SURPASS-3 Results in Detail: Numbers, Subgroups, and Time Course

At a glance

• Trial: SURPASS-3 (NCT03882970)
• N: 1,444 randomized
• Population: Adults with type 2 diabetes inadequately controlled on metformin (with or without SGLT2 inhibitors)
• Intervention: Tirzepatide 5 mg, 10 mg, or 15 mg subcutaneous once weekly
• Comparator: Insulin degludec, titrated to fasting glucose <90 mg/dL
• Duration: 52 weeks
• Primary endpoint: Change from baseline in HbA1c at week 52
• Key result: All tirzepatide doses were superior to insulin degludec for A1C reduction (p<0.001 for all comparisons)

Why This Trial Matters Beyond the Headline

Most trial summaries stop at "tirzepatide beat insulin." That framing misses critical details. SURPASS-3 used an active comparator, insulin degludec, that was titrated aggressively to a fasting glucose target below 90 mg/dL. This was not a straw-man insulin arm. The mean insulin dose at week 52 reached approximately 49 units/day, reflecting genuine optimization. Tirzepatide still won on every glycemic and weight metric, which makes the effect sizes clinically meaningful rather than a product of a weak comparator.

Primary Endpoint: HbA1c Change at Week 52

The primary analysis used a mixed model for repeated measures (MMRM) with the estimand targeting the treatment effect regardless of treatment discontinuation or rescue medication use (the "treatment-regimen estimand").

HealthRX.com Dose-Response Interpretation Grid

The table below organizes the primary and weight data by dose tier, highlighting both absolute change and between-group separation. This framework clarifies that the glycemic benefit plateaus between 10 and 15 mg while the weight effect continues to scale.

| Outcome | Tirzepatide 5 mg (n=358) | Tirzepatide 10 mg (n=360) | Tirzepatide 15 mg (n=359) | Insulin degludec (n=360) | |---|---|---|---|---| | A1C change from baseline (%) | −1.93 | −2.20 | −2.37 | −1.34 | | Difference vs insulin (95% CI) | −0.59 (−0.73 to −0.45) | −0.86 (−1.00 to −0.72) | −1.04 (−1.17 to −0.90) |, | | P-value (superiority) | <0.001 | <0.001 | <0.001 |, | | Weight change, kg | −7.5 | −10.7 | −12.9 | +2.3 | | Difference vs insulin, kg (95% CI) | −9.8 (−10.8 to −8.8) | −13.0 (−14.0 to −12.0) | −15.2 (−16.2 to −14.2) |, |

Baseline A1C across groups was approximately 8.17%, and baseline body weight was roughly 94 kg. The published primary analysis confirmed superiority for all three doses on both co-primary endpoints.

A1C Target Achievement Rates

The percentage of participants reaching standard glycemic thresholds underscores the separation between arms.

| Target | 5 mg | 10 mg | 15 mg | Insulin degludec | |---|---|---|---|---| | A1C <7.0% | 82.4% | 89.7% | 92.6% | 61.3% | | A1C <6.5% | 71.4% | 81.1% | 86.1% | 44.4% | | A1C <5.7% | 26.8% | 39.2% | 48.4% | 9.1% |

Nearly half of participants on tirzepatide 15 mg reached an A1C below 5.7%, the threshold for normoglycemia. For context, the ADA Standards of Care define <5.7% as a non-diabetic range. Achieving this level with a once-weekly injectable, even transiently, was uncommon in prior add-on trials against active comparators.

Time-Course Pattern: When Did Separation Emerge?

A1C reductions were apparent by week 12 across all tirzepatide arms and continued to widen through week 24. Between weeks 24 and 52, the curves largely flattened, suggesting that the maximal glycemic effect is established within the first six months. The insulin arm continued a slower, more linear descent through week 52 as dose titration progressed.

Weight trajectories showed a different pattern. Tirzepatide-treated participants lost weight continuously through the full 52-week period without a clear plateau at any dose. Insulin degludec recipients gained weight steadily. The divergence in body weight between the 15 mg group and insulin exceeded 10 kg by week 28 and reached 15.2 kg by week 52.

This dissociation between glycemic plateau and ongoing weight loss is consistent with the dual GIP/GLP-1 mechanism. GLP-1 receptor agonism drives early glucose control; the GIP component appears to contribute additional and sustained effects on adiposity, as described in the tirzepatide FDA label.

Secondary Endpoints and Cardiometabolic Markers

SURPASS-3 collected a broad set of secondary outcomes. Select results at week 52:

| Marker | 5 mg | 10 mg | 15 mg | Insulin degludec | |---|---|---|---|---| | Fasting serum glucose change (mg/dL) | −43.6 | −55.4 | −58.7 | −50.8 | | Triglycerides change (%) | −15.6 | −21.0 | −24.0 | +3.7 | | Systolic BP change (mmHg) | −4.9 | −7.0 | −8.3 | +1.3 | | Waist circumference change (cm) | −6.4 | −8.5 | −10.2 | +1.4 |

Fasting glucose reductions were comparable between tirzepatide 15 mg and insulin degludec (both achieved meaningful lowering), but the lipid, blood pressure, and anthropometric data consistently favored tirzepatide. The triglyceride reduction of 24% in the 15 mg arm versus a 3.7% increase in the insulin arm reflects an effect size relevant to cardiovascular risk stratification.

Hypoglycemia: The Safety Trade-Off

Despite producing deeper A1C reductions, tirzepatide was associated with less clinically significant hypoglycemia (glucose <54 mg/dL) than insulin degludec.

| Event type | 5 mg | 10 mg | 15 mg | Insulin degludec | |---|---|---|---|---| | Clinically significant hypoglycemia (rate/patient-year) | 0.05 | 0.06 | 0.05 | 0.43 | | Any hypoglycemia (% of participants) | 7% | 8% | 7% | 19% |

The roughly 7-to-8-fold lower rate of clinically significant hypoglycemia with tirzepatide is notable because the tirzepatide arms achieved substantially lower A1C values. In conventional thinking, lower A1C means higher hypoglycemia risk. Tirzepatide breaks that correlation, consistent with glucose-dependent insulin secretion via incretin pathways. This finding aligns with observations in the SURPASS-2 trial comparing tirzepatide to semaglutide.

Gastrointestinal Side Effects

GI events were the most common adverse effects with tirzepatide and the primary reason for discontinuation in the active arms.

| Event | 5 mg | 10 mg | 15 mg | Insulin degludec | |---|---|---|---|---| | Nausea | 12.2% | 18.1% | 22.5% | 1.7% | | Diarrhea | 15.1% | 16.4% | 13.4% | 3.6% | | Vomiting | 4.5% | 5.6% | 8.4% | 0.6% | | Decreased appetite | 5.3% | 7.8% | 9.5% | 0.6% | | Discontinuation due to AE | 3.1% | 5.3% | 7.5% | 1.1% |

GI symptoms were generally transient, peaking during dose escalation (weeks 4 through 12) and subsiding by week 20 in most participants. The dose-escalation protocol (2.5 mg starting dose with monthly 2.5 mg increments) was designed to mitigate early nausea. Even so, the 15 mg arm had a discontinuation rate of 7.5%, roughly three times the insulin arm.

Subgroup Consistency

The primary publication reported subgroup analyses across baseline A1C (<8.5% vs ≥8.5%), BMI (<30 vs ≥30 kg/m²), age (<65 vs ≥65 years), sex, race, diabetes duration, and SGLT2 inhibitor use. Treatment effects on A1C were directionally consistent across all subgroups. Participants with higher baseline A1C experienced larger absolute reductions, a pattern typical of glucose-lowering trials. The addition of an SGLT2 inhibitor at baseline (used in approximately 32% of participants) did not attenuate tirzepatide's efficacy.

Weight-loss subgroup data showed slightly larger reductions in participants with higher baseline BMI, but confidence intervals overlapped. No subgroup showed a reversal of the treatment effect.

Limitations the Authors Acknowledged

The open-label design is the primary methodological limitation. Participants and investigators knew which treatment was assigned, which can influence reporting of subjective adverse events (nausea, appetite) and may affect insulin titration behavior. The protocol mandated insulin dose adjustments based on fasting glucose, but real-world insulin management might differ.

The 52-week duration, while longer than many phase 3 trials, does not address durability beyond one year. Weight loss had not plateaued in the tirzepatide arms at week 52, making it unclear whether further reductions would occur or a rebound would follow.

The study population was predominantly white (approximately 79%), with limited representation from Black and Asian participants. Generalizability to these populations remains uncertain.

Participants with an eGFR below 45 mL/min/1.73m² or a history of pancreatitis were excluded, consistent with standard GLP-1 RA trial design but limiting applicability to those groups.

How SURPASS-3 Fits in the Broader Program

SURPASS-3 was one of five key trials in the tirzepatide registration program. It was the only trial comparing tirzepatide directly to basal insulin and was the primary basis for the FDA label claim of superiority over insulin therapy. SURPASS-1 tested tirzepatide as monotherapy. SURPASS-2 compared it to semaglutide 1 mg. SURPASS-4 enrolled patients with established cardiovascular disease on one to three oral agents. SURPASS-5 added tirzepatide to background insulin glargine.

The weight-loss magnitude observed in SURPASS-3 (up to 12.9 kg with 15 mg) anticipated the results of the SURMOUNT-1 obesity trial, where tirzepatide produced up to 22.5% body weight reduction in participants without diabetes.

Frequently asked questions

›

›

›

›

›

›

›

›

›

›

References

1. Ludvik B, Giorgino F, Jódar E, et al. Once-weekly tirzepatide versus once-daily insulin degludec as add-on to metformin with or without SGLT2 inhibitors in patients with type 2 diabetes (SURPASS-3): a randomised, open-label, parallel-group, phase 3 trial. Lancet. 2021;398(10300):583-598. PubMed
2. Frías JP, Davies MJ, Rosenstock J, et al. Tirzepatide versus semaglutide once weekly in patients with type 2 diabetes. N Engl J Med. 2021;385(6):503-515. PubMed
3. Jastreboff AM, Aronne LJ, Ahmad NN, et al. Tirzepatide once weekly for the treatment of obesity. N Engl J Med. 2022;387(4):327-340. PubMed
4. U.S. Food and Drug Administration. Mounjaro (tirzepatide) prescribing information. 2022. FDA Label
5. American Diabetes Association Professional Practice Committee. Standards of Medical Care in Diabetes, 2022. Diabetes Care. 2022;45(Suppl 1). PubMed