Chen et al. investigated whether maintaining maternal systolic blood pressure at 90% vs. 80% of baseline using noradrenaline boluses influenced neonatal outcomes 1. The large sample size and multicentre design are commendable and provide valuable insights. However, after careful review, we have identified several methodological issues in the study that may compromise the interpretation of its conclusions. The issue of intervention strategy is of concern. The trial used a reactive bolus regimen of noradrenaline, whereas the international consensus statement on the management of hypotension during caesarean delivery under spinal anaesthesia recommends a prophylactic continuous infusion strategy for vasopressors 2. Reactive bolus administration is associated with wider blood pressure fluctuations, which may account for the relatively high incidence of early hypotension observed in both cohorts and which the authors have acknowledged as a study limitation. The study employed a crystalloid preload rather than the crystalloid co-load regimen currently recognised more widely. These differences from contemporary practice may limit the generalisability of the findings. The choice of the primary outcome and the statistical analysis of secondary outcomes warrant discussion. The study set umbilical artery pH as the primary outcome and there was no statistically significant difference between the two groups. The positive finding that favoured the 90% target, namely the lower incidence of umbilical artery pH < 7.2, was a secondary outcome. The authors did not pre-specify any key secondary outcome in the methodology section, nor did they mention that adjustment for multiple comparisons would be performed. Unadjusted analysis of multiple secondary outcomes can increase the risk of false-positive results, that is, statistically significant differences may be found only due to chance. In addition, base excess is widely recognised as a more reliable indicator than pH alone to reflect the fetal metabolic status and the level of anaerobic metabolism 3, while the base excess values of the two groups in this study were nearly identical, which further weakens the clinical significance of the difference in the incidence of pH < 7.2. Although the sample size calculation was mathematically rigorous, it may lack clinical relevance, which may in turn prevent the study from answering the core clinical question adequately. The sample size of this trial was calculated based on the hypothesis that maintaining systolic blood pressure at 90% of baseline could increase umbilical artery pH by 0.005. A difference of 0.005 in pH value is clinically undetectable and cannot have a substantive impact on the health status or prognosis of any neonate. Clinicians are more concerned about pathological thresholds such as pH < 7.2 or 7.0 4, and an increase in mean pH from 7.340 to 7.345 will not change the risk stratification and clinical management plan for any individual. The authors did not provide justification for these extremely small minimal clinically important differences and the calculation was only based on the mean (SD) of historical data. Although the resulting sample size (566 per group) is sufficient mathematically to detect this small difference, the study is severely underpowered for detecting the clinically meaningful outcome of neonatal acidaemia (pH < 7.20). If this trial had taken the incidence of neonatal acidaemia (pH < 7.20) as the primary outcome to detect the reduction in acidaemia incidence from 2.2% to 0.5% observed in this study, 719 patients per group would be required. A trial designed around this outcome would yield results with greater clinical significance.
Zhao et al. (Wed,) studied this question.