Honest Criticisms and Limitations of the SUSTAIN-7 Trial

By
Published Updated Last reviewed
GLP-1 medication and metabolic health image for Honest Criticisms and Limitations of the SUSTAIN-7 Trial

At a glance

| Parameter | Detail | |---|---| | N | 1,201 adults with type 2 diabetes | | Intervention | Semaglutide 0.5 mg or 1.0 mg once weekly | | Comparator | Dulaglutide 0.75 mg or 1.5 mg once weekly | | Duration | 40 weeks | | Primary endpoint | Change in HbA1c from baseline | | Key confirmatory endpoint | Change in body weight from baseline | | Key result | Semaglutide superior to dulaglutide at both dose levels for A1C and weight |

What SUSTAIN-7 Actually Showed

Published in Lancet Diabetes & Endocrinology in 2018, the SUSTAIN-7 trial randomized 1,201 adults with inadequately controlled type 2 diabetes on metformin to one of four arms: semaglutide 0.5 mg, semaglutide 1.0 mg, dulaglutide 0.75 mg, or dulaglutide 1.5 mg, all injected once weekly. At 40 weeks, the semaglutide arms achieved statistically superior reductions in both HbA1c and body weight compared to the corresponding dulaglutide dose groups.

The low-dose comparison (semaglutide 0.5 mg vs dulaglutide 0.75 mg) showed an estimated treatment difference of -0.40 percentage points for A1C and -2.26 kg for body weight. The high-dose comparison (semaglutide 1.0 mg vs dulaglutide 1.5 mg) showed -0.41 percentage points for A1C and -3.55 kg for body weight. All differences favored semaglutide with p < 0.0001.

These numbers made headlines. But the trial design choices that produced them deserve scrutiny.

The HealthRX Limitation Framework for SUSTAIN-7

We organize SUSTAIN-7's limitations into six categories: comparator selection, trial duration, population generalizability, open-label design, statistical methodology, and financial conflicts. Each category is scored for how much it could shift a clinician's confidence in the headline result.

1. Comparator Dose Selection: Was Dulaglutide Given a Fair Fight?

This is the most debated limitation. At the time SUSTAIN-7 was designed, dulaglutide was approved at 0.75 mg and 1.5 mg weekly. Those were the doses tested. Novo Nordisk, the sponsor, argued the comparison was fair because it used each drug's approved dose range.

Critics pushed back. The 1.5 mg dulaglutide dose was the standard maintenance dose, while the 1.0 mg semaglutide dose was the higher of two options. More importantly, Eli Lilly later developed dulaglutide 3.0 mg and 4.5 mg doses (approved in 2020 for the Trulicity label extension), which showed considerably larger A1C and weight reductions than the 1.5 mg dose. SUSTAIN-7 tested dulaglutide at what turned out to be the lower third of its eventual dose range.

This does not invalidate the results. It does mean the trial answered a narrower question than "Is semaglutide better than dulaglutide?" It answered: "Is semaglutide at its 2017 approved doses better than dulaglutide at its 2017 approved doses?" That distinction matters when clinicians are choosing between the two drugs today.

2. Forty Weeks Is Not Long Enough

Type 2 diabetes is managed over decades. SUSTAIN-7 lasted 40 weeks. That is enough time to capture A1C nadir and early weight loss, but it cannot address durability of effect, long-term tolerability, or cardiovascular outcomes.

For context, the REWIND trial followed dulaglutide patients for a median of 5.4 years and demonstrated cardiovascular benefit. The SUSTAIN-6 trial followed semaglutide patients for about 2 years and also showed cardiovascular benefit, though it was a preapproval safety trial not powered for superiority on MACE. Neither of those trials was a head-to-head comparison, so we still lack long-term comparative data between these two drugs.

A 40-week trial tells you which drug drops A1C and weight faster. It cannot tell you which drug keeps a patient healthier at year five. Clinicians extrapolating SUSTAIN-7 to long-term formulary decisions are working beyond the evidence.

3. Who Was Enrolled (and Who Was Not)

The SUSTAIN-7 population had these baseline characteristics:

| Characteristic | Mean/Median | |---|---| | Age | ~56 years | | Duration of diabetes | ~7 years | | Baseline HbA1c | ~8.2% | | Baseline BMI | ~33 kg/m² | | Background therapy | Metformin only | | eGFR | >60 mL/min (required) | | Cardiovascular disease | Not specifically enriched |

Several populations were excluded or underrepresented:

  • Patients on insulin or sulfonylureas. The trial required metformin monotherapy as background. Many real-world patients starting a GLP-1 are on two or three oral agents, or are already on basal insulin.
  • Patients with renal impairment. eGFR < 60 was excluded. GLP-1 prescribing in CKD stages 3b-4 is clinically common and relevant.
  • Older adults. Mean age was 56. Patients over 75 were underrepresented. GI tolerability and weight loss risk differ substantially in older populations.
  • Racial/ethnic diversity. The trial enrolled across 19 countries, but the published demographics show a predominantly White and Asian cohort, with limited Black and Hispanic representation relative to disease prevalence in those groups.

These exclusions are standard for phase 3 trials, but they constrain generalizability. A patient on metformin plus empagliflozin plus basal insulin, with an eGFR of 45, is a common clinical scenario that SUSTAIN-7 cannot directly address.

4. Open-Label Design and Its Consequences

SUSTAIN-7 was open-label. Patients and investigators knew which drug was being administered. The authors justified this as acceptable because the primary endpoints (A1C and weight) are objective laboratory and scale measurements, not patient-reported outcomes.

That argument holds partially. HbA1c is indeed objective. Body weight measured on a calibrated scale is objective. But open-label awareness can influence behavior in ways that affect both endpoints:

  • Dietary and exercise effort. Patients who know they are on the "newer" or "stronger" drug may be more motivated. Patients who know they are on the comparator may feel less optimistic. These psychological effects can shift weight and, to a lesser extent, glycemic control.
  • Investigator titration behavior. Open-label awareness can subtly influence how aggressively investigators manage concomitant medications, rescue therapy thresholds, and dose escalation timing.
  • Adverse event reporting. GI side effects (nausea, vomiting, diarrhea) are subjective. Open-label knowledge of which drug a patient is taking may color both patient reporting and investigator attribution.

The magnitude of these effects is likely small for A1C. For weight, the influence is harder to dismiss entirely, given the known impact of expectation on eating behavior.

5. Statistical Design: Hierarchical Testing and Multiplicity

SUSTAIN-7 used a hierarchical testing procedure to control for multiplicity across the two dose comparisons and two endpoints (A1C and weight). The low-dose comparison for A1C was tested first. Only if that was significant could the high-dose A1C comparison proceed, and so on.

Because all four tests achieved p < 0.0001, the hierarchy did not create any practical issue in this trial. But it is worth noting that the trial was powered primarily for noninferiority on A1C, with superiority as a confirmatory step. The weight endpoint was confirmatory, not co-primary in the traditional sense. This means the trial's statistical architecture was designed to ensure it would "win" on A1C noninferiority even if superiority failed, giving the sponsor a regulatory-friendly result under multiple scenarios.

The estimated treatment differences, while statistically significant, should be evaluated clinically. A 0.40 percentage point A1C difference is meaningful. A 2.26 kg weight difference at the low dose comparison is modest and sits near the threshold of clinical significance for a 40-week intervention.

6. Conflict of Interest and Sponsorship

SUSTAIN-7 was funded by Novo Nordisk A/S, the manufacturer of semaglutide (Ozempic prescribing information). Novo Nordisk employees were involved in trial design, data collection, data analysis, and manuscript preparation. Several academic authors received consulting fees, speaker fees, or research funding from Novo Nordisk and other pharmaceutical companies.

This does not mean the data are fabricated. Industry-sponsored trials undergo regulatory scrutiny, and the SUSTAIN program was reviewed by the FDA and EMA. But sponsorship introduces structural incentives:

  • The sponsor chose the comparator doses (see Section 1).
  • The sponsor chose the 40-week duration (see Section 2).
  • The sponsor chose open-label design (see Section 4).
  • The sponsor controlled the statistical analysis plan.

Each of these choices, individually defensible, collectively favored the sponsor's product. This pattern is common in pharmaceutical research and is not unique to Novo Nordisk, but it should be acknowledged by clinicians interpreting the results.

What Post-Publication Commentary Added

Several letters to the editor and editorials following the SUSTAIN-7 publication raised similar points. The most common criticisms in published correspondence focused on:

  • The absence of higher dulaglutide doses as a comparator
  • Whether the weight differences were clinically meaningful or merely statistically significant
  • The lack of patient-reported outcome data beyond standard safety reporting
  • The need for longer-term comparative effectiveness data

The 2023 ADA Standards of Care recommend GLP-1 receptor agonists with proven cardiovascular benefit for patients with established ASCVD or high cardiovascular risk. Both semaglutide and dulaglutide appear in that recommendation. The guidelines do not cite SUSTAIN-7 as a basis for preferring one over the other, relying instead on each drug's individual cardiovascular outcomes trial.

The Bottom Line for Clinicians

SUSTAIN-7 is a well-executed, adequately powered, phase 3 head-to-head trial that demonstrated semaglutide's glycemic and weight superiority over dulaglutide at matched 2017-era doses over 40 weeks. Those results are real. They are also bounded by design choices (comparator dosing, duration, population, open-label structure) that limit how far clinicians should extend the conclusions. Prescribing decisions between semaglutide and dulaglutide should weigh SUSTAIN-7 alongside each drug's cardiovascular outcomes data, updated dosing options, formulary access, patient preference, and cost.

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
  • 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
  • Marso SP, Bain SC, Consoli A, et al. Semaglutide and cardiovascular outcomes in patients with type 2 diabetes. N Engl J Med. 2016;375(19):1834-1844. PubMed
  • Ozempic (semaglutide) prescribing information. Novo Nordisk. FDA Label
  • Trulicity (dulaglutide) prescribing information. Eli Lilly. FDA Label
  • American Diabetes Association Professional Practice Committee. Standards of Care in Diabetes, 2023. Diabetes Care. 2023;46(Suppl 1). PubMed