A living network meta-analysis of 869 trials and 493,168 participants confirms tirzepatide produces the most weight loss (-8.6 kg), GLP-1 RAs and SGLT-2 inhibitors provide cardiovascular and kidney protection, while gastrointestinal side effects remain the main trade-off for peptide-based therapies.
-8.63 kg with tirzepatideThe largest weight reduction among all 63 diabetes drugs compared across 869 randomized trials with 493,168 participants
What the researchers found
From 869 trials with 493,168 participants covering 63 drugs across 13 classes:
Weight loss leaders: tirzepatide (-8.63 kg, moderate certainty) and orforglipron (-7.87 kg, low certainty), followed by 8 other GLP-1 RAs (high to moderate certainty).
Cardiovascular/kidney benefits: SGLT-2 inhibitors, GLP-1 RAs, and finerenone confirmed with moderate to high certainty evidence.
Key harms: tirzepatide and GLP-1 RAs probably increase severe gastrointestinal events (tirzepatide OR 4.21, moderate certainty). SGLT-2 inhibitors increase genital infections (OR 3.29) and ketoacidosis (OR 2.08). Sulfonylureas and insulin increase hypoglycemia.
Uncertainty remains about whether GLP-1 RAs reduce dementia (OR 0.92, low certainty) and effects on neuropathy and visual impairment.
Why it matters
With millions of people taking diabetes medications and billions in annual drug spending, clinicians need reliable comparative evidence to guide treatment decisions. This living NMA provides the most comprehensive drug comparison available, with risk-stratified absolute effects that allow doctors to match drug choice to each patient's cardiovascular and kidney risk profile. The confirmation that peptide-based therapies (GLP-1 RAs, tirzepatide) lead for weight loss while providing cardiovascular benefits — offset by gastrointestinal harms — is essential information for the growing population using these drugs.
How the study worked
Living systematic review and network meta-analysis using frequentist random effects models and GRADE evidence assessment. Medline and Embase searched through July 2024. Included 869 randomized controlled trials (≥24 weeks) comparing diabetes medications with standard treatment, placebo, or each other. 493,168 participants, 13 drug classes (63 drugs), and 26 outcomes analyzed. Risk-stratified absolute effects generated via interactive tool. Registered on PROSPERO (CRD42022325948). Updates planned at least twice yearly.
What this study cannot tell us
Despite 869 trials, evidence for some outcomes (neuropathy, visual impairment, dementia) remains low certainty. The NMA methodology relies on indirect comparisons where head-to-head data are unavailable. The search through July 2024 may miss recent trials. Not all drug combinations were evaluable. The gastrointestinal harm analysis may not fully capture the dose-titration strategies that mitigate side effects in clinical practice. Cost-effectiveness was not assessed.
How to read the evidence
This is a living systematic review and network meta-analysis published in BMJ — the highest tier of evidence synthesis. GRADE certainty ratings range from high to very low depending on the specific outcome, with cardiovascular and weight outcomes supported by moderate to high certainty evidence.
When this study was published
Published in 2025 with evidence through July 2024 and planned biannual updates, this is a living review that will remain current as new trials report. It represents the definitive comparative evidence base for diabetes pharmacotherapy.
The bigger picture
This landmark living review, linked to BMJ Rapid Recommendations, represents the gold standard for comparative diabetes pharmacotherapy evidence. Its confirmation that peptide-based incretin therapies (GLP-1 RAs and tirzepatide) are among the most effective drugs for weight loss while providing cardiovascular benefits positions these peptides as cornerstone therapies. The living review format ensures recommendations stay current as new trials report — critical in a field where evidence evolves rapidly.
Questions still open
- Will future updates incorporating more tirzepatide and oral GLP-1 RA data change the relative ranking of these drugs?
- Can the suggested dementia reduction with GLP-1 RAs be confirmed with high-certainty evidence in dedicated trials?
- How should clinicians weigh the weight loss advantage of tirzepatide against its 4-fold increase in severe GI events?
Common questions
Which diabetes drug causes the most weight loss?
What are the main downsides of GLP-1 drugs and tirzepatide?
Read the original research
Medications for adults with type 2 diabetes: a living systematic review and network meta-analysis.
BMJ (Clinical research ed.), 390, e083039
Citation
Nong, Kailei; Jeppesen, Britta Tendal; Shi, Qingyang; Agoritsas, Thomas; Guyatt, Gordon H; White, Heath; Gao, Yiyuan; Agarwal, Arnav; Macdonald, Helen; Zou, Xinyu; Millard, Tanya; Schnell, Oliver; Marx, Nikolaus; Brosius, Frank C; McDonald, Steve; Quigley, Matthew; Tian, Xin; Fan, Qinlin; White, Barbara; Mao, Yunhe; Pan, Xiaohui; Liu, Changhai; Zhai, Chunjuan; Yuan, Chi; Li, Qiang; An, Jing; Gan, Yu; Wang, Yanyan; Jin, Yinghui; Sun, Feng; Zhu, Zhiming; Rydén, Lars; Standl, Eberhard; Turner, Tari; Vandvik, Per Olav; Li, Sheyu. (2025). Medications for adults with type 2 diabetes: a living systematic review and network meta-analysis.. BMJ (Clinical research ed.), 390, e083039. https://doi.org/10.1136/bmj-2024-083039