Introduction

Obesity is a complex, multifactorial, relapsing disease influenced by genetic, epigenetic, socioeconomic, and cultural factors [1,2,3]. Obesity leads to numerous related complications such as type 2 diabetes (T2D), cardiovascular disease (CVD), hypertension, osteoarthritis, obstructive sleep apnea (OSA), and premature death [4, 5]. Global obesity prevalence has more than doubled since 1990; in 2022, >890 million adults were diagnosed with obesity, representing around 16% of adults globally [6]. Obesity is defined by body mass index (BMI) in clinical trials, though this does not adequately reflect the burden of adiposity-related disease [7].

Obesity treatment options include a combination of behavioral interventions, dietary changes, and physical activity programs. For patients requiring additional therapeutic measures, obesity management medications (OMMs) or bariatric/metabolic surgery may be considered [8,9,10]. Among OMMs, semaglutide 2.4 mg once-weekly (QW) is a glucagon-like peptide-1 (GLP-1) receptor agonist (RA) and tirzepatide QW is a glucose-dependent insulinotropic polypeptide (GIP) and GLP-1 RA. Both semaglutide and tirzepatide are approved in several regions adjunct to a reduced-calorie diet and increased physical activity for weight management in adults with obesity (BMI ≥ 30 kg/m2), or overweight (BMI ≥ 27 kg/m2) with ≥1 obesity-related complication (ORC) (hypertension, dyslipidemia, OSA, or CVD) [11,12,13,14,15,16].

Tirzepatide 5, 10 and 15 mg QW and semaglutide 2.4 mg QW (both adjunct to diet and exercise) demonstrated statistically significant and sustained reductions versus placebo in body weight in their pivotal trials for in this indication, SURMOUNT-1 (72 weeks) and STEP 1 (68 weeks), respectively [17, 18].

The recently published head-to-head open-label study SURMOUNT-5 directly compared tirzepatide maximum tolerated dose (MTD) (10 or 15 mg) with semaglutide MTD (1.7 or 2.4 mg) for obesity treatment, where tirzepatide had a statistically significantly greater reduction in body weight and waist circumference versus semaglutide [19].

As for indirect evidence, a retrospective cohort study comparing tirzepatide and semaglutide focused on weight reduction and glycemic parameters, but did not distinguish between different tirzepatide doses [20]. Similarly, an indirect treatment comparison (ITC) evaluating the two OMMs for obesity treatment focused only on weight reduction outcomes (weight reduction [%] and patients achieving ≥5% weight reduction) [21]. Network meta-analyses (NMAs) of GLP-1 randomized controlled trials (RCTs) have also been performed, including an NMA additionally considering semaglutide 0.4 mg once-daily (QD) and liraglutide 3 mg QD, and another considering GLP-1 RAs and other OMMs [22, 23]. However, the focus on a wider set of treatments in these NMAs increases the heterogeneity between trials and therefore the uncertainty in results.

Study objective

This study indirectly compared the efficacy and safety of tirzepatide 5, 10 and 15 mg QW for obesity treatment versus semaglutide 2.4 mg QW, with a focus on individual licensed doses to distinguish from previous direct and indirect comparisons [19, 20]. Although weight reduction is important for obesity treatment, this analysis included further outcomes beyond the weight reduction outcomes previously investigated (percent weight reduction and patients achieving ≥5% weight reduction) [21], with a focus on cardiometabolic risk factors and body composition.

Materials and methods

Unadjusted Bucher ITCs, and matching-adjusted indirect comparisons (MAICs) adjusting for differences in patient characteristics, compared tirzepatide 5, 10 and 15 mg versus semaglutide 2.4 mg via placebo, all adjunct to diet and exercise, using data from the pivotal trials SURMOUNT-1 and STEP 1. The patient population of interest was adults (≥18 years of age) with obesity (BMI ≥ 30 kg/m2), or overweight (BMI ≥ 27 kg/m2) with ≥1 ORC (hypertension, dyslipidemia, OSA, or CVD), all without T2D. Comparisons of body composition outcomes were based on subgroups that underwent dual-energy X-ray absorptiometry (DEXA), with BMI ≤ 40 kg/m2 for STEP 1 and no upper BMI limit for SURMOUNT-1. Outcomes from the whole trial population informed the remaining analyses [17, 18]. All analyses were conducted using the statistical software R version 4.3.0 [24].

Heterogeneity assessment

The feasibility of comparing SURMOUNT-1 versus STEP 1 was assessed by evaluating heterogeneity in trial design, patient baseline characteristics and outcome and estimand definitions of both trials. Treatment effect modifiers (TEMs) identified from published literature and clinical input were also evaluated.

Both studies employed two estimands for efficacy outcomes: the treatment regimen estimand, which captured outcomes regardless of patients’ adherence to the assigned treatment, and the efficacy estimand, which reflected outcomes only for patients who remained on their assigned treatment for the entire study. These definitions aligned with the European Medicines Agency’s definitions of estimands (treatment policy and on-treatment, respectively) [25].

The primary timepoints for each trial were used (Week 72 SURMOUNT-1, Week 68 STEP 1) as these were deemed comparable given their close temporal proximity and the plateau observed across key outcomes prior to trial completion [17, 18]. Additionally, this approach ensured that the participants had a similar duration of exposure to the maximum dose, with SURMOUNT-1 requiring four additional weeks to reach tirzepatide 15 mg (20 weeks versus 16 weeks for semaglutide 2.4 mg in STEP 1).

ITC

Bucher ITCs were conducted to compare each tirzepatide dose (5/10/15 mg QW) with injectable semaglutide 2.4 mg QW via the anchored common comparator, placebo, all adjunct to diet and exercise (Fig. 1) [26]. Data were compared for the efficacy estimand and treatment regimen estimand for efficacy outcomes, and for the safety population for safety outcomes. Percent weight reduction and patients achieving ≥5% weight reduction were not analyzed as these outcomes were studied in a previous ITC [21].

Fig. 1: Bucher method diagram.
Fig. 1: Bucher method diagram.
Full size image

The solid straight lines correspond to a study, and the dashed straight line corresponds to the indirect comparison. ITC indirect treatment comparison, QW once-weekly.

The methodology was in line with key guidelines [27, 28], with detail provided in the supplementary materials.

MAIC

After inspection of the aggregate baseline characteristics across SURMOUNT-1 and STEP 1, while the baseline characteristics were largely homogenous between the two trials, some differences were found in the TEMs sex and ethnicity. Population-adjusted methodology, MAIC, was conducted to assess whether unadjusted Bucher ITC results were sensitive to differences in TEMs. Separate MAICs were conducted adjusting for sex (% female) alone, and for both sex (% female) and ethnicity (% Hispanic). MAICs were not deemed feasible for body composition outcomes due to the smaller sample sizes of the DEXA subgroups (Table 1) and the very small proportion of Hispanic patients for STEP 1 (5.7%).

Table 1 Descriptive statistics of participant baseline characteristics across trials for the randomized whole trial population.
Full size table

Frequentist MAICs were conducted in line with key guidelines [28, 29], with details provided in the supplementary materials. For each comparison, weights were generated based on patient-level data from SURMOUNT-1 and applied such that patient populations of SURMOUNT-1 matched the aggregate characteristics of patients from STEP 1 in the selected adjustment variables. The generated weights were deemed suitably spread and the effective sample sizes suitably large for the MAICs to be interpreted.

Results

Heterogeneity assessment

As both SURMOUNT-1 and STEP 1 were double-blind, parallel-group, randomized, placebo-controlled, phase 3, multinational trials, they were considered comparable in terms of study design.

TEMs considered to be of primary importance—sex, glycated hemoglobin (HbA1c) and weight—were identified through previous analyses [30], and TEMs considered to be of secondary importance—race/ethnicity, prediabetes and OSA—were identified through clinical opinion.

Both trials had generally similar patient baseline characteristics, including TEMs (Table 1). Both trials focused on participants with obesity or overweight without T2D. In particular, baseline HbA1c and the proportion of participants with prediabetes were very similar across trials. Baseline weight, waist circumference, BMI and OSA were comparable across both trials. Baseline age and race of participants were also generally comparable across trials. However, there were some differences in baseline sex (67.5% female SURMOUNT-1; 74.1% female STEP 1) and the proportion of Hispanic participants (47.8% SURMOUNT-1; 12.0% STEP 1). Therefore, MAICs were conducted to adjust for differences in sex and ethnicity. Patient characteristics of the DEXA subgroups were broadly comparable to the full trial populations, albeit with smaller differences in the proportion of female participants (72.2% SURMOUNT-1; 75.7% STEP 1) and larger differences in the proportion of Hispanic participants (52.9% SURMOUNT-1; 5.7% STEP 1). Furthermore, due to the upper BMI limit of ≤40 kg/m2 for STEP 1, the STEP 1 DEXA subgroup had a lower mean BMI (34.8 kg/m2) compared with SURMOUNT-1 (38.0 kg/m2).

The background treatment in both trials was similar: ≥150 minutes of physical activity per week, a diet of the participant’s total estimated energy expenditure minus 500 kcal/day, and regular lifestyle counseling (every four weeks for STEP 1, every 4 weeks until Week 12, then every 12 weeks for SURMOUNT-1) [18, 31].

Placebo was used as the common comparator, and results across placebo arms were generally similar for the efficacy outcomes of SURMOUNT-1 and STEP 1 for both efficacy and treatment regimen estimands (Table S1). However, some differences in placebo outcomes were observed in low-density lipoprotein (LDL) and total cholesterol (Table S1). For the safety population, the number of placebo arm total gastrointestinal (GI) and nausea events were higher for STEP 1 than SURMOUNT-1, which indirectly favors the safety profile of semaglutide and could bias the results of the analyses (Table S1).

ITC and MAIC results

With some differences in statistical significance, the treatment regimen and efficacy estimand analyses arrived at broadly similar conclusions regarding relative differences between tirzepatide and semaglutide. The main difference observed was for ≥10/15% weight reduction, where the efficacy estimand ITC showed no association with statistically significant differences between treatments. However, tirzepatide 5 mg was associated with statistically significantly lower odds compared with semaglutide in the treatment regimen ITC (the only statistically significant result), suggesting a greater treatment benefit for semaglutide versus tirzepatide. As the efficacy estimand better represents the maximum potential benefit of the intervention, efficacy estimand and safety population results are given in the text, in Table 2 and Table 3, while treatment regimen estimand results are given in Table S3 and Table S4.

Table 2 Summary of efficacy estimand (weight reduction and cardiometabolic risk factors) and safety population ITC results.
Full size table
Table 3 Summary of efficacy estimand ITC results for DEXA subgroup body composition outcomes.
Full size table

Weight reduction outcomes

In the ITC, tirzepatide 10 and 15 mg were associated with statistically significant greater reductions in body weight (mean difference [MD] [95% CI]: −5.10 [ − 6.63, −3.57] kg and −6.50 [ − 8.03, −4.97] kg, respectively) and BMI (MD [95% CI]: −1.87 [ − 2.43, −1.31] kg/m2 and −2.37 [ − 2.93, −1.81] kg/m2, respectively) versus semaglutide. Tirzepatide 5 mg was associated with non-significant trends of smaller reductions in body weight and BMI versus semaglutide (Table 2, Figure S1). These results align with the individual trial results, where tirzepatide 10 and 15 mg had greater reductions in body weight (−22.20 kg and −23.60 kg, respectively) compared to semaglutide (−17.40 kg), while tirzepatide 5 mg had the smallest reduction of all active treatments (−16.10 kg) (Table S2).

Furthermore, ITC results showed that tirzepatide 10 and 15 mg were associated with statistically significantly greater odds of achieving ≥10/15/20% weight reductions versus semaglutide. In contrast, tirzepatide 5 mg was associated with non-significant trends of smaller odds of achieving ≥10/15% weight reductions and a non-significant trend of greater odds of achieving ≥20% weight reduction (Table 2, Fig. 2, Figure S2). Similar results were observed for the MAICs, except tirzepatide 10 mg was associated with non-significant trends of greater odds of achieving ≥10/15% weight reduction versus semaglutide when adjusting for sex (Table 2, Figure S2), and of achieving ≥20% weight reduction when adjusting for sex and ethnicity (Table 2, Fig. 2).

Fig. 2: Waist circumference and ≥20% weight reduction ITC and MAIC results, efficacy estimand.
Fig. 2: Waist circumference and ≥20% weight reduction ITC and MAIC results, efficacy estimand.
Full size image

ITC and MAIC tirzepatide vs semaglutide 2.4 mg results for CfB in waist circumference (cm) and patients achieving ≥20% weight reduction. Wide 95% CIs for the MAICs due to the smaller ESS of SURMOUNT-1 in these analyses. BMI body mass index, CfB change from baseline, CI confidence interval, ESS effective sample size, ITC indirect treatment comparison, MAIC matching-adjusted indirect comparison, MD mean difference, OR odds ratio.

Cardiometabolic risk factors and waist circumference

In the ITC and both MAICs, tirzepatide 10 and 15 mg were associated with statistically significant greater reductions in waist circumference (cm) versus semaglutide (ITC MD [95% CI]: −5.25 [ − 6.69, −3.81] kg and −5.75 [ − 7.19, −4.31] kg, respectively), while tirzepatide 5 mg was associated with a non-significant trend of greater reductions in waist circumference versus semaglutide (Table 2, Fig. 2).

In the ITCs, tirzepatide 10 and 15 mg were associated with statistically significant greater reductions in HbA1c (%) and fasting plasma glucose (FPG, mg/dL) versus semaglutide, while tirzepatide 5 mg was associated with non-significant trends of smaller reductions in HbA1c and FPG versus semaglutide. MAIC results for FPG were similar to the ITCs; however, MAICs of HbA1c did not show an association with statistical significance for tirzepatide 10 and 15 mg versus semaglutide (Table 2, Fig. 3, Fig. S3).

Fig. 3: Glycemia-related parameter results, ITC and MAIC results, efficacy estimand.
Fig. 3: Glycemia-related parameter results, ITC and MAIC results, efficacy estimand.
Full size image

ITC and MAIC tirzepatide vs semaglutide 2.4 mg results for CfB in HbA1c (%) and patients achieving prediabetes reversal. Wide 95% CIs for the MAICs due to the smaller ESS of SURMOUNT-1 in these analyses. aFor prediabetes reversal, treatment regimen estimand results are presented, given efficacy estimand results were not available for STEP 1. CfB change from baseline, CI confidence interval, ESS effective sample size, FPG fasting plasma glucose, HbA1c glycated hemoglobin, ITC indirect treatment comparison, MAIC matching-adjusted indirect comparison, MD mean difference, OR odds ratio.

As efficacy estimand results were not available for STEP 1, results for prediabetes with reversal (patients with HbA1c of 5.7–6.4% at baseline, reverting to <5.7%) are presented for the treatment regimen estimand. In the ITC and MAICs, all tirzepatide doses were associated with non-significant trends of greater odds of participants with prediabetes achieving reversal versus semaglutide, except for the MAICs where tirzepatide 15 mg was associated with statistically significant greater odds versus semaglutide (Table S3, Fig. 3).

In the ITC and MAICs, tirzepatide 15 mg was associated with a statistically significant greater decrease in triglycerides (%) versus semaglutide, while tirzepatide 10 mg was associated with a non-significant trend of greater and tirzepatide 5 mg of similar (ITC) or greater (MAICs) decreases compared to semaglutide. In the ITC, tirzepatide 10 and 15 mg were associated with non-significant trends of greater decreases in LDL (%) and total cholesterol (%) versus semaglutide, while tirzepatide 5 mg was associated with non-significant trends of smaller decreases in both versus semaglutide. MAIC results were similar, but with tirzepatide 15 mg associated with statistically significant greater reductions in total cholesterol versus semaglutide. In the ITC and MAICs, all tirzepatide doses were associated with statistically significant greater increases (improvement) in HDL (%) versus semaglutide (Table 2).

In the ITC and MAICs, all tirzepatide doses were associated with non-significant trends of greater reductions in systolic blood pressure (SBP, mmHg) versus semaglutide, except tirzepatide 5 mg was associated with a non-significant trend of smaller reduction in SBP when adjusting for sex and ethnicity. All tirzepatide doses were associated with statistically significant greater reductions in diastolic blood pressure (DBP, mmHg) versus semaglutide in the ITCs; however, the MAIC results were only associated with statistical significance for tirzepatide 5 mg (MAIC adjusting for sex) and tirzepatide 10 mg (both MAICs) (Table 2).

Body composition outcomes

In the ITC, tirzepatide 15 mg was associated with a statistically significant greater reduction in total body fat mass, while tirzepatide 5 and 10 mg were associated with non-significant trends of greater or similar reductions in total body fat mass compared with semaglutide. Similarly, tirzepatide 15 mg was associated with a statistically significant greater reduction in the percentage of body fat mass versus semaglutide, while tirzepatide 5 and 10 mg were associated with non-significant trends of greater or similar reductions (Table 3, Fig. S4).

Furthermore, all tirzepatide doses were associated with non-significant trends of smaller or similar reductions in total body lean mass (Fig. S5). While participants taking all treatments lost lean mass, those taking tirzepatide 15 mg were associated with a statistically significant greater increase in the percentage of body lean mass versus semaglutide; whereas tirzepatide 5 and 10 mg were associated with non-significant trends of greater increases, likely due to the greater reductions in absolute fat mass seen for these treatments. All tirzepatide doses showed similar reductions in regional visceral fat mass compared with semaglutide. (Table 3, Fig. S5).

Safety outcomes

In the ITC, tirzepatide 10 and 15 mg were associated with non-significant trends of greater odds of participants experiencing GI adverse events (AEs), nausea AEs, and discontinuations due to AEs versus semaglutide, while tirzepatide 5 mg was associated with non-significant trends of smaller odds versus semaglutide (Table 2). MAIC results were comparable, except tirzepatide 15 mg was associated with a non-significant trend of smaller odds of participants experiencing GI AEs versus semaglutide in the MAIC, adjusting for sex and ethnicity. For all-cause discontinuation, tirzepatide 10 and 15 mg were associated with non-significant trends of smaller odds and tirzepatide 5 mg with statistically significant smaller odds versus semaglutide in the Bucher ITC. In contrast, MAIC results showed statistically significant smaller odds of all-cause discontinuation versus semaglutide for all tirzepatide doses (MAIC adjusting for sex and ethnicity) and for tirzepatide 5 and 15 mg (MAIC adjusting for sex) (Table 2).

Discussion

Previously published direct and indirect comparisons of tirzepatide and semaglutide have either studied the MTD, focused solely on weight reduction outcomes or included multiple other treatments in their comparisons [21,22,23]. In contrast, this analysis compared the licensed treatment doses of tirzepatide and explored outcomes in obesity and overweight beyond weight loss outcomes, with a focus on cardiometabolic risk factors and body composition, providing a more holistic comparison of the two treatments at the optimal use for each licensed dose while reducing the heterogeneity in results.

The comprehensive results from this ITC provide evidence of the comparative efficacy of tirzepatide 10 and 15 mg versus semaglutide in weight reduction, cardiometabolic risk factors, and body composition (notably unique, as no other studies have presented a comparison of body composition for tirzepatide and semaglutide to date), as well as their generally comparable safety profiles. Compared with semaglutide 2.4 mg, tirzepatide was associated with a greater improvement of factors contributing to metabolic syndrome, including HbA1c, HDL, triglycerides, waist circumference, and blood pressure, and therefore could reduce the risk of developing serious cardiac conditions associated with metabolic syndrome [32]. However, not all improvements were statistically significant; for example, results for tirzepatide 5 mg were often only numerically improved compared to semaglutide 2.4 mg (Table 2). Results were presented across both efficacy and treatment regimen estimands to provide comparisons both in participants who adhered to treatment, and in all randomized participants. Additionally, where feasible, analyses were conducted both adjusting for, and not adjusting for, differences in key TEMs (sex; sex and ethnicity). Results were generally similar for both adjusted and unadjusted analyses, and across both estimands.

Although SURMOUNT-5 provides a direct head-to-head comparison between the MTD of tirzepatide and semaglutide [19], this ITC using SURMOUNT-1 and STEP 1 provides novel insights as it focuses on the individual licensed doses of tirzepatide: 5, 10 and 15 mg. Although SURMOUNT-1, SURMOUNT-5 and STEP 1 reported on similar body weight and metabolic risk factor outcomes (with SURMOUNT-5 additionally reporting on ≥30% weight reduction), SURMOUNT-5 did not report on the 5 mg dose of tirzepatide nor on body composition outcomes, highlighting the added value of this ITC. Furthermore, SURMOUNT-1 and STEP 1’s double-blind methodology minimizes potential bias, which may enhance the robustness of the ITC findings in comparison to the open-label SURMOUNT-5.

Results in SURMOUNT-5 generally aligned with results from this ITC. In SURMOUNT-5 at Week 72 for the efficacy estimand, tirzepatide MTD had greater reductions in weight versus semaglutide MTD (mean [95% CI] –6.2% [–7.8, –4.6]) and waist circumference (–5.4 cm [−7.1, –3.7]), which are similar in magnitude to the reductions observed for tirzepatide 10 and 15 mg versus semaglutide 2.4 mg in this ITC (weight reduction: −5.10% [−6.63, −3.57], −6.50% [−8.03, −4.97]; waist circumference: −5.25 cm [−6.69, −3.81], −5.75 cm [−7.19, −4.31]). Differences between SURMOUNT-5 results and this ITC are likely primarily attributable to the differences in dosing. Overall, the results of this ITC support SURMOUNT-5 in confirming that tirzepatide demonstrated greater efficacy relative to semaglutide.

The findings of this ITC expand upon a previous SURMOUNT-1 and STEP 1 ITC, which demonstrated statistically greater reductions in body weight (%) and odds of achieving ≥5% weight reduction with tirzepatide 10 and 15 mg versus semaglutide 2.4 mg [21]. Furthermore, findings generally align with other indirect comparisons of tirzepatide and semaglutide in patients with obesity or overweight. An NMA including additional GLP-1 RAs found that tirzepatide 10 and 15 mg were associated with greater weight reductions than semaglutide 2.4 mg. The safety profiles showed that tirzepatide and GLP-1 RAs were associated with more GI AEs than placebo; however, there were no significant differences between the number of events for tirzepatide versus GLP-1 RAs [22]. In another NMA comparing tirzepatide, GLP-1 RAs and other classes of OMM, tirzepatide 15 mg ranked in the top three for efficacy in weight reduction, glycemia, lipid, and blood pressure parameters. Additionally, tirzepatide 15 mg was associated with the highest efficacy for achieving ≥15% weight reduction. However, tirzepatide and GLP-1 RAs were associated with significant increases in GI AEs compared with placebo [23]. Finally, compared with a cohort study conducted in real-world setting, this ITC showed a greater difference in weight (kg), which may be explained by the pooling of tirzepatide doses in the cohort study [20].

Strengths and generalizability

Two pivotal trials were identified for the ITC, SURMOUNT-1 for tirzepatide, and STEP 1 for semaglutide. As both were RCTs, within-trial bias was reduced and randomization preserved in the analysis.

A heterogeneity assessment ensured that SURMOUNT-1 and STEP 1 were comparable in terms of study design, patient populations, placebo response and the reported outcomes and timepoints. The trials were considered generally homogenous in terms of alignment of eligibility criteria and patient population, other than for potential TEMs identified through literature and clinical relevance, which were adjusted for in MAICs. Both trials were multinational, with representation from North America, South America and Asia in both, implying that the ITC findings are generalizable across multiple geographies.

This ITC utilized two statistical methods to strengthen the conclusions: standard Bucher ITC analyses, and MAICs where feasible to account for differences in the proportion of female and Hispanic patients between the two trials. The methodology aligned with key guidelines [27,28,29]. The MAIC results were generally similar to the Bucher ITC results. In addition, efficacy results were produced for both efficacy and treatment regimen estimands to provide comparisons both in patients adhering to, and not adhering to treatment. There was a generally similar direction of results for both estimands, and these consistent findings support the generalizability of this indirect comparison between tirzepatide and semaglutide.

Limitations

Limitations of this work include the focus on two pivotal trials, SURMOUNT-1 and STEP 1. This resulted in a relatively small evidence base of two key and comparable trials used to inform the ITC. As a result, treatment comparisons were only possible for outcomes reported by both trials. Due to a lack of data reported in STEP 1, analyses were not possible for progression to T2D for both efficacy and treatment regimen estimands, and severe GI AEs for the safety population. Furthermore, ≥20% weight loss was an exploratory outcome in STEP 1, while all outcomes were gated in SURMOUNT-1. Although the two trials were generally well-balanced in their TEMs, facilitating a robust comparison, there were some differences in the percentage of female and Hispanic patients, and STEP 1 included European patients while SURMOUNT-1 did not. While these differences were accounted for in MAICs for most outcomes, MAICs were not deemed feasible for body composition outcomes due to the smaller sample sizes of the DEXA subgroups and the very small proportion of Hispanic patients for STEP 1.

In addition, analyses were restricted to one timepoint for each trial, which may not capture fluctuations in relative treatment effect over the course of follow-up and did not align exactly between SURMOUNT-1 (Week 72) and STEP 1 (Week 68) due to the trials having different primary timepoints. However, this difference was small, with four additional weeks required to reach tirzepatide 15 mg in SURMOUNT-1 compared to semaglutide 2.4 mg in STEP 1, and given that results were observed to plateau before Week 68, the results from both trials were considered clinically comparable. While there was no Week 68 visit in SURMOUNT-1 at which to analyze data to determine the impact of different timepoints, a previous ITC modeled Week 68 SURMOUNT-1 data and had similar results to analyses using Week 72 SURMOUNT-1 data for weight reduction (%) [21].

Despite the use of population-adjusted analyses, care should be taken when interpreting the body composition outcomes as these were based on subgroup results and limited by smaller sample sizes. However, in the absence of head-to-head trial data, these results should be considered informative within the current evidence base. These results should be validated in future head-to-head trials when such data become available to confirm robustness and refine effect estimates. Care should also be taken when interpreting the safety results due to possible differences in the two trials during the collection of safety data, with varying rates of GI AEs, nausea AEs and all-cause discontinuation observed in the placebo arms of STEP 1 and SURMOUNT-1.

Conclusion

In this ITC based on SURMOUNT-1 and STEP 1, tirzepatide 10 and 15 mg were associated with meaningful improvements across efficacy outcomes compared to semaglutide, while tirzepatide 5 mg was associated with a similar profile to semaglutide 2.4 mg. Furthermore, all doses of tirzepatide were associated with a generally similar safety profile to semaglutide 2.4 mg.

MAICs adjusting for the differences in the percentage of female and Hispanic participants between the two trials showed generally similar results to the Bucher ITC, and efficacy results were similar across both efficacy and treatment regimen estimand analyses.

In conclusion, this rigorous comparison provides evidence of an association with comparative or greater clinical effectiveness and similar safety for tirzepatide 10 and 15 mg relative to semaglutide 2.4 mg for weight management in participants without T2D, and with obesity, or overweight with ≥1 ORC. These results align with findings from the head-to-head SURMOUNT-5 trial and support the relative strength of tirzepatide 10 and 15 mg versus semaglutide 2.4 mg in achieving weight reduction, improvements in body composition and addressing multiple cardiometabolic risk factors for ORCs.