Because no large head-to-head trial had previously compared every licensed dose of the two drugs across a full panel of outcomes, the researchers turned to an established statistical technique known as the Bucher indirect treatment comparison. The method anchors two randomized controlled trials that share a common comparator, in this case placebo, and mathematically derives the relative effect of the two active treatments. The team drew on SURMOUNT-1, the 72-week trial that tested tirzepatide at 5, 10 and 15 milligrams per week, and STEP 1, the 68-week trial that tested semaglutide at 2.4 milligrams per week. Both trials enrolled adults with obesity, defined as a body mass index of 30 or higher, or overweight with at least one weight-related complication such as hypertension, dyslipidemia, obstructive sleep apnea or cardiovascular disease, and both excluded people with type 2 diabetes.

The statistical machinery behind the comparison was more elaborate than a simple subtraction. Before running any numbers, the researchers assessed whether the two trials were similar enough to be compared at all, examining trial design, baseline patient characteristics, placebo responses and outcome definitions. They identified sex and ethnicity as potential treatment effect modifiers, since SURMOUNT-1 enrolled 67.5 percent women and 47.8 percent Hispanic participants while STEP 1 enrolled 74.1 percent women and only 12.0 percent Hispanic participants. To test whether these imbalances distorted the results, they repeated the analysis using matching-adjusted indirect comparisons, a technique that reweights individual patient data from one trial so that its population matches the aggregate characteristics of the other. Reassuringly, the adjusted and unadjusted analyses told broadly the same story.

The headline finding concerns weight itself. Tirzepatide at 10 and 15 milligrams produced statistically significantly greater weight reductions than semaglutide 2.4 milligrams, with mean differences of 5.10 and 6.50 kilograms respectively in favor of tirzepatide. Body mass index fell correspondingly further, by 1.87 and 2.37 kilograms per square meter. Patients on the higher tirzepatide doses were also significantly more likely to cross clinically meaningful thresholds, achieving weight losses of at least 10, 15 and 20 percent of their starting body weight more often than those on semaglutide. The lowest tirzepatide dose, 5 milligrams, showed only non-significant trends in the same direction, effectively performing at a level similar to semaglutide. These numbers align closely with the raw trial results, where tirzepatide 10 and 15 milligrams produced losses of 22.2 and 23.6 kilograms against semaglutide’s 17.4 kilograms.

What sets this analysis apart from earlier comparisons is its reach beyond the bathroom scale. Tirzepatide 10 and 15 milligrams were associated with significantly greater reductions in waist circumference, roughly 5.25 and 5.75 centimeters, a marker of visceral adiposity closely tied to cardiovascular risk. Glycemic control also favored the dual agonist, with small but statistically significant additional reductions in HbA1c and fasting plasma glucose even though all participants had normal glucose metabolism at baseline. Blood pressure told a similar story: every tirzepatide dose produced significantly greater drops in diastolic pressure than semaglutide, ranging from 1.30 to 2.40 millimeters of mercury, while systolic improvements trended in the same direction without reaching significance.

The lipid panel added further nuance. All three tirzepatide doses were associated with significantly greater increases in HDL cholesterol, the so-called good cholesterol, ranging from 3.80 to 5.40 percent relative to semaglutide. The highest dose also produced a significantly greater reduction in triglycerides, and in the adjusted analyses it significantly outperformed semaglutide on total cholesterol as well. Taken together, the authors argue, these improvements across waist circumference, glycemia, HDL, triglycerides and blood pressure suggest tirzepatide could meaningfully shift the constellation of risk factors that define metabolic syndrome, the cluster of abnormalities that precedes type 2 diabetes and cardiovascular disease in millions of people worldwide.

Perhaps the most novel contribution involves body composition, an outcome no previous tirzepatide-versus-semaglutide comparison has examined. In the subgroup of participants who underwent dual-energy X-ray absorptiometry scanning, tirzepatide 15 milligrams was associated with a significantly greater reduction in the percentage of body fat mass, 3.50 percent beyond semaglutide, and a significantly greater increase in the percentage of lean mass, 3.40 percent. All participants on both drugs lost some lean tissue along with fat, an expected consequence of substantial weight loss, but the proportion of the body composed of lean mass shifted more favorably with tirzepatide, likely because absolute fat losses were larger. Visceral fat reductions were similar between the drugs.

Safety, the perennial concern with these gut-hormone mimetics, showed no decisive winner. Gastrointestinal adverse events, nausea and discontinuations due to side effects all trended slightly higher with tirzepatide 10 and 15 milligrams, but none of these differences reached statistical significance. All-cause discontinuation actually favored tirzepatide, with significantly lower odds in several of the adjusted analyses. The authors caution that the placebo arms of the two trials showed different rates of gastrointestinal events, which could subtly bias safety conclusions, and they note that the open-label design of the recent head-to-head SURMOUNT-5 trial may introduce more bias than the double-blind trials used here.

The findings arrive with important caveats. The evidence base rests on just two trials, comparisons were only possible at outcomes both trials reported, and the body composition analyses relied on smaller DEXA subgroups that could not be population-adjusted. The timepoints differed by four weeks, though outcomes had plateaued by then, and STEP 1 included European patients while SURMOUNT-1 did not. The study is also industry-sponsored, with most authors employed by Eli Lilly. Still, the results mirror the direct head-to-head SURMOUNT-5 trial, which found tirzepatide superior on weight and waist circumference, and they extend that picture to individual licensed doses and a far broader set of metabolic outcomes. For clinicians and patients weighing the two injectables, the emerging consensus is becoming hard to ignore: at 10 or 15 milligrams, tirzepatide appears to offer more weight loss, better body composition and broader cardiometabolic benefit than semaglutide 2.4 milligrams, with a safety profile that is, on the evidence so far, essentially equivalent.

Subject of Research: Indirect comparison of tirzepatide versus semaglutide efficacy and safety for obesity treatment in adults without type 2 diabetes

Article Title: Indirect comparative efficacy and safety of tirzepatide vs semaglutide for the treatment of obesity or overweight in patients without type 2 diabetes

Article References: Ciudin, A., Johansson, E., Zimner-Rapuch, S., Dimitriadis, G. K., Bertrand, M., Curteis, T., Clark, L. J., Fan, L., Sapin, H., & Bergmann, J.-F. (2026). Indirect comparative efficacy and safety of tirzepatide vs semaglutide for the treatment of obesity or overweight in patients without type 2 diabetes. International Journal of Obesity. https://doi.org/10.1038/s41366-026-02229-6

Image Credits: AI Generated

DOI: 10.1038/s41366-026-02229-6

Keywords: tirzepatide, semaglutide, obesity, weight loss, GLP-1 receptor agonist, GIP, indirect treatment comparison, SURMOUNT-1, STEP 1, cardiometabolic risk factors, body composition, International Journal of Obesity

Colorectal cancer is the third most commonly diagnosed cancer worldwide, accounting for roughly 9.6 percent of cancer cases and 9.3 percent of cancer deaths in 2022. Around 22 percent of patients already have metastatic disease at diagnosis, and about 70 percent will eventually relapse with metastatic spread, making metastatic colorectal cancer the leading cause of death from the disease. Within this population, an estimated 4 to 5 percent harbor microsatellite instability-high or mismatch repair deficient tumors, a molecular signature associated with worse outcomes than microsatellite stable disease but also with striking sensitivity to immune checkpoint inhibitors, because the defective DNA repair machinery generates abundant mutations that the immune system can recognize.

Until recently, pembrolizumab was the recommended first-line treatment for these patients, a position supported by the pivotal phase III KEYNOTE-177 trial, in which median progression-free survival reached 16.5 months with pembrolizumab compared with 8.2 months with chemotherapy at five years of follow-up. The efficacy of nivolumab plus ipilimumab in the same setting has been demonstrated in the ongoing phase III CheckMate-8HW trial, in which patients were randomized to the combination, nivolumab monotherapy, or investigator’s choice of chemotherapy. At a median follow-up of 31.5 months, 79 percent of first-line patients receiving the combination were progression-free at 12 months, compared with just 21 percent of those on chemotherapy, and median progression-free survival was not reached in the combination arm.

What has been missing is direct comparative evidence between the two immunotherapy regimens, since no trial has yet pitted nivolumab plus ipilimumab against pembrolizumab. To bridge that gap, the research team, led by investigators at OPEN Health with colleagues at Bristol Myers Squibb, constructed a three-state semi-Markov model simulating patients’ movement between progression-free health, progressed disease, and death over a lifetime horizon of approximately 33 years, from the perspective of a United States healthcare payer. A semi-Markov structure was chosen deliberately: because mature overall survival data for first-line patients in CheckMate-8HW were not yet available, the model needed to capture the dependency of post-progression survival on the time a patient spends before progressing, an approach previously accepted in multiple health technology assessments for metastatic colorectal cancer.

The technical heart of the study lies in its indirect treatment comparisons. Since pembrolizumab was not studied in CheckMate-8HW, the researchers anchored their comparison on chemotherapy, the common comparator in both CheckMate-8HW and KEYNOTE-177. Because the proportional hazards assumption was violated, a constant hazard ratio would have biased the results, so the base case used time-varying hazard ratios derived from parametric survival extrapolations fitted independently to each trial arm, with a generalized gamma distribution providing the best statistical fit. Published Kaplan-Meier curves from KEYNOTE-177 were digitized and reconstructed into pseudo-individual patient data using validated methodology. To test robustness, the team also applied three alternative methods: a fractional polynomial network meta-analysis, an anchored matching-adjusted indirect comparison adjusting for differences in age, performance status, mutation status, tumor sidedness, and metastatic sites, and an unanchored version of the same technique.

The base case results were striking. Nivolumab plus ipilimumab dominated pembrolizumab, meaning it produced both lower total costs and greater quality-adjusted life years, with cost savings of 58,819 dollars and a gain of 2.32 quality-adjusted life years. Compared with chemotherapy, the combination added 3.49 quality-adjusted life years at an incremental cost of 58,783 dollars, yielding an incremental cost-effectiveness ratio of 16,849 dollars per quality-adjusted life year, far below commonly accepted United States willingness-to-pay thresholds of 50,000 and 100,000 dollars per quality-adjusted life year. In the model, patients on the combination accrued 7.04 discounted quality-adjusted life years and 7.80 life years, versus 4.72 and 3.58 for pembrolizumab and 3.55 and 1.48 for chemotherapy. Total costs were 517,548 dollars for the combination, 576,367 dollars for pembrolizumab, and 458,765 dollars for chemotherapy, with the combination’s higher drug acquisition costs offset by lower disease management and monitoring costs, driven by its substantially reduced risk of progression.

The findings proved robust across the alternative comparison methods, with quality-adjusted life year gains for nivolumab plus ipilimumab over pembrolizumab ranging from 1.65 to 2.35 depending on the methodology employed. Scenario analyses varied survival distributions, patient populations, drug dosing assumptions, and cost inputs, and while dominance over pembrolizumab was not maintained in every cost variation, all resulting incremental cost-effectiveness ratios remained below accepted United States thresholds. In the probabilistic sensitivity analysis of 1,000 iterations, nivolumab plus ipilimumab had an 88 percent probability of being the most cost-effective option at a 50,000 dollar threshold. The parameters most influential on the results were the frequency of inpatient admissions with progressed disease, drug dosing, age at model onset, and the hazard ratio between the two immunotherapies.

The authors acknowledge important limitations. Mature overall survival data for first-line patients in CheckMate-8HW were unavailable, so post-progression survival was modeled using overall survival data from the phase II CheckMate-142 trial, which enrolled second-line patients and may have slightly overestimated mortality; assuming equal post-progression survival across treatments was a deliberately conservative choice that attributes all benefit to prolonged time to progression. Drug prices were based on list prices from the Redbook database, which do not reflect confidential discounts or rebates, and background mortality was age- and sex-adjusted from United States life tables, potentially underestimating deaths in the progression-free state. The model was nevertheless validated against external survival data and prior health technology assessment submissions, and its structure was recently accepted by the United Kingdom’s National Institute for Health and Care Excellence, which approved nivolumab plus ipilimumab as a first-line treatment for this population. Similar positive reimbursement decisions followed in Canada and the Netherlands.

The economic context amplifies the significance of these findings. Colorectal cancer carries the second highest global economic burden of all cancers, with total annual treatment costs in the United States of 10.5 billion dollars, and spending projected to rise by more than one billion dollars between 2017 and 2030 as the population ages and novel therapies enter practice. The study was funded by Bristol Myers Squibb, whose employees co-authored the work alongside OPEN Health researchers, a common arrangement in pharmacoeconomic research that readers should weigh when interpreting results. Nonetheless, the analysis concludes that first-line nivolumab plus ipilimumab represents a compelling and potentially cost-saving option for patients with microsatellite instability-high or mismatch repair deficient metastatic colorectal cancer, and the authors call for further research into comparative effectiveness against pembrolizumab and the costs of progressive disease as longer-term trial data mature.

Subject of Research: Cost-effectiveness of nivolumab plus ipilimumab versus pembrolizumab and chemotherapy as first-line treatment for microsatellite instability-high metastatic colorectal cancer

Article Title: Assessment of the Value of Nivolumab Plus Ipilimumab for Treating Microsatellite Instability-High/Mismatch Repair Deficiency Metastatic Colorectal Cancer: Indirect Treatment Comparison and Cost-Effectiveness Analysis

Article References: Bakker, L. J., Nickel, K., Cheah, Z., Sigworth, E. A., Blomaard, B. P. M., Villacorta, R., Jin, L., Davé, K., Simons, C. L., & Bennison, C. (2026). Assessment of the Value of Nivolumab Plus Ipilimumab for Treating Microsatellite Instability-High/Mismatch Repair Deficiency Metastatic Colorectal Cancer: Indirect Treatment Comparison and Cost-Effectiveness Analysis. Advances in Therapy. https://doi.org/10.1007/s12325-026-03787-x

Image Credits: AI Generated

DOI: 10.1007/s12325-026-03787-x

Keywords: metastatic colorectal cancer, microsatellite instability-high, mismatch repair deficiency, nivolumab, ipilimumab, pembrolizumab, immunotherapy, cost-effectiveness, indirect treatment comparison, semi-Markov model, quality-adjusted life years, CheckMate-8HW