Effect of semaglutide add‐on therapy for people living with type 1 diabetes and excess weight

The coexistence of type 1 diabetes (T1D) and obesity, termed ‘double diabetes,’ is increasingly prevalent worldwide and presents unique therapeutic challenges.1, 2 This phenotype combines autoimmune insulin deficiency with insulin resistance characteristic of type 2 diabetes (T2D), resulting in poor glycaemic control, increased complications and elevated mortality risk.3, 4 Glucagon-like peptide-1 receptor agonists (GLP-1RAs) improve insulin sensitivity, reduce body weight and provide cardiorenal protection in T2D.5, 6 While these benefits remain unproven in T1D, GLP-1RAs represent potentially valuable adjunct therapy for double diabetes.7 Early studies with liraglutide raised safety concerns regarding hypoglycaemia and hyperglycaemia with ketosis.8, 9 However, emerging evidence suggests semaglutide may offer superior benefits with acceptable safety in selected individuals with T1D.7, 10, 11 A critical knowledge gap exists regarding sustained semaglutide effectiveness on continuous glucose monitoring (CGM) parameters in real-world T1D populations with established obesity, particularly among hybrid closed-loop (HCL) system users.10-12 We conducted a retrospective observational study evaluating metabolic outcomes of adjunctive once-weekly semaglutide over 1 year in adults with T1D and excess weight. We searched electronic medical records for patients with T1D followed in a hospital-based adult endocrinology clinic who received physician recommendations to initiate semaglutide between January 2020 and September 2024. Of 1681 records identified, after manual review 1632 were excluded for absence of T1D or excess weight, 13 for missing glycated haemoglobin (HbA1c) values, and two for premature semaglutide discontinuation, yielding 34 subjects. Following institutional review board approval (protocol 6282-19 SMC), data were retrieved using the MDClone™ ADAMS platform. CGM and insulin pump data were extracted from Tidepool, LibreView™, CareLink™ and Dexcom Clarity™ platforms. The primary outcome was HbA1c. Secondary outcomes included body weight and CGM time-in-range (TIR, 70–180 mg/dL). Exploratory outcomes encompassed time above range (TAR) and below range (TBR), total daily insulin dose (TDD), mean sensor glucose (MSG) and lipid profiles. Venous blood samples laboratory results were obtained within 3 months of each timepoint. CGM data required ≥14 days of continuous availability. Categorical variables were reported as frequencies and percentages; continuous variables as means ± standard deviations. Paired t tests assessed treatment effects, with Mann–Whitney U tests for subgroup analyses. Analyses were performed using Microsoft Excel 365, with two-tailed p < 0.05 considered significant. The cohort comprised 34 subjects (22 males) with mean age 49.0 ± 10.7 years and diabetes duration 23.3 ± 9.6 years (Table 1). Baseline HbA1c was 7.7 ± 1.1%, body mass index (BMI) 31.1 ± 4.3 kg/m2. Among 32 pump users, 11 (32%) employed HCL algorithms at baseline, with one additional transition during follow-up. Thirteen subjects (38%) had documented microvascular complications and five (15%) had macrovascular disease. Two patients had prior bariatric surgery. Most subjects (79%, n = 27) achieved the maximal prescribed weekly dose (1 mg) during follow-up. Treatment interruptions (≤1 month) occurred due to drug shortages or gastrointestinal adverse effects. Over 12 months, semaglutide treatment was associated with significant HbA1c reduction (7.7% ± 1.1% to 7.3% ± 0.8%, p = 0.001), increased TIR (60.3% ± 18.9% to 69.1 ± 11.9%, p = 0.02) and decreased TAR >250 mg/dL (10.3% ± 9.3% to 6.4% ± 5.7%, p = 0.02) (Figure S1). MSG showed a non-significant decline (172.4 ± 35.5 to 157.3 ± 20.8 mg/dL, p = 0.11). TDD trended downward (63.1 ± 21.9 to 58.2 ± 20.0 units, p = 0.07). Similar improvements were observed at 6 months (Table 2). Body weight decreased by 6.6% after 1 year (93.5 ± 16.2 to 87.4 ± 15.8 kg, p < 0.001) (Table 2). Non-high-density lipoprotien (HDL) cholesterol showed a trend towards reduction (102.9 ± 30.9 to 91.2 ± 36.2 mg/dL, p = 0.08) without changes in lipid-lowering medications or changes in other blood lipids. One hospitalisation related to diabetic ketoacidosis (DKA) occurred due to insulin pump infusion set failure in a patient with prior bariatric surgery and recurrent DKA history. One patient with established ischemic heart disease presented with reduced exercise capacity and a positive nuclear stress test just prior to starting semaglutide and underwent elective angioplasty 2 months after drug initiation. There was no documentation of hospital or emergency room care for hyperglycaemia, hypoglycaemia or pancreatitis. Of two excluded patients who discontinued semaglutide after 6 months, one stopped due to gastrointestinal symptoms and another, elderly with paraplegia, due to recurrent decubitus ulcers, which later healed. Subjects using HCL systems, compared to those using other insulin delivery methods, exhibited a strong trend for smaller reductions in MSG and HbA1c after 12 months of semaglutide treatment (−9.1 ± 6.3 mg/dL vs. −17.6 ± 32.7 mg/dL, p = 0.053 and −0.3 ± 0.4% vs. −0.5 ± 0.8%, p = 0.07, respectively) (Table S1). This retrospective analysis demonstrates that adjunctive weekly semaglutide in adults with T1D and excess weight is associated with sustained improvements in glycaemic control and body weight over 1 year. Our real-world design captures semaglutide effects under routine care with higher uptake of advanced technology compared to previous real-world studies,10, 11 with most utilising CGM and insulin pumps and one-third HCL systems. Our findings of 0.4% HbA1c reduction, 9% TIR increase and 6.6% weight loss align with previous studies.13, 14 Post-treatment HbA1c (7.3%) and TIR (69.1%) approximate American Diabetes Association targets.15 Ongoing trials will determine if higher-dose semaglutide can further exceed these thresholds. The trend towards reduced insulin requirements is consistent with most semaglutide studies in T1D,10, 11, 14, 16 and differences in outcomes may reflect diabetes duration and sample size variations. The non-HDL cholesterol reduction trend aligns with the only other long-term real-world study examining lipid effects of semaglutide specifically in T1D, where there was a significant reduction in triglycerides and low-density lipoprotein (LDL).11 Notably, our subgroup analysis is the first to reveal a trend for attenuation of semaglutide's glycaemic benefits among HCL users, consistent with this technology's inherent advantages.17-19 This parallels findings where TIR increased 23% in a real-world T1D study with 73% multiple daily injection users and no HCL users,11 versus only 5%–11% rises in HCL-only studies.14, 16 Regarding safety, the absence of hypoglycaemia increase and single unrelated DKA incident echo previous favourable safety signals for semaglutide in T1D,7, 10, 11, 14 though our study is underpowered for adequate DKA detection. Key limitations of this study include its retrospective design, single-centre setting and relatively small sample size, which introduce selection bias. These limitations were compounded by the high cost of semaglutide and its limited availability during the observation period, which may have further influenced patient selection and treatment patterns. The absence of a control group precludes causal inference. Missing data, particularly for CGM metrics and insulin dosing, may introduce further bias. CGM data pooled from multiple systems (Medtronic Guardian Connect, FreeStyle Libre 2, Dexcom G6/G7) could reflect small systematic accuracy differences despite all being factory calibrated, meeting quality standards and mean absolute relative difference (MARD) ranging 8.2%–10.2%.20 Future work should examine whether our findings are consistent within individual CGM platforms and prospectively compare the benefits of GLP-1RAs in those using HCL systems to other insulin delivery technologies. Given rising obesity and insulin resistance prevalence among individuals with T1D, our findings provide encouraging evidence that semaglutide appears associated with meaningful metabolic benefits. Broader clinical adoption requires accumulation of long-term safety data and development of clear guidelines, but the potential to expand treatment options is both timely and significant. Design: NCM and GA-H. Conduct/data collection: DO. Analysis: DO and NCM. Writing manuscript: DO and NCM. Revisions: GA-H. We would like to thank Prof. Amir Tirosh for his support and suggestions, as well as Dr. Amna Jabarin, Dr. Nicole Morozov and Ms. Liat Arnon-Klug for assistance with data extraction. The authors received no funding from an external source. Noga Chlamtac Minsky and Genya Aharon-Hananel report receiving lecture fees from Elii Lilly and Novo Nordisk. Both authors received support to attend academic conferences from Novo Nordisk, and Noga Chlamtac Minsky additionally received this form of support from Eli Lilly. Lastly, Genya Aharon-Hananel has served on an advisory board for Eli Lilly and has received a research grant from Novo Nordisk. Dmitriy Olshanskiy declares no conflict of interest. The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/dom.70389. The data that support the findings of this study are available from the corresponding author upon reasonable request. Figure S1. Time in different glycaemic ranges at baseline, 6 and 12 months. Table S1. Subgroup analyses of differences on semaglutide treatment over 1 year. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.