Systolic blood pressure reduction with tirzepatide in patients with type 2 diabetes: insights from SURPASS clinical program

A database was created using 40/42-week clinical data from five randomized controlled trials, SURPASS-1, -2, -3, -4 and -5. A common 40/42-week primary time point was selected for consistent assessment across the five SURPASS studies. The study design for each trial is described in detail in Additional file 1. Key eligibility criteria, and primary efficacy and safety results have been published previously for all five trials [17‐21] (ClinicalTrial.gov Identifiers: NCT03954834, NCT03987919, NCT03882970, NCT03730662, NCT04039503). Two of the trials were placebo-controlled (SURPASS-1 and -5) while the remaining three trials compared tirzepatide (5 mg, 10 mg and 15 mg) to semaglutide 1 mg, titrated insulin degludec and titrated insulin glargine (SURPASS-2, -3 and -4, respectively). Participants randomized to tirzepatide started at a 2.5 mg dose once weekly and escalated the dose by 2.5 mg every 4 weeks until they reached their assigned dose. Participants continued their baseline antihypertensive medications and were permitted to adjust during the study.

Body weight measurements were carried out in a consistent manner using a calibrated electronic scale in kilograms. All weights for a given patient were measured using the same scale and patients wore light clothes with no shoes while their weight was measured. All laboratory parameters were assessed in a central laboratory. Blood pressure and pulse rate were measured after the participant sat quietly for 5 min. For each parameter, two measurements were taken using the same arm with the recordings taken at least 1 min apart. BP was taken with an automated blood pressure machine.

Post-hoc mediation analyses were conducted to evaluate weight-loss dependent (WL-D) and weight-loss independent (WL-IND) effects of tirzepatide on SBP reductions across the five SURPASS studies individually and also pooled doses per study. The model for the estimation of WL-D and WL-IND effects on SBP at Week 40/42 included the interaction between treatment and weight change, with the baseline variable for SBP, use of antihypertensive drug, country and HbA1c category ([≤ 8.0%, > 8.0%] for SURPASS-5, [≤ 8.5%, > 8.5%] for other studies) as covariates in the model. The safety population (all randomly assigned patients who took at least one dose of study drug) of each study was used in this analysis which included data regardless of adherence to study drug or initiation, modification or discontinuation of antihypertensive medications. With the integrated database from the five studies, subgroup analyses of change from baseline in SBP by baseline antihypertensive drug use (Yes, No) was performed. Correlation between SBP change from baseline and weight change from baseline were performed and the Pearson correlation coefficient was presented. SBP change from baseline was summarized by the baseline SBP quartile.

Patient demographics and baseline characteristics from participants randomized in SURPASS-1 to -5 (n = 4199; n = 1394 receiving tirzepatide 5 mg, n = 1397 receiving tirzepatide 10 mg, n = 1408 receiving tirzepatide 15 mg and n = 2064 receiving placebo or active comparators) are shown in Table 1. Blood pressure at baseline is also shown in Additional file 2. Baseline characteristics and demographics were well balanced between tirzepatide and comparators for each study and pooled dataset.

Across the SURPASS studies, mean age was 54–64 years and 47–94% of participants were using antihypertensive medications at baseline. At baseline, SURPASS-1 participants had the lowest SBP and duration of diabetes (127.6 mmHg and 5 years) compared to SURPASS-5 participants (137.9 mmHg and 13 years). As expected, SURPASS-4 participants had the highest prevalence of CV disease (87%), use of antihypertensive medication at baseline (94%) and lowest estimated glomerular filtration rate (81.3 mL/min/1.73m²), as this study enrolled patients with a high CV risk (coronary heart disease, peripheral arterial disease, cerebrovascular disease, chronic kidney disease or congestive heart failure). SURPASS-4 participants were also older and 63% were male.

Across the SURPASS program, SBP reductions ranged from − 4.2 to − 12.6 mmHg in participants receiving tirzepatide. In each study, SBP reductions were greater with tirzepatide than with placebo or active comparator groups at Week 40/42.

In the monotherapy placebo-controlled study (SURPASS-1), treatment with tirzepatide 10 mg resulted in significantly greater SBP reductions compared with placebo (estimated treatment difference [ETD] [95% CI] − 3.1 [− 6.2, 0.1] mmHg; P = 0.04). In the add-on to basal insulin placebo-controlled study (SURPASS-5), all doses (5, 10 and 15 mg) resulted in significantly greater SBP reductions compared with placebo (ETD [95% CI] − 4.4 [− 7.8, − 1.0], − 6.6 [− 9.9, − 3.2] and − 10.9 [− 14.3, − 7.5] mmHg; P = 0.01, P < 0.001 and P < 0.001, respectively) (Fig. 1).

In the SURPASS-2 study, 10 mg and 15 mg dose groups of tirzepatide demonstrated significantly greater SBP reductions than semaglutide 1 mg (ETD [95% CI] − 1.8 [− 3.4, − 0.1] and − 3.0 [− 4.6, − 1.3] mmHg; P = 0.03 and P < 0.001, respectively), while in SURPASS-3 and SURPASS-4 studies, SBP reductions were greater with all tirzepatide doses compared with insulin degludec (ETD [95% CI] − 4.9 [− 6.8, − 3.0], − 6.3 [− 8.2, − 4.4], − 6.9 [− 8.8, − 5.0] mmHg) and insulin glargine (ETD [95% CI] − 4.7 [− 6.5, − 3.0], − 5.6 [− 7.4, − 3.9], − 7.2 [− 8.9, − 5.5] mmHg) for tirzepatide 5, 10 and 15 mg, respectively: P < 0.001) (Fig. 1).

Overall, SBP reductions were greater with tirzepatide than with placebo or active comparator groups and were dose dependent with the greatest SBP reductions observed in the tirzepatide 15 mg treatment groups. Similarly, pooled analysis across the five SURPASS trials at Week 40/42 indicated dose dependent SBP reductions of − 4.8, − 5.8 and − 7.0-mmHg for tirzepatide 5, 10 and 15 mg treatment groups, respectively. (Fig. 1).

There were similar reductions in body weight for all tirzepatide doses across the five SURPASS studies. Similar to SBP, mean body weight reductions at Week 40/42 were dose dependent, with the greatest reductions observed in the tirzepatide 15 mg treatment groups across the SURPASS program. (− 7.0, − 9.1 and − 10.8 kg for tirzepatide 5, 10 and 15 mg treatment groups, respectively). Body weight reductions did not reach a plateau.

The mediation analysis showed contribution of WL-D and WL-IND effects on total effect of SBP reductions presented as difference between tirzepatide and comparator group for each study (Fig. 2). For WL-D effects between tirzepatide and comparator groups, the ETD (95% CI) in mean SBP change from baseline ranged from − 1.0 (− 1.6, − 0.5) to − 4.5 (− 6.7, − 2.4) mmHg (tirzepatide 5 mg), − 2.0 (− 2.8, − 1.4) to − 6.2 (− 8.9, − 3.6) mmHg (tirzepatide 10 mg) and − 2.4 (− 3.3, − 1.6) to − 7.5 (− 10.6, − 4.4) mmHg (tirzepatide 15 mg) (Fig. 2). WL-IND effects contributed to a lesser extent as the ETD (95% CI) in mean SBP change from baseline between tirzepatide and comparator groups ranged from − 0.3 (− 1.8, 1.3) to − 3.8 (− 8.4, 0.9) mmHg (tirzepatide 5 mg), + 2.5 (− 1.4, 6.7) to − 2.9 (− 5.9, 0.1) mmHg (tirzepatide 10 mg) and + 0.7 (− 3.3, 4.8) to − 6.5 (− 10.8, − 1.9) mmHg (tirzepatide 15 mg) (Fig. 2). Mediation analysis conducted by pooling patients across all doses of tirzepatide within each study also showed consistent results (Additional file 3).

In the SURPASS-4 study where patients with high CV risk were enrolled, and in the SURPASS-5 study where patients had the longest duration of T2D, WL-IND effects explained 33–57% and 26–73% of the difference in SBP change between tirzepatide versus the insulin glargine and placebo group, respectively (Fig. 2).

In the pooled analyses, there was a significant (p < 0.001) but weak correlation (r = 0.18 to 0.22) between change in body weight and SBP from baseline at Week 40/42 in tirzepatide-treated patients (Fig. 3).

At baseline, in the pooled tirzepatide treatment groups, 72.4% of participants were receiving antihypertensive medications. Similarly, in the pooled comparator group, 78.1% were receiving antihypertensive medications at baseline. In the SURPASS-4 study which enrolled participants with established CV disease, 94% were receiving antihypertensive medications at baseline. (Table 1).

Reductions in SBP with tirzepatide were not dependent on concomitant antihypertensive medications as similar reductions were observed whether participants were receiving them or not (tirzepatide 5 mg, − 5.0 vs − 4.3 mmHg; tirzepatide 10 mg, − 5.7 vs − 5.8 mmHg; tirzepatide 15 mg, − 7.0 vs − 7.0 mmHg) with a non-significant treatment by antihypertensive medication (Yes, No) interaction (P = 0.77) (Fig. 4).

The fourth quartile of SBP baseline value was > 140 mmHg, the median was 132.0 mmHg and the first quartile was ≤ 122.5 mmHg. Quartile of baseline SBP value significantly influenced the SBP change at Week 40/42 (p < 0.0001) (Fig. 5). The greatest SBP change with tirzepatide doses, which ranged from − 14.0 to − 17.5 mmHg, was observed in participants with the highest SBP values at baseline (Q4, > 140 mmHg). SBP was significantly reduced in all categories, except for in the lowest SBP baseline value category (Q1, ≤ 122.5 mmHg), where no clinically meaningful changes were observed.

Body weight reduction was similar across all the quartiles of SBP baseline value (tirzepatide 5 mg − 6.8 kg to − 7.3 kg; tirzepatide 10 mg − 8.9 kg to − 9.4 kg; tirzepatide 15 mg − 10.3 kg to − 11.6 kg) (Fig. 5).

Tirzepatide had a safety profile consistent with that of GLP-1 RAs with the majority of gastrointestinal AEs noted during the dose escalation period and decreasing overtime [17‐21]. The most commonly reported gastrointestinal AEs were nausea and diarrhea. The percentage of patients reporting ≥ 1 treatment-emergent gastrointestinal AE by preferred term ranged from 3 to 16% for tirzepatide 5 mg, 3–24% for tirzepatide 10 mg and 6–24% for tirzepatide 15 mg across the SURPASS studies.

Treatment with tirzepatide resulted in a mean increase in heart rate of 1–4, 2–4 and 3–6 beats per minute (bpm) for 5-, 10- and 15-mg groups, respectively, at the end of study treatment (Week 40/42 for SURPASS-1, 2 and 5 and Week 52 for SURPASS-3 and 4). For the two placebo-controlled studies, heart rate increased by 0–2 bpm on average. Treatment with active comparators semaglutide 1 mg, insulin degludec and insulin glargine resulted in mean heart rate increases of 4, 1 and 1 bpm, respectively.

Tirzepatide demonstrated clinically significant improvements in SBP ranging from 4 to 13 mmHg reduction across the three doses (5 mg, 10 mg and 15 mg) at 40 weeks in the SURPASS clinical program. Tirzepatide 10 mg and 15 mg demonstrated statistically significant reductions in SBP compared to GLP-1 RA, semaglutide 1 mg (5.3 mmHg and 6.5 mmHg vs 3.6 mmHg, respectively). The magnitude of effect was also consistent in SURPASS 4 which enrolled patients with established cardiovascular disease.

We observed that the effect of tirzepatide on SBP was mediated through weight loss with different degrees of contributions from weight-loss independent effects across the different SURPASS trials. Furthermore, SBP reduction was not dependent on antihypertensive medication use but dependent on baseline SBP value.

These findings are in agreement with a pooled analysis of six randomized phase 3 clinical trials in GLP-1 RA liraglutide [25]. Similarly, the authors reported that SBP was weakly correlated with weight loss at 26 weeks and SBP reductions were observed in the presence and absence of antihypertensive medication. However, mediation analyses were not carried out across the trials to determine the extent of the weight-loss contribution, nor was the SBP change according to baseline values evaluated. In a meta-analysis conducted across 33 randomized studies, the authors evaluated GLP-1 RAs, liraglutide and exenatide, using random-effect analysis and concluded that these GLP-1 RAs induced a small but significant change in SBP that appeared to be independent of the degree of weight loss and SBP values at baseline [26].

Weight loss of about 5–10% is expected to improve SBP by > 5 mmHg [27]. Across the SURPASS program, weight loss associated with tirzepatide explained the majority of treatment effect on SBP. Excess adiposity, insulin resistance, inflammation and higher oxidative stress are hallmarks of type 2 diabetes and obesity, which are proven to alter endothelial dysfunction and affect haemodynamics resulting in elevated blood pressure [28, 29]. It is not surprising that robust weight loss, reduction in liver and abdominal fat [30] and improvement in insulin sensitivity [31] associated with tirzepatide treatment may likely affect this key pathophysiological state associated with elevated BP in this population.

Nevertheless, there was some heterogeneity in this observation in the studies with higher baseline age, SBP, use of antihypertensive medications and duration of diabetes (SURPASS-4 and SURPASS-5) where weight loss explained roughly half of the total effect on SBP. Reduction in insulin usage in the SURPASS-5 study, particularly in the 15-mg dose group, may have also played a role in SBP reduction. There was a weak correlation between weight loss and SBP in the pooled analyses across SURPASS trials which further raises curiosity to explore potential weight-loss independent mechanisms that could be driving the reduction in SBP. These mechanisms are potentially not related to the study treatment. While the mediation analysis indicated the treatment effect on SBP reduction were mainly through the weight loss, the results of the two analyses are not contradictory. The lack of robust association between weight loss and improvement in SBP has been reported with GLP-1 RAs [24, 25]. Natriuresis [32], direct vasodilation [33], reductions in sympathetic nervous system activity [34], extracellular volume and midregional-pro-atrial natriuretic peptide (proANP) [33, 35] are potential direct mechanisms for GLP-1 RAs leading to SBP lowering.

To ascertain whether use of antihypertensive mediation impacts the degree of SBP reduction with tirzepatide, analyses were conducted for subgroups of patients using or not using antihypertensive medication at treatment initiation. These findings may be relevant for health care professionals initiating tirzepatide in patients with T2D using antihypertensive medications, as this could be a potential opportunity to adjust antihypertensive medications if target blood pressure is reached. SBP reduction with tirzepatide was highest (mean of 14–18 mmHg) in patients with a baseline value greater than 140 mmHg while there was minimal impact on SBP for patients with a baseline value of less than 123 mmHg. This is a clinically relevant finding from a patient safety perspective as this effect minimizes any potential risk of hypotension or syncope. This finding is also consistent with that reported with liraglutide [36].

Although much is known about the effects of GLP-1 RAs on blood pressure, little clinical data are available on the effects of GIP agonism. After a 6-day subcutaneous GIP infusion in patients with type 1 diabetes, a 4.6 mmHg reduction in SBP was noted [37]. In a separate study on patients with T2D already using GLP-1 RA and metformin, continuous acute infusion of GIP (6 pmol/kg/min) resulted in significant reduction in SBP compared to placebo. This effect was hypothesized to be due to elevated proANP and suggests an additive haemodynamic effect of GIP and GLP-1 receptor co-agonism [38]. Superior SBP lowering with tirzepatide compared to semaglutide may be due to greater weight loss or partly due to GIP specific mechanisms that warrant further exploration.

Across SURPASS studies, tirzepatide was associated with an increase in heart rate of 1 to 6 beats per minute at the end of study treatment. There was no significant difference between tirzepatide and semaglutide in the SURPASS-2 study at Week 40 in change in heart rate compared to baseline despite higher reduction in HbA1c, weight and SBP [18]. Several mechanisms have been postulated for elevation in heart rate with GLP-1 RAs, such as reflex tachycardia, increase in sympathetic nervous system activity and direct sino-atrial node action but none have been clinically proven yet [39]. Several long-acting GLP-1 RAs have demonstrated CV protection in a dedicated CV outcome trial (CVOT) [7‐10, 38] while the SURPASS-CVOT study (NCT04255433) for tirzepatide is ongoing and will provide further insights into whether these effects on blood pressure lowering, improvements in metabolic parameters and elevation in heart rate would combine to produce any meaningful impact on hard CV outcomes. To date, tirzepatide has demonstrated CV safety when compared with pooled comparators in a meta-analysis of phase 2 and phase 3 studies with the hazard ratio of 0.80 (95% CI: 0.57–1.11) for MACE-4 [24]. In a post-hoc analysis of SURPASS-4 data, tirzepatide slowed the rate of decline in eGFR, showed clinically meaningful improvement in albuminuria and significantly lowered occurrence of the composite kidney outcomes compared to insulin glargine which may be related to SBP lowering [40].

The limitations of this study include its post-hoc nature and the fact we did not systematically collect indications for antihypertensive medications as these could have been used for other co-morbidities. Studies were not designed to systematically assess blood pressure and randomization was not stratified based on baseline status of hypertension, SBP and other relevant parameters that could have affected the outcomes. Weight loss with tirzepatide did not plateau at 40/42 weeks and therefore longer term data could provide more robust assessments in future.

In conclusion, tirzepatide has demonstrated clinically relevant reductions in SBP across the SURPASS program. This effect was primarily mediated through weight loss, with different degrees of contributions from weight-loss independent effects across the different SURPASS trials. Furthermore, SBP reduction was not dependent on antihypertensive medication use and dependent on baseline SBP value, alleviating theoretical concerns of hypotension in patients with lower baseline SBP.