GLP-1 Receptor Agonist Prescribing Practices in a Pediatric Weight Management Program

Introduction

Glucagon-like protein receptor agonists (GLP-1 RA) have been shown to improve weight status and reduce cardiometabolic risk. ¹ Additionally, they have been established as safe and effective for children ages 12 to 19.^2,3 Recognizing the potential benefit for youth, the American Academy of Pediatrics, in their most recent guidelines, suggests that anti-obesity medications, including the GLP-1 RA be offered to youths with obesity ages 12 and up in conjunction with intensive lifestyle therapy. ⁴

In adult populations, there are concerns about pharmaco-inequity as GLP-1 RA use is lower in patients identifying as Asian, Black, and Hispanic.^5,6 There is little information as to whether pharmaco-inequity is an issue for GLP-1 RA use in pediatrics. Potential barriers to GLP-1 RA use in children include medication cost, insurance coverage, and provider knowledge and comfort with this therapeutic modality. With this study, we analyze the prescribing practices of a large, midwestern pediatric weight management center (PWMC) to gain insight into provider practices that may contribute to pharmaco-inequity.

Methods

Our study was observational and retrospective. The EpicCare Electronic Health Records (Epic Systems Corporation, Verona, WI) of all patients attending the Center for Better Health and Nutrition (CBHN) clinics of Cincinnati Children’s Hospital were reviewed from 7/1/2021 to 6/30/2023. Patients seen in the CBHN are seen by one of four pediatric obesity specialists, a dietitian, and an exercise physiologist. Demographic data extracted included age, race, ethnicity, sex, and insurance type. For anthropometric data, we extracted the percent of the 95^th percentile body mass index for age and sex (p95%BMI). Body mass index (BMI) measurements were subsequently grouped by BMI percentile for age and sex: 95th to <120% of 95th percentile for BMI (Class 1 obesity), 120 to <140% (Class 2 obesity), and 140% and above (Class 3 obesity). Laboratory data included the maximum value for hemoglobin A1C (HbA1c) and the maximum value of alanine aminotransferase (ALT) while the patient was actively enrolled in our PWMC, as patients could have multiple measurements over the 2-year time period. Lastly, whether the patient was prescribed a GLP1-RA agonist (dulaglutide, liraglutide, and semaglutide) and whether metformin was prescribed during the 2-year study period while enrolled in our PWMP were extracted. Patients on metformin only were chosen as a comparison group, as they are high-risk, but prescribing metformin is unlikely to be limited by economic factors, as the cost is low and coverage by insurance is virtually universal. Statistical Analysis Software (SAS®) was used to clean, process, and analyze data. A chi-square test or Kruskal–Wallis rank sum test was done when appropriate to compare unadjusted variables between groups: those prescribed GLP1-RA agonists with or without metformin, those prescribed metformin, and those prescribed neither a GLP1-RA nor metformin. A multivariate logistic regression model was constructed with a prescribed GLP1-RA as the dependent variable and race, ethnicity, sex, age, insurance type (private vs. Medicaid), BMI Class, maximum HbA1c, maximum ALT, and p95% BMI as covariates. The study was done with the approval of the Cincinnati Children’s Hospital Institutional Review Board.

Results

In all, there were 2,563 unique patients seen by the CBHN during the 2-year study period, of whom 73 received a prescription for a GLP1-RA with or without metformin, 380 a prescription for metformin alone, and 1,839 did not receive a prescription for either drug. Table 1 summarizes the characteristics of the three groups. Of note, there were no significant differences between the three groups concerning race, ethnicity, sex, or insurance type. There were significant differences between the groups, however, concerning age, maximum HbA1c, maximum ALT, and BMI status.

The results of the multivariate logistic regression model are summarized in Table 2. There was no significant effect of race, ethnicity, sex, or insurance type on whether a patient was prescribed a GLP1-RA agonist. There was, however, a significant association with Class 3 BMI (odds ratio: 7.65), older age (odds ratio: 1.22.), increasing maximum HbA1c (odds ratio: 3.17,) and higher ALT (odds ratio: 1.01).

When comparing those who received a GLP1-RA with or without metformin to all others, the GLP1-RA had a higher maximum HbA1c (6.5 vs. 5.6%), a higher maximum ALT (79 IU vs. 37 IU), and a higher maximum p95%BMI (169 vs. 141).

Discussion

With our study, we document the GLP1-RA prescribing practices of the four pediatric obesity specialists in our weight management over 2 years beginning 6 months after liraglutide was first approved for weight loss in youth 12 to 18 years. In our retrospective review, race, ethnicity, and insurance status were not associated with the likelihood of a child being prescribed a GLP1-RA. This finding suggests that obesity practitioners at our site are not biased by these factors. Rather, prescribing seems to be influenced by disease burden, as GLP1-RA prescription was associated with a 17% greater HgA1C, 120% greater ALT, and 22% greater p95%BMI than those with no Metformin or GLP1-RA. Interestingly, in our multivariate analysis, there was no association with sex. The finding is inconsistent with previous studies in adults, which show women are over three times more likely to be prescribed anti-obesity medications than men.^7,8

Our study has several limitations. The study was retrospective, and thus data collection was not under uniform circumstances. The study is from one practice, and the results may not be generalized to other geographic regions or other practice types and settings. Most importantly, we only have information on what was prescribed and have no information on whether insurance covered the prescriptions, whether the drug was available in pharmacies, whether the patient/family filled the prescription, and what the adherence was if the drug was obtained. Further, important socioeconomic factors such as household income, housing, and food insecurity were not available for this analysis. Additionally, it is not clear whether previous laboratory evaluations, such as an elevated HbA1c. or other practitioner input may have played a role in the decision to prescribe a GLP1-RA. Still, with our study, we show that patients in our PWMP were prescribed GLP1-RA agonists without regard to race, ethnicity, or insurance type but rather based on the extent of disease burden. It appears that the physician prescribing practices at our midwestern pediatric weight management program do not contribute to inequitable prescribing of GLP1-RA medications.

Impact Statement

This study is among the first on prescribing practices of obesity medicine pediatricians. This type of study is critical in addressing pharmaco-equity issues. We demonstrate that physicians in the pediatric weight management setting do not appear biased by race or other demographic factors other than age, which is clinically appropriate.