed to the normal decline of eGFR with age in individuals without intrinsic CKD requires GFR ≥ 60 mL/min/1.73 m² with kidney damage (proteinuria or radiological abnormalities, etc) or GFR < 135 to 140 mL/min/1.73 m²minus the patient’s age in years.²³ The patient’s Cr peak at 1.83 mg/dL in 2018 led to an inappropriate diagnosis of kidney disease stage 3a based on an eGFR_Cr (2021 equation) of 52 mL/min/1.73 m² when not corrected for body surface area.²⁰ However, using the new 2021 equation using both Cr and CysC, the patient’s eGFR_Cr-Cyswas 92 mL/min/1.73 m² after a correction for body surface area.

The 2009 CKD-EPI recommended the calculation of eGFR based on SCr concentration using age, sex, and race while the 2021 CKD-EPI recommended the exclusion of race.³ Both equations are less accurate in African American patients, individuals taking medications that interfere with Cr secretion and assay, and patients taking creatine supplements, high daily protein intake, or with high muscle mass.⁷ These settings result in a decreased eGFR_Cr without corresponding eGFR_Cys changes. Using SCr and CysC together, the eGFR_Cr-Cys yields improved concordance to measured GFR across race groups compared to GFR estimation based on Cr alone, which can avoid unnecessary expensive diagnostic workup, inappropriate kidney disease diagnosis, incorrect dosing of drugs, and accurately represent the military readiness of patients. Interestingly, in African American patients with recently diagnosed HIV, CKD-EPI using both Cr and CysC without race inclusion led to only a 2.9% overestimation of GFR and was the only equation with no statistically significant bias compared with measured GFR.²⁴

A March 2023 case involving an otherwise healthy 26-year-old male active-duty US Navy member with a history of excessive protein supplement intake and intense exercise < 24 hours before laboratory work was diagnosed with CKD after a measured Cr of 16 mg/dL and an eGFR_Cr of 4 mL/min/1.73 m² without any other evidence of kidney disease. His CysC remained within normal limits, resulting in a normal eGFR_Cys of 121 mL/min/1.73 m², indicating no CKD. His Cr and eGFR recovered 10 days after his clinic visit and cessation of his supplement intake. These findings may not be uncommon given that 65% of active-duty military use protein supplements and 38% use other performance-enhancing supplements, such as creatine, according to a study.²⁵

Unfortunately, the BMP/CMP traditionally used at VA centers use the eGFR_Cr equation, and it is unknown how many primary care practitioners recognize the limitations of these metabolic panels on accurate estimation of kidney function. However, in 2022 an expert panel including VA physicians recommended the immediate use of eGFR_Cr-Cys or eGFR_Cys for confirmatory testing and potentially screening of CKD.²⁶ A small number of VAs have since adopted this recommendation, which should lead to fewer misdiagnoses among US military members as clinicians should now have access to more accurate measurements of GFR.

The VA spends about $18 billion (excluding dialysis) for care for 1.1 to 2.5 million VA patients with CKD.²⁷ The majority of these diagnoses were undoubtedly made using the eGFR_Cr equation, raising the question of how many may be misdiagnosed. Assessment with CysC is currently relatively expensive, but it will likely become more affordable as the use of CysC as a confirmatory test increases.⁵ The cost of a sCr test is about $2.50, while CysC costs about $10.60, with variation from laboratory to laboratory.²⁸ By comparison, a renal ultrasound costs $99 to $140 for uninsured patients.²⁹ Furthermore, the cost of CysC testing is likely to trend downward as more facilities adopt the use of CysC measurements, which can be run on the same analytical equipment currently used for Cr measurements. Currently, most laboratories do not have established assays to use in-house and thus require CysC to be sent out to a laboratory, which increases result time and makes Cr a more attractive option. As more laboratories adopt assays for CysC, the cost of reagents will further decrease.

Given such considerations, confirmation testing of kidney function with CysC in specific patient populations with decreased eGFR_Cr without other features of CKD can offer great medical and financial benefits. A 2023 KDIGO report noted that many individuals may be mistakenly diagnosed with CKD when using eGFR_Cr.³ KDIGO noted that a 2013 meta-analysis of 90,000 individuals found that with a Cr-based eGFR of 45 to 59 mL/min/1.73 m² (42%) had a CysC-based eGFR of ≥ 60 mL/min/1.73 m². An eGFR_Cr of 45 to 59 represents 54% of all patients with CKD, amounting to millions of people (including current and former military personnel).^3,29-31 Correcting a misdiagnosis of CKD would bring significant relief to patients and save millions in health care spending.

Conclusions

In patients who meet CKD criteria using eGFR_Cr but without other features of CKD, we recommend using confirmatory CysC levels and the eGFR_Cr-Cys equation. This will align care with the KDIGO guidelines and could be a cost-effective step toward improving military patient care. Further work in this area should focus on determining the knowledge gaps in primary care practitioners’ understanding of the limits of eGFR_Cr, the potential mitigation of concomitant CysC testing in equivocal CKD cases, and the cost-effectiveness and increased utilization of CysC.

Clinicians usually measure renal function by using surrogate markers because directly measuring glomerular filtration rate (GFR) is not routinely feasible in a clinical setting.^1,2 Creatinine (Cr) and cystatin C (CysC) are the 2 main surrogate molecules used to estimate GFR.³

Creatine is a molecule nonenzymatically converted into Cr, weighing only 113 Da in skeletal muscles.⁴ It is then filtered at the glomeruli and secreted at the proximal tubules of the kidneys. However, serum Cr (sCr) levels are affected by several factors, including age, biological sex, liver function, diet, and muscle mass.⁵ Historically, sCr levels also are affected by race.⁵ In an early study of factors affecting accurate GFR, researchers reported that self-identified African American patients had a 16% higher GFR than those who did not when using Cr.⁶ Despite this, the inclusion of Cr on a basic metabolic panel has allowed automatic reporting of an estimated GFR using sCr (eGFR_Cr) to be readily available.⁷

In comparison to Cr, CysC is an endogenous protein weighing 13 kDa produced by all nucleated cells.^8,9 CysC is filtered by the kidney at the glomeruli and completely reabsorbed and catabolized by epithelial cells at the proximal tubule.⁹ Since production is not dependent on skeletal muscle, there are fewer physiological impacts on serum concentration of CysC. Levels of CysC may be elevated by factors shown in the Table.

Estimating Glomerular Filtration Rates

Multiple equations were developed to mitigate the impact of extraneous factors on the accuracy of an eGFR_Cr. The first widely used equation that included a variable adjustment for race was the Modification of Diet in Renal Disease study, presented in 2006.¹⁰ The equation increased the accuracy of eGFR_Cr further by adjusting for sex and age. It was followed by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation in 2009, which was more accurate at higher GFR levels.¹¹

CysC was simultaneously studied as an alternative to Cr with multiple equation iterations shown to be viable in various populations as early as 2003.^12-15 However, it was not until 2012 that an equation for the use of CysC was offered for widespread use as an alternative to Cr alongside further refinement of the CKD-EPI equation for Cr.¹⁶ A new formula was presented in 2021 to use both sCr and serum CysC levels to obtain a more accurate estimation of GFR.¹⁷ Research continues its effort to accurately estimate GFR for diagnosing kidney disease and assessing comorbidities relating to decreased kidney function.³

All historical equations attempted to mitigate the potential impact of race on sCr level when calculating eGFR_Crby assigning a separate variable for African American patients. As an unintended adverse effect, these equations may have led to discrimination by having a different equation for African American patients.¹⁸ Moreover, these Cr-based equations remain less accurate in patients with varied muscle mass, such as older patients, bodybuilders, athletes, and individuals with varied extremes of daily protein intake.^1,8,9,19Several medications can also directly affect Cr clearance, reducing its ability to act as a surrogate for kidney function.¹In this case report, we discuss an African American patient with high muscle mass and protein intake who was initially diagnosed with kidney disease based on an elevated Cr and found to be misdiagnosed based on the use of CysC for a more accurate GFR estimation.

A 35-year-old African American man serving in the military and recently diagnosed with HIV was referred to a nephrology clinic for further evaluation of an acute elevation in sCr. Before treatment for HIV, a brief record review showed a baseline Cr of about 1.3 mg/dL, with an eGFR_Cr of 75 mL/min/1.73 m².²⁰ In the same month, the patient was prescribed bictegravir/emtricitabine/tenofovir alafenamide, an HIV drug with nephrotoxic potential.²¹ The patient's total viral load remained low, and CD4 count remained > 500 after initiation of the HIV treatment. He was in his normal state of health and had no known contributory history before his HIV diagnosis. Cr readings peaked at 1.83 mg/dL after starting the HIV treatment and remained elevated to 1.73 mg/dL over the next few months, corresponding to CKD stage 3A. Because bictegravir/emtricitabine/tenofovir alafenamide is cleared by the kidneys and has a nephrotoxic profile, the clinical care team considered dosage adjustment or a medication switch given his observed elevated eGFR_Cr based on the CKD-EPI 2021 equation for Cr alone. It was also noted that the patient had a similar Cr spike to 1.83 mg/dL in 2018 without any identifiable renal insult or symptoms (Figure).

Diagnostic Evaluation

The primary care team ordered a renal ultrasound and referred the patient to the nephrology clinic. The nephrologist ordered the following laboratory studies: urine microalbumin to Cr ratio, basic metabolic panel (BMP), comprehensive metabolic panel (CMP), urinalysis, urine protein, urine Cr, parathyroid hormone level, hemoglobin A_1c, complement component 3/4 panels, antinuclear and antineutrophil cytoplasmic antibodies titers, glomerular basement membrane antibody titer, urine light chains, serum protein electrophoresis, κ/λ ratio, viral hepatitis panel, and rapid plasma reagin testing. Much of this laboratory evaluation served to rule out any secondary causes of kidney disease, including autoimmune disease, monoclonal or polyclonal gammopathies, diabetic nephropathy or glomerulosclerosis, and nephrotic or nephritic syndromes.

All laboratory studies returned within normal limits; no proteinuria was discovered on urinalysis, and no abnormalities were visualized on renal ultrasound. Bictegravir/emtricitabine/tenofovir alafenamide nephrotoxicity was highest among the differential diagnoses due to the timing of Cr elevation coinciding with the initiation of the medications. The patient's CysC level was 0.85 mg/dL with a calculated eGFR_Cys of 125 mL/min/1.73 m². The calculated sCR and serum cystatin C (eGFR_Cr-Cys) using the new 2021 equation and when adjusting for body surface area placed his eGFR at 92 mL/min/1.73 m².²⁰

The patient’s eGFR_Cysreassured the care team that the patient’s renal function was not acutely or chronically impacted by bictegravir/emtricitabine/tenofovir alafenamide, resulting in avoidance of unnecessary dosage adjustment or discontinuation of the HIV treatment. The patient reported a chronic habit of protein and creatine supplementation and bodybuilding, which likely further compounded the discrepancy between eGFR_Cr and eGFR_Cys and explained his previous elevation in Cr in 2018.

Follow-up

The patient underwent serial monitoring that revealed a stable Cr and unremarkable eGFR, ruling out CKD. There has been no evidence of worsening kidney disease to date, and the patient remained on his initial HIV regimen.

This case shows the importance of using CysC as an alternative or confirmatory marker compared with sCr to estimate GFR in patients with high muscle mass and/or high creatine intake, such as many in the US Department of Defense (DoD) and US Department of Veterans Affairs (VA) patient populations. In the presented case, recorded Cr levels climbed from baseline Cr with the initiation of bictegravir/emtricitabine/tenofovir alafenamide. This raised the concern that HIV treatment was leading to the development of kidney damage.²²

Diagnosis of kidney disease as opposed to the normal decline of eGFR with age in individuals without intrinsic CKD requires GFR ≥ 60 mL/min/1.73 m² with kidney damage (proteinuria or radiological abnormalities, etc) or GFR < 135 to 140 mL/min/1.73 m²minus the patient’s age in years.²³ The patient’s Cr peak at 1.83 mg/dL in 2018 led to an inappropriate diagnosis of kidney disease stage 3a based on an eGFR_Cr (2021 equation) of 52 mL/min/1.73 m² when not corrected for body surface area.²⁰ However, using the new 2021 equation using both Cr and CysC, the patient’s eGFR_Cr-Cyswas 92 mL/min/1.73 m² after a correction for body surface area.

The 2009 CKD-EPI recommended the calculation of eGFR based on SCr concentration using age, sex, and race while the 2021 CKD-EPI recommended the exclusion of race.³ Both equations are less accurate in African American patients, individuals taking medications that interfere with Cr secretion and assay, and patients taking creatine supplements, high daily protein intake, or with high muscle mass.⁷ These settings result in a decreased eGFR_Cr without corresponding eGFR_Cys changes. Using SCr and CysC together, the eGFR_Cr-Cys yields improved concordance to measured GFR across race groups compared to GFR estimation based on Cr alone, which can avoid unnecessary expensive diagnostic workup, inappropriate kidney disease diagnosis, incorrect dosing of drugs, and accurately represent the military readiness of patients. Interestingly, in African American patients with recently diagnosed HIV, CKD-EPI using both Cr and CysC without race inclusion led to only a 2.9% overestimation of GFR and was the only equation with no statistically significant bias compared with measured GFR.²⁴

A March 2023 case involving an otherwise healthy 26-year-old male active-duty US Navy member with a history of excessive protein supplement intake and intense exercise < 24 hours before laboratory work was diagnosed with CKD after a measured Cr of 16 mg/dL and an eGFR_Cr of 4 mL/min/1.73 m² without any other evidence of kidney disease. His CysC remained within normal limits, resulting in a normal eGFR_Cys of 121 mL/min/1.73 m², indicating no CKD. His Cr and eGFR recovered 10 days after his clinic visit and cessation of his supplement intake. These findings may not be uncommon given that 65% of active-duty military use protein supplements and 38% use other performance-enhancing supplements, such as creatine, according to a study.²⁵

Unfortunately, the BMP/CMP traditionally used at VA centers use the eGFR_Cr equation, and it is unknown how many primary care practitioners recognize the limitations of these metabolic panels on accurate estimation of kidney function. However, in 2022 an expert panel including VA physicians recommended the immediate use of eGFR_Cr-Cys or eGFR_Cys for confirmatory testing and potentially screening of CKD.²⁶ A small number of VAs have since adopted this recommendation, which should lead to fewer misdiagnoses among US military members as clinicians should now have access to more accurate measurements of GFR.

The VA spends about $18 billion (excluding dialysis) for care for 1.1 to 2.5 million VA patients with CKD.²⁷ The majority of these diagnoses were undoubtedly made using the eGFR_Cr equation, raising the question of how many may be misdiagnosed. Assessment with CysC is currently relatively expensive, but it will likely become more affordable as the use of CysC as a confirmatory test increases.⁵ The cost of a sCr test is about $2.50, while CysC costs about $10.60, with variation from laboratory to laboratory.²⁸ By comparison, a renal ultrasound costs $99 to $140 for uninsured patients.²⁹ Furthermore, the cost of CysC testing is likely to trend downward as more facilities adopt the use of CysC measurements, which can be run on the same analytical equipment currently used for Cr measurements. Currently, most laboratories do not have established assays to use in-house and thus require CysC to be sent out to a laboratory, which increases result time and makes Cr a more attractive option. As more laboratories adopt assays for CysC, the cost of reagents will further decrease.

Given such considerations, confirmation testing of kidney function with CysC in specific patient populations with decreased eGFR_Cr without other features of CKD can offer great medical and financial benefits. A 2023 KDIGO report noted that many individuals may be mistakenly diagnosed with CKD when using eGFR_Cr.³ KDIGO noted that a 2013 meta-analysis of 90,000 individuals found that with a Cr-based eGFR of 45 to 59 mL/min/1.73 m² (42%) had a CysC-based eGFR of ≥ 60 mL/min/1.73 m². An eGFR_Cr of 45 to 59 represents 54% of all patients with CKD, amounting to millions of people (including current and former military personnel).^3,29-31 Correcting a misdiagnosis of CKD would bring significant relief to patients and save millions in health care spending.

Conclusions

In patients who meet CKD criteria using eGFR_Cr but without other features of CKD, we recommend using confirmatory CysC levels and the eGFR_Cr-Cys equation. This will align care with the KDIGO guidelines and could be a cost-effective step toward improving military patient care. Further work in this area should focus on determining the knowledge gaps in primary care practitioners’ understanding of the limits of eGFR_Cr, the potential mitigation of concomitant CysC testing in equivocal CKD cases, and the cost-effectiveness and increased utilization of CysC.

As emergency medicine doctors, we regularly give medical advice to family and close friends when they get sick or are injured and don’t know what to do. In a matter of moments, we triage, diagnose, and assemble a logical plan, whatever the issue may be. This skill comes from our training and years of experience in treating emergencies and also routine medical matters. The value proposition is clear.

Frankly, it’s a service everyone should have. Think about the potential time and money saved if this option for medical care and triage was broadly available. Overtriage would plummet. That’s when people run to the emergency department (ED) and wait endless hours, only to be reassured or receive limited treatment. Undertriage would also decline. That’s when people should go to the ED but, unwisely, wait. For example, this may occur when symptoms of dizziness end up being a stroke.

Why doesn’t everyone have an ED doctor they can call? The primary reason is that the current system mostly doesn’t support it. The most common scenario is that insurance companies pay us to see patients in an expensive box called the ED. Most EDs are situated within an even more expensive box, called a hospital.

Here’s the good news: Better access to emergency care and people who are formally trained in emergency medicine and routine matters of urgent care is increasing.

One example is telemedicine, where a remote doctor — either your own or a doctor through an app — conducts a visit. Telemedicine is more common since the pandemic, now that insurance pays for it. In emergency situations, it’s rare that your own doctor can see you immediately by telemedicine. By contrast, direct-to-consumer telemedicine (eg, Teladoc, Doctor On Demand, and others) connects you with a random doctor.

In many apps, it’s unclear not only who the doctor is, but more importantly, what their specific medical specialty or training is. It may be an ED doctor evaluating your child’s fever, or it may be a retired general surgeon or an adult rheumatology specialist in the midst of their fellowship, making an extra buck, who may have no pediatric training.
 

Training Matters

Clinical training and whether the doctor knows you matters. A recent JAMA study from Ontario, Canada, found that patients with virtual visits who saw outside family physicians (whom they had never met) compared with their own family physicians were 66% more likely to visit an ED within 7 days after the visit. This illustrates the importance of understanding your personal history in assessing acute symptoms.

Some healthcare systems do use ED physicians for on-demand telehealth services, such as Thomas Jefferson’s JeffConnect. Amazon Clinic recently entered this space, providing condition-specific acute or chronic care to adults aged 18-64 years for a fee that is, notably, not covered by insurance.

A second innovative approach, albeit not specifically in the realm of a personal emergency medicine doctor, is artificial intelligence (AI)–powered kiosks. A concierge medicine company known as Forward recently unveiled an innovative concept known as CarePods that are now available in Sacramento, California; Chandler, Arizona; and Chicago. For a membership fee, you swipe into what looks like an oversize, space-age porta-potty. You sit in a chair and run through a series of health apps, which includes a biometric body scan along with mental health screenings. It even takes your blood (without a needle) and sequences your DNA. Results are reviewed by a doctor (not yours) who talks to you by video. They advertise that AI helps make the diagnosis. Although diagnostic AI is emerging and exciting, its benefit is not clear in emergency conditions. Yet, one clear value in a kiosk over telemedicine is the ability to obtain vital signs and lab results, which are useful for diagnosis.

Another approach is the telehealth offerings used in integrated systems of care, such as Kaiser Permanente. Kaiser is both an insurance company and a deliverer of healthcare services. Kaiser maintains a nurse call center and can handle urgent e-visits. Integrated systems not only help triage patients’ acute issues but also have access to their personal health histories. They can also provide a definitive plan for in-person treatment or a specific referral. A downside of integrated care is that it often limits your choice of provider.

Insurance companies also maintain call-in lines such as HumanaFirst, which is also staffed by nurses. We have not seen data on the calls such services receive, but we doubt people that want to call their insurance company when sick or injured, knowing that the insurer benefits when you receive less care. Additionally, studies have found that nurse-only triage is not as effective as physician triage and results in higher ED referral rates.
 

The Concierge Option

Probably the closest thing to having your own personal emergency medicine doctor is concierge medicine, which combines personalized care and accessibility. Concierge doctors come in many forms, but they usually charge a fixed fee for 24/7 availability and same-day appointments. A downside of concierge medicine is its expense ($2000–$3500 per year), and that many don’t take insurance. Concierge medicine is also criticized because, as doctors gravitate toward it, people in the community often lose their physician if they can’t afford the fees.

Ultimately, remote medical advice for emergency care is clearly evolving in new ways. The inability of traditional care models to achieve this goal will lead to innovation to improve the available options that have led us to think outside of the proverbial “box” we refer to as the ED-in-the-case.

At this time, will any option come close to having a personal emergency medicine physician willing to answer your questions, real-time, as with family and close friends? We think not.

But the future certainly holds promise for alternatives that will hopefully make payers and the Centers for Medicare & Medicaid Services take notice. Innovations in personalized care that reduce costs will be critical in our current healthcare landscape.
 

Dr. Pines is clinical professor of emergency medicine at George Washington University in Washington, DC, and chief of clinical innovation at US Acute Care Solutions in Canton, Ohio. He disclosed ties with CSL Behring and Abbott Point-of-Care. Dr. Glatter is assistant professor of emergency medicine at Zucker School of Medicine at Hofstra/Northwell in Hempstead, New York. He is a medical advisor for Medscape and hosts the Hot Topics in EM series. He has disclosed no relevant financial relationships.

A version of this article appeared on Medscape.com.

As emergency medicine doctors, we regularly give medical advice to family and close friends when they get sick or are injured and don’t know what to do. In a matter of moments, we triage, diagnose, and assemble a logical plan, whatever the issue may be. This skill comes from our training and years of experience in treating emergencies and also routine medical matters. The value proposition is clear.

Frankly, it’s a service everyone should have. Think about the potential time and money saved if this option for medical care and triage was broadly available. Overtriage would plummet. That’s when people run to the emergency department (ED) and wait endless hours, only to be reassured or receive limited treatment. Undertriage would also decline. That’s when people should go to the ED but, unwisely, wait. For example, this may occur when symptoms of dizziness end up being a stroke.

Why doesn’t everyone have an ED doctor they can call? The primary reason is that the current system mostly doesn’t support it. The most common scenario is that insurance companies pay us to see patients in an expensive box called the ED. Most EDs are situated within an even more expensive box, called a hospital.

Here’s the good news: Better access to emergency care and people who are formally trained in emergency medicine and routine matters of urgent care is increasing.

One example is telemedicine, where a remote doctor — either your own or a doctor through an app — conducts a visit. Telemedicine is more common since the pandemic, now that insurance pays for it. In emergency situations, it’s rare that your own doctor can see you immediately by telemedicine. By contrast, direct-to-consumer telemedicine (eg, Teladoc, Doctor On Demand, and others) connects you with a random doctor.

In many apps, it’s unclear not only who the doctor is, but more importantly, what their specific medical specialty or training is. It may be an ED doctor evaluating your child’s fever, or it may be a retired general surgeon or an adult rheumatology specialist in the midst of their fellowship, making an extra buck, who may have no pediatric training.
 

Training Matters

Clinical training and whether the doctor knows you matters. A recent JAMA study from Ontario, Canada, found that patients with virtual visits who saw outside family physicians (whom they had never met) compared with their own family physicians were 66% more likely to visit an ED within 7 days after the visit. This illustrates the importance of understanding your personal history in assessing acute symptoms.

Some healthcare systems do use ED physicians for on-demand telehealth services, such as Thomas Jefferson’s JeffConnect. Amazon Clinic recently entered this space, providing condition-specific acute or chronic care to adults aged 18-64 years for a fee that is, notably, not covered by insurance.

A second innovative approach, albeit not specifically in the realm of a personal emergency medicine doctor, is artificial intelligence (AI)–powered kiosks. A concierge medicine company known as Forward recently unveiled an innovative concept known as CarePods that are now available in Sacramento, California; Chandler, Arizona; and Chicago. For a membership fee, you swipe into what looks like an oversize, space-age porta-potty. You sit in a chair and run through a series of health apps, which includes a biometric body scan along with mental health screenings. It even takes your blood (without a needle) and sequences your DNA. Results are reviewed by a doctor (not yours) who talks to you by video. They advertise that AI helps make the diagnosis. Although diagnostic AI is emerging and exciting, its benefit is not clear in emergency conditions. Yet, one clear value in a kiosk over telemedicine is the ability to obtain vital signs and lab results, which are useful for diagnosis.

Another approach is the telehealth offerings used in integrated systems of care, such as Kaiser Permanente. Kaiser is both an insurance company and a deliverer of healthcare services. Kaiser maintains a nurse call center and can handle urgent e-visits. Integrated systems not only help triage patients’ acute issues but also have access to their personal health histories. They can also provide a definitive plan for in-person treatment or a specific referral. A downside of integrated care is that it often limits your choice of provider.

Insurance companies also maintain call-in lines such as HumanaFirst, which is also staffed by nurses. We have not seen data on the calls such services receive, but we doubt people that want to call their insurance company when sick or injured, knowing that the insurer benefits when you receive less care. Additionally, studies have found that nurse-only triage is not as effective as physician triage and results in higher ED referral rates.
 

The Concierge Option

Probably the closest thing to having your own personal emergency medicine doctor is concierge medicine, which combines personalized care and accessibility. Concierge doctors come in many forms, but they usually charge a fixed fee for 24/7 availability and same-day appointments. A downside of concierge medicine is its expense ($2000–$3500 per year), and that many don’t take insurance. Concierge medicine is also criticized because, as doctors gravitate toward it, people in the community often lose their physician if they can’t afford the fees.

Ultimately, remote medical advice for emergency care is clearly evolving in new ways. The inability of traditional care models to achieve this goal will lead to innovation to improve the available options that have led us to think outside of the proverbial “box” we refer to as the ED-in-the-case.

At this time, will any option come close to having a personal emergency medicine physician willing to answer your questions, real-time, as with family and close friends? We think not.

But the future certainly holds promise for alternatives that will hopefully make payers and the Centers for Medicare & Medicaid Services take notice. Innovations in personalized care that reduce costs will be critical in our current healthcare landscape.
 

Dr. Pines is clinical professor of emergency medicine at George Washington University in Washington, DC, and chief of clinical innovation at US Acute Care Solutions in Canton, Ohio. He disclosed ties with CSL Behring and Abbott Point-of-Care. Dr. Glatter is assistant professor of emergency medicine at Zucker School of Medicine at Hofstra/Northwell in Hempstead, New York. He is a medical advisor for Medscape and hosts the Hot Topics in EM series. He has disclosed no relevant financial relationships.

A version of this article appeared on Medscape.com.

As emergency medicine doctors, we regularly give medical advice to family and close friends when they get sick or are injured and don’t know what to do. In a matter of moments, we triage, diagnose, and assemble a logical plan, whatever the issue may be. This skill comes from our training and years of experience in treating emergencies and also routine medical matters. The value proposition is clear.

Frankly, it’s a service everyone should have. Think about the potential time and money saved if this option for medical care and triage was broadly available. Overtriage would plummet. That’s when people run to the emergency department (ED) and wait endless hours, only to be reassured or receive limited treatment. Undertriage would also decline. That’s when people should go to the ED but, unwisely, wait. For example, this may occur when symptoms of dizziness end up being a stroke.

Why doesn’t everyone have an ED doctor they can call? The primary reason is that the current system mostly doesn’t support it. The most common scenario is that insurance companies pay us to see patients in an expensive box called the ED. Most EDs are situated within an even more expensive box, called a hospital.

Here’s the good news: Better access to emergency care and people who are formally trained in emergency medicine and routine matters of urgent care is increasing.

One example is telemedicine, where a remote doctor — either your own or a doctor through an app — conducts a visit. Telemedicine is more common since the pandemic, now that insurance pays for it. In emergency situations, it’s rare that your own doctor can see you immediately by telemedicine. By contrast, direct-to-consumer telemedicine (eg, Teladoc, Doctor On Demand, and others) connects you with a random doctor.

In many apps, it’s unclear not only who the doctor is, but more importantly, what their specific medical specialty or training is. It may be an ED doctor evaluating your child’s fever, or it may be a retired general surgeon or an adult rheumatology specialist in the midst of their fellowship, making an extra buck, who may have no pediatric training.
 

Training Matters

Clinical training and whether the doctor knows you matters. A recent JAMA study from Ontario, Canada, found that patients with virtual visits who saw outside family physicians (whom they had never met) compared with their own family physicians were 66% more likely to visit an ED within 7 days after the visit. This illustrates the importance of understanding your personal history in assessing acute symptoms.

Some healthcare systems do use ED physicians for on-demand telehealth services, such as Thomas Jefferson’s JeffConnect. Amazon Clinic recently entered this space, providing condition-specific acute or chronic care to adults aged 18-64 years for a fee that is, notably, not covered by insurance.

A second innovative approach, albeit not specifically in the realm of a personal emergency medicine doctor, is artificial intelligence (AI)–powered kiosks. A concierge medicine company known as Forward recently unveiled an innovative concept known as CarePods that are now available in Sacramento, California; Chandler, Arizona; and Chicago. For a membership fee, you swipe into what looks like an oversize, space-age porta-potty. You sit in a chair and run through a series of health apps, which includes a biometric body scan along with mental health screenings. It even takes your blood (without a needle) and sequences your DNA. Results are reviewed by a doctor (not yours) who talks to you by video. They advertise that AI helps make the diagnosis. Although diagnostic AI is emerging and exciting, its benefit is not clear in emergency conditions. Yet, one clear value in a kiosk over telemedicine is the ability to obtain vital signs and lab results, which are useful for diagnosis.

Another approach is the telehealth offerings used in integrated systems of care, such as Kaiser Permanente. Kaiser is both an insurance company and a deliverer of healthcare services. Kaiser maintains a nurse call center and can handle urgent e-visits. Integrated systems not only help triage patients’ acute issues but also have access to their personal health histories. They can also provide a definitive plan for in-person treatment or a specific referral. A downside of integrated care is that it often limits your choice of provider.

Insurance companies also maintain call-in lines such as HumanaFirst, which is also staffed by nurses. We have not seen data on the calls such services receive, but we doubt people that want to call their insurance company when sick or injured, knowing that the insurer benefits when you receive less care. Additionally, studies have found that nurse-only triage is not as effective as physician triage and results in higher ED referral rates.
 

The Concierge Option

Probably the closest thing to having your own personal emergency medicine doctor is concierge medicine, which combines personalized care and accessibility. Concierge doctors come in many forms, but they usually charge a fixed fee for 24/7 availability and same-day appointments. A downside of concierge medicine is its expense ($2000–$3500 per year), and that many don’t take insurance. Concierge medicine is also criticized because, as doctors gravitate toward it, people in the community often lose their physician if they can’t afford the fees.

Ultimately, remote medical advice for emergency care is clearly evolving in new ways. The inability of traditional care models to achieve this goal will lead to innovation to improve the available options that have led us to think outside of the proverbial “box” we refer to as the ED-in-the-case.

At this time, will any option come close to having a personal emergency medicine physician willing to answer your questions, real-time, as with family and close friends? We think not.

But the future certainly holds promise for alternatives that will hopefully make payers and the Centers for Medicare & Medicaid Services take notice. Innovations in personalized care that reduce costs will be critical in our current healthcare landscape.
 

Dr. Pines is clinical professor of emergency medicine at George Washington University in Washington, DC, and chief of clinical innovation at US Acute Care Solutions in Canton, Ohio. He disclosed ties with CSL Behring and Abbott Point-of-Care. Dr. Glatter is assistant professor of emergency medicine at Zucker School of Medicine at Hofstra/Northwell in Hempstead, New York. He is a medical advisor for Medscape and hosts the Hot Topics in EM series. He has disclosed no relevant financial relationships.

A version of this article appeared on Medscape.com.

To the Editor:

Phototherapy regularly is utilized in the outpatient setting to address various skin pathologies, including atopic dermatitis, psoriasis, pruritus, vitiligo, and mycosis fungoides.^1,2 Phototherapy is broadly defined by the measured administration of nonionizing radiation within the UV range including wavelengths within the UVA (eg, psoralen sensitizer plus UVA-1) and UVB (eg, broadband UVB, narrowband UVB) spectrums.^1,3 Generally, the mechanism of action is derived from effects on inflammatory components of cutaneous disorders and the induction of apoptosis, both precipitating numerous downstream events.⁴

From 2015 to 2018, there were more than 1.3 million outpatient phototherapy visits in the United States, with the most common procedural indications being dermatitis not otherwise specified, atopic dermatitis, and pruritus.⁵ From 2000 to 2015, the quantity of phototherapy services billed to Medicare trended upwards by an average of 5% per year, increasing from 334,670 in the year 2000 to 692,093 in 2015.⁶ Therefore, an illustration of associated costs would be beneficial. Additionally, because total cost and physician reimbursement fluctuate from year to year, studies demonstrating overall trends can inform both US policymakers and physicians. There is a paucity of research on geographical trends for procedural reimbursements in dermatology for phototherapy. Understanding geographic trends of reimbursement could duly serve to optimize dermatologist practice patterns involving access to viable and quality care for patients seeking treatment as well as draw health policymakers’ attention to striking adjustments in physician fees. Therefore, in this study we aimed to illustrate the most recent regional payment trends in phototherapy procedures for Medicare B patients.

We queried the Centers for Medicare & Medicaid Services Medicare Physician Fee Schedule (MPFS) database (https://www.cms.gov/medicare/payment/fee-schedules/physician/lookup-tool) for the years 2010 to 2023 for Current Procedural Terminology (CPT) codes common to phototherapy procedures: actinotherapy (96900); photochemotherapy by Goeckerman treatment or using petrolatum and UVB (96910); photochemotherapy using psoralen plus UVA (96912); and photochemotherapy of severe dermatoses requiring a minimum of 4 hours of care under direct physician supervision (96913). Nonfacility prices for these procedures were analyzed. For 2010, due to midyear alterations to Medicare reimbursement (owed to bills HR 3962 and HR 4872), the mean price data of MPFS files 2010A and 2010B were used. All dollar values were converted to January 2023 US dollars using corresponding consumer price index inflation data. The Medicare Administrative Contractors were used to group state pricing information by region in accordance with established US Census Bureau subdivisions (https://www.census.gov/programs-surveys/economic-census/guidance-geographies/levels.html). Weighted percentage change in reimbursement rate was calculated using physician (MD or DO) utilization (procedure volume) data available in the 2020 Physician and Other Practitioners Public Use File (https://data.cms.gov/provider-summary-by-type-of-service/medicare-physician-other-practitioners/medicare-physician-other-practitioners-by-provider-and-service). All descriptive statistics and visualization were generated using R software (v4.2.2)(R Development Core Team).

Table 1 provides physician utilization data and the corresponding number of Part B beneficiaries for phototherapy procedures in 2020. There were 65,045 services of actinotherapy provided to a total of 6855 unique Part B beneficiaries, 173,979 services of photochemotherapy by Goeckerman treatment or using petrolatum and UVB provided to 13,122 unique Part B beneficiaries, 2524 services of photochemotherapy using psoralen plus UVA provided to a total of 357 unique Part B beneficiaries, and 37 services of photochemotherapy of severe dermatoses requiring a minimum of 4 hours of care under direct physician supervision provided to a total of 27 unique Part B beneficiaries.

On average (unweighted), phototherapy reimbursement rates in the North increased by 0.68% between 2010 and 2023 (Table 2). After weighting for 2020 physician utilization, the average change in reimbursement rate was +19.37%. During this time period, CPT code 96910 reported the greatest adjusted increase in reimbursement (+31.45%)($98.12 to $128.98; compound annual growth rate [CAGR], +0.0213), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−12.76%)($126.09 to $109.97; CAGR, −0.0105). For CPT code 96900, the reported adjusted decrease in reimbursement was −11.68% ($30.21 to $26.68; CAGR, −0.0095), and for CPT code 96913, the reported adjusted decrease in reimbursement was −4.27% ($174.03 to $166.60; CAGR, −0.0034).

On average (unweighted), phototherapy reimbursement rates in the Midwest increased by 8.40% between 2010 and 2023 (Table 3). After weighting for 2020 physician utilization, the average change in reimbursement rate was +28.53%. During this time period, CPT code 96910 reported the greatest adjusted change in reimbursement (+41.48%)($80.42 to $113.78; CAGR, +0.0270), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−6.14%)($103.28 to $97.03; CAGR, −0.0049). For CPT code 96900, the reported adjusted decrease in reimbursement was −4.73% ($24.69 to $23.52; CAGR, −0.0037), and for CPT code 96913, the reported adjusted increase in reimbursement was +2.99% ($142.72 to $146.99; CAGR, +0.0023).

On average (unweighted), phototherapy reimbursement rates in the South decreased by 2.62% between 2010 and 2023 (Table 4). After weighting for 2020 physician utilization, the average change in reimbursement rate was +15.41%. During this time period, CPT code 96910 reported the greatest adjusted change in reimbursement (+27.26%)($90.40 to $115.04 USD; CAGR, +0.0187), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−15.50%)($116.08 to $98.09; CAGR, −0.0129). For CPT code 96900, the reported adjusted decrease in reimbursement was −15.06% ($28.02 to $23.80; CAGR, −0.0125), and for CPT code 96913, the reported adjusted decrease in reimbursement was −7.19% ($160.11 to $148.61; CAGR, −0.0057).

On average (unweighted), phototherapy reimbursement rates in the West increased by 27.53% between 2010 and 2023 (Table 5). After weighting for 2020 physician utilization, the average change in reimbursement rate was +51.16%. Reimbursement for all analyzed procedures increased in the western United States. During this time period, CPT code 96910 reported the greatest adjusted increase in reimbursement (+66.56%)($80.84 to $134.65; CAGR, +0.0400), and CPT code 96912 reported the lowest adjusted increase in reimbursement (+10.64%)($103.88 to $114.93; CAGR, +0.0078). For CPT code 96900, the reported adjusted increase in reimbursement was 11.54% ($24.88 to $27.75; CAGR, +0.0084), and for CPT code 96913, the reported adjusted increase in reimbursement was 21.38% ($143.39 to $174.04; CAGR, +0.0150).

In this study evaluating geographical payment trends for phototherapy from 2010 to 2023, we demonstrated regional inconsistency in mean inflation-adjusted Medicare reimbursement rates. We found that all phototherapy procedures had increased reimbursement in the western United States, whereas all other regions reported cuts in reimbursement rates for at least half of the analyzed procedures. After adjusting for procedure utilization by physicians, weighted mean reimbursement for phototherapy increased in all US regions.

In a cross-sectional study that explored trends in the geographic distribution of dermatologists from 2012 to 2017, dermatologists in the northeastern and western United States were more likely to be located in higher-income zip codes, whereas dermatologists in the southern United States were more likely to be located in lower-income zip codes,⁷ suggesting that payment rate changes are not concordant with cost of living. Additionally, Lauck and colleagues⁸ observed that 75% of the top 20 most common procedures performed by dermatologists had decreased reimbursement (mean change, −10.8%) from 2011 to 2021. Other studies on Medicare reimbursement trends over the last 2 decades have reported major decreases within other specialties, suggesting that declining Medicare reimbursements are not unique to dermatology.^9,10 It is critical to monitor these developments, as the Centers for Medicare & Medicaid Services emphasized health care policy changes aimed at increasing reimbursements for evaluation and management services with compensatory payment cuts in billing for procedural services.¹¹

Mazmudar et al¹² previously reported a mean reimbursement decrease of −6.6% for laser/phototherapy procedures between 2007 and 2021, but these data did not include the heavily utilized Goeckerman treatment. Changes in reimbursement pose major ramifications for dermatologists—for practice size, scope, and longevity—as rates influence changes in commercial insurance reimbursements.¹³ Medicare plays a major role in the US health care system as the second largest expenditure¹⁴; indeed, between 2000 and 2015, Part B billing volume for phototherapy procedures increased 5% annually. However, phototherapy remains inaccessible in many locations due to unequal regional distribution of phototherapy clinics.⁶ Moreover, home phototherapy units are not yet widely utilized because of safety and efficacy concerns, lack of physician oversight, and difficulty obtaining insurance coverage.¹⁵ Acknowledgment and consideration of these geographical trends may persuasively allow policymakers, hospitals, and physicians to facilitate cost-effective phototherapy reimbursements that ensure continued access to quality and sustainable dermatologic care in the United States that tailor to regional needs.

In sum, this analysis reveals regional trends in Part B physician reimbursement for phototherapy procedures, with all US regions reporting a mean increase in phototherapy reimbursement after adjusting for utilization, albeit to varying degrees. Mean reimbursement for photochemotherapy by Goeckerman treatment or using petrolatum and UVB increased most among phototherapy procedures. Mean reimbursement for both actinotherapy and photochemotherapy using psoralen plus UVA decreased in all regions except the western United States.

Limitations include the restriction to Part B MPFS and the reliance on single-year (2020) physician utilization data to compute weighted changes in average reimbursement across a multiyear range, effectively restricting sweeping conclusions. Still, this study puts forth actionable insights for dermatologists and policymakers alike to appreciate and consider.

To the Editor:

Phototherapy regularly is utilized in the outpatient setting to address various skin pathologies, including atopic dermatitis, psoriasis, pruritus, vitiligo, and mycosis fungoides.^1,2 Phototherapy is broadly defined by the measured administration of nonionizing radiation within the UV range including wavelengths within the UVA (eg, psoralen sensitizer plus UVA-1) and UVB (eg, broadband UVB, narrowband UVB) spectrums.^1,3 Generally, the mechanism of action is derived from effects on inflammatory components of cutaneous disorders and the induction of apoptosis, both precipitating numerous downstream events.⁴

From 2015 to 2018, there were more than 1.3 million outpatient phototherapy visits in the United States, with the most common procedural indications being dermatitis not otherwise specified, atopic dermatitis, and pruritus.⁵ From 2000 to 2015, the quantity of phototherapy services billed to Medicare trended upwards by an average of 5% per year, increasing from 334,670 in the year 2000 to 692,093 in 2015.⁶ Therefore, an illustration of associated costs would be beneficial. Additionally, because total cost and physician reimbursement fluctuate from year to year, studies demonstrating overall trends can inform both US policymakers and physicians. There is a paucity of research on geographical trends for procedural reimbursements in dermatology for phototherapy. Understanding geographic trends of reimbursement could duly serve to optimize dermatologist practice patterns involving access to viable and quality care for patients seeking treatment as well as draw health policymakers’ attention to striking adjustments in physician fees. Therefore, in this study we aimed to illustrate the most recent regional payment trends in phototherapy procedures for Medicare B patients.

We queried the Centers for Medicare & Medicaid Services Medicare Physician Fee Schedule (MPFS) database (https://www.cms.gov/medicare/payment/fee-schedules/physician/lookup-tool) for the years 2010 to 2023 for Current Procedural Terminology (CPT) codes common to phototherapy procedures: actinotherapy (96900); photochemotherapy by Goeckerman treatment or using petrolatum and UVB (96910); photochemotherapy using psoralen plus UVA (96912); and photochemotherapy of severe dermatoses requiring a minimum of 4 hours of care under direct physician supervision (96913). Nonfacility prices for these procedures were analyzed. For 2010, due to midyear alterations to Medicare reimbursement (owed to bills HR 3962 and HR 4872), the mean price data of MPFS files 2010A and 2010B were used. All dollar values were converted to January 2023 US dollars using corresponding consumer price index inflation data. The Medicare Administrative Contractors were used to group state pricing information by region in accordance with established US Census Bureau subdivisions (https://www.census.gov/programs-surveys/economic-census/guidance-geographies/levels.html). Weighted percentage change in reimbursement rate was calculated using physician (MD or DO) utilization (procedure volume) data available in the 2020 Physician and Other Practitioners Public Use File (https://data.cms.gov/provider-summary-by-type-of-service/medicare-physician-other-practitioners/medicare-physician-other-practitioners-by-provider-and-service). All descriptive statistics and visualization were generated using R software (v4.2.2)(R Development Core Team).

Table 1 provides physician utilization data and the corresponding number of Part B beneficiaries for phototherapy procedures in 2020. There were 65,045 services of actinotherapy provided to a total of 6855 unique Part B beneficiaries, 173,979 services of photochemotherapy by Goeckerman treatment or using petrolatum and UVB provided to 13,122 unique Part B beneficiaries, 2524 services of photochemotherapy using psoralen plus UVA provided to a total of 357 unique Part B beneficiaries, and 37 services of photochemotherapy of severe dermatoses requiring a minimum of 4 hours of care under direct physician supervision provided to a total of 27 unique Part B beneficiaries.

On average (unweighted), phototherapy reimbursement rates in the North increased by 0.68% between 2010 and 2023 (Table 2). After weighting for 2020 physician utilization, the average change in reimbursement rate was +19.37%. During this time period, CPT code 96910 reported the greatest adjusted increase in reimbursement (+31.45%)($98.12 to $128.98; compound annual growth rate [CAGR], +0.0213), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−12.76%)($126.09 to $109.97; CAGR, −0.0105). For CPT code 96900, the reported adjusted decrease in reimbursement was −11.68% ($30.21 to $26.68; CAGR, −0.0095), and for CPT code 96913, the reported adjusted decrease in reimbursement was −4.27% ($174.03 to $166.60; CAGR, −0.0034).

On average (unweighted), phototherapy reimbursement rates in the Midwest increased by 8.40% between 2010 and 2023 (Table 3). After weighting for 2020 physician utilization, the average change in reimbursement rate was +28.53%. During this time period, CPT code 96910 reported the greatest adjusted change in reimbursement (+41.48%)($80.42 to $113.78; CAGR, +0.0270), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−6.14%)($103.28 to $97.03; CAGR, −0.0049). For CPT code 96900, the reported adjusted decrease in reimbursement was −4.73% ($24.69 to $23.52; CAGR, −0.0037), and for CPT code 96913, the reported adjusted increase in reimbursement was +2.99% ($142.72 to $146.99; CAGR, +0.0023).

On average (unweighted), phototherapy reimbursement rates in the South decreased by 2.62% between 2010 and 2023 (Table 4). After weighting for 2020 physician utilization, the average change in reimbursement rate was +15.41%. During this time period, CPT code 96910 reported the greatest adjusted change in reimbursement (+27.26%)($90.40 to $115.04 USD; CAGR, +0.0187), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−15.50%)($116.08 to $98.09; CAGR, −0.0129). For CPT code 96900, the reported adjusted decrease in reimbursement was −15.06% ($28.02 to $23.80; CAGR, −0.0125), and for CPT code 96913, the reported adjusted decrease in reimbursement was −7.19% ($160.11 to $148.61; CAGR, −0.0057).

On average (unweighted), phototherapy reimbursement rates in the West increased by 27.53% between 2010 and 2023 (Table 5). After weighting for 2020 physician utilization, the average change in reimbursement rate was +51.16%. Reimbursement for all analyzed procedures increased in the western United States. During this time period, CPT code 96910 reported the greatest adjusted increase in reimbursement (+66.56%)($80.84 to $134.65; CAGR, +0.0400), and CPT code 96912 reported the lowest adjusted increase in reimbursement (+10.64%)($103.88 to $114.93; CAGR, +0.0078). For CPT code 96900, the reported adjusted increase in reimbursement was 11.54% ($24.88 to $27.75; CAGR, +0.0084), and for CPT code 96913, the reported adjusted increase in reimbursement was 21.38% ($143.39 to $174.04; CAGR, +0.0150).

In this study evaluating geographical payment trends for phototherapy from 2010 to 2023, we demonstrated regional inconsistency in mean inflation-adjusted Medicare reimbursement rates. We found that all phototherapy procedures had increased reimbursement in the western United States, whereas all other regions reported cuts in reimbursement rates for at least half of the analyzed procedures. After adjusting for procedure utilization by physicians, weighted mean reimbursement for phototherapy increased in all US regions.

In a cross-sectional study that explored trends in the geographic distribution of dermatologists from 2012 to 2017, dermatologists in the northeastern and western United States were more likely to be located in higher-income zip codes, whereas dermatologists in the southern United States were more likely to be located in lower-income zip codes,⁷ suggesting that payment rate changes are not concordant with cost of living. Additionally, Lauck and colleagues⁸ observed that 75% of the top 20 most common procedures performed by dermatologists had decreased reimbursement (mean change, −10.8%) from 2011 to 2021. Other studies on Medicare reimbursement trends over the last 2 decades have reported major decreases within other specialties, suggesting that declining Medicare reimbursements are not unique to dermatology.^9,10 It is critical to monitor these developments, as the Centers for Medicare & Medicaid Services emphasized health care policy changes aimed at increasing reimbursements for evaluation and management services with compensatory payment cuts in billing for procedural services.¹¹

Mazmudar et al¹² previously reported a mean reimbursement decrease of −6.6% for laser/phototherapy procedures between 2007 and 2021, but these data did not include the heavily utilized Goeckerman treatment. Changes in reimbursement pose major ramifications for dermatologists—for practice size, scope, and longevity—as rates influence changes in commercial insurance reimbursements.¹³ Medicare plays a major role in the US health care system as the second largest expenditure¹⁴; indeed, between 2000 and 2015, Part B billing volume for phototherapy procedures increased 5% annually. However, phototherapy remains inaccessible in many locations due to unequal regional distribution of phototherapy clinics.⁶ Moreover, home phototherapy units are not yet widely utilized because of safety and efficacy concerns, lack of physician oversight, and difficulty obtaining insurance coverage.¹⁵ Acknowledgment and consideration of these geographical trends may persuasively allow policymakers, hospitals, and physicians to facilitate cost-effective phototherapy reimbursements that ensure continued access to quality and sustainable dermatologic care in the United States that tailor to regional needs.

In sum, this analysis reveals regional trends in Part B physician reimbursement for phototherapy procedures, with all US regions reporting a mean increase in phototherapy reimbursement after adjusting for utilization, albeit to varying degrees. Mean reimbursement for photochemotherapy by Goeckerman treatment or using petrolatum and UVB increased most among phototherapy procedures. Mean reimbursement for both actinotherapy and photochemotherapy using psoralen plus UVA decreased in all regions except the western United States.

Limitations include the restriction to Part B MPFS and the reliance on single-year (2020) physician utilization data to compute weighted changes in average reimbursement across a multiyear range, effectively restricting sweeping conclusions. Still, this study puts forth actionable insights for dermatologists and policymakers alike to appreciate and consider.

To the Editor:

Phototherapy regularly is utilized in the outpatient setting to address various skin pathologies, including atopic dermatitis, psoriasis, pruritus, vitiligo, and mycosis fungoides.^1,2 Phototherapy is broadly defined by the measured administration of nonionizing radiation within the UV range including wavelengths within the UVA (eg, psoralen sensitizer plus UVA-1) and UVB (eg, broadband UVB, narrowband UVB) spectrums.^1,3 Generally, the mechanism of action is derived from effects on inflammatory components of cutaneous disorders and the induction of apoptosis, both precipitating numerous downstream events.⁴

From 2015 to 2018, there were more than 1.3 million outpatient phototherapy visits in the United States, with the most common procedural indications being dermatitis not otherwise specified, atopic dermatitis, and pruritus.⁵ From 2000 to 2015, the quantity of phototherapy services billed to Medicare trended upwards by an average of 5% per year, increasing from 334,670 in the year 2000 to 692,093 in 2015.⁶ Therefore, an illustration of associated costs would be beneficial. Additionally, because total cost and physician reimbursement fluctuate from year to year, studies demonstrating overall trends can inform both US policymakers and physicians. There is a paucity of research on geographical trends for procedural reimbursements in dermatology for phototherapy. Understanding geographic trends of reimbursement could duly serve to optimize dermatologist practice patterns involving access to viable and quality care for patients seeking treatment as well as draw health policymakers’ attention to striking adjustments in physician fees. Therefore, in this study we aimed to illustrate the most recent regional payment trends in phototherapy procedures for Medicare B patients.

We queried the Centers for Medicare & Medicaid Services Medicare Physician Fee Schedule (MPFS) database (https://www.cms.gov/medicare/payment/fee-schedules/physician/lookup-tool) for the years 2010 to 2023 for Current Procedural Terminology (CPT) codes common to phototherapy procedures: actinotherapy (96900); photochemotherapy by Goeckerman treatment or using petrolatum and UVB (96910); photochemotherapy using psoralen plus UVA (96912); and photochemotherapy of severe dermatoses requiring a minimum of 4 hours of care under direct physician supervision (96913). Nonfacility prices for these procedures were analyzed. For 2010, due to midyear alterations to Medicare reimbursement (owed to bills HR 3962 and HR 4872), the mean price data of MPFS files 2010A and 2010B were used. All dollar values were converted to January 2023 US dollars using corresponding consumer price index inflation data. The Medicare Administrative Contractors were used to group state pricing information by region in accordance with established US Census Bureau subdivisions (https://www.census.gov/programs-surveys/economic-census/guidance-geographies/levels.html). Weighted percentage change in reimbursement rate was calculated using physician (MD or DO) utilization (procedure volume) data available in the 2020 Physician and Other Practitioners Public Use File (https://data.cms.gov/provider-summary-by-type-of-service/medicare-physician-other-practitioners/medicare-physician-other-practitioners-by-provider-and-service). All descriptive statistics and visualization were generated using R software (v4.2.2)(R Development Core Team).

Table 1 provides physician utilization data and the corresponding number of Part B beneficiaries for phototherapy procedures in 2020. There were 65,045 services of actinotherapy provided to a total of 6855 unique Part B beneficiaries, 173,979 services of photochemotherapy by Goeckerman treatment or using petrolatum and UVB provided to 13,122 unique Part B beneficiaries, 2524 services of photochemotherapy using psoralen plus UVA provided to a total of 357 unique Part B beneficiaries, and 37 services of photochemotherapy of severe dermatoses requiring a minimum of 4 hours of care under direct physician supervision provided to a total of 27 unique Part B beneficiaries.

On average (unweighted), phototherapy reimbursement rates in the North increased by 0.68% between 2010 and 2023 (Table 2). After weighting for 2020 physician utilization, the average change in reimbursement rate was +19.37%. During this time period, CPT code 96910 reported the greatest adjusted increase in reimbursement (+31.45%)($98.12 to $128.98; compound annual growth rate [CAGR], +0.0213), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−12.76%)($126.09 to $109.97; CAGR, −0.0105). For CPT code 96900, the reported adjusted decrease in reimbursement was −11.68% ($30.21 to $26.68; CAGR, −0.0095), and for CPT code 96913, the reported adjusted decrease in reimbursement was −4.27% ($174.03 to $166.60; CAGR, −0.0034).

On average (unweighted), phototherapy reimbursement rates in the Midwest increased by 8.40% between 2010 and 2023 (Table 3). After weighting for 2020 physician utilization, the average change in reimbursement rate was +28.53%. During this time period, CPT code 96910 reported the greatest adjusted change in reimbursement (+41.48%)($80.42 to $113.78; CAGR, +0.0270), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−6.14%)($103.28 to $97.03; CAGR, −0.0049). For CPT code 96900, the reported adjusted decrease in reimbursement was −4.73% ($24.69 to $23.52; CAGR, −0.0037), and for CPT code 96913, the reported adjusted increase in reimbursement was +2.99% ($142.72 to $146.99; CAGR, +0.0023).

On average (unweighted), phototherapy reimbursement rates in the South decreased by 2.62% between 2010 and 2023 (Table 4). After weighting for 2020 physician utilization, the average change in reimbursement rate was +15.41%. During this time period, CPT code 96910 reported the greatest adjusted change in reimbursement (+27.26%)($90.40 to $115.04 USD; CAGR, +0.0187), and CPT code 96912 reported the greatest adjusted decrease in reimbursement (−15.50%)($116.08 to $98.09; CAGR, −0.0129). For CPT code 96900, the reported adjusted decrease in reimbursement was −15.06% ($28.02 to $23.80; CAGR, −0.0125), and for CPT code 96913, the reported adjusted decrease in reimbursement was −7.19% ($160.11 to $148.61; CAGR, −0.0057).

On average (unweighted), phototherapy reimbursement rates in the West increased by 27.53% between 2010 and 2023 (Table 5). After weighting for 2020 physician utilization, the average change in reimbursement rate was +51.16%. Reimbursement for all analyzed procedures increased in the western United States. During this time period, CPT code 96910 reported the greatest adjusted increase in reimbursement (+66.56%)($80.84 to $134.65; CAGR, +0.0400), and CPT code 96912 reported the lowest adjusted increase in reimbursement (+10.64%)($103.88 to $114.93; CAGR, +0.0078). For CPT code 96900, the reported adjusted increase in reimbursement was 11.54% ($24.88 to $27.75; CAGR, +0.0084), and for CPT code 96913, the reported adjusted increase in reimbursement was 21.38% ($143.39 to $174.04; CAGR, +0.0150).

In this study evaluating geographical payment trends for phototherapy from 2010 to 2023, we demonstrated regional inconsistency in mean inflation-adjusted Medicare reimbursement rates. We found that all phototherapy procedures had increased reimbursement in the western United States, whereas all other regions reported cuts in reimbursement rates for at least half of the analyzed procedures. After adjusting for procedure utilization by physicians, weighted mean reimbursement for phototherapy increased in all US regions.

In a cross-sectional study that explored trends in the geographic distribution of dermatologists from 2012 to 2017, dermatologists in the northeastern and western United States were more likely to be located in higher-income zip codes, whereas dermatologists in the southern United States were more likely to be located in lower-income zip codes,⁷ suggesting that payment rate changes are not concordant with cost of living. Additionally, Lauck and colleagues⁸ observed that 75% of the top 20 most common procedures performed by dermatologists had decreased reimbursement (mean change, −10.8%) from 2011 to 2021. Other studies on Medicare reimbursement trends over the last 2 decades have reported major decreases within other specialties, suggesting that declining Medicare reimbursements are not unique to dermatology.^9,10 It is critical to monitor these developments, as the Centers for Medicare & Medicaid Services emphasized health care policy changes aimed at increasing reimbursements for evaluation and management services with compensatory payment cuts in billing for procedural services.¹¹

Mazmudar et al¹² previously reported a mean reimbursement decrease of −6.6% for laser/phototherapy procedures between 2007 and 2021, but these data did not include the heavily utilized Goeckerman treatment. Changes in reimbursement pose major ramifications for dermatologists—for practice size, scope, and longevity—as rates influence changes in commercial insurance reimbursements.¹³ Medicare plays a major role in the US health care system as the second largest expenditure¹⁴; indeed, between 2000 and 2015, Part B billing volume for phototherapy procedures increased 5% annually. However, phototherapy remains inaccessible in many locations due to unequal regional distribution of phototherapy clinics.⁶ Moreover, home phototherapy units are not yet widely utilized because of safety and efficacy concerns, lack of physician oversight, and difficulty obtaining insurance coverage.¹⁵ Acknowledgment and consideration of these geographical trends may persuasively allow policymakers, hospitals, and physicians to facilitate cost-effective phototherapy reimbursements that ensure continued access to quality and sustainable dermatologic care in the United States that tailor to regional needs.

In sum, this analysis reveals regional trends in Part B physician reimbursement for phototherapy procedures, with all US regions reporting a mean increase in phototherapy reimbursement after adjusting for utilization, albeit to varying degrees. Mean reimbursement for photochemotherapy by Goeckerman treatment or using petrolatum and UVB increased most among phototherapy procedures. Mean reimbursement for both actinotherapy and photochemotherapy using psoralen plus UVA decreased in all regions except the western United States.

Limitations include the restriction to Part B MPFS and the reliance on single-year (2020) physician utilization data to compute weighted changes in average reimbursement across a multiyear range, effectively restricting sweeping conclusions. Still, this study puts forth actionable insights for dermatologists and policymakers alike to appreciate and consider.

Hypertension is a major risk factor for heart disease, stroke, and kidney disease.The prevalence of hypertension increases with age, primarily due to age-related changes in arterial physiology.¹ For older adults, current guidelines regarding blood pressure (BP) treatment goals vary. The American Heart Association/American College of Cardiology 2017 clinical practice guidelines recommend a systolic BP (SBP) treatment goal of < 130 mm Hg for community-dwelling, ambulatory, noninstitutionalized adults aged ≥ 65 years; whereas the American College of Physicians/American Academy of Family Physicians recommend a goal of < 150 mm Hg for those aged ≥ 60 years without comorbidities and < 140 mm Hg for those with increased cardiovascular risk.^1-3 Regardless of the specific threshold, agreement that some degree of BP control even in those with advanced age improves outcomes.²

First-line therapy for uncomplicated hypertension includes thiazide diuretics, long-acting calcium channel blockers, and renin-angiotensin system inhibitors. When choosing between these options, it is recommended to engage in shared decision making and to consider the patient’s comorbidities. Among patients who are likely to require a second agent (eg, if initial BP is > 20/10 mm Hg above goal), it is recommended to begin both drugs at the same time, preferably benazepril plus amlodipine due to the reduction in cardiovascular events reported in the ACCOMPLISH trial.⁴ If BP remains elevated despite 2 agents at moderate to maximum doses, it is important to investigate for secondary hypertension causes and to explore medication adherence as possible etiologies of treatment failure. Older adults are often at higher risk of adverse drug events due to age-related changes in pharmacodynamics. Despite this, there are no guidelines for choosing between different classes of antihypertensives in this population. We present a case of thiazide-induced hyponatremia in an older adult and review the risks of thiazide use in this population.

Case Presentation

A man aged > 90 years was admitted to the hospital after a syncopal episode. His history was significant for hypertension, hyperlipidemia, and vitamin D deficiency. At the time, his home medications were amlodipine 5 mg daily, atorvastatin 40 mg daily, ergocalciferol 50,000 IU weekly, and polyethylene glycol 17 g daily as needed. His syncope workup was unremarkable and included negative orthostatic vital signs, normal serial troponins, an electrocardiogram without ischemic changes, normal serum creatinine, sodium, and glucose, and a head computed tomography without any acute abnormality. Throughout the patient’s hospital stay, he had multiple elevated SBP readings, including many > 200 mm Hg. On discharge, in addition to continuing his home medications, he was started on valsartan 20 mg daily and enrolled in a remote BP monitoring program.

Three weeks later, the patient was seen by their primary care practitioner for follow-up. He reported adherence to his antihypertensive regimen. However, his remote BP monitoring revealed persistently elevated BPs, with an average of 179/79 mm Hg, a high of 205/85 mm Hg, and a low of 150/67 mm Hg over the previous 7 days. Laboratory tests obtained at the visit were notable for serum sodium of 138 mmol/L and potassium of 4.1 mmol/L. His weight was 87 kg. Given persistently elevated BP readings, in addition to continuing his amlodipine 5 mg daily and valsartan 20 mg daily, he was started on hydrochlorothiazide 25 mg daily, with plans to repeat a basic metabolic panel in 2 weeks.

Two weeks later, he fell after getting out of his bed. On examination, he was noted to have dry mucous membranes, and although no formal delirium screening was performed, he was able to repeat the months of the year backward. Vital signs were notable for positive postural hypertension, and his laboratory tests revealed a normal serum creatinine, serum sodium of 117 mmol/L (reference range, 135-145 mmol/L), serum potassium of 3.2 mmol/L (reference range, 3.0-5.0 mmol/L), a low serum osmolality, and urine sodium of 35 mmol/L most consistent with hypovolemic hypoosmotic hyponatremia secondary to thiazide initiation. The patient’s hydrochlorothiazide was discontinued, and he was admitted to the hospital for close monitoring. His sodium levels gradually normalized over the next 2 weeks without any other intervention.

Discussion

Although thiazide diuretics are recommended as first-line therapy for uncomplicated hypertension, they are known to cause electrolyte abnormalities, including hypomagnesemia, hypokalemia, and hyponatremia.⁴ These metabolic derangements are more likely to occur in older adults. One study of adults aged ≥ 65 years found that at 9 months of follow-up, 14.3% of new thiazide users had developed a thiazide-related metabolic adverse event (hyponatremia < 135 mmol/L, hypokalemia < 3.5 mmol/L, and decrease in estimated glomerular filtration rate by > 25%) compared with 6.0% of nonusers (P < .001; number needed to harm [NNH] = 12).⁵ In addition, 3.8% of new thiazide users had an emergency department visit or were hospitalized for complications related to thiazides compared with only 2.0% of nonusers (P = .02; NNH = 56).⁵ Independent risk factors for thiazide-induced hyponatremia include high-comorbidity burden, low body weight, low-normal or unmeasured serum sodium, low potassium, and aged > 70 years.^5-7 Each 10-year increment in age is associated with a 2-fold increase in risk, suggesting that older adults are at a much higher risk for hyponatremia than their younger peers.⁶

Despite their designation as a first-line option for uncomplicated hypertension, thiazide diuretics may cause more harm than good in some older adults, especially those with additional risk factors for thiazide-induced hyponatremia. In this population, these adverse effects should be discussed before starting thiazides for the treatment of hypertension. If thiazides are initiated, they should be started at the lowest possible dose, and plans made to monitor bloodwork within 1 to 2 weeks of initiation or dose change and periodically thereafter while the patient remains on the therapy.

Medication Management in Older Adults

Due to the risks of medication use in older adults, the phrase “start low, go slow” is commonly used in geriatric medicine to describe the optimal method for initiation and up-titration of new medication with the hope of mitigating adverse drug events. In our case, we started valsartan at 20 mg daily—one-fourth the recommended initial dose. Although this strategy is reasonable to “start low,” we were not surprised to find that the patient’s BP did not markedly improve on such a low dose. The team could have increased the valsartan dose to a therapeutically efficacious dose before choosing to add another hypertensive agent. In alignment with geriatric prescribing principles, starting at the lowest possible dose of hydrochlorothiazide is recommended.⁵ However, the clinician started hydrochlorothiazide at 25 mg daily, potentially increasing this patient’s risk of electrolyte abnormalities and eventual fall.

Managing hypertension also invites a discussion of polypharmacy and medication adherence. Older adults are at risk of polypharmacy, defined as the prescription of 5 or more medications.⁸ Polypharmacy is associated with increased hospitalizations, higher costs of care for individuals and health care systems, increased risks of adverse drug events, medication nonadherence, and lower quality of life for patients.⁹ In some situations, the risks of polypharmacy may outweigh the benefits of using multiple antihypertensives with different mechanisms of action if patients can reach their BP goal on the maximum dose of a single agent. For patients taking multiple antihypertensives, it is important to routinely monitor BP and assess whether deprescribing is indicated. Cognitive impairment and decreased social support may affect medication adherence for older adults.⁶ Clinicians should be aware of strategies, such as medication reminders and pillboxes, to increase antihypertensive medication adherence. Polypills that contain 2 antihypertensives can be another tool used to manage older adults to increase adherence and decrease health care costs.¹⁰

A current strategy that encompasses discussing many, if not all, of these noted elements is the Institute for Healthcare Improvement’s Age-Friendly Health System. This framework uses evidence-based tools to provide care for older adults across all clinical settings and highlights the 4Ms: what matters, medication, mentation, and mobility.¹¹ Medication considers whether a medication is necessary, whether its use has benefits that outweigh the risks, and how it interacts with what matters, mentation, and mobility. In particular, what matters plays an important role in hypertension management in older adults given the recommended target BP differs, depending on which specialty organization guideline is followed. By better understanding what matters to patients, including their goals and priorities, clinicians can engage patients in shared decision making and provide individualized recommendations based on geriatric principles (eg, start low, go slow, principles of medication adherence) and patient comorbidities (eg, medical history and risk factors for hyponatremia) to help patients make a more informed choice about their antihypertensive treatment regimen (Figure).

Conclusions

This case illustrates the need for a specialized approach to hypertension management in older adults and the risks of thiazide diuretics in this population. Clinicians should consider BP goals, patient-specific factors, and principles of medication management in older adults. If initiating thiazide therapy, discuss the risks associated with use, start at the lowest possible dose, and monitor bloodwork within 1 to 2 weeks of initiation/dose change and periodically thereafter while the patient remains on the therapy to decrease the risk of adverse events. Finally, the Institute for Healthcare Improvement’s Age-Friendly Health System framework can be a useful when considering the addition of a new medication in an older adult’s treatment plan.

Acknowledgments

This material is the result of work supported with resources and the use of facilities at the New England Geriatrics Research, Education, and Clinical Center, Veterans Affairs Boston Healthcare System, and the Cincinnati VeteransAffairs Medical Center.

Hypertension is a major risk factor for heart disease, stroke, and kidney disease.The prevalence of hypertension increases with age, primarily due to age-related changes in arterial physiology.¹ For older adults, current guidelines regarding blood pressure (BP) treatment goals vary. The American Heart Association/American College of Cardiology 2017 clinical practice guidelines recommend a systolic BP (SBP) treatment goal of < 130 mm Hg for community-dwelling, ambulatory, noninstitutionalized adults aged ≥ 65 years; whereas the American College of Physicians/American Academy of Family Physicians recommend a goal of < 150 mm Hg for those aged ≥ 60 years without comorbidities and < 140 mm Hg for those with increased cardiovascular risk.^1-3 Regardless of the specific threshold, agreement that some degree of BP control even in those with advanced age improves outcomes.²

First-line therapy for uncomplicated hypertension includes thiazide diuretics, long-acting calcium channel blockers, and renin-angiotensin system inhibitors. When choosing between these options, it is recommended to engage in shared decision making and to consider the patient’s comorbidities. Among patients who are likely to require a second agent (eg, if initial BP is > 20/10 mm Hg above goal), it is recommended to begin both drugs at the same time, preferably benazepril plus amlodipine due to the reduction in cardiovascular events reported in the ACCOMPLISH trial.⁴ If BP remains elevated despite 2 agents at moderate to maximum doses, it is important to investigate for secondary hypertension causes and to explore medication adherence as possible etiologies of treatment failure. Older adults are often at higher risk of adverse drug events due to age-related changes in pharmacodynamics. Despite this, there are no guidelines for choosing between different classes of antihypertensives in this population. We present a case of thiazide-induced hyponatremia in an older adult and review the risks of thiazide use in this population.

Case Presentation

A man aged > 90 years was admitted to the hospital after a syncopal episode. His history was significant for hypertension, hyperlipidemia, and vitamin D deficiency. At the time, his home medications were amlodipine 5 mg daily, atorvastatin 40 mg daily, ergocalciferol 50,000 IU weekly, and polyethylene glycol 17 g daily as needed. His syncope workup was unremarkable and included negative orthostatic vital signs, normal serial troponins, an electrocardiogram without ischemic changes, normal serum creatinine, sodium, and glucose, and a head computed tomography without any acute abnormality. Throughout the patient’s hospital stay, he had multiple elevated SBP readings, including many > 200 mm Hg. On discharge, in addition to continuing his home medications, he was started on valsartan 20 mg daily and enrolled in a remote BP monitoring program.

Three weeks later, the patient was seen by their primary care practitioner for follow-up. He reported adherence to his antihypertensive regimen. However, his remote BP monitoring revealed persistently elevated BPs, with an average of 179/79 mm Hg, a high of 205/85 mm Hg, and a low of 150/67 mm Hg over the previous 7 days. Laboratory tests obtained at the visit were notable for serum sodium of 138 mmol/L and potassium of 4.1 mmol/L. His weight was 87 kg. Given persistently elevated BP readings, in addition to continuing his amlodipine 5 mg daily and valsartan 20 mg daily, he was started on hydrochlorothiazide 25 mg daily, with plans to repeat a basic metabolic panel in 2 weeks.

Two weeks later, he fell after getting out of his bed. On examination, he was noted to have dry mucous membranes, and although no formal delirium screening was performed, he was able to repeat the months of the year backward. Vital signs were notable for positive postural hypertension, and his laboratory tests revealed a normal serum creatinine, serum sodium of 117 mmol/L (reference range, 135-145 mmol/L), serum potassium of 3.2 mmol/L (reference range, 3.0-5.0 mmol/L), a low serum osmolality, and urine sodium of 35 mmol/L most consistent with hypovolemic hypoosmotic hyponatremia secondary to thiazide initiation. The patient’s hydrochlorothiazide was discontinued, and he was admitted to the hospital for close monitoring. His sodium levels gradually normalized over the next 2 weeks without any other intervention.

Discussion

Although thiazide diuretics are recommended as first-line therapy for uncomplicated hypertension, they are known to cause electrolyte abnormalities, including hypomagnesemia, hypokalemia, and hyponatremia.⁴ These metabolic derangements are more likely to occur in older adults. One study of adults aged ≥ 65 years found that at 9 months of follow-up, 14.3% of new thiazide users had developed a thiazide-related metabolic adverse event (hyponatremia < 135 mmol/L, hypokalemia < 3.5 mmol/L, and decrease in estimated glomerular filtration rate by > 25%) compared with 6.0% of nonusers (P < .001; number needed to harm [NNH] = 12).⁵ In addition, 3.8% of new thiazide users had an emergency department visit or were hospitalized for complications related to thiazides compared with only 2.0% of nonusers (P = .02; NNH = 56).⁵ Independent risk factors for thiazide-induced hyponatremia include high-comorbidity burden, low body weight, low-normal or unmeasured serum sodium, low potassium, and aged > 70 years.^5-7 Each 10-year increment in age is associated with a 2-fold increase in risk, suggesting that older adults are at a much higher risk for hyponatremia than their younger peers.⁶

Despite their designation as a first-line option for uncomplicated hypertension, thiazide diuretics may cause more harm than good in some older adults, especially those with additional risk factors for thiazide-induced hyponatremia. In this population, these adverse effects should be discussed before starting thiazides for the treatment of hypertension. If thiazides are initiated, they should be started at the lowest possible dose, and plans made to monitor bloodwork within 1 to 2 weeks of initiation or dose change and periodically thereafter while the patient remains on the therapy.

Medication Management in Older Adults

Due to the risks of medication use in older adults, the phrase “start low, go slow” is commonly used in geriatric medicine to describe the optimal method for initiation and up-titration of new medication with the hope of mitigating adverse drug events. In our case, we started valsartan at 20 mg daily—one-fourth the recommended initial dose. Although this strategy is reasonable to “start low,” we were not surprised to find that the patient’s BP did not markedly improve on such a low dose. The team could have increased the valsartan dose to a therapeutically efficacious dose before choosing to add another hypertensive agent. In alignment with geriatric prescribing principles, starting at the lowest possible dose of hydrochlorothiazide is recommended.⁵ However, the clinician started hydrochlorothiazide at 25 mg daily, potentially increasing this patient’s risk of electrolyte abnormalities and eventual fall.

Managing hypertension also invites a discussion of polypharmacy and medication adherence. Older adults are at risk of polypharmacy, defined as the prescription of 5 or more medications.⁸ Polypharmacy is associated with increased hospitalizations, higher costs of care for individuals and health care systems, increased risks of adverse drug events, medication nonadherence, and lower quality of life for patients.⁹ In some situations, the risks of polypharmacy may outweigh the benefits of using multiple antihypertensives with different mechanisms of action if patients can reach their BP goal on the maximum dose of a single agent. For patients taking multiple antihypertensives, it is important to routinely monitor BP and assess whether deprescribing is indicated. Cognitive impairment and decreased social support may affect medication adherence for older adults.⁶ Clinicians should be aware of strategies, such as medication reminders and pillboxes, to increase antihypertensive medication adherence. Polypills that contain 2 antihypertensives can be another tool used to manage older adults to increase adherence and decrease health care costs.¹⁰

A current strategy that encompasses discussing many, if not all, of these noted elements is the Institute for Healthcare Improvement’s Age-Friendly Health System. This framework uses evidence-based tools to provide care for older adults across all clinical settings and highlights the 4Ms: what matters, medication, mentation, and mobility.¹¹ Medication considers whether a medication is necessary, whether its use has benefits that outweigh the risks, and how it interacts with what matters, mentation, and mobility. In particular, what matters plays an important role in hypertension management in older adults given the recommended target BP differs, depending on which specialty organization guideline is followed. By better understanding what matters to patients, including their goals and priorities, clinicians can engage patients in shared decision making and provide individualized recommendations based on geriatric principles (eg, start low, go slow, principles of medication adherence) and patient comorbidities (eg, medical history and risk factors for hyponatremia) to help patients make a more informed choice about their antihypertensive treatment regimen (Figure).

Conclusions

This case illustrates the need for a specialized approach to hypertension management in older adults and the risks of thiazide diuretics in this population. Clinicians should consider BP goals, patient-specific factors, and principles of medication management in older adults. If initiating thiazide therapy, discuss the risks associated with use, start at the lowest possible dose, and monitor bloodwork within 1 to 2 weeks of initiation/dose change and periodically thereafter while the patient remains on the therapy to decrease the risk of adverse events. Finally, the Institute for Healthcare Improvement’s Age-Friendly Health System framework can be a useful when considering the addition of a new medication in an older adult’s treatment plan.

Acknowledgments

This material is the result of work supported with resources and the use of facilities at the New England Geriatrics Research, Education, and Clinical Center, Veterans Affairs Boston Healthcare System, and the Cincinnati VeteransAffairs Medical Center.

Hypertension is a major risk factor for heart disease, stroke, and kidney disease.The prevalence of hypertension increases with age, primarily due to age-related changes in arterial physiology.¹ For older adults, current guidelines regarding blood pressure (BP) treatment goals vary. The American Heart Association/American College of Cardiology 2017 clinical practice guidelines recommend a systolic BP (SBP) treatment goal of < 130 mm Hg for community-dwelling, ambulatory, noninstitutionalized adults aged ≥ 65 years; whereas the American College of Physicians/American Academy of Family Physicians recommend a goal of < 150 mm Hg for those aged ≥ 60 years without comorbidities and < 140 mm Hg for those with increased cardiovascular risk.^1-3 Regardless of the specific threshold, agreement that some degree of BP control even in those with advanced age improves outcomes.²

First-line therapy for uncomplicated hypertension includes thiazide diuretics, long-acting calcium channel blockers, and renin-angiotensin system inhibitors. When choosing between these options, it is recommended to engage in shared decision making and to consider the patient’s comorbidities. Among patients who are likely to require a second agent (eg, if initial BP is > 20/10 mm Hg above goal), it is recommended to begin both drugs at the same time, preferably benazepril plus amlodipine due to the reduction in cardiovascular events reported in the ACCOMPLISH trial.⁴ If BP remains elevated despite 2 agents at moderate to maximum doses, it is important to investigate for secondary hypertension causes and to explore medication adherence as possible etiologies of treatment failure. Older adults are often at higher risk of adverse drug events due to age-related changes in pharmacodynamics. Despite this, there are no guidelines for choosing between different classes of antihypertensives in this population. We present a case of thiazide-induced hyponatremia in an older adult and review the risks of thiazide use in this population.

Case Presentation

A man aged > 90 years was admitted to the hospital after a syncopal episode. His history was significant for hypertension, hyperlipidemia, and vitamin D deficiency. At the time, his home medications were amlodipine 5 mg daily, atorvastatin 40 mg daily, ergocalciferol 50,000 IU weekly, and polyethylene glycol 17 g daily as needed. His syncope workup was unremarkable and included negative orthostatic vital signs, normal serial troponins, an electrocardiogram without ischemic changes, normal serum creatinine, sodium, and glucose, and a head computed tomography without any acute abnormality. Throughout the patient’s hospital stay, he had multiple elevated SBP readings, including many > 200 mm Hg. On discharge, in addition to continuing his home medications, he was started on valsartan 20 mg daily and enrolled in a remote BP monitoring program.

Three weeks later, the patient was seen by their primary care practitioner for follow-up. He reported adherence to his antihypertensive regimen. However, his remote BP monitoring revealed persistently elevated BPs, with an average of 179/79 mm Hg, a high of 205/85 mm Hg, and a low of 150/67 mm Hg over the previous 7 days. Laboratory tests obtained at the visit were notable for serum sodium of 138 mmol/L and potassium of 4.1 mmol/L. His weight was 87 kg. Given persistently elevated BP readings, in addition to continuing his amlodipine 5 mg daily and valsartan 20 mg daily, he was started on hydrochlorothiazide 25 mg daily, with plans to repeat a basic metabolic panel in 2 weeks.

Two weeks later, he fell after getting out of his bed. On examination, he was noted to have dry mucous membranes, and although no formal delirium screening was performed, he was able to repeat the months of the year backward. Vital signs were notable for positive postural hypertension, and his laboratory tests revealed a normal serum creatinine, serum sodium of 117 mmol/L (reference range, 135-145 mmol/L), serum potassium of 3.2 mmol/L (reference range, 3.0-5.0 mmol/L), a low serum osmolality, and urine sodium of 35 mmol/L most consistent with hypovolemic hypoosmotic hyponatremia secondary to thiazide initiation. The patient’s hydrochlorothiazide was discontinued, and he was admitted to the hospital for close monitoring. His sodium levels gradually normalized over the next 2 weeks without any other intervention.

Discussion

Although thiazide diuretics are recommended as first-line therapy for uncomplicated hypertension, they are known to cause electrolyte abnormalities, including hypomagnesemia, hypokalemia, and hyponatremia.⁴ These metabolic derangements are more likely to occur in older adults. One study of adults aged ≥ 65 years found that at 9 months of follow-up, 14.3% of new thiazide users had developed a thiazide-related metabolic adverse event (hyponatremia < 135 mmol/L, hypokalemia < 3.5 mmol/L, and decrease in estimated glomerular filtration rate by > 25%) compared with 6.0% of nonusers (P < .001; number needed to harm [NNH] = 12).⁵ In addition, 3.8% of new thiazide users had an emergency department visit or were hospitalized for complications related to thiazides compared with only 2.0% of nonusers (P = .02; NNH = 56).⁵ Independent risk factors for thiazide-induced hyponatremia include high-comorbidity burden, low body weight, low-normal or unmeasured serum sodium, low potassium, and aged > 70 years.^5-7 Each 10-year increment in age is associated with a 2-fold increase in risk, suggesting that older adults are at a much higher risk for hyponatremia than their younger peers.⁶

Despite their designation as a first-line option for uncomplicated hypertension, thiazide diuretics may cause more harm than good in some older adults, especially those with additional risk factors for thiazide-induced hyponatremia. In this population, these adverse effects should be discussed before starting thiazides for the treatment of hypertension. If thiazides are initiated, they should be started at the lowest possible dose, and plans made to monitor bloodwork within 1 to 2 weeks of initiation or dose change and periodically thereafter while the patient remains on the therapy.

Medication Management in Older Adults

Due to the risks of medication use in older adults, the phrase “start low, go slow” is commonly used in geriatric medicine to describe the optimal method for initiation and up-titration of new medication with the hope of mitigating adverse drug events. In our case, we started valsartan at 20 mg daily—one-fourth the recommended initial dose. Although this strategy is reasonable to “start low,” we were not surprised to find that the patient’s BP did not markedly improve on such a low dose. The team could have increased the valsartan dose to a therapeutically efficacious dose before choosing to add another hypertensive agent. In alignment with geriatric prescribing principles, starting at the lowest possible dose of hydrochlorothiazide is recommended.⁵ However, the clinician started hydrochlorothiazide at 25 mg daily, potentially increasing this patient’s risk of electrolyte abnormalities and eventual fall.

Managing hypertension also invites a discussion of polypharmacy and medication adherence. Older adults are at risk of polypharmacy, defined as the prescription of 5 or more medications.⁸ Polypharmacy is associated with increased hospitalizations, higher costs of care for individuals and health care systems, increased risks of adverse drug events, medication nonadherence, and lower quality of life for patients.⁹ In some situations, the risks of polypharmacy may outweigh the benefits of using multiple antihypertensives with different mechanisms of action if patients can reach their BP goal on the maximum dose of a single agent. For patients taking multiple antihypertensives, it is important to routinely monitor BP and assess whether deprescribing is indicated. Cognitive impairment and decreased social support may affect medication adherence for older adults.⁶ Clinicians should be aware of strategies, such as medication reminders and pillboxes, to increase antihypertensive medication adherence. Polypills that contain 2 antihypertensives can be another tool used to manage older adults to increase adherence and decrease health care costs.¹⁰

A current strategy that encompasses discussing many, if not all, of these noted elements is the Institute for Healthcare Improvement’s Age-Friendly Health System. This framework uses evidence-based tools to provide care for older adults across all clinical settings and highlights the 4Ms: what matters, medication, mentation, and mobility.¹¹ Medication considers whether a medication is necessary, whether its use has benefits that outweigh the risks, and how it interacts with what matters, mentation, and mobility. In particular, what matters plays an important role in hypertension management in older adults given the recommended target BP differs, depending on which specialty organization guideline is followed. By better understanding what matters to patients, including their goals and priorities, clinicians can engage patients in shared decision making and provide individualized recommendations based on geriatric principles (eg, start low, go slow, principles of medication adherence) and patient comorbidities (eg, medical history and risk factors for hyponatremia) to help patients make a more informed choice about their antihypertensive treatment regimen (Figure).

Conclusions

This case illustrates the need for a specialized approach to hypertension management in older adults and the risks of thiazide diuretics in this population. Clinicians should consider BP goals, patient-specific factors, and principles of medication management in older adults. If initiating thiazide therapy, discuss the risks associated with use, start at the lowest possible dose, and monitor bloodwork within 1 to 2 weeks of initiation/dose change and periodically thereafter while the patient remains on the therapy to decrease the risk of adverse events. Finally, the Institute for Healthcare Improvement’s Age-Friendly Health System framework can be a useful when considering the addition of a new medication in an older adult’s treatment plan.

Acknowledgments

This material is the result of work supported with resources and the use of facilities at the New England Geriatrics Research, Education, and Clinical Center, Veterans Affairs Boston Healthcare System, and the Cincinnati VeteransAffairs Medical Center.

Hypoglycemia and weight gain are well-known adverse effects that can result from insulin and sulfonylureas in patients with type 2 diabetes mellitus (T2DM).^1,2 Insulin and sulfonylurea medications can cause additional weight gain in patients who are overweight or obese, which can increase the burden of diabetes therapy with added medications, raise the risk of hypoglycemia complications, and raise atherosclerotic cardiovascular disease risk factors.³ Although increasing the insulin or sulfonylurea dose is an option health care practitioners or pharmacists have, this approach can increase the risk of hypoglycemia, especially in older adults, such as the veteran population, which could lead to complications, such as falls.²

Previous studies focusing on hypoglycemic events in patients with T2DM showed that glucagon-like peptide-1 (GLP-1) agonist monotherapy has a low incidence of a hypoglycemic events. However, when a GLP-1 agonist is combined with insulin or sulfonylureas, patients have an increased chance of a hypoglycemic event.^3-8 According to the prescribing information for semaglutide, 1.6% to 3.8% of patients on a GLP-1 agonist monotherapy reported a documented symptomatic hypoglycemic event (blood glucose ≤ 70 mg/dL), based on semaglutide dosing. ⁹ Patients on combination therapy of a GLP-1 agonist and basal insulin and a GLP-1 agonist and a sulfonylurea reported a documented symptomatic hypoglycemic event ranging from 16.7% to 29.8% and 17.3% to 24.4%, respectively.⁹ The incidences of hypoglycemia thus dramatically increase with combination therapy of a GLP-1 agonist plus insulin or a sulfonylurea.

When adding a GLP-1 agonist to insulin or a sulfonylurea, clinicians must be mindful of the increased risk of hypoglycemia. Per the warnings and precautions in the prescribing information of GLP-1 agonists, concomitant use with insulin or a sulfonylurea may increase the risk of hypoglycemia, and reducing the dose of insulin or a sulfonylurea may be necessary.^9-11 According to the American College of Cardiology guidelines, when starting a GLP-1 agonist, the insulin dose should be decreased by about 20% in patients with a well-controlled hemoglobin A_1c (HbA_1c).¹²

This study aimed to determine the percentage of patients who required dose reductions or discontinuations of insulin and sulfonylureas with the addition of a GLP-1 agonist. Understanding necessary dose reductions or discontinuations of these concomitant diabetes agents can assist pharmacists in preventing hypoglycemia and minimizing weight gain.

Methods

This clinical review was a single-center, retrospective chart review of patients prescribed a GLP-1 agonist while on insulin or a sulfonylurea between January 1, 2019, and September 30, 2022, at the Wilkes-Barre Veterans Affairs Medical Center (WBVAMC) in Pennsylvania and managed in a pharmacist-led patient aligned care team (PACT) clinic. It was determined by the US Department of Veterans Affairs Office of Research and Development that an institutional review board or other review committee approval was not needed for this nonresearch Veterans Health Administration quality assurance and improvement project. Patients aged ≥ 18 years were included in this study. Patients were excluded if they were not on insulin or a sulfonylurea when starting a GLP-1 agonist, started a GLP-1 agonist outside of the retrospective chart review dates, or were prescribed a GLP-1 agonist by anyone other than a pharmacist in their PACT clinic. This included if a GLP-1 agonist was prescribed by a primary care physician, endocrinologist, or someone outside the VA system.

The primary study outcomes were to determine the percentage of patients with a dose reduction of insulin or sulfonylurea and discontinuation of insulin or a sulfonylurea at intervals of 0 (baseline), 3, 6, and 12 months. Secondary outcomes included changes in HbA_1c and body weight measured at the same intervals of 0 (baseline), 3, 6, and 12 months.
Data were collected using the VA Computerized Patient Record System (CPRS) and stored in a locked spreadsheet. Descriptive statistics were used to analyze the data. Patient data included the number of patients on insulin or a sulfonylurea when initiating a GLP-1 agonist, the percentage of patients started on a certain GLP-1 agonist (dulaglutide, liraglutide, exenatide, and semaglutide), and the percentage of patients with a baseline HbA_1c of < 8%, 8% to 10%, and > 10%. The GLP-1 agonist formulary was adjusted during the time of this retrospective chart review. Patients who were not on semaglutide were switched over if they were on another GLP-1 agonist as semaglutide became the preferred GLP-1 agonist.

Patients were considered to have a dose reduction or discontinuation of insulin or a sulfonylurea if the dose or medication they were on decreased or was discontinued permanently within 12 months of starting a GLP-1 agonist. For example, if a patient who was administering 10 units of insulin daily was decreased to 8 but later increased back to 10, this was not counted as a dose reduction. If a patient discontinued insulin or a sulfonylurea and then restarted it within 12 months of initiating a GLP-1 agonist, this was not counted as a discontinuation.

This retrospective review included 136 patients; 96 patients taking insulin and 54 taking a sulfonylurea when they started a GLP-1 agonist. Fourteen patients were on both. Criteria for use, which are clinical criteria to determine if a patient is eligible for the use of a given medication, are used within the VA. The inclusion criteria for a patient initiating a GLP-1 agonist is that the patient must have atherosclerotic cardiovascular disease or chronic kidney disease with the patient receiving metformin (unless unable to use metformin) and empagliflozin (unless unable to use empagliflozin).

The baseline mean age and weight for the patient population in this retrospective chart review was 70.7 years and 238.2 lb, respectively. Ninety-six patients (70.6%) were started on semaglutide, 27 (19.9%) on dulaglutide, 12 (8.8%) on liraglutide, and 1 (0.7%) on exenatide. The mean HbA_1c when patients initiated a GLP-1 agonist was 8.6%. When starting a GLP-1 agonist, 34 patients (25.0%) had an HbA_1c < 8%, 89 (65.4%) had an HbA_1c between 8% to 10%, and 13 (9.6%) had an HbA_1c > 10% (Table).

For the primary results, 25 patients (26.0%) had a dose reduction of insulin when they started a GLP-1 agonist, and 55 patients (57.3%) had at least 1 insulin dose reduction within the year follow-up. Seven patients (13.0%) had a dose reduction of a sulfonylurea when they started a GLP-1 agonist, and 16 patients (29.6%) had at least 1 dose reduction of a sulfonylurea within the year follow-up. Six patients (6.3%) discontinued insulin use when they initially started a GLP-1 agonist, and 14 patients (14.6%) discontinued insulin use within the year follow-up. Eleven patients (20.4%) discontinued sulfonylurea use when they initially started a GLP-1 agonist, and 21 patients (38.9%) discontinued sulfonylurea use within the year follow-up (Figure).

Fourteen patients were on both insulin and a sulfonylurea. Two patients (14.3%) had a dose reduction of insulin when they started a GLP-1 agonist, and 5 (35.7%) had ≥ 1 insulin dose reduction within the year follow-up. Three patients (21.4%) had a dose reduction of a sulfonylurea when they started a GLP-1 agonist, and 6 (42.9%) had ≥ 1 dose reduction of a sulfonylurea within the year follow-up. Seven patients (50.0%) discontinued sulfonylurea and 3 (21.4%) discontinued insulin at any time throughout the year. The majority of the discontinuations were at the initial start of GLP-1 agonist therapy.

The mean HbA_1c for patients on GLP-1 agonist was 8.6% at baseline, 8.0% at 0 to 3 months, 7.6% at 3 to 6 months, and 7.5% at 12 months. Patients experienced a mean HbA_1c reduction of 1.1%. The mean weight when a GLP-1 agonist was started was 238.2 lb, 236.0 lb at 0 to 3 months, 223.8 lb at 3 to 6 months, and 224.3 lb after 12 months. Study participants lost a mean weight of 13.9 lb while on a GLP-1 agonist.

Discussion

While this study did not examine why there were dose reductions or discontinuations, we can hypothesize that insulin or sulfonylureas were reduced or discontinued due to a myriad of reasons, such as prophylactic dosing per guidelines, patients having a hypoglycemic event, or pharmacists anticipating potential low blood glucose trends. Also, there could have been numerous reasons GLP-1 agonists were started in patients on insulin or a sulfonylurea, such as HbA_1c not being within goal range, cardiovascular benefits (reduce risk of stroke, heart attack, and death), weight loss, and renal protection, such as preventing albuminuria.^13,14

This retrospective chart review found a large proportion of patients had a dose reduction of insulin (57.3%) or sulfonylurea (29.6%). The percentage of patients with a dose reduction was potentially underestimated as patients were not counted if they discontinued insulin or sulfonylurea. Concomitant use of GLP-1 agonists with insulin or a sulfonylurea may increase the risk of hypoglycemia and reducing the dose of insulin or a sulfonylurea may be necessary.^9-11 The dose reductions in this study show that pharmacists within pharmacy-led PACT clinics monitor for or attempt to prevent hypoglycemia, which aligns with the prescribing information of GLP-1 agonists. While increasing the insulin or sulfonylurea dose is an option for patients, this approach can increase the risk of hypoglycemia, especially in an older population, like this one with a mean age > 70 years. The large proportions of patients with dose reductions or insulin and sulfonylurea discontinuations suggest that pharmacists may need to take a more cautious approach when initiating a GLP-1 agonist to prevent adverse health outcomes related to low blood sugar for older adults, such as falls and fractures.

Insulin was discontinued in 20.4% of patients and sulfonylurea was discontinued in 38.9% of patients within 12 months after starting a GLP-1 agonist. When a patient was on both insulin and a sulfonylurea, the percentage of patients who discontinued insulin (21.4%) or a sulfonylurea (50.0%) was higher compared with patients just on insulin (14.6%) or a sulfonylurea (38.9%) alone. Patients on both insulin and a sulfonylurea may need closer monitoring due to a higher incidence of discontinuations when these diabetes agents are administered in combination.

Limitations

Several limitations of this study should be considered when evaluating the results. This review was comprised of a mostly older, male population, which results in a low generalizability to organizations other than VA medical centers. In addition, this study only evaluated patients on a GLP-1 agonist followed in a pharmacist-led PACT clinic. This study excluded patients who were prescribed a GLP-1 agonist by an endocrinologist or a pharmacist at one of the community-based outpatient clinics affiliated with WBVAMC, or a pharmacist or clinician outside the VA. The sole focus of this study was patients in a pharmacist-led VAMC clinic. Not all patient data may have been included in the study. If a patient did not have an appointment at baseline, 3, 6, and 12 months or did not obtain laboratory tests, HbA_1c and weights were not recorded. Data were collected during the COVID-19 pandemic and in-person appointments were potentially switched to phone or video appointments. There were many instances during this chart review where a weight was not recorded at each time interval. Also, this study did not consider any other diabetes medications the patient was taking. There were many instances where the patient was taking metformin and/or sodium-glucose cotransporter-2 (SGLT-2) inhibitors. These medications along with diet could have affected the weight results as metformin is weight neutral and SGLT-2 inhibitors promote weight loss.¹⁵ Lastly, this study did not evaluate the amount of insulin reduced, only if there was a dose reduction or discontinuation of insulin and/or a sulfonylurea.

Conclusions

Dose reductions and a discontinuation of insulin or a sulfonylurea with the addition of a GLP-1 agonist may be needed. Patients on both insulin and a sulfonylurea may need closer monitoring due to the higher incidences of discontinuations compared with patients on just 1 of these agents. Dose reductions or discontinuations of these diabetic agents can promote positive patient outcomes, such as preventing hypoglycemia, minimizing weight gain, increasing weight loss, and reducing HbA_1c levels.

Acknowledgments

This material is the result of work supported with resources and the use of facilities at the Wilkes-Barre Veterans Affairs Medical Center in Pennsylvania.

Hypoglycemia and weight gain are well-known adverse effects that can result from insulin and sulfonylureas in patients with type 2 diabetes mellitus (T2DM).^1,2 Insulin and sulfonylurea medications can cause additional weight gain in patients who are overweight or obese, which can increase the burden of diabetes therapy with added medications, raise the risk of hypoglycemia complications, and raise atherosclerotic cardiovascular disease risk factors.³ Although increasing the insulin or sulfonylurea dose is an option health care practitioners or pharmacists have, this approach can increase the risk of hypoglycemia, especially in older adults, such as the veteran population, which could lead to complications, such as falls.²

Previous studies focusing on hypoglycemic events in patients with T2DM showed that glucagon-like peptide-1 (GLP-1) agonist monotherapy has a low incidence of a hypoglycemic events. However, when a GLP-1 agonist is combined with insulin or sulfonylureas, patients have an increased chance of a hypoglycemic event.^3-8 According to the prescribing information for semaglutide, 1.6% to 3.8% of patients on a GLP-1 agonist monotherapy reported a documented symptomatic hypoglycemic event (blood glucose ≤ 70 mg/dL), based on semaglutide dosing. ⁹ Patients on combination therapy of a GLP-1 agonist and basal insulin and a GLP-1 agonist and a sulfonylurea reported a documented symptomatic hypoglycemic event ranging from 16.7% to 29.8% and 17.3% to 24.4%, respectively.⁹ The incidences of hypoglycemia thus dramatically increase with combination therapy of a GLP-1 agonist plus insulin or a sulfonylurea.

When adding a GLP-1 agonist to insulin or a sulfonylurea, clinicians must be mindful of the increased risk of hypoglycemia. Per the warnings and precautions in the prescribing information of GLP-1 agonists, concomitant use with insulin or a sulfonylurea may increase the risk of hypoglycemia, and reducing the dose of insulin or a sulfonylurea may be necessary.^9-11 According to the American College of Cardiology guidelines, when starting a GLP-1 agonist, the insulin dose should be decreased by about 20% in patients with a well-controlled hemoglobin A_1c (HbA_1c).¹²

This study aimed to determine the percentage of patients who required dose reductions or discontinuations of insulin and sulfonylureas with the addition of a GLP-1 agonist. Understanding necessary dose reductions or discontinuations of these concomitant diabetes agents can assist pharmacists in preventing hypoglycemia and minimizing weight gain.

Methods

This clinical review was a single-center, retrospective chart review of patients prescribed a GLP-1 agonist while on insulin or a sulfonylurea between January 1, 2019, and September 30, 2022, at the Wilkes-Barre Veterans Affairs Medical Center (WBVAMC) in Pennsylvania and managed in a pharmacist-led patient aligned care team (PACT) clinic. It was determined by the US Department of Veterans Affairs Office of Research and Development that an institutional review board or other review committee approval was not needed for this nonresearch Veterans Health Administration quality assurance and improvement project. Patients aged ≥ 18 years were included in this study. Patients were excluded if they were not on insulin or a sulfonylurea when starting a GLP-1 agonist, started a GLP-1 agonist outside of the retrospective chart review dates, or were prescribed a GLP-1 agonist by anyone other than a pharmacist in their PACT clinic. This included if a GLP-1 agonist was prescribed by a primary care physician, endocrinologist, or someone outside the VA system.

The primary study outcomes were to determine the percentage of patients with a dose reduction of insulin or sulfonylurea and discontinuation of insulin or a sulfonylurea at intervals of 0 (baseline), 3, 6, and 12 months. Secondary outcomes included changes in HbA_1c and body weight measured at the same intervals of 0 (baseline), 3, 6, and 12 months.
Data were collected using the VA Computerized Patient Record System (CPRS) and stored in a locked spreadsheet. Descriptive statistics were used to analyze the data. Patient data included the number of patients on insulin or a sulfonylurea when initiating a GLP-1 agonist, the percentage of patients started on a certain GLP-1 agonist (dulaglutide, liraglutide, exenatide, and semaglutide), and the percentage of patients with a baseline HbA_1c of < 8%, 8% to 10%, and > 10%. The GLP-1 agonist formulary was adjusted during the time of this retrospective chart review. Patients who were not on semaglutide were switched over if they were on another GLP-1 agonist as semaglutide became the preferred GLP-1 agonist.

Patients were considered to have a dose reduction or discontinuation of insulin or a sulfonylurea if the dose or medication they were on decreased or was discontinued permanently within 12 months of starting a GLP-1 agonist. For example, if a patient who was administering 10 units of insulin daily was decreased to 8 but later increased back to 10, this was not counted as a dose reduction. If a patient discontinued insulin or a sulfonylurea and then restarted it within 12 months of initiating a GLP-1 agonist, this was not counted as a discontinuation.

This retrospective review included 136 patients; 96 patients taking insulin and 54 taking a sulfonylurea when they started a GLP-1 agonist. Fourteen patients were on both. Criteria for use, which are clinical criteria to determine if a patient is eligible for the use of a given medication, are used within the VA. The inclusion criteria for a patient initiating a GLP-1 agonist is that the patient must have atherosclerotic cardiovascular disease or chronic kidney disease with the patient receiving metformin (unless unable to use metformin) and empagliflozin (unless unable to use empagliflozin).

The baseline mean age and weight for the patient population in this retrospective chart review was 70.7 years and 238.2 lb, respectively. Ninety-six patients (70.6%) were started on semaglutide, 27 (19.9%) on dulaglutide, 12 (8.8%) on liraglutide, and 1 (0.7%) on exenatide. The mean HbA_1c when patients initiated a GLP-1 agonist was 8.6%. When starting a GLP-1 agonist, 34 patients (25.0%) had an HbA_1c < 8%, 89 (65.4%) had an HbA_1c between 8% to 10%, and 13 (9.6%) had an HbA_1c > 10% (Table).

For the primary results, 25 patients (26.0%) had a dose reduction of insulin when they started a GLP-1 agonist, and 55 patients (57.3%) had at least 1 insulin dose reduction within the year follow-up. Seven patients (13.0%) had a dose reduction of a sulfonylurea when they started a GLP-1 agonist, and 16 patients (29.6%) had at least 1 dose reduction of a sulfonylurea within the year follow-up. Six patients (6.3%) discontinued insulin use when they initially started a GLP-1 agonist, and 14 patients (14.6%) discontinued insulin use within the year follow-up. Eleven patients (20.4%) discontinued sulfonylurea use when they initially started a GLP-1 agonist, and 21 patients (38.9%) discontinued sulfonylurea use within the year follow-up (Figure).

Fourteen patients were on both insulin and a sulfonylurea. Two patients (14.3%) had a dose reduction of insulin when they started a GLP-1 agonist, and 5 (35.7%) had ≥ 1 insulin dose reduction within the year follow-up. Three patients (21.4%) had a dose reduction of a sulfonylurea when they started a GLP-1 agonist, and 6 (42.9%) had ≥ 1 dose reduction of a sulfonylurea within the year follow-up. Seven patients (50.0%) discontinued sulfonylurea and 3 (21.4%) discontinued insulin at any time throughout the year. The majority of the discontinuations were at the initial start of GLP-1 agonist therapy.

The mean HbA_1c for patients on GLP-1 agonist was 8.6% at baseline, 8.0% at 0 to 3 months, 7.6% at 3 to 6 months, and 7.5% at 12 months. Patients experienced a mean HbA_1c reduction of 1.1%. The mean weight when a GLP-1 agonist was started was 238.2 lb, 236.0 lb at 0 to 3 months, 223.8 lb at 3 to 6 months, and 224.3 lb after 12 months. Study participants lost a mean weight of 13.9 lb while on a GLP-1 agonist.

Discussion

While this study did not examine why there were dose reductions or discontinuations, we can hypothesize that insulin or sulfonylureas were reduced or discontinued due to a myriad of reasons, such as prophylactic dosing per guidelines, patients having a hypoglycemic event, or pharmacists anticipating potential low blood glucose trends. Also, there could have been numerous reasons GLP-1 agonists were started in patients on insulin or a sulfonylurea, such as HbA_1c not being within goal range, cardiovascular benefits (reduce risk of stroke, heart attack, and death), weight loss, and renal protection, such as preventing albuminuria.^13,14

This retrospective chart review found a large proportion of patients had a dose reduction of insulin (57.3%) or sulfonylurea (29.6%). The percentage of patients with a dose reduction was potentially underestimated as patients were not counted if they discontinued insulin or sulfonylurea. Concomitant use of GLP-1 agonists with insulin or a sulfonylurea may increase the risk of hypoglycemia and reducing the dose of insulin or a sulfonylurea may be necessary.^9-11 The dose reductions in this study show that pharmacists within pharmacy-led PACT clinics monitor for or attempt to prevent hypoglycemia, which aligns with the prescribing information of GLP-1 agonists. While increasing the insulin or sulfonylurea dose is an option for patients, this approach can increase the risk of hypoglycemia, especially in an older population, like this one with a mean age > 70 years. The large proportions of patients with dose reductions or insulin and sulfonylurea discontinuations suggest that pharmacists may need to take a more cautious approach when initiating a GLP-1 agonist to prevent adverse health outcomes related to low blood sugar for older adults, such as falls and fractures.

Insulin was discontinued in 20.4% of patients and sulfonylurea was discontinued in 38.9% of patients within 12 months after starting a GLP-1 agonist. When a patient was on both insulin and a sulfonylurea, the percentage of patients who discontinued insulin (21.4%) or a sulfonylurea (50.0%) was higher compared with patients just on insulin (14.6%) or a sulfonylurea (38.9%) alone. Patients on both insulin and a sulfonylurea may need closer monitoring due to a higher incidence of discontinuations when these diabetes agents are administered in combination.

Limitations

Several limitations of this study should be considered when evaluating the results. This review was comprised of a mostly older, male population, which results in a low generalizability to organizations other than VA medical centers. In addition, this study only evaluated patients on a GLP-1 agonist followed in a pharmacist-led PACT clinic. This study excluded patients who were prescribed a GLP-1 agonist by an endocrinologist or a pharmacist at one of the community-based outpatient clinics affiliated with WBVAMC, or a pharmacist or clinician outside the VA. The sole focus of this study was patients in a pharmacist-led VAMC clinic. Not all patient data may have been included in the study. If a patient did not have an appointment at baseline, 3, 6, and 12 months or did not obtain laboratory tests, HbA_1c and weights were not recorded. Data were collected during the COVID-19 pandemic and in-person appointments were potentially switched to phone or video appointments. There were many instances during this chart review where a weight was not recorded at each time interval. Also, this study did not consider any other diabetes medications the patient was taking. There were many instances where the patient was taking metformin and/or sodium-glucose cotransporter-2 (SGLT-2) inhibitors. These medications along with diet could have affected the weight results as metformin is weight neutral and SGLT-2 inhibitors promote weight loss.¹⁵ Lastly, this study did not evaluate the amount of insulin reduced, only if there was a dose reduction or discontinuation of insulin and/or a sulfonylurea.

Conclusions

Dose reductions and a discontinuation of insulin or a sulfonylurea with the addition of a GLP-1 agonist may be needed. Patients on both insulin and a sulfonylurea may need closer monitoring due to the higher incidences of discontinuations compared with patients on just 1 of these agents. Dose reductions or discontinuations of these diabetic agents can promote positive patient outcomes, such as preventing hypoglycemia, minimizing weight gain, increasing weight loss, and reducing HbA_1c levels.

Acknowledgments

This material is the result of work supported with resources and the use of facilities at the Wilkes-Barre Veterans Affairs Medical Center in Pennsylvania.

Hypoglycemia and weight gain are well-known adverse effects that can result from insulin and sulfonylureas in patients with type 2 diabetes mellitus (T2DM).^1,2 Insulin and sulfonylurea medications can cause additional weight gain in patients who are overweight or obese, which can increase the burden of diabetes therapy with added medications, raise the risk of hypoglycemia complications, and raise atherosclerotic cardiovascular disease risk factors.³ Although increasing the insulin or sulfonylurea dose is an option health care practitioners or pharmacists have, this approach can increase the risk of hypoglycemia, especially in older adults, such as the veteran population, which could lead to complications, such as falls.²

Previous studies focusing on hypoglycemic events in patients with T2DM showed that glucagon-like peptide-1 (GLP-1) agonist monotherapy has a low incidence of a hypoglycemic events. However, when a GLP-1 agonist is combined with insulin or sulfonylureas, patients have an increased chance of a hypoglycemic event.^3-8 According to the prescribing information for semaglutide, 1.6% to 3.8% of patients on a GLP-1 agonist monotherapy reported a documented symptomatic hypoglycemic event (blood glucose ≤ 70 mg/dL), based on semaglutide dosing. ⁹ Patients on combination therapy of a GLP-1 agonist and basal insulin and a GLP-1 agonist and a sulfonylurea reported a documented symptomatic hypoglycemic event ranging from 16.7% to 29.8% and 17.3% to 24.4%, respectively.⁹ The incidences of hypoglycemia thus dramatically increase with combination therapy of a GLP-1 agonist plus insulin or a sulfonylurea.

When adding a GLP-1 agonist to insulin or a sulfonylurea, clinicians must be mindful of the increased risk of hypoglycemia. Per the warnings and precautions in the prescribing information of GLP-1 agonists, concomitant use with insulin or a sulfonylurea may increase the risk of hypoglycemia, and reducing the dose of insulin or a sulfonylurea may be necessary.^9-11 According to the American College of Cardiology guidelines, when starting a GLP-1 agonist, the insulin dose should be decreased by about 20% in patients with a well-controlled hemoglobin A_1c (HbA_1c).¹²

This study aimed to determine the percentage of patients who required dose reductions or discontinuations of insulin and sulfonylureas with the addition of a GLP-1 agonist. Understanding necessary dose reductions or discontinuations of these concomitant diabetes agents can assist pharmacists in preventing hypoglycemia and minimizing weight gain.

Methods

This clinical review was a single-center, retrospective chart review of patients prescribed a GLP-1 agonist while on insulin or a sulfonylurea between January 1, 2019, and September 30, 2022, at the Wilkes-Barre Veterans Affairs Medical Center (WBVAMC) in Pennsylvania and managed in a pharmacist-led patient aligned care team (PACT) clinic. It was determined by the US Department of Veterans Affairs Office of Research and Development that an institutional review board or other review committee approval was not needed for this nonresearch Veterans Health Administration quality assurance and improvement project. Patients aged ≥ 18 years were included in this study. Patients were excluded if they were not on insulin or a sulfonylurea when starting a GLP-1 agonist, started a GLP-1 agonist outside of the retrospective chart review dates, or were prescribed a GLP-1 agonist by anyone other than a pharmacist in their PACT clinic. This included if a GLP-1 agonist was prescribed by a primary care physician, endocrinologist, or someone outside the VA system.

The primary study outcomes were to determine the percentage of patients with a dose reduction of insulin or sulfonylurea and discontinuation of insulin or a sulfonylurea at intervals of 0 (baseline), 3, 6, and 12 months. Secondary outcomes included changes in HbA_1c and body weight measured at the same intervals of 0 (baseline), 3, 6, and 12 months.
Data were collected using the VA Computerized Patient Record System (CPRS) and stored in a locked spreadsheet. Descriptive statistics were used to analyze the data. Patient data included the number of patients on insulin or a sulfonylurea when initiating a GLP-1 agonist, the percentage of patients started on a certain GLP-1 agonist (dulaglutide, liraglutide, exenatide, and semaglutide), and the percentage of patients with a baseline HbA_1c of < 8%, 8% to 10%, and > 10%. The GLP-1 agonist formulary was adjusted during the time of this retrospective chart review. Patients who were not on semaglutide were switched over if they were on another GLP-1 agonist as semaglutide became the preferred GLP-1 agonist.

Patients were considered to have a dose reduction or discontinuation of insulin or a sulfonylurea if the dose or medication they were on decreased or was discontinued permanently within 12 months of starting a GLP-1 agonist. For example, if a patient who was administering 10 units of insulin daily was decreased to 8 but later increased back to 10, this was not counted as a dose reduction. If a patient discontinued insulin or a sulfonylurea and then restarted it within 12 months of initiating a GLP-1 agonist, this was not counted as a discontinuation.

This retrospective review included 136 patients; 96 patients taking insulin and 54 taking a sulfonylurea when they started a GLP-1 agonist. Fourteen patients were on both. Criteria for use, which are clinical criteria to determine if a patient is eligible for the use of a given medication, are used within the VA. The inclusion criteria for a patient initiating a GLP-1 agonist is that the patient must have atherosclerotic cardiovascular disease or chronic kidney disease with the patient receiving metformin (unless unable to use metformin) and empagliflozin (unless unable to use empagliflozin).

The baseline mean age and weight for the patient population in this retrospective chart review was 70.7 years and 238.2 lb, respectively. Ninety-six patients (70.6%) were started on semaglutide, 27 (19.9%) on dulaglutide, 12 (8.8%) on liraglutide, and 1 (0.7%) on exenatide. The mean HbA_1c when patients initiated a GLP-1 agonist was 8.6%. When starting a GLP-1 agonist, 34 patients (25.0%) had an HbA_1c < 8%, 89 (65.4%) had an HbA_1c between 8% to 10%, and 13 (9.6%) had an HbA_1c > 10% (Table).

For the primary results, 25 patients (26.0%) had a dose reduction of insulin when they started a GLP-1 agonist, and 55 patients (57.3%) had at least 1 insulin dose reduction within the year follow-up. Seven patients (13.0%) had a dose reduction of a sulfonylurea when they started a GLP-1 agonist, and 16 patients (29.6%) had at least 1 dose reduction of a sulfonylurea within the year follow-up. Six patients (6.3%) discontinued insulin use when they initially started a GLP-1 agonist, and 14 patients (14.6%) discontinued insulin use within the year follow-up. Eleven patients (20.4%) discontinued sulfonylurea use when they initially started a GLP-1 agonist, and 21 patients (38.9%) discontinued sulfonylurea use within the year follow-up (Figure).

Fourteen patients were on both insulin and a sulfonylurea. Two patients (14.3%) had a dose reduction of insulin when they started a GLP-1 agonist, and 5 (35.7%) had ≥ 1 insulin dose reduction within the year follow-up. Three patients (21.4%) had a dose reduction of a sulfonylurea when they started a GLP-1 agonist, and 6 (42.9%) had ≥ 1 dose reduction of a sulfonylurea within the year follow-up. Seven patients (50.0%) discontinued sulfonylurea and 3 (21.4%) discontinued insulin at any time throughout the year. The majority of the discontinuations were at the initial start of GLP-1 agonist therapy.

The mean HbA_1c for patients on GLP-1 agonist was 8.6% at baseline, 8.0% at 0 to 3 months, 7.6% at 3 to 6 months, and 7.5% at 12 months. Patients experienced a mean HbA_1c reduction of 1.1%. The mean weight when a GLP-1 agonist was started was 238.2 lb, 236.0 lb at 0 to 3 months, 223.8 lb at 3 to 6 months, and 224.3 lb after 12 months. Study participants lost a mean weight of 13.9 lb while on a GLP-1 agonist.

Discussion

While this study did not examine why there were dose reductions or discontinuations, we can hypothesize that insulin or sulfonylureas were reduced or discontinued due to a myriad of reasons, such as prophylactic dosing per guidelines, patients having a hypoglycemic event, or pharmacists anticipating potential low blood glucose trends. Also, there could have been numerous reasons GLP-1 agonists were started in patients on insulin or a sulfonylurea, such as HbA_1c not being within goal range, cardiovascular benefits (reduce risk of stroke, heart attack, and death), weight loss, and renal protection, such as preventing albuminuria.^13,14

This retrospective chart review found a large proportion of patients had a dose reduction of insulin (57.3%) or sulfonylurea (29.6%). The percentage of patients with a dose reduction was potentially underestimated as patients were not counted if they discontinued insulin or sulfonylurea. Concomitant use of GLP-1 agonists with insulin or a sulfonylurea may increase the risk of hypoglycemia and reducing the dose of insulin or a sulfonylurea may be necessary.^9-11 The dose reductions in this study show that pharmacists within pharmacy-led PACT clinics monitor for or attempt to prevent hypoglycemia, which aligns with the prescribing information of GLP-1 agonists. While increasing the insulin or sulfonylurea dose is an option for patients, this approach can increase the risk of hypoglycemia, especially in an older population, like this one with a mean age > 70 years. The large proportions of patients with dose reductions or insulin and sulfonylurea discontinuations suggest that pharmacists may need to take a more cautious approach when initiating a GLP-1 agonist to prevent adverse health outcomes related to low blood sugar for older adults, such as falls and fractures.

Insulin was discontinued in 20.4% of patients and sulfonylurea was discontinued in 38.9% of patients within 12 months after starting a GLP-1 agonist. When a patient was on both insulin and a sulfonylurea, the percentage of patients who discontinued insulin (21.4%) or a sulfonylurea (50.0%) was higher compared with patients just on insulin (14.6%) or a sulfonylurea (38.9%) alone. Patients on both insulin and a sulfonylurea may need closer monitoring due to a higher incidence of discontinuations when these diabetes agents are administered in combination.

Limitations

Several limitations of this study should be considered when evaluating the results. This review was comprised of a mostly older, male population, which results in a low generalizability to organizations other than VA medical centers. In addition, this study only evaluated patients on a GLP-1 agonist followed in a pharmacist-led PACT clinic. This study excluded patients who were prescribed a GLP-1 agonist by an endocrinologist or a pharmacist at one of the community-based outpatient clinics affiliated with WBVAMC, or a pharmacist or clinician outside the VA. The sole focus of this study was patients in a pharmacist-led VAMC clinic. Not all patient data may have been included in the study. If a patient did not have an appointment at baseline, 3, 6, and 12 months or did not obtain laboratory tests, HbA_1c and weights were not recorded. Data were collected during the COVID-19 pandemic and in-person appointments were potentially switched to phone or video appointments. There were many instances during this chart review where a weight was not recorded at each time interval. Also, this study did not consider any other diabetes medications the patient was taking. There were many instances where the patient was taking metformin and/or sodium-glucose cotransporter-2 (SGLT-2) inhibitors. These medications along with diet could have affected the weight results as metformin is weight neutral and SGLT-2 inhibitors promote weight loss.¹⁵ Lastly, this study did not evaluate the amount of insulin reduced, only if there was a dose reduction or discontinuation of insulin and/or a sulfonylurea.

Conclusions

Dose reductions and a discontinuation of insulin or a sulfonylurea with the addition of a GLP-1 agonist may be needed. Patients on both insulin and a sulfonylurea may need closer monitoring due to the higher incidences of discontinuations compared with patients on just 1 of these agents. Dose reductions or discontinuations of these diabetic agents can promote positive patient outcomes, such as preventing hypoglycemia, minimizing weight gain, increasing weight loss, and reducing HbA_1c levels.

Acknowledgments

This material is the result of work supported with resources and the use of facilities at the Wilkes-Barre Veterans Affairs Medical Center in Pennsylvania.