Spinal bupivacaine dose and recorded blood pressure during caesarean delivery without intrathecal opioids: a retrospective cohort study

Abstract Background The association between spinal bupivacaine dose and recorded blood pressure within commonly used dose ranges remains uncertain, particularly without intrathecal opioids. We examined this association in historical clinical practice. Methods This single-centre retrospective study included 999 caesarean deliveries in 955 women given single-shot spinal anaesthesia without preceding epidural labour analgesia, 2016–2020. Hyperbaric bupivacaine doses were 9–16 mg, with 85.9% receiving 11–13 mg. The primary outcome was the lowest recorded systolic pressure between spinal injection and delivery. Doses ≥ 12 and < 12 mg, after exclusion of conversions to general anaesthesia, were compared using multivariable regression, propensity-score weighting and matching. Outcomes and covariates were chosen after preliminary results. A separate exploratory cohort of 207 records after epidural labour analgesia used fixed 0–25-minute measurements. Eight published definitions of hypotension were also applied. Results The adjusted mean difference in the primary outcome was − 0.19 mmHg (95% confidence interval [CI] − 2.43 to 2.05); the association per additional milligram was − 0.14 mmHg (− 1.02 to 0.74). Weighted and matched estimates were also near zero. Post-delivery intravenous fentanyl was given in 31.6% versus 24.7% of deliveries at < 12 versus ≥ 12 mg (adjusted odds ratio for ≥ 12 mg, 0.71; 95% CI 0.49 to 1.03). Hypotension (≤ 90 mmHg) occurred in 42.3% and 43.1% of lower- and higher-dose deliveries (adjusted odds ratio 1.19; 0.83 to 1.69). In the secondary cohort, the adjusted difference in the lowest recorded pressure was 1.55 mmHg (− 2.82 to 5.92). Across the eight definitions the measured incidence ranged from 20.9% to 88.8% and the direction of the association with baseline systolic pressure reversed, while the dose estimate remained compatible with the null under every definition. Conclusions Observed adjusted differences in recorded systolic pressure were small in this historical cohort. Hypotension incidence and its association with baseline pressure varied by definition, so outcome definitions need to be specified explicitly when studies are interpreted or compared. Non-random dose selection, intermittent measurements and incompletely recorded vasopressor treatment limit interpretation; these findings do not establish equivalent clinical safety, the optimal dose or the benefit of changing dose in contemporary practice.