The wall-sit blood-pressure headline did not survive a 24-hour monitoring test

Exercise lowers blood pressure. Which kind lowers it most is a question the evidence has answered twice, differently. A 2023 network meta-analysis of 270 randomised trials ranked isometric exercise — the wall sit, in particular — first by a wide margin [s1], and the finding travelled widely. Two network meta-analyses published in May 2026 do not reproduce it: one, using 24-hour ambulatory monitoring, concludes that the evidence for isometric resistance training remains uncertain [s2]; the other, pooling 105 trials, places isometric exercise among the relatively weaker modalities [s3].

What the 2023 analysis found

Researchers searched PubMed, the Cochrane Library and Web of Science for randomised controlled trials published between 1990 and February 2023 that tested an exercise intervention of at least two weeks against a non-intervention control and reported resting blood pressure [s1]. They included 270 trials with a pooled sample of 15,827 participants [s1].

In pairwise analysis, every mode worked. Reductions in systolic and diastolic pressure were −4.49/−2.53 mm Hg for aerobic training, −4.55/−3.04 for dynamic resistance training, −6.04/−2.54 for combined training, −4.08/−2.50 for high-intensity interval training, and −8.24/−4.00 for isometric exercise training, all at p < 0.001 [s1].

The network analysis then ranked them by surface under the cumulative ranking curve. For systolic pressure the order was isometric exercise (SUCRA 98.3%), combined training (75.7%), dynamic resistance (46.1%), aerobic (40.5%) and high-intensity intervals (39.4%) [s1]. A secondary analysis of sub-modes identified the isometric wall squat as most effective for systolic pressure (90.4%) and running for diastolic (91.3%) [s1]. The authors argued the results should inform future exercise guideline recommendations for hypertension [s1].

What happened when the outcome changed

Resting blood pressure is measured in a clinic. Twenty-four-hour ambulatory monitoring, which records pressure repeatedly through a normal day and night, is the better predictor of cardiovascular outcomes and a harder test.

A network meta-analysis published in May 2026 searched MEDLINE, Embase, Cochrane Central and the Regional Portal of the Virtual Health Library from November 2024 to August 2025 for randomised trials of at least four weeks reporting 24-hour ambulatory pressure in adults with hypertension [s2]. It included 31 trials with 67 arms and 1,345 participants [s2].

Against control, three modalities reduced 24-hour systolic pressure: combined training (mean difference −6.18 mm Hg, 95% credible interval −11.45 to −1.21), aerobic training (−4.73, −7.53 to −2.01) and high-intensity interval training (−5.71, −11.31 to −0.002) [s2]. For 24-hour diastolic pressure, combined training (−3.94, −6.47 to −1.34), aerobic (−2.76, −4.21 to −1.34), high-intensity intervals (−4.64, −8.21 to −0.72) and pilates (−4.18, −7.18 to −1.17) reached significance [s2].

Isometric training is absent from both lists. The authors state plainly that evidence for dynamic and isometric resistance training remains uncertain, and that exercise-versus-exercise comparisons were inconclusive with respect to superiority between modalities [s2].

And when the trial set grew

A separate Bayesian network meta-analysis published in the Journal of the American Heart Association two days later pooled 105 randomised trials published up to April 2025, in adults with prehypertension or established hypertension [s3].

Combined training produced the largest reduction — systolic −12.05 mm Hg (95% CrI −15.08 to −9.05), diastolic −6.20 (−7.79 to −4.62) — followed by high-intensity interval training, systolic −10.97 (−14.97 to −6.95) and diastolic −6.42 (−8.68 to −4.16) [s3]. Yoga and tai chi had moderate effects, while aerobic exercise, isometric exercise training and resistance training showed relatively weaker effects [s3].

Dose-response modelling found a non-linear, U-shaped relationship between exercise volume and blood pressure, with the greatest benefit at approximately 830 metabolic equivalents per minute per week, and optimal doses varying by modality [s3].

How three analyses of the same question disagree

They are not analysing the same question. The 2023 analysis pooled resting clinic blood pressure from 1990 onward across a very large trial set, including many small isometric trials [s1]. The ambulatory analysis pooled a much smaller set — 31 trials, 1,345 participants — but on a stricter outcome [s2]. The JAHA analysis used a third trial set and a Bayesian framework [s3]. Network meta-analyses rank treatments partly through indirect comparison, and rankings are notoriously unstable when the underlying trials are small, unblinded and heterogeneous, which describes most exercise trials.

Wide credible intervals are visible in the newer analyses — the interval training estimate for 24-hour systolic pressure runs from −11.31 to −0.002 mm Hg [s2], which barely excludes zero.

What all three agree on is more useful than the rankings they disagree about. Every modality tested lowered blood pressure in every analysis [s1] [s2] [s3]. Combined training — aerobic plus resistance work — placed at or near the top of all three. And the newest analysis finds the relationship with exercise volume is U-shaped rather than ever-increasing [s3].

The specific claim that a wall sit is the best exercise for blood pressure rests on one ranking, from one outcome measure, that two subsequent analyses did not reproduce.

This article is informational and is not medical advice. Blood pressure treatment, and any change to exercise in the presence of hypertension or cardiovascular disease, is a matter for a clinician.