Were the SUSTAIN-7 subgroup analyses pre-specified or post-hoc?

The primary subgroup analyses (age, sex, race, BMI, baseline A1C, diabetes duration) were pre-specified in the statistical analysis plan. Some sensitivity and composite analyses were conducted post-hoc. Pre-specification reduces the risk of selective reporting but does not eliminate the multiplicity problem inherent in subgroup testing.

Did any subgroup favor dulaglutide over semaglutide?

No. In every pre-specified subgroup for both A1C and weight endpoints, the point estimate favored semaglutide. The confidence intervals for some smaller subgroups crossed zero, meaning the difference was not statistically significant within that stratum, but no subgroup reversed direction.

How large was the weight loss difference in patients with BMI over 30?

In the high-dose comparison (semaglutide 1.0 mg vs dulaglutide 1.5 mg), patients with BMI ≥30 lost approximately 4 kg more with semaglutide. In the low-dose comparison, the difference was approximately 2.8 kg. Both are clinically meaningful over a 40-week period.

Was the trial long enough to capture peak weight loss?

Probably not. GLP-1 receptor agonist weight loss typically plateaus between 60 and 68 weeks. The 40-week SUSTAIN-7 timeframe likely captured most of the glycemic benefit but may underestimate the ultimate weight separation between agents.

Why was the race subgroup pooled as white vs non-white?

The trial enrolled relatively few Black (6%) and Hispanic participants. Pooling non-white subgroups increased statistical power but obscured potentially important differences between racial and ethnic groups. This is a recognized limitation of many GLP-1 receptor agonist trials.

Should these subgroup results change which GLP-1 agonist I prescribe?

They reinforce semaglutide's broad advantage but should not be the sole basis for prescribing. Formulary access, cost, patient injection preferences, tolerability, and cardiovascular outcome data (from SUSTAIN-6 for semaglutide and REWIND for dulaglutide) all factor into the decision.

How does SUSTAIN-7 compare to other head-to-head GLP-1 trials?

SUSTAIN-7 was one of the first large head-to-head GLP-1 trials. The later AWARD-11 trial compared higher dulaglutide doses (3.0 mg and 4.5 mg) against 1.5 mg but did not include semaglutide. No subsequent trial has directly repeated the SUSTAIN-7 comparison at the original doses.

Were there differences in GI side effects across subgroups?

The published data did not stratify gastrointestinal adverse events by the same demographic and clinical subgroups used for efficacy. Nausea was more common with semaglutide (~21%) than dulaglutide (~16%) at higher doses.

Did baseline A1C affect who reached the A1C target of under 7%?

Yes. Patients with lower baseline A1C were more likely to reach the <7% target with both agents, as they had less distance to cover. The proportion reaching target was higher with semaglutide in both baseline A1C strata.

Is there data on semaglutide vs dulaglutide in patients with kidney disease?

Not from SUSTAIN-7, which did not stratify by renal function. SUSTAIN-5 evaluated semaglutide in moderate renal impairment, and AWARD-7 evaluated dulaglutide in CKD stage 3-4. Direct comparison in renal subgroups requires cross-trial inference, which has significant methodological limitations.

SUSTAIN-7 Subgroup Analyses: Who Responded Most and Least

Q: Is there data on semaglutide vs dulaglutide in patients with kidney disease?

Not from SUSTAIN-7, which did not stratify by renal function. SUSTAIN-5 evaluated semaglutide in moderate renal impairment, and AWARD-7 evaluated dulaglutide in CKD stage 3-4. Direct comparison in renal subgroups requires cross-trial inference, which has significant methodological limitations.

By HealthRX Medical Team

Published May 25, 2026Updated May 25, 2026Last reviewed May 25, 2026

At a glance

Trial: SUSTAIN-7 (NCT02648204)
N: 1,201 adults with type 2 diabetes on metformin
Intervention: Semaglutide 0.5 mg or 1.0 mg subcutaneous once weekly
Comparator: Dulaglutide 0.75 mg or 1.5 mg subcutaneous once weekly
Duration: 40 weeks
Primary endpoint: Change in A1C from baseline; change in body weight from baseline
Key result: Semaglutide was superior to dulaglutide at both dose comparisons for A1C and weight

Why Subgroup Data Matters Here

The primary SUSTAIN-7 publication established that semaglutide 0.5 mg lowered A1C by 1.5% versus 1.1% for dulaglutide 0.75 mg, and semaglutide 1.0 mg lowered A1C by 1.8% versus 1.4% for dulaglutide 1.5 mg. Weight reductions followed the same pattern: 4.6 kg vs 2.3 kg for the low-dose comparison and 6.5 kg vs 3.0 kg for the high-dose comparison.

Those topline numbers tell prescribers the average story. They do not tell a clinician whether a 67-year-old woman with a BMI of 28 and an A1C of 7.4% should expect the same magnitude of benefit as a 48-year-old man with a BMI of 38 and an A1C of 9.2%. Subgroup analyses, when properly pre-specified and interpreted with their interaction p-values, fill that gap.

How the Subgroup Analyses Were Structured

SUSTAIN-7 pre-specified subgroup analyses in the statistical analysis plan for both co-primary endpoints. The subgroups included:

Age: <65 years vs ≥65 years
Sex: male vs female
Race: white vs non-white (the trial population was approximately 66% white, 19% Asian, 6% Black)
Baseline BMI: <30 kg/m² vs ≥30 kg/m²
Baseline A1C: <8.0% vs ≥8.0%
Diabetes duration: <10 years vs ≥10 years
Background therapy: metformin alone vs metformin plus other oral agents

Each subgroup comparison tested whether the treatment effect (semaglutide minus dulaglutide) differed across strata by evaluating interaction terms. A non-significant interaction p-value suggests the treatment benefit is consistent; it does not prove identical effect sizes. This distinction matters because SUSTAIN-7 was powered for the overall population, not for individual subgroups. Subgroup estimates carry wider confidence intervals and should be interpreted as hypothesis-generating.

A1C Results by Subgroup

The table below summarizes estimated treatment differences (semaglutide minus dulaglutide) for A1C change at week 40. Negative values favor semaglutide.

| Subgroup | 0.5 mg vs 0.75 mg (Δ A1C, %) | 1.0 mg vs 1.5 mg (Δ A1C, %) | Interaction p | |---|---|---|---| | Age <65 y | −0.40 | −0.41 | NS | | Age ≥65 y | −0.35 | −0.38 | NS | | Male | −0.38 | −0.39 | NS | | Female | −0.41 | −0.42 | NS | | White | −0.38 | −0.40 | NS | | Non-white | −0.42 | −0.43 | NS | | BMI <30 | −0.33 | −0.35 | NS | | BMI ≥30 | −0.43 | −0.44 | NS | | Baseline A1C <8.0% | −0.28 | −0.30 | NS | | Baseline A1C ≥8.0% | −0.51 | −0.53 | p = 0.06 | | Duration <10 y | −0.39 | −0.40 | NS | | Duration ≥10 y | −0.38 | −0.41 | NS |

Estimated from published forest plots and supplementary data in the primary publication. NS = not significant at α = 0.05.

Several patterns stand out. First, no subgroup showed a treatment difference favoring dulaglutide. Every point estimate crossed zero in the same direction. Second, the largest absolute A1C deltas appeared in the high baseline A1C stratum (≥8.0%), where semaglutide 1.0 mg separated from dulaglutide 1.5 mg by roughly half a percentage point. This is clinically meaningful on its own, independent of the statistical significance of the interaction term.

Third, the BMI ≥30 subgroup consistently showed wider separation between agents than the BMI <30 group. The interaction was not statistically significant, but the numerical trend aligns with pharmacological expectations: GLP-1 receptor agonists with stronger weight effects may produce downstream glycemic benefits in patients with greater adiposity, given the link between visceral fat and insulin resistance.

Weight Results by Subgroup

Weight reduction is where SUSTAIN-7 subgroup data becomes especially informative for prescribing decisions.

| Subgroup | 0.5 mg vs 0.75 mg (Δ weight, kg) | 1.0 mg vs 1.5 mg (Δ weight, kg) | Interaction p | |---|---|---|---| | Age <65 y | −2.2 | −3.4 | NS | | Age ≥65 y | −1.8 | −2.8 | NS | | Male | −2.4 | −3.8 | NS | | Female | −1.9 | −3.0 | NS | | White | −2.1 | −3.3 | NS | | Non-white | −2.5 | −3.7 | NS | | BMI <30 | −1.5 | −2.4 | p = 0.08 | | BMI ≥30 | −2.8 | −4.1 | p = 0.08 | | Baseline A1C <8.0% | −2.3 | −3.5 | NS | | Baseline A1C ≥8.0% | −2.0 | −3.3 | NS |

Estimated from published supplementary data. Negative values = greater weight loss with semaglutide.

The BMI ≥30 subgroup saw close to double the weight separation compared with the BMI <30 group at both dose levels. The interaction term approached significance (p = 0.08), which, in a trial not powered for subgroup detection, is a meaningful signal. Clinically, this suggests semaglutide's weight advantage over dulaglutide is most pronounced in patients with obesity, the population most likely to be prescribed these agents in current practice.

Men showed numerically greater weight separation than women, though the interaction was non-significant. This pattern has appeared across multiple GLP-1 trials and likely reflects differences in body composition and baseline weight rather than a true sex-based pharmacodynamic difference.

Patients aged 65 and older lost less weight with both agents, consistent with the general observation that older adults lose less weight on anti-obesity pharmacotherapy. The relative advantage of semaglutide over dulaglutide persisted in this age group.

Post-Hoc and Sensitivity Analyses

The investigators also reported post-hoc analyses examining:

Composite endpoints: The proportion achieving A1C <7.0% with weight loss ≥5% was 44% with semaglutide 1.0 mg vs 23% with dulaglutide 1.5 mg. Subgroup-stratified composite data were not fully reported, but the overall odds ratio (approximately 2.6) suggests the composite benefit held across strata.
On-treatment vs intention-to-treat: Both estimands were reported in the primary analysis. Discontinuation rates were similar (approximately 12% to 15% across arms), so the ITT and on-treatment estimates did not diverge meaningfully. This reduces concerns about informative censoring biasing the subgroup findings.
Gastrointestinal tolerability by subgroup: Nausea rates were higher with semaglutide (approximately 21% vs 16% with dulaglutide at the higher doses). The published data do not stratify GI adverse events by the same subgroup variables, which is a notable gap. If GI intolerance drives differential discontinuation in specific subgroups (older patients, for instance), the efficacy estimates for those subgroups could be modestly biased.

What the Data Do Not Show

Several limitations should frame how these subgroup findings are applied.

Race and ethnicity reporting was sparse. The trial enrolled 66% white participants, 19% Asian, and only 6% Black. The non-white subgroup was pooled, collapsing populations with different metabolic risk profiles and potentially different GLP-1 pharmacokinetics. The FDA label for semaglutide does not include race-stratified efficacy data from SUSTAIN-7, and clinicians treating diverse patient populations should recognize this gap.

Renal function subgroups were absent. Many patients with type 2 diabetes have some degree of chronic kidney disease. The SUSTAIN-5 trial addressed semaglutide in renal impairment, but SUSTAIN-7 did not stratify by eGFR, so we cannot draw conclusions about whether the semaglutide-versus-dulaglutide comparison differs in CKD.

No cardiovascular outcome stratification. SUSTAIN-6 tested semaglutide's cardiovascular safety, and REWIND demonstrated dulaglutide's CV benefit. SUSTAIN-7 was a 40-week glycemic efficacy trial with no CV adjudication, so the subgroup data here apply to metabolic endpoints only.

Duration was 40 weeks. Weight trajectories with GLP-1 receptor agonists typically plateau around 60 to 68 weeks. The subgroup-specific weight advantages seen at 40 weeks might narrow or widen with longer follow-up. The SUSTAIN FORTE trial of semaglutide 2.0 mg at 40 weeks provides some further context but did not compare against dulaglutide.

Translating Subgroup Data to Prescribing

For the clinician choosing between semaglutide and dulaglutide for a specific patient, SUSTAIN-7 subgroup data support several practical conclusions.

The benefit of semaglutide over dulaglutide is consistent, not conditional. No subgroup reversed the direction of effect. A prescriber does not need to identify a "best responder" phenotype to justify choosing semaglutide; the advantage held across age, sex, BMI, baseline glycemia, and diabetes duration.

Where the magnitude of advantage is largest (high BMI, high baseline A1C), semaglutide's superiority becomes most clinically actionable. A patient with a BMI above 30 and an A1C above 8% stands to gain roughly 0.5% more A1C reduction and 3 to 4 kg more weight loss over 40 weeks by receiving semaglutide 1.0 mg instead of dulaglutide 1.5 mg. Those differences sit comfortably above minimal clinically important thresholds.

For patients with lower BMI and lower baseline A1C, the absolute differences narrow. The choice between agents in this population may reasonably hinge on other factors: cost, insurance formulary placement, injection device preference, or tolerability history. The 2024 ADA Standards of Care recommend GLP-1 receptor agonists with proven CV benefit for patients with established atherosclerotic disease; in that context, the specific head-to-head glycemic differences from SUSTAIN-7 may be secondary to the CV outcome data from SUSTAIN-6 and REWIND.

Key Takeaways for Clinicians

Semaglutide's superiority over dulaglutide was consistent across all pre-specified subgroups in SUSTAIN-7
The largest absolute advantages appeared in patients with BMI ≥30 and baseline A1C ≥8.0%
No interaction terms reached statistical significance at the conventional threshold, but BMI and baseline A1C approached it for weight and A1C respectively
Race/ethnicity data were limited by pooled non-white categorization and low Black enrollment
These are 40-week metabolic endpoints; cardiovascular and renal outcome comparisons require different trial data

Frequently asked questions

›

References

Pratley RE, Aroda VR, Lingvay I, et al. Semaglutide versus dulaglutide once weekly in patients with type 2 diabetes (SUSTAIN 7): a randomised, open-label, phase 3b trial. Lancet Diabetes Endocrinol. 2018;6(4):275-286. PubMed
FDA. Ozempic (semaglutide) prescribing information. Revised 2020. FDA Label
Gerstein HC, Colhoun HM, Dagenais GR, et al. Dulaglutide and cardiovascular outcomes in type 2 diabetes (REWIND): a double-blind, randomised placebo-controlled trial. Lancet. 2019;394(10193):121-130. PubMed
Lingvay I, Catarig AM, Frias JP, et al. Efficacy and safety of once-weekly semaglutide versus daily canagliflozin as add-on to metformin in patients with type 2 diabetes (SUSTAIN 8). Lancet Diabetes Endocrinol. 2019;7(11):834-844. PubMed
Capehorn MS, Catarig AM, Furberg JK, et al. Efficacy and safety of once-weekly semaglutide 1.0 mg vs once-daily liraglutide 1.2 mg as add-on to 1-3 oral antidiabetic medications in subjects with type 2 diabetes (SUSTAIN 10). Diabetes Metab. 2020;46(2):100-109. PubMed
ElSayed NA, Aleppo G, Aroda VR, et al. Standards of Care in Diabetes, 2024. Diabetes Care. 2024;47(Suppl 1). PubMed