Tirzepatide

Why this rating

Direct evidence from a large, pre-registered, well-designed RCT (SURMOUNT-1, n=2,539) with a clear dose–response and a clinically meaningful effect size at all three doses (16.0% / 21.4% / 22.5% vs 2.4% placebo). About 9 in 10 participants on tirzepatide lost weight. No serious GRADE downgrade factors apply.

Why this rating

Two pre-registered randomized withdrawal RCTs now converge. In SURMOUNT-4 (n=670), continued tirzepatide produced a further −5.5% weight change from week 36 to 88 while placebo regained +14.0%, and 89.5% of those who continued maintained ≥80% of their lead-in loss vs 16.6% on placebo. SURMOUNT-MAINTAIN (Lancet 2026, n=378 randomized to maintenance) replicated the pattern: after a 60-week open-label lead-in, weight change to week 112 was −21.9% on the maximum tolerated dose and −16.6% on 5mg, vs −9.9% on placebo (treatment differences −12.0 and −6.6 percentage points, both p<0.0001). The consistent direction and magnitude across two independent trials support a high-certainty rating with no downgrades.

Why this rating

Two independent 52-week phase 3, double-blind, placebo-controlled RCTs (SURMOUNT-OSA) showed consistent, large reductions in the apnea-hypopnea index, the standard measure of OSA severity. In adults not using PAP, AHI fell 25.3 events/hour on tirzepatide vs 5.3 on placebo (treatment difference −20.0, p<0.001); in adults on PAP, 29.3 vs 5.5 (treatment difference −23.8, p<0.001). Built-in replication across two trials, a large direct effect, and concurrent improvements in body weight, hypoxic burden, and blood pressure. This evidence supported the FDA's 2024 approval of tirzepatide (Zepbound) for OSA in adults with obesity. No serious GRADE downgrade applies.

Why this rating

Direct evidence from a large, pre-registered, double-blind, placebo-controlled phase 3 RCT (SURMOUNT-2, n=938). At 72 weeks, mean body-weight change was −12.8% (10mg) and −14.7% (15mg) vs −3.2% on placebo (treatment differences −9.6 and −11.6 percentage points, both p<0.0001). Weight loss in people with type 2 diabetes runs somewhat smaller than in people without diabetes (compare the up-to −22.5% in SURMOUNT-1), a consistent pattern across the GLP-1 class. The effect is large, direct, and consistent with the broader SURMOUNT and SURPASS program, so no serious downgrade applies.

Why this rating

Glycemic benefit is established across the five-trial SURPASS phase 3 program, all consistently positive. In the head-to-head SURPASS-2 RCT (n=1,879, 40 weeks), tirzepatide reduced HbA1c by 2.01% (5mg), 2.24% (10mg), and 2.30% (15mg) versus 1.86% for semaglutide 1mg, meeting both noninferiority and superiority at every dose (treatment differences −0.15 to −0.45 percentage points). HbA1c is the standard, objective measure of glycemic control, so the open-label design of SURPASS-2 poses little risk of bias for this lab endpoint. The effect is large, direct, and heavily replicated, so no serious GRADE downgrade applies.