Abstract: Supported by the National Institute of Diabetes and Digestive and Kidney Diseases (NIH/NIDDK) R01DK133465, this study evaluates the safety and healthcare utilization outcomes of GLP-1 receptor agonists (GLP-1RAs) in adults with Type 1 Diabetes (T1D) and obesity. Using a target trial emulation framework, the research finds no significant increase in acute complications and a marked reduction in all-cause hospitalizations.
Introduction
While GLP-1 receptor agonists (GLP-1RAs) such as semaglutide have transformed the management of Type 2 Diabetes (T2D) and obesity, their application in Type 1 Diabetes (T1D) remains controversial and largely off-label. Historically characterized as a disease of lean individuals, T1D is increasingly intersecting with the global obesity epidemic. However, the medical community has remained cautious regarding GLP-1RA use in this population due to the theoretical risks of "euglycemic" diabetic ketoacidosis (DKA) and severe hypoglycemia.
Because T1D patients are frequently excluded from major GLP-1RA clinical trials, a significant gap exists in real-world safety data. This study addresses this gap by investigating whether these medications increase the risk of acute complications in T1D patients who meet clinical criteria for weight-loss intervention.
Methodology
To approximate the evidence quality of a randomized controlled trial (RCT) using observational data, the researchers employed a target trial emulation framework. The study utilized longitudinal Electronic Health Records (EHR) from the OneFlorida+ clinical research network, spanning 2014 to 2024 and covering approximately 21 million individuals across Florida, Georgia, and Alabama.
Cohort Selection and Matching
From an initial cohort of 15,585 T1D patients eligible for anti-obesity medication, the team identified "true" T1D cases using a validated computable phenotype algorithm with >90% accuracy. To minimize selection bias, 1:1 time-conditional propensity score matching was used to pair 651 GLP-1RA initiators with 651 non-initiators based on identical baseline characteristics, including age, BMI, and comorbidities.
Advanced Causal Learning
Beyond standard Cox proportional hazards models, the researchers utilized a "doubly-robust meta-learner" (incorporating XGBoost) to explore Heterogeneous Treatment Effects (HTE). This allowed for the identification of specific patient subgroups that might experience varying levels of risk or benefit from the treatment.
Key Results
Safety Profile: DKA and Hypoglycemia
The primary safety concern—an increase in DKA—was not supported by the data. The incidence rate of DKA was actually lower in the GLP-1RA group (13.5 vs. 21.8 per 1,000 person-years), resulting in a Hazard Ratio (HR) of 0.62 [95% CI 0.33–1.17]. While this trend toward benefit did not reach statistical significance, it effectively provided a "green light" regarding the safety of these drugs in T1D. Similarly, the risk of severe hypoglycemia showed no significant increase (HR 0.52 [95% CI 0.17–1.55]).
Healthcare Utilization
The most significant findings related to general healthcare utilization. GLP-1RA users experienced a substantial reduction in both all-cause hospitalizations and emergency department (ED) visits:
- All-Cause Hospitalizations: HR 0.74 [95% CI 0.62–0.90]
- ED Visits: HR 0.73 [95% CI 0.57–0.92]
These results suggest that GLP-1RAs may provide broader metabolic or cardiovascular benefits that translate into reduced acute care needs.
Subgroup Insights and Heterogeneous Effects
The causal learning analysis revealed critical interactions between comorbidities and treatment outcomes:
- The Hypothyroidism Interaction: Patients without hypothyroidism experienced a significant reduction in DKA risk (HR 0.47), whereas those with hypothyroidism saw no such benefit. This suggests that thyroid dysfunction might interfere with the metabolic advantages of GLP-1RAs or exacerbate issues related to gastric motility.
- Metformin Synergy: The reduction in DKA risk was most pronounced in patients not using metformin at baseline. This indicates that for patients already on multiple insulin-sensitizing agents, the addition of a GLP-1RA may require more precise insulin titration to mitigate ketogenic risks.
- Optimal Subgroup: The most dramatic benefit was observed in patients without hypothyroidism but with dyslipidemia, who saw a massive reduction in DKA risk (HR 0.12).
Discussion and Limitations
This study provides robust real-world evidence that GLP-1RAs are likely safe for adults with T1D and obesity, potentially offering significant protective benefits against hospitalization. However, the study's observational nature means it cannot fully account for unmeasured confounders such as diet, exercise, or medication adherence. Additionally, while the total cohort was substantial, the relatively small number of DKA events (16 vs. 24) led to wide confidence intervals in subgroup analyses.
Future prospective RCTs are necessary to validate these findings and to further explore the physiological mechanisms behind the observed interactions with hypothyroidism and other comorbidities. For now, these results offer reassurance to clinicians considering GLP-1RAs for their T1D patients with obesity.
Summary updated