FormalPara Key Summary Points

Why carry out this study?

Early glycaemic control and body weight loss efficacy are clinically relevant in people with type 2 diabetes, provided they can be achieved safely.

Tirzepatide has demonstrated significantly greater HbA1c and body weight reductions than semaglutide 1 mg (SURPASS-2) and insulin degludec (SURPASS-3) at 40 weeks and 52 weeks, respectively.

We explored the time to achieve glycaemic control and body weight loss in these studies.

What was learned from the study?

More participants achieved early glycaemic control, and glycaemic thresholds (HbA1c < 7.0% and ≤ 6.5%) were achieved faster with tirzepatide compared with semaglutide 1 mg and titrated insulin degludec.

Early body weight reductions were significantly greater with tirzepatide than with semaglutide 1 mg and the body weight loss goal of ≥ 5% was achieved significantly faster.

These findings will inform patients and healthcare professionals regarding the early efficacy profile of tirzepatide and guide in treatment decisions.

Introduction

Tirzepatide is a once-weekly glucose-dependent insulinotropic polypeptide (GIP)/glucagon-like peptide-1 (GLP-1) receptor agonist approved in major geographies including the USA, Europe, and Japan for the treatment of people with type 2 diabetes and under investigation for chronic weight management. Across the SURPASS phase 3 clinical trial programme, tirzepatide demonstrated clinically meaningful reductions in glycated haemoglobin (HbA1c) and body weight as monotherapy and when added to different oral glucose lowering medications and titrated basal insulin [1,2,3,4,5]. The decreases in HbA1c and body weight achieved with tirzepatide were significantly greater than with placebo, the basal insulins degludec and glargine, and the selective GLP-1 receptor agonist semaglutide 1 mg at treatment completion [1,2,3,4,5].

GLP-1 receptor agonists are associated with gastrointestinal adverse events, which are often most frequent during treatment initiation and dose escalation [6]. To mitigate this and improve gastrointestinal tolerability, treatment is initiated at a lower dose and gradually increased. Within these dose escalation schemes, the initial doses may not be intended for glycaemic control and the time it takes to reach the more efficacious (e.g., maintenance) dose varies between treatments. For example, the 5, 10, and 15 mg doses of tirzepatide can be reached 4 weeks, 12 weeks, and 20 weeks, respectively, after treatment initiation, while the 1 mg semaglutide dose can be reached after 8 weeks [7, 8]. Although these differences make indirect comparisons of early efficacy between treatments challenging, the dose escalation scheme may become a clinically relevant factor when selecting agents for patients with varied glycaemic control needs. It is therefore clinically important to understand the early efficacy profile and time to achieve clinical targets, and is particularly relevant given that treatment guidelines suggest therapy re-evaluation every 3 to 6 months based on HbA1c levels [6]. In an exploratory pre-planned analysis of the published SURPASS-2 and SURPASS-3 studies [2, 3], we evaluated time to achieve glycaemic control and weight loss targets with tirzepatide once weekly (QW) compared with semaglutide 1 mg QW and insulin degludec once daily.

Methods

Primary Study Designs

SURPASS-2 was an open-label trial in which participants were randomised 1:1:1:1 to tirzepatide 5, 10, or 15 mg QW (with double-blinded doses) or semaglutide 1 mg QW for a 40-week treatment period [2]. Participants had type 2 diabetes and inadequate glycaemic control (HbA1c of 7.0–10.5%) with metformin ≥ 1500 mg per day, body mass index (BMI) ≥ 25 kg/m2, and stable weight during the previous 3 months.

In the open-label SURPASS-3 trial, participants were randomised 1:1:1:1 to tirzepatide 5, 10, or 15 mg or once-daily titrated insulin degludec for 52 weeks of treatment. Participants were insulin-naïve and had type 2 diabetes with inadequate glycaemic control (HbA1c 7.0–10.5%) with metformin ≥ 1500 mg per day alone or in combination with a sodium-glucose co-transporter-2 (SGLT2) inhibitor, BMI ≥ 25 kg/m2, and stable weight during the previous 3 months [3].

In both studies, tirzepatide was initiated at 2.5 mg QW for 4 weeks, and the dose was increased by 2.5 mg every 4 weeks until the assigned dose was reached. In SURPASS-2, semaglutide was initiated following its label at 0.25 mg, with the dose doubled every 4 weeks until 1 mg was reached and maintained for the rest of the study. In SURPASS-3, insulin degludec (100 U/ml) was initiated at 10 U per day and titrated weekly to a fasting blood glucose target of < 90 mg/dl (5 mmol/l), using a treat-to-target algorithm based on self-monitored blood glucose (Supplementary Material p2).

Full study design and methodological details for SURPASS-2 and SURPASS-3 have been published previously [2, 3]. The protocols were approved by local institutional review boards and the studies were conducted in accordance with the principles of the Declaration of Helsinki and Good Clinical Practice guidelines of the International Council for Harmonisation. Both are registered at ClinicalTrials.gov, NCT03987919 and NCT03882970. All participants provided written informed consent.

Outcomes for Analyses

In this exploratory pre-planned analysis, change from baseline in HbA1c and the proportion of participants achieving glycaemic thresholds were assessed between tirzepatide and comparators at Weeks 4, 12, and 24 in both studies. Week 4 was selected for analyses as it was the earliest measurement in the studies. Weeks 12 and 24 were selected as HbA1c is typically measured every 3 months in clinical practice. The time to first reach the HbA1c thresholds of < 7.0% and ≤ 6.5% was assessed. Additionally, since weight loss is increasingly recognised as important in the treatment of type 2 diabetes, the change from baseline in body weight and the proportion of patients achieving body weight loss of ≥ 5% were analysed in SURPASS-2, along with time to first reach weight loss of ≥ 5%.

Statistical Analysis

Analyses were conducted on the efficacy analysis set in the modified intent-to-treat (mITT) population. The mITT population included all randomised participants who took at least one dose of study drug. The efficacy analysis set included data from the mITT population excluding patients who discontinued the study drug because of inadvertent enrolment and data after initiating rescue medication (for severe persistent hyperglycaemia) or early discontinuation of study drug. HbA1c and body weight were analysed using a mixed model for repeated measures in the efficacy analysis set with treatment, visit, treatment-visit interaction, stratification factors, and corresponding baseline value as covariates [2, 3]. For categorical end points, longitudinal logistic regression was used with the same factors and covariates. A Cox proportional hazard model was used to compare hazard rates among treatment groups. Time-to-event analyses and median were estimated using the Kaplan-Meier method. When thresholds were not met, patients were censored at last date on study drug or initiation date of rescue medication. All P values are two-sided, with a level of 0.05 considered to be statistically significant.

Results

Baseline Characteristics

In SURPASS-2, 1878 participants were included in the mITT population and 1783 completed the study [2]. Mean baseline HbA1c was 8.28%, fasting serum glucose was 172.9 mg/dl (9.6 mmol/l), body weight was 93.7 kg, and BMI was 34.2 kg/m2 (Supplementary Material p3).

In SURPASS-3, the mITT population included 1437 participants and 1325 completed the study [3]. Mean HbA1c was 8.17%, fasting serum glucose was 169.3 mg/dl (9.4 mmol/l), body weight was 94.3 kg, BMI was 33.5 kg/m2, and all participants used metformin with 32% also using an SGLT2-inhibitor (Supplementary Material p3).

In both studies, the mean duration of diabetes was approximately 8.5 years. Baseline characteristics were similar across the treatment groups in each study [2, 3].

Early Glycaemic Efficacy

In SURPASS-2, treatment with tirzepatide 5, 10, and 15 mg QW and semaglutide 1 mg QW led to clinically relevant reductions from baseline in HbA1c across all time points (Table 1). Reductions with all tirzepatide doses were significantly greater (− 0.76% to − 0.80% vs. − 0.59% at Week 4, − 1.84% to − 1.97% vs. − 1.50% at Week 12, and − 2.14% to − 2.43% vs. − 1.88% at Week 24, respectively) than with semaglutide (all P < 0.001). In both studies, findings were also consistent across other time points at which HbA1c was measured (Weeks 8, 16, and 20, data not shown).

Table 1 HbA1c and body weight changes at early time points in SURPASS-2 and SURPASS-3
Full size table

A higher proportion of tirzepatide-treated participants across the dosage range achieved HbA1c < 7.0% at Week 4 (22.7–23.5% vs. 15.3%), Week 12 (91.7–93.3% vs 78.7%), and Week 24 (94.7–97.4% vs. 93.0%) compared with semaglutide 1 mg-treated participants (P = 0.003 for the 15 mg dose at Week 24 and all others P ≤ 0.001, except the 5 mg dose at Week 24 where P = 0.331) (Fig. 1a). A similar pattern was observed for HbA1c ≤ 6.5% (P = 0.034 for the 5 mg dose and P = 0.003 for the 10 mg dose at Week 4, P < 0.001 for others), except for tirzepatide 5 mg at Week 24, which was not significant (P = 0.08) (Fig. 1a).

Fig. 1
figure 1

Proportion of patients achieving glycaemic and weight loss goals at early time points in SURPASS-2 and SURPASS-3. Proportion of patients achieving glycaemic goals in SURPASS-2 (a) and SURPASS-3 (b) and weight loss ≥ 5% in SURPASS-2 (c). Data are least-squares mean (standard error) percentage. †P = 0.034, §P = 0.003, P = 0.005, ‡P = 0.006, and *P ≤ 0.001 vs. semaglutide 1 mg in SURPASS-2 and insulin degludec in SURPASS-3. HbA1c glycated haemoglobin, TZP tirzepatide

Full size image

In SURPASS-3, the mean (standard deviation) dose of insulin degludec received was 21.2 U (6.7) at Week 4, 33.3 U (14.4) at Week 12, and 41.1 U (21.2) at Week 24. At Weeks 4, 12, and 24, HbA1c decreases from baseline were significantly greater with each dose of tirzepatide compared with insulin degludec (all P < 0.001) (Table 1). Additionally, the proportion of patients achieving HbA1c < 7.0% and ≤ 6.5% was significantly greater with each dose of tirzepatide compared with insulin degludec at Weeks 4, 12, and 24 (P = 0.005 for HbA1c ≤ 6.5% with tirzepatide 10 mg at 4 weeks and P < 0.001 for others), except for HbA1c ≤ 6.5% with tirzepatide 10 mg at Week 4 (P = 0.068) (Fig. 1b).

Time to Achieve Glycaemic Control

In SURPASS-2, the median time to first reach HbA1c < 7.0% was 8.1 weeks for each tirzepatide dose compared with 12.0 weeks for semaglutide (HR [95% CI] 1.25 [1.08, 1.45] for tirzepatide 5 mg, 1.40 [1.21, 1.62] for 10 mg, and 1.42 [1.23, 1.64] for the 15 mg dose; P = 0.0023 for 5 mg and P < 0.001 for 10 mg and 15 mg) (Fig. 2a). The median time to first reach HbA1c ≤ 6.5% was 12.1 weeks for each dose of tirzepatide compared with 15.7 weeks for semaglutide 1 mg (HR [95% CI] 1.24 [1.07, 1.45] for 5 mg, 1.57 [1.35, 1.83] for 10 mg, and 1.60 [1.38, 1.86] for 15 mg; P = 0.0047 for 5 mg and P < 0.001 for 10 mg and 15 mg) (Fig. 2b).

Fig. 2
figure 2

Time to reach glycaemic targets in SURPASS-2 and SURPASS-3. Kaplan-Meier plots of time to first reach glycaemic targets. Labels indicate median time to achieve target in each group. †P = 0.0023, §P = 0.0047, and *P < 0.001 vs. semaglutide 1 mg in SURPASS-2 and insulin degludec in SURPASS-3. HbA1c glycated haemoglobin, iDeg insulin degludec, Sema semaglutide 1 mg, TZP tirzepatide

Full size image

In SURPASS-3, the median time to first reach HbA1c < 7.0% was 8.1 weeks for each dose of tirzepatide compared with 12.1 weeks for insulin degludec (HR [95% CI] 1.75 [1.47, 2.08] for tirzepatide 5 mg, 1.79 [1.50, 2.13] for 10 mg, and 1.93 [1.62, 2.29] for the 15 mg dose; all P < 0.001) (Fig. 2c). The median time to first reach HbA1c ≤ 6.5% was 12.1 weeks for each tirzepatide dose compared with 24.1 weeks for insulin degludec (HR [95% CI] 2.21 [1.84, 2.65], 2.36 [1.96, 2.84], and 2.59 [2.15, 3.12], respectively, for the 5, 10, and 15 mg doses of tirzepatide; all P < 0.001] (Fig. 2d).

Early Bodyweight Loss Efficacy

Tirzepatide 5, 10, and 15 mg and semaglutide 1 mg significantly reduced body weight at Weeks 4, 12, and 24. At each time point, the magnitude of reduction was significantly greater with each tirzepatide dose than with semaglutide 1 mg (all P < 0.001) (Table 1). Furthermore, significantly greater proportions (4.9–6.7% vs. 1.6% at Week 4, 48.5–55.1% vs. 29.0% at Week 12, and 64.7–82.7% vs. 51.5% at Week 24) of participants who received tirzepatide achieved body weight loss ≥ 5% than participants who received semaglutide 1 mg at these time points (P = 0.006 and P = 0.005 for 10 mg and 15 mg, respectively, at Week 4 and P < 0.001 for all others) (Fig. 1c).

The median time to first reach weight loss of ≥ 5% was 16.0 weeks for tirzepatide 5 mg and 12.4 weeks for both tirzepatide 10 mg and 15 mg compared with 24.0 weeks for semaglutide 1 mg (HR [95% CI] 1.51 [1.28, 1.78], 2.08 [1.77, 2.45], and 2.23 [1.90, 2.62], respectively, for the 5, 10, and 15 mg doses; all P < 0.001) (Fig. 3).

Fig. 3
figure 3

Time to reach body weight loss of ≥ 5% in SURPASS-2. Kaplan-Meier plot of time to first reach body weight loss of ≥ 5%. Labels indicate median time to achieve target in each group. *P < 0.001 vs. semaglutide 1 mg. Sema semaglutide 1 mg, TZP tirzepatide

Full size image

Discussion

In this pre-planned exploratory analysis, people who received tirzepatide achieved the glycaemic thresholds of HbA1c < 7.0% and ≤ 6.5% faster than those treated with semaglutide 1 mg and titrated insulin degludec. Treatment with tirzepatide was also associated with reaching the body weight loss goal of ≥ 5% faster than with semaglutide 1 mg. In addition, more people achieved these treatment thresholds, and the magnitude of HbA1c and body weight reductions was significantly greater with tirzepatide at the early time points of 4, 12, and 24 weeks after treatment initiation.

Semaglutide and tirzepatide treatments are initiated at non-therapeutic doses for the first 4 weeks (0.25 mg and 2.5 mg, respectively) [7, 8]. Despite this, in SURPASS-2 both agents led to clinically significant reductions in HbA1c of − 0.76% to − 0.80% vs. − 0.59% at Week 4, respectively, with a statistically significant difference favouring tirzepatide. By 12 weeks of new treatment initiation, when HbA1c is typically assessed in clinical settings, semaglutide 1 mg, the tirzepatide 5 mg dose, and the 7.5 mg dose for participants escalating to 10 or 15 mg tirzepatide had reached steady state. At this time point, HbA1c reductions were statistically significantly greater with tirzepatide (− 1.84% to − 1.97% vs. 1.50%) than with semaglutide 1 mg. Overall, our findings indicate that the slower dose escalation scheme for tirzepatide compared with semaglutide had no negative impact on the time to reach glycaemic control, given that patients who received tirzepatide reached glycaemic control (i.e., HbA1c < 7.0% and ≤ 6.5%) about 4 weeks faster compared with those who received semaglutide 1 mg. This effect was more pronounced on weight loss, with the median time to lose ≥5% of body weight being half with two higher doses of tirzepatide (12 vs. 24 weeks) compared with semaglutide 1 mg. GLP-1 receptor agonists are one of the most efficacious agents for glycaemic control in patients with type 2 diabetes and can also provide weight loss benefit. However, the real-world persistence of use of these agents is approximately 75% at the end of 3 months and 50% 1 year after treatment initiation [9, 10]. Starting treatment at non-therapeutic doses, longer dose escalation schemes, and gastrointestinal adverse events could be potential reasons for this poor persistence of use in the real world. Understanding the comparative early efficacy of tirzepatide and semaglutide may help clinicians to appropriately set expectations with patients and adjust dose/alter treatments per each patient’s needs.

Glycaemic control was achieved faster in SURPASS-3 with tirzepatide, which had a fixed dose escalation scheme, than with insulin degludec, which was titrated to a fasting blood glucose of < 90 mg/dl. At 24 weeks, by which time the majority of patients in the insulin degludec arm had completed titration, HbA1c changes were significantly greater with tirzepatide 5 mg (− 1.91%), 10 mg (− 2.14%), and 15 mg (− 2.24%) compared with insulin degludec (− 1.36%). Basal insulins are titrated gradually and to different fasting blood glucose targets per patient needs in clinical practice. A similar analysis conducted using LixiLan-O study data compared insulin glargine and a fixed ratio combination of insulin glargine and lixisenatide and reported the median time to reach HbA1c < 7.0% was 24 weeks and 12 weeks, respectively [11]. Faster median time to HbA1c < 7.0% (8 vs. 12 weeks) and ≤ 6.5% (12 vs. 24 weeks) with tirzepatide compared with insulin degludec may be relevant for patients requiring faster glycaemic control. This finding may also have a greater relevance in real-world settings, where it is known that up-titration of insulins is poor compared with a clinical trial [12].

Safety and tolerability are important considerations when initiating or intensifying a treatment. The dose escalation schemes needed for incretin agonists to mitigate gastrointestinal adverse events must be balanced against speed of efficacy. As previously reported, gastrointestinal adverse events, which were primarily mild to moderate in severity, occurred in 40–46% of tirzepatide-treated participants and 41% of semaglutide-treated participants in SURPASS-2 [2]. Our findings, in conjunction with these tolerability data, indicate that tirzepatide can provide faster glycaemic control and body weight reductions, with a similar gastrointestinal adverse event profile to semaglutide 1 mg. This is especially important considering the longer dose escalation period for tirzepatide 10 mg and 15 mg compared with semaglutide 1 mg. Regarding safety, understanding the early glycaemic profile of tirzepatide may also be important for clinicians and patients when considering adjustment of other glucose-lowering medication, especially insulin and sulfonylurea, to manage the risk of hypoglycaemia.

While achieving glycaemic control quickly is beneficial, there may be certain risks. For example, although long-term glycaemic control reduces the risk of diabetic retinopathy, a rapid improvement in glycaemic control may increase retinopathy risk [13]. A higher risk of retinopathy complications was reported for semaglutide 1 mg vs. placebo in SUSTAIN-6, which appeared to be attributable to a marked early HbA1c reduction with semaglutide treatment, particularly in participants with pre-existing retinopathy at baseline [14, 15]. In the tirzepatide SURPASS phase 3 studies, including SURPASS-2 and -3, patients with a history of proliferative diabetic retinopathy, diabetic maculopathy, or nonproliferative diabetic retinopathy that required acute treatment were excluded. The total number of retinopathy cases in SURPASS-2 and SURPASS-3 were low, 2 and 3, respectively. The SURPASS-CVOT study (NCT04255433) includes an addendum to investigate the diabetic retinopathy progression on tirzepatide, which will provide more information in this regard.

Although pre-planned, this study was exploratory in nature and the trial sample sizes were not specifically powered to address some of the questions in these analyses. The higher 2 mg dose of semaglutide for the treatment of type 2 diabetes could not be assessed in these analyses, since it was not approved until after the SURPASS-2 study was completed. We have reported significantly greater body weight loss, which was achieved faster, with tirzepatide compared to semaglutide 1 mg. Indeed, in people living with obesity without type 2 diabetes, the doses of tirzepatide studied here (5, 10, and 15 mg) have shown weight loss efficacy with mean weight changes of − 16.0%, − 21.4%, and − 22.5%, respectively, after 72 weeks of treatment [16]. In contrast, a higher 2.4 mg dose of semaglutide than used in SURPASS-2 has been approved for the treatment of people living with obesity; this showed a 16.9% reduction in body weight after 68 weeks of treatment [17]. Data are from two clinical trials and these populations cannot be fully representative of everyday clinical practice. For example, while in clinical practice fasting blood glucose targets for insulin titration are individualised, in SURPASS-3 all participants in the insulin degludec arm were titrated to a common fasting blood glucose target of < 90 mg/dl (5 mmol/l).

Conclusion

In these exploratory analyses, more patients who received tirzepatide achieved glycaemic thresholds compared with those who received semaglutide 1 mg or insulin degludec, and these thresholds were achieved earlier in treatment. Additionally, tirzepatide-treated participants achieved body weight loss of ≥ 5% significantly faster than participants treated with semaglutide 1 mg.