Type 2 diabetes and obesity often travel together, and over the past several years the number of ways to treat both conditions at once has grown considerably. Two injectable peptide therapies and a surgical procedure have each attracted significant attention, but until recently there were very few head-to-head comparisons drawn from everyday clinical practice rather than tightly controlled trials.
A retrospective cohort study published in a major diabetes and endocrinology journal set out to fill that gap. Researchers pulled de-identified electronic health records from more than 1,600 hospitals and roughly 37,900 clinics across 280 health systems in the United States, ultimately analyzing data from just over 45,000 adults. The goal was straightforward: after one year, how often did each treatment group reach a combined target of losing at least 20 percent of body weight and bringing a key blood-sugar marker down to a level associated with no diabetes diagnosis?
The three groups studied were adults who maintained a stable dose of a GLP-1 receptor agonist (semaglutide), adults who maintained a stable dose of a dual GIP and GLP-1 receptor agonist (tirzepatide), and adults who underwent sleeve gastrectomy, a form of bariatric surgery. All participants entered the study with a body mass index of at least 35 and an HbA1c above 6.4 percent, a standard threshold for a type 2 diabetes diagnosis.
Study design and patient groups
The research team used a retrospective design, meaning they looked backward through existing records rather than randomly assigning participants to treatments. This approach captures real-world prescribing patterns but also comes with limitations the authors were careful to flag.
After applying their inclusion and exclusion rules, 33,482 patients fell into the semaglutide group, 4,178 into the tirzepatide group, and 7,433 into the sleeve gastrectomy group. Researchers adjusted their statistical models for factors including age, sex, race, baseline HbA1c, BMI, a standard measure of how many other health conditions each patient had, a social vulnerability index, and how many medications each person was already taking for diabetes, cholesterol, and blood pressure.
One important context point the authors raised: the sleeve gastrectomy group started with a higher average BMI, was younger on average, and had a somewhat lower baseline HbA1c than the two medication groups. Those differences matter when interpreting the results, because they could independently influence outcomes.
Primary outcome results
The composite target, reaching at least 20 percent body weight loss and an HbA1c below 5.7 percent at one year, was met at very different rates across the three groups.
For the semaglutide group, the adjusted probability of hitting both targets was 3.0 percent. For the tirzepatide group it was 13.2 percent, roughly four times higher. For the sleeve gastrectomy group it was 24.0 percent, the highest of the three. Each comparison came with confidence intervals that did not overlap, suggesting the differences were statistically meaningful rather than due to chance.
The authors describe the composite target as demanding. Reaching 20 percent weight loss and normalizing blood sugar simultaneously within a single year is a high bar, which is part of why even the best-performing group reached it in roughly one in four patients rather than the majority.
Safety signals observed
Researchers also tracked three prespecified safety outcomes: emergency department visits within one year, new prescriptions for gastro-esophageal reflux disease (GERD), and new prescriptions for nausea.
Emergency department visits were more frequent in the sleeve gastrectomy group than in either medication group. New prescriptions for both GERD and nausea appeared across all three groups, but the absolute rates were higher after surgery. The authors note that these safety findings were analyzed descriptively, meaning they describe what happened in the data rather than rigorously testing causal links.
Neither peptide treatment was without side-effect signals. Nausea is a well-documented feature of GLP-1 class medications, and the prescription data reflected that. However, the frequency of emergency department use was lower in the two medication groups, which the authors highlighted as a relevant consideration when weighing treatment paths.
Why the numbers carry caveats
Retrospective observational data from electronic health records is powerful in scale but limited in precision. The study cannot fully account for every factor that led a clinician to choose one treatment over another for a given patient, a problem statisticians call residual confounding.
The baseline differences between groups are also worth keeping in mind. Patients who underwent surgery were younger and heavier, which could make weight loss easier even without surgery, while their lower starting HbA1c could make the blood-sugar target easier to reach. The authors acknowledge these points explicitly in their interpretation section.
Additionally, the study only tracked safety through new prescriptions and emergency department visits. Longer-term outcomes, quality-of-life measures, and a wider range of potential complications were outside the scope of this particular analysis. The study period captured the first year, so what happens to each group beyond 12 months remains an open question.
What the peptide mechanisms look like
Both semaglutide and tirzepatide work by mimicking hormones the gut releases after a meal. Semaglutide targets the GLP-1 receptor, which slows stomach emptying, reduces appetite signals in the brain, and prompts the pancreas to release insulin when blood sugar rises. Tirzepatide adds a second mechanism by also activating the GIP receptor, another incretin hormone pathway that appears to work in concert with GLP-1 to produce larger effects on both weight and blood sugar.
The difference in outcome rates between the two compounds, roughly 3 percent versus 13 percent on the composite target, is consistent with clinical trial data suggesting that dual-receptor activation produces more pronounced metabolic changes than single-receptor activation alone. The retrospective study adds real-world context to those controlled-trial findings.
Sleeve gastrectomy works through a different set of mechanisms entirely, involving permanent anatomical changes to the stomach that reduce its volume and alter gut hormone release patterns over the long term.
Broader context for researchers and readers
A study of this scale, drawing on records from 280 health systems, is relatively rare in the metabolic disease literature. The sheer number of patients included gives the findings more statistical weight than smaller single-center studies, even accounting for the inherent limits of retrospective design.
The literature suggests that the gap between semaglutide and tirzepatide in this dataset is among the larger real-world comparisons yet published. Prior phase-3 trial data pointed in the same direction, but enrolling more than 45,000 patients who were actually prescribed these compounds in routine care adds a layer of generalizability.
For researchers studying peptide-based metabolic interventions, the findings highlight how receptor selectivity, the number of hormonal pathways a compound engages, appears to translate into meaningfully different clinical outcomes even when both compounds belong to the same broad therapeutic class. The data also reinforce that surgical and pharmacological options occupy different points on the spectrum of efficacy and risk, and that neither approach is uniformly superior across all dimensions measured.
