Document resource
WHAT IS ALREADY KNOWN ON THIS TOPIC In the SURPASS-SWITCH trial, switching to tirzepatide from dulaglutide versus escalating dulaglutide resulted in significantly greater improvements in hemoglobin A1c (HbA1c) and body weight in participants with T2D on 0–3 oral antihyperglycemic medications.WHAT THIS STUDY ADDS This study explored the consistency of the treatment effect across a range of baseline characteristics that may be relevant to clinical decision-making and setting treatment goals.HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY This study indicates that switching to tirzepatide may represent a clinical option for patients across a variety of baseline characteristics not achieving clinical goals on current therapy with submaximal dulaglutide.Introduction Type 2 diabetes (T2D) is a chronic and progressive disease that often leads to the development of microvascular and macrovascular complications when glycemic control is not achieved or maintained. With cardiovascular (CV) disease being a major cause of death in people with T2D, it is important to manage people with T2D effectively and as early as possible to delay or prevent these adverse outcomes. 1 2 A patient-centered approach for managing hyperglycemia is recommended, as patients with unique characteristics have different treatment needs and goals.3 One of the classes of medicines for treating T2D is glucagon-like peptide-1 (GLP-1) receptor agonists (RAs), which are incretin-based therapies that improve glycemic control, have a low risk of hypoglycemia, and have weight loss potential.4Dulaglutide, a once-weekly GLP-1 RA, is approved for the treatment of T2D and also the reduction of CV risk in people with T2D. It demonstrated glycemic efficacy and a generally well-tolerated safety profile across studies in the Assessment of Weekly AdministRation of LY2189265 (dulaglutide) in Diabetes (AWARD) trial program,5–15 and showed consistent results across baseline subgroups of sex, duration of T2D, and baseline HbA1c.16Tirzepatide, another class of incretin-based therapy, is a dual glucose-dependent insulinotropic polypeptide and GLP-1 RA approved for the treatment of T2D and obesity. It has shown greater glycemic improvement and weight loss than GLP-1 RAs in two head-to-head trials.17 18 In the SURPASS clinical trial program, tirzepatide demonstrated significant reductions in HbA1c and body weight in participants with T2D.18–23 In a post hoc analysis of the SURPASS-1 to SURPASS-5 trials, a composite outcome of glycemic control (HbA1c ≤6.5%) and weight loss (≥10%) without hypoglycemia was achieved across baseline subgroups of T2D duration, sex, HbA1c, age, and body mass index (BMI), with a consistent safety profile.24In SURPASS-SWITCH, switching to tirzepatide from dulaglutide versus escalating dulaglutide resulted in significantly greater improvements in HbA1c and body weight in participants with T2D not achieving glycemic control on a stable dose of dulaglutide (0.75 mg or 1.5 mg) with 0–3 oral antihyperglycemic medications (OAMs).25 When individualizing glycemic management, clinicians may query the applicability of these findings to patients with varied demographic and clinical backgrounds. To address this gap in knowledge, we conducted this subgroup analysis to assess the consistency of the intervention effect for changes in HbA1c and weight across a range of baseline characteristics that may be relevant to clinical decision-making and setting treatment goals: age, HbA1c, duration of T2D, baseline dulaglutide dose, duration of dulaglutide dose, BMI, ethnicity, and sex.Methods SURPASS-SWITCH Design overview This analysis was conducted in prespecified subgroups of participants in SURPASS-SWITCH (ClinicalTrials.gov registration NCT05564039), a phase IV, randomized, open-label, active-controlled, parallel-group, multicenter, multinational trial in adults with T2D with inadequate glycemic control.25Setting and participants SURPASS-SWITCH enrolled adults ≥18 years of age with T2D, HbA1c ≥7.0% to ≤9.5%, BMI ≥25 kg/m 2, on a stable, non-maximal dose of once-weekly dulaglutide (0.75 mg or 1.5 mg) for at least 6 months, and treatment with a stable dose of 0–3 OAMs for at least 3 months (any combination of metformin, sulfonylurea, and sodium-glucose co-transporter 2 inhibitor). The study comprised a 3-week screening period, a 40-week treatment period, and a 4-week safety follow-up period.25Randomization and interventions Using an interactive web response system, participants were randomly and sequentially assigned 1:1 through block randomization (block size of 2) to continue with and escalate to dulaglutide 4.5 mg or maximum tolerated dose (MTD) or switch to tirzepatide 2.5 mg and escalate to 15 mg or MTD. Dose escalations for both study medications were performed every 4 weeks using a single-dose pen. 25Blinding SURPASS-SWITCH was an open-label study to allow for dose escalation and modification. 25 To avoid the introduction of bias when detecting events in the study, clinicians used open-ended and non-leading verbal questioning when asking participants about event occurrences. A blinding and unblinding plan was developed to minimize bias and limit access to the unblinded data during the course of the study.Blinding was not applicable to the subgroup analyses.Compliance data All patients were compliant with taking their study medications and were included in the subgroup analyses. Treatment compliance was defined as taking at least 75% of the required doses of study medication.Procedures Study procedures have been previously described and can be found in the study protocol and Statistical Analysis Plan (SAP) ( online supplemental file 2). Briefly, participants attended study visits for physical examination, collection of vitals including weight, collection of blood samples for scheduled laboratory assessments including HbA1c, assessment of patient-reported outcomes, and assessment of tolerability and safety.25SP210.1136/bmjdrc-2025-005711.supp2Supplementary dataOutcomes and follow-up The primary end point was the change from baseline (CFB) in HbA1c at week 40. The key secondary end point was CFB in body weight at week 40. 25 These two end points were further assessed for the subgroups that were prespecified in the protocol and SAP. Safety end points were treatment-emergent adverse events (TEAEs) collected throughout the duration of the study and during safety follow-up (4 weeks after the last dose).Statistical analyses Overall study population Statistical analyses for the overall study population have been described. 25 Briefly, changes in HbA1c and weight were assessed using both the efficacy estimand (all randomized participants who received at least one dose of study drug, with data included up to treatment discontinuation, initiation of rescue antihyperglycemic medication, prohibited treatment, or change in background OAM) and treatment-regimen estimand (all randomized participants who received at least one dose of study drug, regardless of treatment discontinuation, initiation of rescue or prohibited treatment, or change in background OAM). Safety was analyzed using data obtained during the treatment period plus the 4-week safety follow-up period for the modified intent-to-treat population (all randomly assigned participants who took at least one dose of study drug). The study was powered to detect a 0.58% difference in HbA1c CFB between treatment groups at week 40 irrespective of adherence to treatment, changes in background OAMs, or introduction of rescue antihyperglycemic medication. The study was not powered to detect differences in safety outcomes between the treatment groups.Prespecified subgroups Subgroup analyses using efficacy estimand were prespecified in the study protocol and SAP. Comparisons within treatment subgroups for continuous measurements over time were conducted using the mixed model for repeated measures (MMRM) method. The independent variables of the MMRM model were treatment (tirzepatide, dulaglutide), visit, subgroup, treatment-by-visit interaction, treatment-by-subgroup interaction, visit-by-subgroup interaction, treatment-by-visit-by-subgroup interaction, number of background OAMs (0–1 or 2–3), dulaglutide dose at screening (0.75 mg or 1.5 mg), and region (North America or European Union) as fixed effects, and baseline measurement as a covariate. Missing data were addressed by the MMRM model. No explicit imputation methods for missing data were employed. An unstructured covariance matrix was used to model the within-patient errors.All tests of treatment effects were conducted at a two-sided alpha level of 0.05, unless otherwise stated, and all CIs were given at a two-sided (95%) level. SAS software (V.9.4) and R statistical software were used for statistical analyses.Results Overall study population In SURPASS-SWITCH, 282 participants were randomized to tirzepatide (n=139) and dulaglutide (n=143). The proportion of participants achieving the maximum dose of dulaglutide (4.5 mg) and tirzepatide (15 mg) was 89.5% and 79.9%, respectively. 25Baseline demographics were similar between the treatment groups, except for a lower proportion of females and a greater mean weight in the tirzepatide group. Participants in the study population (n=282) had a mean overall age of 57.6 years (SD 9.8) at baseline and duration of T2D of 11.2 years (SD 7.0), with a BMI of 34.3 kg/m2 (SD 6.3) and HbA1c of 7.82% (SD 0.71). All study participants were on a stable baseline dose of dulaglutide 0.75 mg (94 participants, 33.3%) or 1.5 mg (188 participants, 66.7%). There were 146 (51.8%) males and 136 (48.2%) females enrolled in the trial.25From baseline to week 40, using the treatment-regimen estimand, the CFB in HbA1c was −1.44% (SE 0.07) and −0.67% (SE 0.08) for the tirzepatide and dulaglutide groups, respectively, with changes in body weight of −10.5 kg (SE 0.5) and −3.6 kg (SE 0.5). At week 40, the estimated treatment difference of tirzepatide compared with dulaglutide was −0.77% (95% CI −0.98 to −0.56, p<0.001) for HbA1c and −6.9 kg (95% CI−8.3 to −5.5, p<0.001) for body weight.25The overall incidence of adverse events (AEs) was similar between the tirzepatide and dulaglutide treatment groups. The most commonly reported TEAEs were gastrointestinal. Nausea occurred in 25.2% of participants in the tirzepatide group and 18.9% in the dulaglutide group. Diarrhea occurred in 22.3% of participants in the tirzepatide group and 13.3% in the dulaglutide group. There was a greater occurrence of hypoglycemia in the tirzepatide treatment group. Level 1 hypoglycemia (blood glucose ≥54 and <70 mg/dL) occurred in 18.0% and 8.4% of participants in the tirzepatide and dulaglutide groups, respectively. Level 2 hypoglycemia (blood glucose <54 mg/dL) occurred in 5.8% and 2.8% of participants in the tirzepatide and dulaglutide groups, respectively. There were 10 serious AEs in each treatment group, and five participants discontinued study treatment due to AEs (four in the tirzepatide group, one in the dulaglutide group).25HbA1c reduction by baseline subgroups The frequency of participants across baseline subgroups is shown in online supplemental table S1. The size of the subgroups varied for the different baseline characteristics but was notably smaller for the subgroup of participants with BMI <27 kg/m2 (tirzepatide: n=12; dulaglutide: n=15) compared with those with BMI ≥27 kg/m2 (tirzepatide: n=127; dulaglutide: n=128).SP110.1136/bmjdrc-2025-005711.supp1Supplementary dataThere was a consistent treatment effect in HbA1c reduction from baseline to week 40 across baseline subgroups using the efficacy estimand (figure 1A). The estimated treatment difference for tirzepatide versus dulaglutide was numerically consistent across subgroups using the efficacy estimand, ranging from −0.7% to −1.2% in favor of tirzepatide (figure 1B).Figure 1(A) Change from baseline to week 40 in HbA1c (%) across baseline subgroups. Data are least squares (LS) mean (SE). (B) Estimated treatment difference for tirzepatide versus dulaglutide in HbA1c reduction across baseline subgroups. Data are LS mean (95% CI). All data are for the modified intent-to-treat efficacy estimand. DU, dulaglutide; HbA1c, hemoglobin A1c; T2D, type 2 diabetes; TZP, tirzepatide.Participants with HbA1c >8.5% had larger HbA1c reductions in both the tirzepatide and dulaglutide treatment groups (2.3% (SE 0.2) and 1.1% (SE 0.2), respectively). The tirzepatide treatment effect was larger in participants with a lower BMI of <27 kg/m2, with HbA1c reduction of 2.0% (SE 0.3) for tirzepatide compared with 0.3% (SE 0.3) for dulaglutide (online supplemental figure S1).In participants with HbA1c >8.5%, the estimated treatment difference was −1.2% (95% CI −1.8 to −0.6) and was similar to the other baseline subgroups. Conversely, participants with BMI <27 kg/m2 had an estimated treatment difference of −1.7% (95% CI −2.5 to −0.9), which was significantly larger than those with BMI ≥27 kg/m2 (interaction p value=0.026) (online supplemental S1).Results were consistent using the treatment-regimen estimand (online supplemental figures S2 and S3).There was a consistent treatment effect for HbA1c reduction across baseline subgroups for both tirzepatide and dulaglutide in participants achieving HbA1c ≤6.5% and <7%. An exception was those treated with dulaglutide with baseline HbA1c >8.5% where there was a small sample size. The number of participants treated with dulaglutide achieving HbA1c <5.7% across baseline subgroups was too small to calculate treatment interaction. Generally, across subgroups of participants achieving HbA1c targets by baseline subgroups, there were more participants in the tirzepatide treatment group than in the dulaglutide treatment group (figure 2).Figure 2Change from baseline to week 40 in HbA1c in participants achieving HbA1c targets of A. <7%, B. ≤6.5%, and C. <5.7%. Data are least squares mean (SE) for the modified intent-to-treat efficacy estimand. DU, dulaglutide; HbA1c, hemoglobin A1c; T2D, type 2 diabetes; TZP, tirzepatide.Weight reduction by baseline subgroups There was a consistent treatment effect for weight reduction across all baseline subgroups using the efficacy estimand ( figure 3A). The estimated treatment difference for tirzepatide versus dulaglutide was numerically consistent across all baseline subgroups using efficacy estimand, ranging from −5.7 kg to −8.6 kg in favor of tirzepatide (figure 3B). In addition, in participants with BMI <27 kg/m2, the estimated treatment difference (95% CI) was −9.6 kg (−13.8 to −5.4) in favor of tirzepatide (online supplemental figure S1).Figure 3(A) Change from baseline to week 40 in body weight (kg) across baseline subgroups. Data are least squares (LS) mean (SE). (B) Estimated treatment difference for tirzepatide versus dulaglutide in body weight reduction across baseline subgroups. Data are LS mean (95% CI). All data are for the modified intent-to-treat efficacy estimand. DU, dulaglutide; HbA1c, hemoglobin A1c; T2D, type 2 diabetes; TZP, tirzepatide.Participants in the non-Hispanic or non-Latino subgroup had a significantly larger tirzepatide treatment effect for reduction in weight compared with those in the Hispanic or Latino subgroup (interaction p value=0.047) (figure 3A). In these subgroups, the estimated treatment difference for tirzepatide versus dulaglutide was −8.6 kg (95% CI −10.3 to −7.0) and −5.9 kg (95% CI −8.1 to −3.7), respectively.Results were consistent using the treatment-regimen estimand (online supplemental figures S4 and S5).The treatment effect for weight reduction was consistent across baseline subgroups in participants achieving weight reductions of ≥5% and ≥10%. The number of participants treated with dulaglutide achieving a weight reduction of ≥15% across baseline subgroups was too small to calculate treatment interaction. Generally, across subgroups of participants achieving weight loss targets by baseline subgroups, there were more participants in the tirzepatide treatment group than in the dulaglutide treatment group (figure 4).Figure 4Change from baseline to week 40 in body weight in participants achieving weight loss targets of A. ≥5%, B. ≥10%, and C. ≥15%. Data are least squares mean (SE) for the modified intent-to-treat efficacy estimand. DU, dulaglutide; HbA1c, hemoglobin A1c; T2D, type 2 diabetes; TZP, tirzepatide.Safety profile by baseline subgroups The frequency of occurrence of TEAEs was numerically similar across baseline subgroups, except for numeric differences in participants aged ≥65 years in both treatment arms (82.9% and 81.1% for participants aged ≥65 years vs 69.2% and 70.8% for participants aged <65 years for tirzepatide and dulaglutide, respectively). The frequency of occurrence of hypoglycemia increased as the duration of diabetes increased and was lower in participants with a longer duration of dulaglutide. In the tirzepatide treatment arm, the frequency of occurrence of hypoglycemia was numerically higher in those with lower BMI and higher baseline HbA1c ( table 1). TEAEs occurring in ≥5% of participants are presented in online supplemental table S2, showing a generally well-tolerated safety profile across baseline subgroups, with nausea and diarrhea usually being the most common TEAEs.Table 1Frequency of TEAEs and hypoglycemia across baseline subgroupsBaseline subgroup≥1 TEAEHypoglycemia <70 mg/dLTZP 15 mg or MTDDU 4.5 mg or MTDTZP 15 mg or MTDDU 4.5 mg or MTDAge (years)<6572 (69.2)75 (70.8)8 (7.7)7 (6.6)≥6529 (82.9)30 (81.1)4 (11.4)0 (0.0)SexMale52 (69.3)52 (73.2)6 (8.0)2 (2.8)Female49 (76.6)53 (73.6)6 (9.4)5 (6.9)EthnicityHispanic or Latino44 (72.1)41 (66.1)6 (9.8)3 (4.8)Non-Hispanic or Latino57 (73.1)64 (79.0)6 (7.7)4 (4.9)Duration of T2D (years)≤522 (73.3)22 (71.0)1 (3.3)0 (0.0)>5 to ≤1024 (63.2)28 (70.0)3 (7.9)2 (5.0)>1055 (77.5)55 (76.4)8 (11.3)5 (6.9)BMI (kg/m2)<278 (66.7)11 (73.3)3 (25.0)0 (0.0)≥2793 (73.2)94 (73.4)9 (7.1)7 (5.5)<3024 (72.7)31 (75.6)4 (12.1)1 (2.4)≥30 to <3532 (71.1)35 (68.6)6 (13.3)2 (3.9)≥3545 (73.8)39 (76.5)2 (3.3)4 (7.8)HbA1c (%)≤8.581 (71.7)87 (71.9)8 (7.1)6 (5.0)>8.520 (76.9)18 (81.8)4 (15.4)1 (4.5)DU dose (mg)0.7534 (72.3)34 (72.3)4 (8.5)3 (6.4)1.567 (72.8)71 (74.0)8 (8.7)4 (4.2)Duration of DU (years)<132 (69.6)30 (75.0)6 (13.0)3 (7.5)≥169 (74.2)75 (72.8)6 (6.5)4 (3.9)Data are presented as n (%). Hypoglycemia analysis includes documented symptomatic hypoglycemia.BMI, body mass index; DU, dulaglutide; HbA1c, hemoglobin A1c; MTD, maximum tolerated dose; T2D, type 2 diabetes; TEAE, treatment-emergent adverse event; TZP, tirzepatide.Discussion Subgroup analyses provide additional information for clinical decision-making by assessing heterogeneity of treatment effects in subgroups of trial participants. In SURPASS-SWITCH, participants who switched from dulaglutide to tirzepatide had significantly greater reductions in HbA1c and body weight across all baseline subgroups, compared with those who escalated dulaglutide dose. For HbA1c reduction, the treatment-by-subgroup interaction p value was only significant in the comparison of BMI <27 kg/m 2 with BMI ≥27 kg/m2. For body weight reduction, the treatment-by-subgroup interaction p value was only significant for Hispanic or Latino versus non-Hispanic or non-Latino subgroups.In previous subgroup analyses of the GLP-1 RAs, dulaglutide, semaglutide, and liraglutide, elevated baseline HbA1c was associated with greater reductions in HbA1c.16 26–29 In one pooled analysis of the Semaglutide Unabated Sustainability in Treatment of Type 2 Diabetes (SUSTAIN) trials, HbA1c and body weight reductions were similar regardless of the duration of T2D, and the magnitude of weight reduction with semaglutide was lower in participants with elevated baseline HbA1c.26 In another pooled analysis of the SUSTAIN trials, weight loss was significant versus placebo and active comparators regardless of baseline BMI.30 In a head-to-head trial comparing exenatide with liraglutide, higher baseline HbA1c was associated with larger reduction in HbA1c and numerically larger reduction in body weight for both treatment groups.31 In the AWARD trials, weight reduction with dulaglutide was similar across baseline HbA1c subgroups except in AWARD-4, where dulaglutide combined with mealtime insulin lispro was not associated with weight loss in participants with higher baseline HbA1c.16 Additionally, reductions in HbA1c were similar between males and females and regardless of duration of T2D; also, females had numerically larger reductions in weight or lower weight gain.16 In AWARD-11, improvements in HbA1c reduction with dulaglutide dose escalation were larger in those with higher baseline HbA1c.32 In Japanese participants, dulaglutide significantly improved HbA1c regardless of age, sex, HbA1c, body weight, BMI, and duration of diabetes.28 HbA1c reductions were larger in older participants and in participants with lower body weight, while weight reductions were larger in females, participants with lower HbA1c, and participants with lower body weight.28 The findings of the current work regarding changes in HbA1c are consistent with those reported for these GLP-1 RAs. When interpreting weight changes compared with the current work, the use of different concomitant T2D medications and their effects on weight should be considered.In the SURPASS-1 to SURPASS-5 studies for tirzepatide, a composite end point of HbA1c ≤6.5% and weight reduction ≥10% without hypoglycemia was similarly achieved with tirzepatide treatment in subgroups of baseline HbA1c, duration of T2D, and BMI (<30, ≥30 and <35, ≥35 kg/m2), and was achieved by significantly more females than males.24 In a post hoc analysis of SURPASS-1 to SURPASS-4, the efficacy and safety of tirzepatide was similar between Hispanic or Latino and non-Hispanic or non-Latino participants except in SURPASS-2, where tirzepatide 15 mg was associated with significantly smaller reductions in body weight in the Hispanic or Latino subgroup, which may have been due to their lower baseline body weight.29 In a post hoc analysis of SURPASS-1 to SURPASS-4, the odds of achieving normoglycemia (HbA1c <5.7%) were reduced with increased age, duration of T2D, and baseline HbA1c.33 Participants who achieved normoglycemia had a larger reduction in body weight.33 In East Asian participants, the efficacy profile of tirzepatide was similar across BMI and age subgroups.34 In pooled Japanese participants from the SURPASS studies, the tirzepatide treatment effect versus dulaglutide for HbA1c reduction was greater in those with baseline HbA1c >8.5% than ≤8.5%, while body weight reduction was similar between the subgroups.35In this current work, HbA1c reduction was numerically larger in participants with HbA1c >8.5% in both the tirzepatide and dulaglutide treatment groups. Weight reduction was similar between HbA1c subgroups. Reductions in HbA1c and body weight were similar regardless of baseline age, duration of T2D, sex, BMI (<30, ≥30 and <35, ≥35 kg/m2), dulaglutide dose, and duration of dulaglutide treatment. The tirzepatide treatment effect for HbA1c reduction was significantly larger in participants with BMI <27 kg/m2 compared with those with BMI ≥27 kg/m2, the significance of which is unclear given the small number of participants in the BMI <27 kg/m2 group. BMI ≥27 kg/m2 is associated with greater obesity-related risks, including hypertension, hypercholesterolemia, and premature death36; thus, BMI is an important consideration for clinical decision-making, with high BMI being associated with worse outcomes in T2D. In a retrospective study, individuals with T2D and class 1 or 2 obesity were 24% more likely to have HbA1c ≥7%.37 Other studies have shown an association between BMI and CV outcomes. Those with T2D and BMI ≥30 kg/m2 were 23% more likely to have microvascular complications and 56% more likely to have a history of CV disease,38 while a separate study showed an increased risk of vascular complications with each 5 kg/m2 increase in BMI.39Additionally, the tirzepatide treatment association with weight reduction was significantly larger in non-Hispanic or non-Latino participants, similar to SURPASS-2. This is relevant because different diets, cultural beliefs, diabetes care, and socioeconomic status may affect disease management and lifestyle.29 The mechanism for these variations in associated weight reduction is unclear, but regional genetic variations in Hispanic/Latino subgroups, and variants in genes such as single-nucleotide polymorphisms for glp1r, gipr, and tcf7l2, may play a role in concert with lifestyle and beliefs and requires further elucidation.40–42In summary, the prespecified subgroups assessed in this work represent some of the patient characteristics clinicians may take into consideration when changes to treatment plans are needed. Regarding mean HbA1c changes, the changes were consistent across subgroups but were greater for those with higher baseline HbA1c (>8.5%), the latter of which is consistent with observations in other published literature. The similarity in reduction by duration of diabetes suggests efficacy can be achieved regardless of time since diagnosis of T2D. Mean weight changes observed were also relatively consistent across subgroups. Participants in the higher BMI subgroups had slightly greater numerical reductions in weight than those in the lower BMI subgroups. In older participants, a large amount of weight reduction may not be desired; in this work, the reduction in weight was only slightly greater numerically in participants aged ≥65 years.More participants in this older subgroup also reported ≥1 TEAE. The frequency of hypoglycemia increased as the duration of diabetes increased, which may be consistent with progressive β-cell dysfunction with longer disease duration/progression. In the tirzepatide treatment arm, the frequency of hypoglycemia was higher in those with lower BMI, which may reflect greater insulin sensitivity and reduced buffering of blood glucose by adipose tissue at lower BMI. Hypoglycemia was also higher in older participants, suggesting that additional counseling may help this population to be aware of signs and symptoms and how to manage it. However, given the limited number of events, subgroup trends in hypoglycemia require further evaluation.These findings should be interpreted keeping in mind the limitations of an open-label trial design and the small sample size for certain subgroup populations. The power calculation was for the overall primary outcome and not for these subgroups or for the safety outcomes. Multiple subgroup testing increases the risk for false positives; hence, caution in interpretation is needed. The study is also limited by the eligibility criteria of the trial. The HbA1c inclusion criterion was 7%–9.5%, and the BMI inclusion criterion was ≥25 kg/m2, meaning the analysis did not include people in the normal HbA1c or BMI range. Additionally, there are other subgroups that may be clinically relevant that were not prespecified and described in this analysis.Conclusion In this subgroup analysis, across clinically relevant baseline subgroups, switching to tirzepatide from dulaglutide was generally well-tolerated with significant and consistent reductions in HbA1c and body weight. While there was statistically significant improvement from baseline in HbA1c and body weight for each treatment group, the improvement from baseline was generally greater when switching to tirzepatide, irrespective of age, sex, ethnicity, duration of diabetes, BMI, HbA1c, dulaglutide dose, or duration of dulaglutide. A greater proportion, numerically, of older participants reported at least one TEAE compared with younger participants in both treatment groups. This analysis indicates that switching to tirzepatide may represent a clinical option for patients with a broad variety of baseline characteristics not achieving clinical goals on current therapy with submaximal dulaglutide, even beyond escalation of current therapy. While these results provide additional information for clinical decision-making, further examination of relevant clinical subgroups in larger and more diverse populations of people with T2D could form the basis for even greater supporting evidence for clinical application.