REF-1670 · Hu et al., 2026 · Sports Medicine · Paper · PubMed

In adults aged 45 and over, every exercise mode lowered resting blood pressure against control in a Bayesian network meta-analysis of 159 RCTs (10,821 participants): circuit-based training showed the largest systolic fall (−13.52 mmHg, 95% CrI −18.59 to −8.44), followed by combined training, HIIT, moderate continuous training, dynamic resistance and, last, isometric exercise; certainty varied, was low to very low for circuit training's lead over combined training, and no comparison was rated high confidence, so the authors call the ranking not definitive

Main caveat: The authors': certainty that varied, low to very low for the circuit-training lead; many trials with some concerns or high risk of bias.
fairly strong evidenceMeasured on non-athletes to trained athletesRead from the abstract only

Hu and colleagues compared exercise modes for lowering resting systolic and diastolic blood pressure in adults aged 45 years and over, and modelled how the systolic fall changes with weekly exercise dose. From six databases searched to May 2025 they included 159 RCTs with 10,821 participants.

Against control, every mode was associated with significantly lower systolic and diastolic pressure.

Circuit-based training had the largest estimated systolic fall, followed in order by combined training, HIIT, moderate-intensity continuous training, dynamic resistance training and isometric exercise; circuit and combined training mostly stayed near the top in sensitivity analyses, which the authors read as support for integrated aerobic-resistance training.

Each 10 mmHg of higher baseline systolic pressure went with a 2.49 mmHg larger fall. The dose-response curve was nonlinear, with the largest modelled systolic fall near 1200 MET-min a week, the upper end of WHO activity recommendations, and the circuit-training curve peaked near 580 MET-min a week.

The authors stress that effect size, certainty and robustness varied, that the circuit-training lead rests on limited evidence of low to very low certainty, that many trials had some concerns or high risk of bias, and no network comparison reached high confidence, so the ranking is not proof that one mode is superior.

Participants

159 RCTs, n = 10,821 adults aged 45 years or more.
Not in the abstract: sex split, baseline pressure range, share with hypertension or on medication, training status, health conditions.

Protocol

Search: six electronic databases, inception to May 2025 (databases not named in the abstract).
Comparison: exercise modes against control and against each other through the network; the control conditions are not described.
Outcomes: resting systolic and diastolic pressure; dose-response for systolic pressure only.
Analysis: Bayesian random-effects NMA; model-based dose-response on MET-min a week; network meta-regression for effect modifiers; sensitivity analyses.
Not in the abstract: the risk-of-bias and certainty tools by name, the number of trials per mode, heterogeneity, the diastolic estimates per mode, the isometric estimate, the circuit-training definition and the registration.

Key Results

FindingValue
Circuit-based training vs control, systolic−13.52 mmHg (95% CrI −18.59 to −8.44), largest estimate
Ranking for systolic fallcircuit > combined > HIIT > moderate continuous > dynamic resistance > isometric
All modes vs controlsignificant falls in systolic and diastolic pressure
Baseline systolic as modifierβ −2.49 mmHg per 10 mmHg higher baseline (95% CrI −3.73 to −1.24)
Dose with the largest modelled systolic fall, all modesabout 1200 MET-min/week
Peak of the circuit-training curveabout 580 MET-min/week
Certaintyvaried across comparisons; low to very low for circuit over combined training; no network comparison rated high confidence
Risk of biasmany included studies with some concerns or high risk (proportions not in the abstract)
The abstract gives no estimate for combined training, HIIT, moderate continuous, dynamic resistance or isometric exercise, and no diastolic value for any mode.
The authors caution that the circuit-over-combined order rests on limited circuit-specific evidence and on sensitivity analyses.

Practical Application

For middle-aged and older adults, every exercise mode was associated with lower resting pressure; integrated aerobic-plus-resistance work (circuits or combined sessions) ranked highest, with circuits estimated at −13.52 mmHg systolic against control — but no network comparison reached high confidence and the circuit-over-combined lead is of low to very low certainty, so no single mode is shown to be best.
Where this differed from [SYNTH-88]: until 2026-11-10 the guide stated that isometric training ranked first of five modes for resting pressure (REF-1150, 270 RCTs of adults of all ages and pressures); it now reports that ranking beside this network's and no longer treats isometric training as the best mode.This network, limited to adults aged 45 and over and with circuit training as its own mode, put isometric exercise last of the six modes the abstract names for systolic pressure. What does not differ: isometric exercise was still associated with a significant fall against control; the abstract gives no isometric estimate and no head-to-head isometric comparison, so the size of its gap to the other modes is unknown, and REF-1150's isometric ranking rested on indirect comparisons only, while here no network comparison reached high confidence.For an athlete over 45 this supports keeping isometric holds as an add-on, not as the first choice for blood pressure; since 2026-11-10 SYNTH-88 says, as the base's inference, not tested, that with raised pressure the first lever is aerobic training together with strength training and holds are the add-on for when time is short.
Inference, not tested here: for a masters endurance athlete with raised pressure, regular aerobic training plus strength work already approximates the combined training this network ranks highly; the dose peak near 1200 MET-min a week is well below what such an athlete usually trains, so more training is not expected to lower pressure further.
Do not quote −13.5 mmHg to someone with normal pressure: each 10 mmHg of higher baseline went with about 2.5 mmHg more fall.
This is not treatment advice: a person with raised blood pressure or on medication agrees any exercise programme, and above all HIIT, with their physician, and exercise is added to medical care, not used instead of it.

Limitations

The authors': certainty that varied, low to very low for the circuit-training lead; many trials with some concerns or high risk of bias; no network comparison rated high confidence; limited circuit-specific evidence; rankings not definitive evidence of superiority.
Two of the authors (Hu, Xin) also wrote a network meta-analysis of exercise and blood pressure in prehypertension and hypertension published in 2026, so the two networks likely share many trials and are not independent confirmations (our reading).
The dose-response uses MET-minutes, which are usually assigned from activity tables rather than measured (whether here is not stated in the abstract).
Written from the abstract and the PubMed record; full text not read: per-mode estimates other than circuits, diastolic results, heterogeneity, the participants' health and training status, and the risk-of-bias and certainty tables are unknown.
Revision 2026-11-10 (decision): the lines on SYNTH-88 follow the guide as it stands — it no longer treats isometric training as the best mode, and with raised pressure it names aerobic plus strength training as the first lever and holds as the add-on, as the base's inference.
How this entry was checked and changed
Why this level

a meta-analysis of RCTs whose abstract reports its appraisal and what it found (many trials with some concerns or high risk of bias, certainty that varied and was low to very low for the circuit-training lead, no network comparison rated high confidence), so the abstract-only cap does not apply, and that appraisal is serious, which holds it at A- at most

In adults aged 45 and over, every exercise mode lowered resting blood pressure against control in a Bayesian network meta-analysis of 159 RCTs (10,821 participants): circuit-based training showed the largest systolic fall (−13.52 mmHg, 95% CrI −18.59 to −8.44), followed by combined training, HIIT, moderate continuous training, dynamic resistance and, last, isometric exercise; certainty varied, was low to very low for circuit training's lead over combined training, and no comparison was rated high confidence, so the authors call the ranking not definitive — BLOCKS Science · Blocks