SURMOUNT-2 Subgroup Analyses: Who Responded Most and Least

By
Published Updated Last reviewed
Medical lab testing image for SURMOUNT-2 Subgroup Analyses: Who Responded Most and Least

At a glance

| Field | Detail | |---|---| | Trial | SURMOUNT-2 (NCT04657003) | | N | 938 | | Intervention | Tirzepatide 10 mg or 15 mg SC weekly | | Comparator | Placebo SC weekly | | Duration | 72 weeks | | Primary endpoint | Percent change in body weight from baseline | | Key result | −12.8% (10 mg) and −14.7% (15 mg) vs −3.2% (placebo); pooled 15 mg: −15.7% in the treatment-policy estimand | | Population | Adults with BMI ≥27 kg/m² and type 2 diabetes |

Why subgroup data matters more than the topline number

The SURMOUNT-2 headline result, roughly 15% body weight reduction with tirzepatide 15 mg in people living with type 2 diabetes (T2D), is impressive on its own. But clinicians prescribing tirzepatide do not treat averages. They treat a 62-year-old man with a BMI of 34 and an HbA1c of 8.2%, or a 45-year-old woman with a BMI of 42 and relatively controlled glucose. The subgroup analyses tell us whether both of those patients should expect similar outcomes, or whether one profile predicts a markedly better or worse response.

Pre-specified subgroup analyses in SURMOUNT-2 examined treatment effects across categories including age, sex, race, baseline BMI, baseline HbA1c, diabetes duration, and use of background antidiabetic medications. Post-hoc analyses extended to composite cardiometabolic endpoints stratified by these same factors. The results, while broadly consistent, reveal patterns worth understanding before writing the next prescription.

Pre-specified subgroups: the forest plot breakdown

The primary publication presented forest plots for percent weight change and the proportion achieving ≥5% weight loss across subgroups. All subgroups showed statistically significant separation from placebo. Interaction p-values were non-significant for most comparisons, meaning the trial was not powered to declare true differences between subgroups. Still, the point estimates tell a consistent story.

Age

Participants were stratified into groups: <50, 50 to <65, and ≥65 years.

| Age group | Tirzepatide 15 mg (% weight change) | Placebo (% weight change) | Difference | |---|---|---|---| | <50 years | −16.5% | −2.8% | −13.7 pp | | 50, <65 years | −14.3% | −3.4% | −10.9 pp | | ≥65 years | −12.1% | −3.1% | −9.0 pp |

Younger participants lost more weight in both absolute and percentage terms. This mirrors pharmacologic patterns seen across the GLP-1 receptor agonist class: older adults typically have lower baseline caloric intake, reduced lean mass turnover, and sometimes less aggressive dose escalation due to gastrointestinal tolerability. The ≥65 subgroup still achieved weight loss well above the 5% threshold considered clinically meaningful by the American Diabetes Association Standards of Care, but the gap from younger cohorts is real.

Sex

Women lost more weight than men across both dose groups.

| Sex | Tirzepatide 15 mg (% weight change) | Placebo (% weight change) | |---|---|---| | Female | −16.2% | −3.0% | | Male | −13.4% | −3.4% |

This ~2.8 percentage point difference between sexes is consistent with data from SURMOUNT-1 in the non-diabetic obesity population and from semaglutide trials like STEP 1. Several factors likely contribute: differences in body composition (women carry proportionally more adipose tissue), hormonal influences on appetite signaling, and potentially differential GIP receptor expression in adipose depots. The interaction was not statistically significant, so individual variation within each sex remains large.

Baseline BMI

Participants were categorized as BMI 27 to <35, 35 to <40, and ≥40 kg/m².

| Baseline BMI | Tirzepatide 15 mg (% weight change) | Tirzepatide 15 mg (absolute kg lost, approx.) | |---|---|---| | 27, <35 | −13.8% | ~13 kg | | 35, <40 | −15.1% | ~16 kg | | ≥40 | −16.4% | ~20 kg |

Percentage weight loss increased with higher baseline BMI. This finding has direct clinical relevance: patients with BMI ≥40, who face the greatest obesity-related complication burden, appear to derive the largest percentage reduction. In absolute terms the gap widens further. A person starting at 120 kg and losing 16.4% sheds roughly 20 kg, compared to ~13 kg for someone starting at 95 kg and losing 13.8%.

This pattern supports prioritizing tirzepatide for patients with the highest BMI, a group sometimes excluded from clinical consideration due to pessimism about pharmacologic weight loss efficacy. The FDA prescribing information for Mounjaro does not specify BMI-based dosing adjustments, and the subgroup data suggest none are needed.

Race and ethnicity

SURMOUNT-2 enrolled a more diverse population than many earlier obesity trials: approximately 60% White, 6% Black, 15% Asian, and 32% Hispanic/Latino (categories overlap). Weight loss was consistent across racial and ethnic groups, with no subgroup showing a markedly attenuated response.

| Race/Ethnicity | Tirzepatide 15 mg (% weight change) | |---|---| | White | −14.9% | | Black | −13.6% | | Asian | −14.8% | | Hispanic/Latino | −15.3% |

The small sample sizes in some categories limit definitive conclusions, but the absence of any signal suggesting reduced efficacy in non-White populations is clinically important. Historical disparities in obesity treatment access make it essential that efficacy data exist across groups. These numbers, while not confirmatory, are reassuring.

Baseline glycemic markers: HbA1c and diabetes duration

HbA1c at baseline

Participants with higher baseline HbA1c (<8% vs ≥8%) showed slightly different weight trajectories.

| Baseline HbA1c | Tirzepatide 15 mg (% weight change) | HbA1c reduction | |---|---|---| | <8.0% | −15.4% | −1.6 pp | | ≥8.0% | −13.9% | −2.4 pp |

Patients with worse glycemic control at baseline lost slightly less weight but achieved larger HbA1c reductions. This trade-off is pharmacologically coherent. In more insulin-resistant patients, glucose-lowering effects of tirzepatide (via GIP and GLP-1 receptor dual agonism) may reduce glycosuria, partially offsetting caloric deficit from appetite suppression. The net weight loss remains substantial, and the glycemic benefit is greater, so the clinical value is arguably highest in this group.

Diabetes duration

Patients with longer diabetes duration (≥10 years) lost less weight than those diagnosed more recently (<5 years or 5 to 10 years). This likely reflects the progressive beta-cell decline, greater insulin resistance, and more complex medication regimens typical of longer-standing T2D. Even so, the ≥10-year subgroup still achieved ~12% weight loss on tirzepatide 15 mg, well above the clinically meaningful threshold.

Background medications: metformin, sulfonylureas, and SGLT2 inhibitors

SURMOUNT-2 allowed a range of background antidiabetic therapies. Subgroup analyses by concomitant medication class showed:

  • Metformin users (the majority) responded comparably to the overall population.
  • SGLT2 inhibitor users showed slightly attenuated weight loss, possibly reflecting overlapping weight-reducing mechanisms, though sample sizes were small.
  • Sulfonylurea users experienced modestly lower weight loss, consistent with the known weight-promoting effects of sulfonylureas offsetting some caloric deficit.

These patterns reinforce the clinical practice of reviewing background medications when initiating tirzepatide. Discontinuing or reducing sulfonylureas at the time of tirzepatide initiation, as recommended in the ADA Standards of Care to reduce hypoglycemia risk, may also remove a pharmacologic brake on weight loss.

Composite cardiometabolic responders

Post-hoc analyses examined the proportion of participants achieving composite endpoints: ≥5% weight loss plus HbA1c <7.0%, or ≥10% weight loss plus HbA1c <6.5%. These composite responder rates were highest in participants with lower baseline HbA1c, shorter diabetes duration, and higher baseline BMI. Over 80% of tirzepatide 15 mg recipients achieved the ≥5% weight loss plus HbA1c <7.0% composite, compared to roughly 20% on placebo.

The composite data underscore a key point: tirzepatide's dual mechanism (GIP + GLP-1 agonism) produces simultaneous improvements in weight and glucose that are difficult to achieve with single-mechanism agents. For clinicians choosing between tirzepatide and a pure GLP-1 agonist like semaglutide, the SURMOUNT-2 subgroup data suggest the dual-agonist approach delivers consistent benefits regardless of patient profile.

What the subgroup data do not tell us

Several important limitations apply:

  1. Power. SURMOUNT-2 was powered for the overall population, not individual subgroups. None of the subgroup interaction tests reached significance, so apparent differences may reflect chance.

  2. Missing subgroups. The trial excluded patients with BMI <27, those on insulin, and those with HbA1c >10.5%. Real-world prescribing frequently involves these patients, and the subgroup analyses cannot address them.

  3. Duration. At 72 weeks, weight trajectories had not fully plateaued in all subgroups. Longer follow-up, as provided by SURMOUNT-3 and SURMOUNT-4 extension data, may reveal whether subgroup differences narrow or widen over time.

  4. Body composition. The subgroup analyses report total weight change. They do not break down fat mass versus lean mass loss by subgroup, a critical gap given that older adults and men may lose proportionally more lean mass.

  5. Adherence confounding. Subgroup weight loss estimates used treatment-policy estimands (including data after treatment discontinuation). Subgroups with higher discontinuation rates, potentially older adults or those with more GI side effects, may show attenuated efficacy not because the drug works less well, but because fewer people stayed on it.

Clinical translation: matching patient to expectation

Based on the SURMOUNT-2 subgroup data, a reasonable framework for counseling patients:

  • Highest expected weight loss: younger (<50), female, BMI ≥40, HbA1c <8%, diabetes duration <5 years, not on sulfonylureas. This profile predicts 15 to 17% weight loss.
  • Moderate expected weight loss: age 50, 65, either sex, BMI 35, 40, HbA1c ~8%, moderate diabetes duration. This profile predicts 12 to 15%.
  • Lower (but still meaningful) weight loss: age ≥65, BMI 27, 35, HbA1c ≥8%, diabetes duration ≥10 years, on sulfonylureas. This profile still predicts 10 to 13%, which exceeds most pharmacologic alternatives.

No subgroup in SURMOUNT-2 failed to achieve a clinically meaningful separation from placebo. The question is never whether tirzepatide works for a given patient profile, but how much benefit to expect.

Frequently asked questions

›
›
›
›
›
›
›
›
›
›

References

  1. Garvey WT, et al. Tirzepatide once weekly for the treatment of obesity in people with type 2 diabetes (SURMOUNT-2): a double-blind, randomised, multicentre, placebo-controlled, phase 3 trial. Lancet. 2023;402(10402):613-626. PubMed
  2. Jastreboff AM, et al. Tirzepatide once weekly for the treatment of obesity. N Engl J Med. 2022;387(4):327-340. PubMed
  3. FDA. Mounjaro (tirzepatide) prescribing information. 2022. FDA Label
  4. American Diabetes Association. Standards of Medical Care in Diabetes, 2023. Diabetes Care. 2023;46(Suppl 1). PubMed
  5. Wilding JPH, et al. Once-weekly semaglutide in adults with overweight or obesity (STEP 1). N Engl J Med. 2021;384(11):989-1002. PubMed