GLP-1 in Resistant Hypertension: Equal BP, 4,153 Pairs
EClinicalMedicine published a retrospective cohort study on 29 August 2026 from the Department of Nephrology and Hypertension at Hannover Medical School, comparing a GLP-1 receptor agonist against the guideline-recommended fourth-line option in resistant hypertension.
Blood pressure fell by almost exactly the same amount in both groups. Cardiovascular events, kidney events and deaths did not.
That combination is either an important finding or an artefact of who received which drug, and the paper is worth reading for how hard it is to tell.
A framing note. GLP-1 receptor agonists are approved medicines and this is an observational study in patients. Peptra Labs supplies Tirzepatide as a reference material for laboratory research only, and nothing below is guidance about human use or about blood pressure.
What the GLP-1 comparison actually compared
Resistant hypertension means blood pressure that stays uncontrolled despite an ACE inhibitor or angiotensin receptor blocker, a calcium antagonist and a diuretic. At that point a fourth drug is added, and guidelines recommend a mineralocorticoid receptor antagonist.
The authors used the TriNetX US Collaborative Network across 67 healthcare organisations, identifying adults with overweight or obesity and resistant hypertension who started a fourth-line therapy between 1 June 2017 and 31 March 2025.
The exposure was a GLP-1 receptor agonist, either semaglutide or tirzepatide, pooled as one arm. The comparator was a mineralocorticoid receptor antagonist, the guideline-recommended choice.
Of 213,309 eligible patients, 22,694 started a GLP-1 receptor agonist and 5,673 started an antagonist. After propensity score matching, 4,153 remained in each group, followed for a median of 1.4 years.
The GLP-1 results, and the one that should stop you
| Outcome | Hazard ratio |
|---|---|
| Major adverse cardiovascular events | 0.63, CI 0.52 to 0.78 |
| All-cause mortality | 0.34, CI 0.21 to 0.55 |
| Cardiovascular events | 0.74, CI 0.59 to 0.92 |
| Major adverse kidney events | 0.64, CI 0.46 to 0.88 |
| Acute kidney injury | 0.62, CI 0.46 to 0.83 |
And the blood pressure, at 12 weeks: a fall of 5.7 mmHg systolic in one arm against 6.3 mmHg in the other. Effectively the same, with the guideline drug marginally ahead.
The mortality figure is the one to examine. A hazard ratio of 0.34 means the comparison group died roughly three times as often. That is an enormous effect for a choice between two fourth-line blood pressure treatments in a population followed for less than eighteen months.
Effects that large in observational data are usually telling you something about the patients rather than about the drugs.
Confounding by indication in the GLP-1 arm
Nearly four times as many people started a GLP-1 receptor agonist as started the guideline drug, from the same pool of 213,309.
Think about what separates those groups before any matching happens. One of these medicines is expensive, often requires prior authorisation, and is prescribed to patients who are engaged with the health system and being actively followed. The other is a cheap generic that any clinician can add.
Propensity score matching adjusts for what is recorded: age, diagnoses, prescriptions, laboratory values. It cannot adjust for whether a patient turns up, whether they can afford a co-payment, or how sick their clinician judged them to be on the day.
Matching 4,153 pairs out of 28,367 initiators also means discarding most of the data to find comparable people. That is the correct procedure and it narrows the population to whoever could be matched.
None of this means the finding is wrong. It means a mortality hazard ratio of 0.34 is the part of the result that would need a randomised trial before anyone should believe it at that size, and the authors close by calling for exactly that.
Why the GLP-1 blood pressure result is the interesting one
Strip out the outcomes for a moment and look at what the drugs did to the thing they were prescribed for.
Both arms produced a systolic fall of around six mmHg. On the measure that defines the indication, they were equivalent, and the guideline-recommended agent was very slightly better.
So whatever produced the difference in events, it was not blood pressure control. The authors say this directly: lower cardiovascular and kidney risk despite smaller blood pressure reductions.
That is the same shape we have written about all month, with a different intermediate marker each time. We covered an earlier analysis of blood pressure with this class, a heart failure trial read by sex, and a matched cohort where hard liver outcomes came out level despite one arm losing more body mass index.
Weight, liver fat, and now blood pressure. Each time the intermediate measure and the outcome fail to move together, and each time the study design cannot say why.
The honest reading is that these GLP-1 observations are accumulating faster than the explanations for them.
What would settle the GLP-1 question
A randomised trial of a GLP-1 receptor agonist against a mineralocorticoid receptor antagonist as fourth-line therapy in this population, with adjudicated cardiovascular and kidney endpoints.
Nothing else will separate the GLP-1 effect from the difference between the people who receive each drug. The authors state that prospective studies are needed to determine whether these agents should be incorporated into treatment strategies here, which is the appropriately modest conclusion for the design they had.
The study is funded by nobody, according to the funding statement, which is worth noting in a field where most large comparisons carry a sponsor.
What it means for the catalogue
The transferable content is about reading effect sizes rather than about the compounds.
An implausibly large effect in observational data is a signal to look at the comparison groups, not a reason for excitement. The rule is simple enough to apply anywhere: when an effect is much bigger than the mechanism can explain, the explanation is usually in the selection.
Retatrutide appears in none of this work, and cannot, since database studies need years of prescribing before they produce anything. Every comparative GLP-1 dataset we have covered this month has the same blind spot for compounds that are not yet approved.
For laboratories, our European research buyer guide covers procurement and documentation, everything is supplied on a research use only basis, and our ranking of the most-studied research peptides counts primary studies for each compound.
References
- Tian Z, Willerding JM, Schmidt-Ott KM, Melk A, Schmidt BMW. Comparative effectiveness of GLP-1 receptor agonists versus mineralocorticoid receptor antagonists as fourth-line pharmacological therapy in patients with resistant hypertension and overweight or obesity: a retrospective multicenter cohort study in the USA. EClinicalMedicine, 29 August 2026, 99:104176. doi 10.1016/j.eclinm.2026.104176
- Nicholls SJ, Pavo I, Bhatt DL, et al. Cardiovascular Outcomes with Tirzepatide versus Dulaglutide in Type 2 Diabetes. New England Journal of Medicine, 18 December 2025, 393(24):2409-2420
- Banerjee M, Roy A, Sharma VM, Das A. Major Adverse Liver Outcomes With Tirzepatide and Injectable Semaglutide in Adults With Overweight or Obesity and Type 2 Diabetes. Obesity (Silver Spring), 19 August 2026
The products referenced on this site are supplied for laboratory research use only. They are not medicines and are not intended for human or veterinary use. This article summarises published research for informational purposes and is not medical advice. Statements about third-party studies belong to their authors.
