SURPASS-6 (NCT04537923) was a randomised, open-label, multicentre, parallel-arm, phase 3b study that compared the effect of the addition of tirzepatide once weekly or insulin lispro three times a day in adults with T2D and inadequate glycaemic control despite insulin glargine, with or without metformin. The primary results of SURPASS-6 have been presented elsewhere [39]. The following is an overview of the SURPASS-6 study design (Fig. 1).

SURPASS-6 study design. aThe screening/insulin standardisation period was a total of 12 weeks for participants who needed insulin glargine optimisation and 5 weeks for those who did not need insulin glargine optimisation. IU international units, QW once weekly, TID three times a day

The study was conducted at 135 medical research centres and hospitals in 15 countries (Argentina, Belgium, Brazil, Czech Republic, Germany, Spain, Greece, Hungary, Italy, Mexico, Romania, Russia, Slovakia, Turkey and the USA). The study protocol was approved by the ethical review board at each site (Supplementary Material Table 1), including the ethical review board at the principal investigator site (Advarra Inc., Columbia, MD). This study was conducted in accordance with consensus ethical principles derived from international guidelines including the Declaration of Helsinki of 1964 and its later amendments, Council for International Organizations of Medical Sciences International Ethical Guidelines, applicable International Council for Harmonisation Good Clinical Practice guidelines and applicable local laws and regulations. All participants provided written informed consent. This article is based on a previously conducted study and does not contain any new studies with human participants or animals performed by any of the authors.

Eligible patients were adults with T2D, baseline HbA1c of ≥ 7.5% to ≤ 11.0% (58–97 mmol/mol), BMI from ≥ 23 kg/m2 to ≤ 45 kg/m2 and stable weight (± 5%) who agreed not to initiate an intensive diet and/or exercise programme during the study other than the lifestyle and dietary measures for diabetes treatment. Patients were receiving stable doses of once or twice daily basal insulin (with doses of ≥ 30 IU/day and between ≥ 0.3 and ≤ 1.0 IU/kg/day within 90 days prior to screening considered stable). Patients could also be receiving oral glucose-lowering agents in any combination of up to two of the following: stable daily dose of metformin ≥ 1500 mg/day up to maximum approved dose per country-specific approved label, sulfonylureas or dipeptidyl peptidase-4 inhibitors.

Key exclusion criteria included the presence of type 1 diabetes, history of pancreatitis, proliferative diabetic retinopathy, non-proliferative diabetic retinopathy requiring acute treatment, severe hypoglycaemia and established cardiovascular disease. All patients provided written informed consent before participation in the study.

Eligible participants were randomised (1:1:1:3) to receive tirzepatide 5, 10 or 15 mg once weekly or insulin lispro (100 IU/mL) three times a day, as subcutaneously administered add-on to standardised insulin glargine (100 IU/mL), with or without metformin (≥ 1500 mg/day), for 52 weeks determined by a computer-generated random sequence using an interactive web response system. Randomisation was stratified by country, pre-randomisation HbA1c level (≤ 8.5%, > 8.5% [≤ 69, > 69 mmol/mol]) and baseline metformin use (yes, no).

SURPASS-6 was an open-label study due to the different method of administration, frequency and dosing requirements of tirzepatide and insulin lispro. However, the study team made every effort to remain blinded to the doses of tirzepatide received by participants.

All participants discontinued oral glucose-lowering agents except metformin (≥ 1500 mg/day) and switched their pre-study basal insulin regimen to insulin glargine once daily at bedtime during insulin standardisation period (Fig. 1). Insulin glargine was administered via subcutaneous injection and titrated to target fasting blood glucose of 100 to 125 mg/dL as per investigator’s clinical discretion.

Participants with HbA1c of ≥ 7.5% (58 mmol/mol) at the end of the insulin standardisation period of up to 10 weeks were randomised to the 52-week treatment period (Fig. 1). Tirzepatide was administered via subcutaneous injection once weekly using single-dose pens, and doses were escalated gradually to improve gastrointestinal tolerability. Insulin lispro was administered via subcutaneous injection three times a day using prefilled pens and titrated to a target pre-meal/bedtime blood glucose of 100 to 125 mg/dL (5.6 to 6.9 mmol/L).

Clinical and safety data were collected throughout the study. PROs were assessed at baseline, endpoint (week 52) and any early discontinuation visit using the instruments described below.

The primary outcome, change in HbA1c from baseline, and secondary clinical outcomes, including changes from baseline in fasting serum glucose and body weight, as well as proportions of participants at HbA1c and body weight reduction goals, for tirzepatide pooled versus insulin lispro at 52 weeks have been reported elsewhere [39]. PROs were measured at baseline and 52 weeks using the SF-36v2 acute form (secondary outcome), EQ-5D-5L, APPADL questionnaire and IW-SP questionnaire (tertiary/exploratory outcomes) and were compared for the tirzepatide-pooled dose group and for each dose of tirzepatide versus insulin lispro at 52 weeks. An overview of the PRO measures included in SURPASS-6 is presented in Table 1. Safety endpoints included the occurrence of hypoglycaemic events. Safety and tolerability data have been disclosed reported elsewhere [39].

The SF-36v2 acute form is a generic, validated, patient-completed measure designed to assess eight domains of functioning and health that are also combined to obtain two component summary scores (physical and mental) (Fig. 2) [40].

Short Form-36 Health Survey version 2 (SF-36v2) acute form component summary measures, domains and items. All health domains contribute to the scoring of both the Physical and Mental Component summaries. Domains contributing most to the scoring of the component summaries are represented by black lines; domains contributing to a lesser degree are represented by grey lines. EVGFP excellent–very good–good–fair–poor health

The EQ-5D-5L is a generic instrument commonly used to measure overall health status [41]. The EQ visual analogue scale (EQ VAS) allows the patient to self-rate their overall health status. The EQ-5D-5L index score is based on five dimensions (mobility, self-care, usual activities, pain/discomfort and anxiety/depression) with each dimension having five response levels. EQ-5D-5L index scores were based on the UK value set [42].

The APPADL questionnaire contains seven items that assess how difficult it is for patients to engage in certain activities considered to be integral to normal daily life, such as walking, standing and climbing stairs [43].

The IW-SP has three items and is designed to assess how self-perceptions of individuals with T2D and obesity are affected by their body weight [44].

Higher scores indicate better outcomes across all PRO measures included in SURPASS-6.

Sample size selection was guided by establishing non-inferiority of each tirzepatide dose and tested against insulin lispro, relative to the primary endpoint [39]. The tirzepatide-pooled treatment group included participants randomised to tirzepatide 5, 10 or 15 mg.

Guided by the efficacy estimand, defined as the average treatment effect of tirzepatide versus insulin lispro had participants taken treatment as intended, analyses were conducted on the modified intention-to-treat efficacy analysis set. This dataset composed of randomised participants exposed to ≥ 1 dose of study drug (tirzepatide or insulin lispro) who did not discontinue the study drug due to inadvertent enrolment, and observations occurring after the start of rescue therapy with another glucose-lowering agent or early treatment discontinuation were excluded. Baseline demographic and clinical characteristics were described only for participants with a non-missing baseline score and at least one non-missing post-baseline score for at least one of the PRO measures analysed. For PROs, the main comparative analyses were the change from baseline to 52 weeks in individual PRO measures for tirzepatide pooled and each dose of tirzepatide versus insulin lispro. Only participants with a non-missing baseline score and at least one non-missing post-baseline score for the response variable were included in the analysis for that PRO measure. Responses were analysed using analysis of covariance (ANCOVA) models for endpoint measures with last observation carried forward imputation, with model terms treatment, pooled country, baseline HbA1c (≤ 8.5%, > 8.5% [≤ 69, > 69 mmol/mol]), and baseline metformin use (yes, no) as fixed effects, and baseline PRO score as a covariate. P values < 0.05 were considered statistically significant, and the values were not adjusted for multiplicity.

Exploratory post hoc analyses, conducted for SF-36v2 acute form Physical and Mental Component Summary scores and EQ VAS in the tirzepatide-pooled population only, modelled six baseline characteristics and/or factors changing over time (percentage of body weight change from baseline; HbA1c change from baseline; and baseline weight, age, sex and geographic region) identified as being correlated to the PRO scores based on Akaike information criterion (AIC) using a stepwise linear regression model.

Additional exploratory post hoc analyses were conducted to explore the relationship between weight reduction and PROs using ANCOVA models with model terms treatment, weight change category, treatment-by-weight-change-category interaction and pooled country as fixed effects and baseline PRO score and baseline weight as covariates. Analyses of all PROs by weight reduction category (≥ 0% to < 5%, ≥ 5% to < 10%, ≥ 10% to < 15% and ≥ 15%) were performed in the tirzepatide-pooled population only. All analyses were carried out using SAS® version 9.4 (SAS Institute Inc., Cary, NC). There was no data monitoring committee.

The funder of the study, Eli Lilly and Company, had a role in the study design, data collection, data analysis, data interpretation and writing of the report.