https://orcid.org/0000-0003-4649-079X
Among trial participants randomized to semaglutide, there was a 4.8 to 5.0 mmHg greater decline in SBP compared to those who received placebo, largely mediated by decline in body weight.
This editorial refers to ‘Semaglutide and blood pressure: an individual patient data meta-analysis’, by C. Kennedy et al., https://doi.org/10.1093/eurheartj/ehae564.
Hypertension and obesity are among the world’s most pervasive public health issues and are considerably inter-related.1 The World Health Organization estimates that an overwhelming 1 in 3 adults worldwide have hypertension, 4 out of 5 of whom are either untreated or under-treated.2 Approximately 1 in 8 people worldwide are living with obesity. Worryingly, international obesity rates among adults have doubled since 1990 and have quadrupled among adolescents during the same time period. Individuals with obesity are at markedly elevated risk of cardiovascular disease, stroke, and chronic kidney disease compared with those with normal weight. Visceral adiposity, in particular, increases sympathetic nervous system and renin–angiotensin aldosterone system activity, resulting in excess fluid and sodium retention, kidney ultrafiltration, and elevated blood pressures (BPs).1 Accordingly, much of the elevated risk from obesity is likely to be mediated through hypertension.
Glucagon-like peptide-1 (GLP-1) receptor agonists are a class of antidiabetic drugs that include semaglutide and tirzepatide (also a gastric inhibitory polypeptide agonist), among others, which can result in significant weight loss by suppressing appetite, slowing food release from the stomach, and increasing feelings of fullness after eating. Multiple outcomes trials have also demonstrated that GLP-1 receptor agonists lower the risk of major cardiovascular, heart failure, and kidney events in several subpopulations of adults with and without diabetes.3,4 Whether some of the protective target organ effects of this medication class occur through BP lowering has not been established.
In the current issue of the European Heart Journal, Kennedy et al. present an analysis of pooled, individual patient data from 3136 participants of three randomized controlled trials (RCTs) that evaluated the effect of semaglutide vs. placebo on body weight among adults without diabetes.5 The authors aimed to determine the effect of semaglutide vs. placebo on BP, particularly among individuals with hypertension and resistant hypertension. At the time of randomization, 65% of participants met the authors’ criteria for hypertension (BP >130/80 mmHg, a diagnosis of hypertension, or treatment with antihypertensive medication), while only 4% had resistant hypertension (BP >130/80 mmHg on at least three antihypertensive agents including a diuretic, or on at least four antihypertensive agents regardless of BP). Over 68 weeks of follow-up, participants in the semaglutide group achieved an average of 14% reduction in body weight compared with a 1% reduction in the placebo group. The magnitude of weight reduction across randomization arms was similar in the subgroup of individuals with hypertension. Adjusting for baseline BP, the authors observed a 5 mmHg greater decline in systolic BP (SBP) at 68 weeks with semaglutide vs. placebo in the overall study population [−4.95 mmHg, 95% confidence interval (CI) −5.86 to −4.05], with similar results after restricting them to only those with hypertension (−4.78 mmHg, 95% CI −5.97 to −3.59; see Graphical Abstract). The decline in SBP with semaglutide was attenuated and no longer significant after restricting the results to the much smaller subgroup of individuals with resistant hypertension (−3.23 mmHg, 95% CI −8.84 to 2.28). Rates of antihypertensive medication de-escalation were overall low, but more than twice as common among hypertensive participants who were randomized to semaglutide (11%) compared with placebo (5%; odds ratio 2.38, 95% CI 1.63–3.56). This amounted to an average reduction of half of an antihypertensive medication in the semaglutide vs. placebo group (difference in therapeutic intensity score, a metric of the proportion of the maximum dose of medications prescribed, −0.54; 95% CI −0.74 to −0.33). Mediation analysis estimated that 89% (95% CI 70–116) of the decline in SBP was due to decline in body weight. However, it is important to note that because of the strong assumptions required for valid mediation analysis, it is unclear if causal conclusions can be drawn in this instance.
To provide context for the effect size of semaglutide observed in the current study, a half-standard dose of antihypertensive medication is expected to produce a reduction of 5–9 mmHg in SBP, depending largely on pre-treatment BP levels.6 For instance, a half-standard dose of antihypertensive medication is expected to reduce SBP by 4.4 mmHg when the pre-treatment SBP is 120 mmHg and by 9.0 mmHg when the pre-treatment SBP is 180 mmHg.6 In contrast, the effect of semaglutide on BP in the current analysis was consistent across baseline BP levels. The reason the effect size of semaglutide was not augmented for those with higher pre-treatment BP levels is unclear but could be attributed to the pattern of de-escalation of antihypertensive medications. Thus, the BP-lowering effect of semaglutide is noteworthy because it is similar to that of traditional antihypertensive medications, albeit on the lower end of the expected range, and is consistent across pre-treatment BP levels.
Of note, participants in both the semaglutide and placebo groups had substantial declines in BP, particularly among those with higher BPs at baseline. For example, among participants whose BP was >140/90 mmHg at baseline, those randomized to semaglutide had a 17 mmHg decline in SBP, while those randomized to placebo had a 13 mmHg decline in SBP. This observation may represent a broader challenge in interpreting the study results. First, several RCTs evaluating BP-lowering interventions have struggled with striking BP reductions among their placebo groups, often attributed to differences in antihypertensive medication adherence following enrolment or regression to the mean,7 and complicating interpretation of the findings. Second, BP was not measured in the standardized manner that would be necessitated in antihypertensive efficacy trials, recognizing that the semaglutide RCTs were not designed for this purpose. BPs were based on a single reading, rather than the average of three readings, with no protocolized requirement for consistent or validated BP measurement devices across readings. While this is common in RCTs that are not focused on BP and would be consistent across randomization arms, it could still substantially impact the results.8
In most adults with obesity and hypertension, weight loss through caloric restriction or metabolic surgery rapidly and persistently lowers BP. For example, the GATEWAY (Gastric Bypass to Treat Obese Patients With Steady Hypertension) Trial randomized 100 individuals with obesity and treated hypertension to Roux-en-Y gastric bypass plus medical therapy vs. medical therapy alone.9 Participants in the gastric bypass group had a 27% decline in body mass, compared with a 2% decline in the medical therapy group. At 1 year, 84% of participants who underwent gastric bypass experienced a ≥ 30% reduction in the total number of antihypertensive medications while maintaining BP control, compared with 13% in the medical therapy group. About half of participants who underwent gastric bypass had remission of their hypertension, whereas none of the participants in the medical therapy group experienced remission. These BP findings were sustained at 5 years.10 Similarly, the Swedish Obese Subjects Study prospectively matched 2010 individuals with obesity who underwent bariatric surgery with contemporaneous control participants receiving medical obesity treatment.11 After 10 years of follow-up, individuals who underwent bariatric surgery had a 16% decline in body weight compared with a 2% increase in body weight among those who did not undergo surgery. Those who underwent bariatric surgery had approximately twice the rate of remission of hypertension compared with those who received medical therapy (19% vs. 11%). This magnitude of hypertension remission relative to surgical weight loss has also been corroborated using real-world data.12
Historically, BP lowering from weight loss medications has not been as effective as surgical weight loss, which is likely to be due to differences in the magnitude and sustainability of achieved weight loss. Orlistat, phentermine/topiramate extended release, and naltrexone/bupropion are medications that were approved for weight loss in several countries prior to the advent of GLP-1 receptor agonists. These classes of weight loss medication reduced body weight by an average of 3%–5% in RCTs, which corresponded to only 1–3 mmHg declines in SBP (and a slight increase in BP in the case of naltrexone/bupropion).1 Semaglutide, on the other hand, results in substantially greater weight loss than these prior weight loss drugs, probably explaining the greater magnitude of BP reduction observed in the current study.
The results of the current study are consistent with a prior, study-level meta-analysis by the same authors that analysed data from six RCTs of semaglutide vs. placebo for weight loss among individuals without diabetes.13 Similar to the present findings, the authors observed a 5 mmHg reduction of SBP with semaglutide vs. placebo among all participants, amidst modest de-escalation of antihypertensive therapy. The current study adds to the prior study by demonstrating similar BP reductions among subgroups of individuals with hypertension and resistant hypertension with an average de-escalation by half of an antihypertensive drug, and that the magnitude of weight loss is a mediator of BP decline. These findings also complement a study-level meta-analysis of RCTs of tirzepatide vs. placebo, which demonstrated 5–6 mmHg dose-dependent reductions in SBP.14
A 5 mmHg reduction in SBP has significant implications for population health, particularly given the high prevalence of both obesity and hypertension. Meta-analyses of RCTs consistently demonstrate that a 5 mmHg reduction in SBP is associated with a 10% relative decrease in the risk of major cardiovascular events.15 The findings of the current study suggest that BP lowering may be an important mediator of the protective target organ effects of GLP-1 receptor agonists, which merits closer investigation.
Declarations
Disclosure of Interest
The authors declare no disclosure of interest for this contribution.
References
Author notes
The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.
