BIGGER PICTURE Better care
A warning before blood pressure falls, with a plan for what to do
An operating-room monitor normally shows what is happening now. The HYPE trial tested whether a learned warning from the arterial pressure waveform could help anaesthesia teams act before prolonged low pressure developed.

- Trial
- 68 randomized; 60 completed
- Median low-pressure time
- 8.0 versus 32.7 minutes
- Setting
- Elective noncardiac surgery in Amsterdam
How the idea works
Sample arterial pulses
An arterial line supplies the pressure waveform.
Predict a pressure drop
Learned waveform patterns trigger an early warning.
Investigate and respond
The team follows diagnostic guidance and a treatment protocol.
More information than a single pressure number
An arterial line provides a continuous pressure waveform. The Hypotension Prediction Index looks for patterns associated with an approaching pressure drop, updating its warning value every twenty seconds in this study. The purpose is to create a window for assessment before the usual threshold is crossed.
Prediction was paired with a clinical response
The intervention did not merely display a score. Anaesthesia staff also had diagnostic guidance and a treatment protocol to investigate the cause and respond. Both groups aimed to maintain mean arterial pressure at least 65 millimetres of mercury; the difference was anticipatory information and its associated workflow.
Sources: [2]
Fewer minutes below the threshold
Of 68 randomized patients, sixty completed the trial. Median time with low blood pressure was eight minutes with the warning-and-response system versus 32.7 minutes with standard care. The primary measure, which combined the depth and duration of hypotension, also improved.
Sources: [1]
The tested unit was the whole response loop
This preliminary, unblinded study came from one centre and selected operations. It did not establish reduced postoperative organ injury or mortality. Its useful design lesson is narrower and concrete: a forecast can change a physiological outcome when it arrives with a workable route to action.
Sources: [1]
The original sources
- JAMA / University of Twente repositoryThe HYPE Randomized Clinical Trial: machine-learning warning for intraoperative hypotension
- TrialsThe use of a machine-learning algorithm that predicts hypotension during surgery in combination with personalized treatment guidance: study protocol for a randomized clinical trial