When every drug in a class comparison gets an odds ratio, the temptation is to read a league table — the same temptation the wider drug menu invites. This one has the oldest and cheapest drug at the top, which is a useful reason not to.
What was pooled
Ninety-three randomized trials of at least 40 weeks, comparing diabetes medications against placebo or an active comparator, with cardiovascular events and deaths formally adjudicated. [1] Both pairwise and network methods were used, so drugs never tested against each other could still be compared indirectly.
In the network analysis against placebo, metformin came in at OR 0.69, 95% CI 0.53 to 0.89; SGLT2 inhibitors at 0.82, 95% CI 0.75 to 0.90; tirzepatide at 0.82, 95% CI 0.72 to 0.93; and GLP-1 agonists at 0.87, 95% CI 0.83 to 0.92. Insulin, DPP-4 inhibitors and sulfonylureas showed no significant effect on events in the network.
On all-cause mortality, tirzepatide came in at 0.72, 95% CI 0.64 to 0.82, with SGLT2 inhibitors and GLP-1 agonists both at 0.85, 95% CI 0.80 to 0.91. In the pairwise analyses, sulfonylureas were associated with increased cardiovascular events — the only class that moved the wrong way.
Why this is not a ranking
Look at the four intervals rather than the four point estimates. Metformin runs 0.53 to 0.89. SGLT2 inhibitors 0.75 to 0.90. Tirzepatide 0.72 to 0.93. GLP-1 agonists 0.83 to 0.92. Almost all of that overlaps.
An analysis can produce an ordering while being unable to distinguish the things it ordered, and the ordering is what gets quoted — the same problem as a pooled average that describes nobody. What these numbers support is that four classes reduce cardiovascular events and three reduce death. What they do not support is a first place.
The population, and who is reading
Everyone in these 93 trials had type 2 diabetes. Someone buying a GLP-1 for weight loss without diabetes is not in this evidence, and the cardiovascular benefit in that group rests on different trials with different populations.
That distinction matters more here than usual, because metformin is not a weight-loss drug and is not sold as one. A table in which it leads on cardiovascular events is a table about diabetes care, not a shopping list — and a comparison on the wrong axis is how a drug ends up credited for something it was never tested for.
What a reader can use
One transferable idea. When a comparison puts an old generic near the top, that is usually a signal about evidence rather than about pharmacology — old drugs have old trials, and old trials have wider intervals and looser methods.
It cuts both ways. It means the generic’s apparent advantage should be discounted. It also means the newer drug’s advantage over it has not been demonstrated, because nobody ran that trial either — an indirect comparison is not a head-to-head one. Almost nothing a seller publishes distinguishes the two, which the disclosure scorecard counts.