Shih, Po-YuanPo-YuanShihTZU-JUNG WEICHEN-TSE LEEJessica KANGShih, Ming-ChiehMing-ChiehShihLee, Yow-ShanYow-ShanLeeChen, Hsin-TingHsin-TingChenYING-HSI CHENCHUN-YU WU2026-03-102026-03-102026-02-15https://scholars.lib.ntu.edu.tw/handle/123456789/736172Introduction: Intra-operative hypotension is common during caesarean section and may result in adverse maternal effects, such as nausea and vomiting. While oscillometric blood pressure and continuous non-invasive arterial pressure monitoring enable reactive treatment, the hypotension prediction index offers a machine learning-based approach that may allow proactive haemodynamic intervention. This study investigated whether hypotension prediction index-guided management could reduce the incidence of intra-operative hypotension compared with oscillometric and continuous non-invasive arterial pressure monitoring. Methods: Patients scheduled for elective caesarean section under spinal anaesthesia were allocated randomly to one of three haemodynamic monitoring strategies: oscillometric blood pressure; continuous non-invasive arterial pressure; or hypotension prediction index-guided monitoring. Hypotension and hypertension were defined as mean arterial pressure < 65 mmHg and ≥ 100 mmHg, respectively. Hypotension was treated using intermittent intravenous boluses of noradrenaline. The primary outcome was the time-weighted average of hypotension. Secondary outcomes included time-weighted average hypertension and maternal adverse effects, including bradycardia, nausea and vomiting. Results: Data from 171 patients were analysed. Median time-weighted average hypotension was significantly higher in patients allocated to the oscillometric and continuous non-invasive arterial pressure groups compared with the hypotension prediction index group (0.89 mmHg and 0.30 mmHg vs. 0.08 mmHg, respectively, p < 0.001). Patients allocated to the oscillometric group had a higher incidence of maternal nausea and vomiting compared with those allocated to the continuous non-invasive arterial pressure and hypotension prediction index groups (nausea: 43/59 vs. 29/55 and 30/57, respectively, p = 0.038; and vomiting: 13/59 vs. 6/55 and 3/57, respectively, p = 0.023). Discussion: Hypotension prediction index-guided management during caesarean section significantly reduced intra-operative hypotension without increasing the risk of hypertension. This approach provides a proactive strategy for haemodynamic optimisation in obstetric anaesthesia.encaesarean sectioncontinuous non‐invasive arterial pressurehypotensionhypotension prediction indexjournal article10.1111/anae.7016141693355