Background: Post-spinal hypotension occurs in 50-80% of cesarean deliveries and causes maternal nausea, vomiting, and reduced neonatal vigor; norepinephrine (NE), as boluses or prophylactic infusion, is the mainstay of treatment. Intermittent cuff monitoring cannot anticipate hypotension, whereas the Hypotension Prediction Index (HPI), derived from the noninvasive ClearSight arterial waveform, predicts hypotension (MAP \< 65 mmHg) up to 5 minutes in advance. A previous randomized trial by the investigators showed that HPI-guided management reduced hypotension burden and maternal nausea and vomiting during cesarean delivery. The investigators hypothesize that HPI-guided NE boluses provide better hypotension control than prophylactic NE infusion, with lower total NE dose, less reactive hypertension, and less maternal discomfort.
Methods: This single-center, randomized, open-label trial at National Taiwan University Hospital enrolls women aged 20-50 years with singleton pregnancies undergoing non-emergent cesarean delivery under neuraxial anesthesia. Exclusions: preeclampsia, NYHA class ≥ II cardiopulmonary disease, arrhythmia, pre-existing cardiovascular disease, BMI \> 40, and failed spinal or block below T6. A total of 180 parturients will be randomized 1:1:1 into three groups of 60 (51 per group required for 80% power, plus allowance for dropouts). All participants receive standard monitoring (ECG, pulse oximetry, cuff every 2.5 min, ClearSight, bilateral cerebral oximetry), a 1000 mL Ringer's preload, and combined spinal-epidural anesthesia with 12 mg of 0.5% hyperbaric bupivacaine. Hypotension is defined as MAP \< 65 mmHg and bradycardia as HR \< 60/min (treated with atropine 0.01 mg/kg). Group allocation is revealed to the anesthesiologist after block completion. The three groups differ in the monitoring used to guide treatment and in the mode of NE administration:
1. HPI bolus group (reactive boluses guided by continuous monitoring plus prediction): blood pressure is monitored with the ClearSight continuous arterial waveform and HPI, and no values are blinded. No prophylactic infusion is given. The target is HPI ≤ 85; when HPI exceeds 85 for 30 s, NE 5-10 mcg is given as a bolus (plus atropine 0.01 mg/kg if HR \< 60/min). If ClearSight MAP falls below 65 mmHg, NE 5-10 mcg is given immediately regardless of HPI.
2. ClearSight infusion group (prophylactic infusion guided by continuous monitoring): blood pressure is monitored with the ClearSight continuous arterial waveform; cuff readings and HPI are blinded. Prophylactic NE infusion begins immediately after spinal anesthesia at 0.04 mcg/kg/min and is titrated to maintain MAP ≥ 65 mmHg. Rescue boluses of NE 5 mcg (MAP 55-65 mmHg for 30 s) or 10 mcg (MAP \< 55 mmHg for 30 s) are given; the infusion is paused when MAP exceeds 100 mmHg for 30 s and resumed when MAP falls below 75 mmHg.
3. NIBP infusion group (control; prophylactic infusion guided by intermittent cuff monitoring): blood pressure is monitored with the conventional oscillometric cuff only; all ClearSight values are blinded. The same prophylactic NE infusion (0.04 mcg/kg/min, titrated to MAP ≥ 65 mmHg), rescue bolus thresholds (5 mcg for MAP 55-65 mmHg; 10 mcg for MAP \< 55 mmHg), and infusion-pause rule (MAP \> 100 mmHg, resume at \< 75 mmHg) are applied based on cuff readings.
The primary outcome is the time-weighted average of MAP \< 65 mmHg (TWA, mmHg) during the first 30 minutes after spinal anesthesia, calculated from ClearSight data in all groups. Secondary outcomes include umbilical arterial and venous blood gases (pH, pCO2, pO2, base excess, lactate); time-adjusted AUC for systolic pressure below 80% of baseline; incidence and total duration of hypotension; area under the 65 mmHg threshold; total NE and atropine doses; maternal symptoms (nausea, vomiting, bradycardia, dizziness, shivering); number of hypotensive events detected by cuff versus ClearSight; MAP and cerebral oximetry at eight time points (baseline through 15 min post-spinal); and breastfeeding effectiveness assessed with the LATCH score on postoperative day 2 and at discharge.
Statistical analysis: Continuous data will be compared by one-way ANOVA or Kruskal-Wallis test with Conover-Iman post hoc tests; categorical data by chi-square or Fisher's exact test; repeated measurements by repeated-measures ANOVA or linear mixed-effects models (two-tailed, p \< 0.05).
Safety and ethics: The intervention is low risk (finger-cuff discomfort, vasopressor-related hypertension or bradycardia, managed per protocol). Informed consent will be obtained and data coded and securely stored. Enrollment is planned from June 2026 to June 2028, with analysis and manuscript preparation through December 2028.