Comparative Effects of Obesity Drugs in Adults With Overweight or Obesity
Key Takeaways
- Tirzepatide and cagrilintide-semaglutide produced the largest one-year weight-loss estimates versus lifestyle modification alone in the network.
- Subcutaneous semaglutide was the only drug associated with lower all-cause mortality and myocardial infarction, while no drug exceeded the prespecified minimally important quality-of-life difference.
- Several agents were associated with more discontinuation and gastrointestinal events, with additional signals for fatigue and gallbladder-related disorders.
The review included 262 randomized trials involving 99,791 participants and evaluated 19 obesity drugs over 12 to 172 weeks. Participants were adults with overweight or obesity, with or without cardiovascular or metabolic complications, and lifestyle modification alone served as the main reference comparator. Median follow-up across trials was 26 weeks, underscoring that many estimates reflected relatively short observation periods. Investigators used frequentist random-effects models, Bayesian dose-response modeling, GRADE, and RoB 2 assessments in a single framework. That approach separated findings across efficacy, cardiovascular outcomes, quality of life, and harms.
At one year, estimated body-weight change was -14.9% with tirzepatide (95% CI -16.0% to -13.9%) and -14.8% with cagrilintide-semaglutide (-16.9% to -12.7%). Corresponding estimates were -10.9% with oral semaglutide (-12.7% to -9.1%), -9.9% with orforglipron (-12.4% to -7.5%), -9.8% with subcutaneous semaglutide (-10.6% to -9.1%), and -8.1% with phentermine-topiramate (-9.7% to -6.5%). Emerging agents including ecnoglutide, mazdutide, and retatrutide also showed large estimated effects, although certainty was low or very low. Among body-composition outcomes, tirzepatide reduced fat mass by 25.7% and lean mass by 8.3%, the largest changes in both categories. Larger weight reductions generally clustered with newer agents, although certainty was stronger for the leading estimates than for several emerging drugs.
Subcutaneous semaglutide was the only drug associated with lower all-cause mortality and myocardial infarction, with risk ratios of 0.81 (95% CI 0.72 to 0.93) and 0.72 (0.61 to 0.85). Heart-failure risk was also lower with subcutaneous semaglutide, at 0.43 (0.21 to 0.84), and with tirzepatide, at 0.49 (0.27 to 0.88). No drug convincingly reduced kidney failure, although subcutaneous semaglutide probably reduced kidney disease progression, and the mortality and myocardial infarction estimates were largely informed by trials in high-risk populations. Quality-of-life results remained below the prespecified 10-point minimally important difference, with all mean differences under 5 points.
Discontinuation because of adverse events was higher among the highest-risk agents, with risk ratios ranging from 1.9 to 4.2. Gastrointestinal adverse-event incidence rate ratios ranged from 3.1 to 4.2, and naltrexone-bupropion showed the clearest fatigue signal at 8.9, or 331 more cases per 1000 over one year. Gallbladder-related disorders increased most with cagrilintide-semaglutide, oral semaglutide, liraglutide, and tirzepatide, and greater weight loss generally came with more adverse events and discontinuation. Most trials were relatively short, few extended beyond two years, evidence for several newer agents was sparse and low certainty, and trial populations may not fully reflect real-world patients.