SURPASS-3

Study record

Population

Insulin-naive adults with type 2 diabetes on stable metformin with or without an SGLT2 inhibitor; HbA1c 7–10.5%, BMI ≥25. 1,444 randomized and 1,437 treated. eGFR below 45 was excluded.

Read source · Full-text reportMethods p585; Results and participant flow

Formulation

Tirzepatide single-dose pen; insulin degludec prefilled 3 mL U100 pen.

Read source · Full-text reportProcedures p585

Route

Weekly subcutaneous tirzepatide versus daily subcutaneous insulin degludec.

Read source · Full-text reportProcedures p585

Dose & frequency

Tirzepatide began at 2.5 mg weekly, increasing by 2.5 mg every four weeks to 5, 10 or 15 mg. During escalation, persistent intolerance could prompt reduction to a tolerated maintenance dose; not permitted in the assigned 5 mg group or after week 24. Degludec began at 10 U/day and was titrated weekly toward fasting glucose below 90 mg/dL.

Read source · Full-text reportProcedures p585

Study design

Open-label randomized active-controlled phase 3 trial, allocation 1:1:1:1, on metformin with or without an SGLT2 inhibitor. Degludec was a comparator, not a tirzepatide combination.

Read source · Full-text reportMethods pp585–586

Duration

52 weeks of treatment followed by four weeks of safety follow-up.

Read source · Full-text reportProcedures p585

Reported findings

At 52 weeks, efficacy-analysis HbA1c changes were −1.93, −2.20 and −2.37 percentage points versus −1.34 with degludec. Weight changes were −7.5, −10.7 and −12.9 kg versus +2.3 kg. These estimates assume continued treatment without rescue; treatment-regimen analyses address discontinuation and rescue separately.

Read source · Full-text reportTable2 pp589–590; Statistical analysis p586

Adverse events

Adverse events led to discontinuation in 25/358, 37/360 and 39/359 versus 5/360. Table3 reports serious events in 29, 20 and 26 versus 22 participants, but flags one event in the 10 mg group as invalid because it preceded randomization. Glucose below 54 mg/dL affected 5, 4 and 7 versus 26; severe hypoglycemia affected 0, 0 and 1 versus 0. Categories must not be summed as unique patients.

Read source · Full-text reportTable3 and footnotes p594

Monitoring

Degludec titration used the median of the last three self-monitored glucose measurements, with investigator deviations allowed for safety. Safety assessments included adverse events, hypoglycemia, vital signs, pancreatic events and laboratory measures.

Read source · Full-text reportProcedures p585; Outcomes p586