Zepbound Evidence Base Graded by GRADE: What the Clinical Data Actually Show

What a GRADE rating actually measures

GRADE (Grading of Recommendations Assessment, Development and Evaluation) is a framework for stating how much confidence to place in an effect estimate. Randomized trials start at high confidence and are downgraded for risk of bias, inconsistency, indirectness, imprecision and publication bias. They can be upgraded for a very large effect or a dose-response gradient. The rating describes the quality of the evidence, not how good the drug is.

SURMOUNT-1 was double-blind and placebo-controlled, with coprimary endpoints and a treatment-regimen estimand [1]. The effect was large, the confidence intervals were narrow, and the three doses produced a stepwise gradient [1].

How each part of the evidence base rates

The ratings below are this article's application of those criteria to the trials described. They are not a rating issued by any regulator.

• Weight reduction in adults with obesity and no diabetes: high. SURMOUNT-1 was double-blind and placebo-controlled, the effect was large, and the three doses produced a stepwise gradient [1].
• Weight reduction in adults with obesity and type 2 diabetes: high, from a 938-person double-blind placebo-controlled trial [5].
• Holding weight loss while treatment continues: high for the double-blind randomized withdrawal comparison through week 88 [6].
• Tirzepatide against semaglutide: moderate, because SURMOUNT-5 was randomized but open-label [3].
• The fat-versus-lean split of the weight lost: low, from one 160-person substudy, and DXA cannot separate muscle from other lean tissue [7].
• Cardiovascular outcomes in adults with obesity and no diabetes: low, because the outcomes trial described here enrolled adults with type 2 diabetes and established cardiovascular disease [8].
• Obstructive sleep apnea: high, from two double-blind placebo-controlled trials with large reductions in the apnea-hypopnea index [2].
• Side effects: high for common gastrointestinal effects from the pooled placebo-controlled trials [4]; low for rare harms: adjudicated pancreatitis was reported in 0.2% of both tirzepatide-treated and placebo-treated patients, and the thyroid C-cell warning rests on rat studies [4].

SURMOUNT-1: the anchor weight-loss trial

SURMOUNT-1 randomized 2,539 adults with a BMI of 30 or higher, or 27 or higher plus at least one weight-related complication, and without diabetes, to weekly tirzepatide 5 mg, 10 mg or 15 mg or placebo for 72 weeks, including a 20-week dose-escalation period [1]. Mean baseline weight was 104.8 kg and mean BMI 38.0 [1]. The coprimary endpoints were percentage weight change and the proportion losing at least 5% [1].

Mean weight change at week 72 was minus 15.0% (95% CI minus 15.9 to minus 14.2) with 5 mg, minus 19.5% (minus 20.4 to minus 18.5) with 10 mg and minus 20.9% (minus 21.8 to minus 19.9) with 15 mg, versus minus 3.1% (minus 4.3 to minus 1.9) with placebo (P<0.001 for all comparisons) [1].

At least 5% weight loss was reached by 85%, 89% and 91% of the three dose groups versus 35% on placebo [1]. At least 20% was reached by 30.0% on 5 mg, 50.1% on 10 mg and 56.7% on 15 mg versus 3.1% on placebo, and 14.3%, 16.4% and 15.1% of the dose groups discontinued study drug versus 26.4% on placebo [4].

Is the weight lost mostly muscle?

A body-composition substudy of SURMOUNT-1 measured 160 participants with DXA scans at baseline and week 72: 124 on pooled tirzepatide doses and 36 on placebo, 73% female, mean weight 102.5 kg [7]. Over 72 weeks, body weight fell 21.3%, fat mass 33.9% and lean mass 10.9% with tirzepatide, versus 5.3%, 8.2% and 2.6% with placebo (p<0.001 for each) [7]. Roughly 75% of the weight lost was fat mass and about 25% lean mass, the same split seen on placebo [7].

That is a 160-person substudy, and DXA does not separate muscle from other lean tissue, so the lean-mass figure covers more than muscle [7].

SURMOUNT-2: obesity with type 2 diabetes

SURMOUNT-2 randomized 938 adults with a BMI of 27 or higher and HbA1c between 7% and 10% to tirzepatide 10 mg, 15 mg or placebo for 72 weeks [5]. Mean weight change was minus 12.8% with 10 mg and minus 14.7% with 15 mg versus minus 3.2% with placebo, and at least 5% weight loss was reached by 79.2%, 82.8% and 32.5% respectively [4][5]. For reference, SURMOUNT-1 reported minus 20.9% at 15 mg; these were different trials in different populations, so this is not a head-to-head comparison [1][5].

Head-to-head with semaglutide

SURMOUNT-5 is a phase 3b, open-label trial of 751 adults with obesity and no type 2 diabetes, randomized to the maximum tolerated dose of tirzepatide (10 mg or 15 mg) or the maximum tolerated dose of semaglutide (1.7 mg or 2.4 mg) for 72 weeks [3]. Mean weight change was minus 20.2% with tirzepatide and minus 13.7% with semaglutide (P<0.001), and waist circumference fell 18.4 cm versus 13.0 cm [3]. Participants on tirzepatide were more likely to lose at least 10%, 15%, 20% and 25% [3].

Earlier, in type 2 diabetes, SURPASS-2 compared tirzepatide 5, 10 and 15 mg with semaglutide 1 mg over 40 weeks in 1,879 adults on metformin [9]. Weight fell 7.6 kg, 9.3 kg and 11.2 kg versus 5.7 kg [4], and HbA1c fell 2.01, 2.24 and 2.30 percentage points versus 1.86 [9]. That trial was also open-label, and its semaglutide dose (1 mg) was lower than the 1.7 or 2.4 mg used in SURMOUNT-5, so it answers a different question [3][9].

SURMOUNT-4: what happens when treatment stops

SURMOUNT-4 was a randomized-withdrawal trial at 70 sites in four countries [6]. All 783 participants received open-label tirzepatide at the maximum tolerated dose (10 mg or 15 mg) for 36 weeks; the 670 who completed that period had a mean weight reduction of 20.9% and were randomized to continue tirzepatide or switch to placebo for 52 weeks [6]. From week 36 to week 88, mean weight change was minus 5.5% for those who continued and plus 14.0% for those switched to placebo, a between-group difference of 19.4 percentage points [6].

Sleep apnea: two placebo-controlled trials

Two phase 3, double-blind trials enrolled adults with moderate-to-severe obstructive sleep apnea and obesity: trial 1 in people not using positive airway pressure at baseline (mean apnea-hypopnea index 51.5 events per hour) and trial 2 in people using it (mean 49.5) [2]. Over 52 weeks, the apnea-hypopnea index fell 25.3 events per hour with tirzepatide versus 5.3 with placebo in trial 1 (difference 20.0) and 29.3 versus 5.5 in trial 2 (difference 23.8) [2]. The sleep apnea indication was approved on 20 December 2024, with maintenance doses of 10 mg or 15 mg [4].

Cardiovascular outcomes: one trial, one population

SURPASS-CVOT was a double-blind noninferiority outcomes trial in 13,165 adults with type 2 diabetes and atherosclerotic cardiovascular disease in the modified intention-to-treat population [8]. Participants received tirzepatide up to 15 mg or dulaglutide 1.5 mg [8]. The composite of cardiovascular death, myocardial infarction or stroke occurred in 12.2% versus 13.1% (hazard ratio 0.92, 95.3% CI 0.83 to 1.01); noninferiority was met (p=0.003) and superiority was not (p=0.09) [8].

Those results belong to adults with diabetes and established cardiovascular disease [8]. The Zepbound indications are weight reduction and obstructive sleep apnea in adults with obesity [4].

Safety evidence and how firm it is

The label pools two placebo-controlled weight-management trials (SURMOUNT-1 and SURMOUNT-2): 958 participants on placebo and 630, 948 and 941 on 5 mg, 10 mg and 15 mg, treated for up to 72 weeks [4]. Nausea occurred in 8% on placebo versus 25%, 29% and 28%; diarrhea in 8% versus 19%, 21% and 23%; vomiting in 2% versus 8%, 11% and 13%; constipation in 5% versus 17%, 14% and 11% [4]. Any gastrointestinal adverse reaction occurred in 56% on every tirzepatide dose versus 30% on placebo [4]. Treatment was stopped for an adverse reaction by 4.8%, 6.3% and 6.7% versus 3.4%, and for a gastrointestinal reaction by 1.9%, 3.3% and 4.3% versus 0.5% [4]. Severe gastrointestinal reactions were reported in 1.7%, 2.5% and 3.1% versus 1% [4]. Most nausea, vomiting and diarrhea occurred during dose escalation and decreased over time [4].

Rarer events in the same pool: cholecystitis 0.7% versus 0.2%, adjudicated acute pancreatitis 0.2% versus 0.2%, and severe hypersensitivity reactions 0.1% versus 0% [4]. Hair loss was reported by 5% versus 1%, was associated with weight reduction, and was far more common in women (7.1%) than men (0.5%); no Zepbound-treated patient stopped treatment for it [4]. Heart rate rose by a mean of 1 to 3 beats per minute and hypotension occurred in 1.6% versus 0.1% [4].

The boxed warning on thyroid C-cell tumors rests on rat studies at clinically relevant exposures; whether it applies to humans, including medullary thyroid carcinoma, is undetermined [4]. Zepbound is contraindicated in people with a personal or family history of medullary thyroid carcinoma or multiple endocrine neoplasia syndrome type 2 [4]. Pregnancy is not a contraindication: the label states that weight loss offers no benefit in pregnancy and may cause fetal harm, and that Zepbound should be discontinued when pregnancy is recognized [4]. The label also notes that adverse reaction rates from clinical trials of one drug cannot be directly compared with those of another [4].

What the trials cover, and for how long

Zepbound has no pediatric indication; its label says safety and effectiveness in pediatric patients have not been established [4].

Treatment periods run from 52 weeks in the two sleep apnea trials [2] to 72 weeks in SURMOUNT-1, SURMOUNT-2 and SURMOUNT-5 [1][5][3], and 88 weeks in SURMOUNT-4, which combined 36 weeks of open-label treatment with a 52-week double-blind comparison [6].