Abstract / Summary
We read with interest the randomized sequential allocation study by Sun et al, which estimated the effective dose in 90% of patients (ED90) of programmed intermittent epidural bolus (PIEB) volume at three ropivacaine concentrations during dural puncture epidural labor analgesia.1 The authors reported ED90 volumes of 13.2, 11.1, and 9.4 mL for ropivacaine concentrations of 0.075%, 0.1%, and 0.125%, respectively.This clinically relevant investigation may inform PIEB programming.However, three issues warrant clarification before these estimates are adopted in practice.First, the allocation rule and the target estimand may not be optimally aligned.After the first patient in each concentration group received 10 mL, the volume was increased by 2 mL after ineffective analgesia and decreased by 2 mL after effective analgesia.This symmetric up-and-down rule concentrates observations near the median effective volume rather than the 90th percentile. 2 Although isotonic regression can estimate an upper quantile from the resulting dose-response data, the precision of ED90 depends on adequate observations in the upper tail.In Table 2, only four patients received 14 mL in the 0.075% group, two received 12 mL in the 0.1% group, and three received 10 mL in the 0.125% group; all patients at these respective uppermost observed volumes had effective analgesia.Consequently, each ED90 estimate depends heavily on two to four observations at the highest administered volume.Moreover, the unadjusted 95% confidence interval for the 0.075% group extended to 16.7 mL, exceeding the protocol-defined 16-mL ceiling and the highest observed volume of 14 mL.The reported number of reversals describes sequence behavior near the allocation target but does not, by itself, establish precision at the 90th percentile.A biased-coin design can instead allocate observations around a prespecified nonmedian quantile and can be combined with isotonic regression.3 We invite the authors to report the fitted isotonic probabilities at every volume, the bootstrap distribution of ED90, and sensitivity analyses using alternative dose-response models.Interpretation of "dose" also requires care.Because the sufentanil concentration was fixed at 0.3 μg/mL, both the ropivacaine and sufentanil doses changed with bolus volume.At the adjusted ED90 values, the corresponding ropivacaine doses were 9.90, 11.10, and 11.75 mg, whereas the sufentanil doses were 3.96, 3.33, and 2.82 μg, respectively.Thus, increasing ropivacaine concentration reduced the required volume but increased the total ropivacaine dose and reduced the accompanying sufentanil dose.The estimates should therefore be interpreted as regimen-specific ED90 volumes for ropivacaine-sufentanil mixtures rather than as isolated effects of ropivacaine concentration or evidence of reduced total drug exposure.Second, the binary definition of effective analgesia incorporated unequal observation periods.A participant was classified as having effective analgesia if neither patient-controlled epidural analgesia nor a clinician-administered rescue bolus was required before delivery.Therefore, a participant who delivered soon after PIEB initiation had less opportunity to experience breakthrough pain than one with prolonged labor.The reported duration of labor analgesia varied substantially, with median values of 265 minutes (interquartile range, 206-460), 240 minutes (210-300), and 300 minutes (210-420) across the three groups.A nonsignificant group-level comparison of these durations does not eliminate Therapy 2026:20 659549 1 Drug Design, Development and