Design: Bayesian random-effects meta-analysis of randomised trials
Population: 15 trials in major noncardiac surgery
Primary outcome: All-cause mortality
Effect: OR 1.00 (95% CI 0.83 to 1.21); AKI OR 0.87 (0.73 to 1.03); myocardial injury OR 1.04 (0.93 to 1.15)

This Bayesian meta-analysis included 15 randomised trials in major noncardiac surgery comparing protocolised arterial pressure targets, often individualised or higher, with usual care or lower targets.

Mortality across 13 trials was unchanged (OR 1.00, 95% CI 0.83 to 1.21). Myocardial injury after noncardiac surgery, from 8 trials, was also unchanged (OR 1.04). Acute kidney injury, from 14 trials, had the most favourable estimate (OR 0.87, 95% CI 0.73 to 1.03), but this crossed no effect, showed moderate heterogeneity and shifted when single trials were removed.

This does not mean blood pressure is unimportant. It means that formal protocols for hitting a target added little over the care anaesthetists already give. The possible kidney signal may depend on fluids, vasopressors and cardiac output as much as on the pressure itself.

The practical response is to keep avoiding sustained hypotension, to individualise for patients with chronic hypertension or kidney disease, and not to invest heavily in elaborate pressure-targeting protocols on the strength of an expected outcome gain.

Continue to avoid sustained intraoperative hypotension
Protocolised pressure targets did not reduce mortality or myocardial injury
Any kidney benefit is uncertain; consider it most in patients with chronic kidney disease
Individualise targets to the patient's usual blood pressure where known
Do not expect a pressure protocol alone to change outcomes

Why it matters

It challenges the assumption that tighter, protocolised pressure control is the next step in perioperative care.

The statistics, in plain English

The mortality odds ratio of 1.00 sits exactly on no effect, with a confidence interval from 0.83 to 1.21. The kidney estimate leans towards benefit, but its interval reaches 1.03, so no effect remains possible.