HYPE

Effect of a Machine Learning-Derived Early Warning System for Intraoperative Hypotension vs Standard Care on Depth and Duration of Intraoperative Hypotension During Elective Noncardiac Surgery: The HYPE Randomized Clinical Trial

Patient / Population Intervention / Exposition Vergleich Endpunkt

In 68 adults undergoing elective non-cardiac surgery with invasive blood-pressure monitoring at a tertiary centre in Amsterdam, a machine-learning hypotension early warning system with a haemodynamic diagnostic and treatment protocol reduced depth and duration of intra-operative hypotension compared with standard care targeting the same MAP ≥65 mmHg.

N
68
Design
Single-centre preliminary unblinded RCT, Netherlands
Endpunkt
Time-weighted average of hypotension (MAP <65 mmHg) during surgery
Relevanz
1
ErgebnisMedian time-weighted average hypotension 0.10 vs 0.44 mmHg (median difference 0.38 mmHg, 95% CI 0.14–0.43; P=.001); median time in hypotension 8.0 vs 32.7 min (difference 16.7 min, 95% CI 7.7–31.0; P<.001). Serious adverse events resulting in death 0 vs 2 (7%).
Wijnberge M, et al. JAMA. 2020;323(11):1052-1060. 10.1001/jama.2020.0592
Diskussion & Kritik

The first randomised trial of the Hypotension Prediction Index, and proof of concept that predicting hypotension shortly before it occurs lets anaesthetists prevent it. Tiny, single-centre and unblinded, with a surrogate endpoint (hypotension burden) rather than organ injury, and 60 of 68 completed the trial. Maheshwari HPI, run with a different protocol, failed to replicate the effect, and Ripollés-Melchor HPI found no benefit on kidney injury.