Epidural Hydromorphone and Pain During Cesarean Delivery

Key Takeaways

• Among women undergoing elective epidural cesarean delivery at a center in China, added hydromorphone was associated with a significantly different pain trajectory across surgical phases than saline.
• Intraoperative nausea or vomiting occurred less often with added epidural hydromorphone than with saline.
• Clinician-rated traction responses favored hydromorphone, whereas the principal reported heart rate variability metric showed no statistically significant longitudinal interaction.
• No maternal respiratory depression was recorded, and neonatal measures did not materially differ between groups.

Visceral traction and uterine handling during cesarean delivery can cause discomfort or nausea despite an established epidural block. For women receiving epidural ropivacaine, whether an opioid adjunct alters discomfort across surgical phases remained an open question that investigators tested against saline.

At a center in Chongqing, China, a randomized trial compared epidural hydromorphone with saline under double blinding; registration preceded recruitment. Investigators randomized 96 women, 48 per arm; all completed follow-up and were analyzed as assigned. Eligible women were aged 20–40, had term singleton pregnancies, and had American Society of Anesthesiologists (ASA) physical status II or III; all enrolled women were ASA II. Epidural loading comprised 8 mL of 0.5% ropivacaine plus either 0.3 mg hydromorphone in 1 mL or 1 mL saline; surgery followed confirmation of a bilateral T6 sensory block. Registered outcomes tracked visual analog scale (VAS) pain and heart rate variability (HRV) before anesthesia (T0), at delivery (T1), during abdominal exploration and uterine handling (T2), at peritoneal closure (T3), and during fundal massage (T4). No single primary time-point estimand was predefined.

Patient-reported pain followed different trajectories across surgical phases in the protocol-specified group-by-time analysis (Greenhouse–Geisser-corrected p=0.008). A participant-level permutation analysis gave p=0.009. An unstructured-covariance sensitivity analysis gave p=0.063, which did not reach the conventional significance threshold.

At T2, the peak-traction phase selected for focal reporting after registration, median VAS pain was 1.65 with hydromorphone versus 1.90 with saline (nominal p=0.039). This pointwise comparison lacked multiplicity adjustment and was distinct from the registered longitudinal analysis; central pain scores overlapped, and the rank-based interval approached zero. For the registered HRV domain, investigators selected the natural logarithm of the low-frequency/high-frequency (LF/HF) power ratio, ln(LF/HF), after registration as the principal reported metric. Its longitudinal group-by-time interaction was not significant (p=0.442), although an exploratory sympathetic-index interaction suggested a different pattern.

In an exploratory comparison unadjusted for multiple testing, intraoperative nausea or vomiting occurred in 3/48 (6.2%) women receiving hydromorphone versus 12/48 (25%) receiving saline (risk ratio 0.25, 95% CI 0.08–0.83; p=0.022). Clinician-rated traction grades favored hydromorphone. Exploratory rhythm-event patterns also differed, without establishing an antiarrhythmic effect. No maternal respiratory depression was recorded, and neonatal Apgar scores and umbilical artery blood gases showed no material between-group differences.

The single-center cohort consisted entirely of women with ASA II status, and the sample-size calculation addressed binary traction-pain incidence rather than the longitudinal interaction. Neither a single primary time-point estimand nor a secondary-outcome hierarchy was prespecified. The post-registration T2 focus and ln(LF/HF) selection, along with unadjusted phase-specific comparisons, complicate interpretation; dispersion, baseline-adjusted HRV, and win-ratio analyses were post hoc. The pain finding changed with covariance specification. HRV is not an objective pain measure, respiratory rate differed during peak traction, and rhythm events were not independently adjudicated. Because the comparator lacked a long-acting neuraxial opioid, incremental benefit with morphine- or diamorphine-containing regimens remains undetermined. The absence of observed respiratory or neonatal harm cannot exclude uncommon events.

The researchers concluded that adding epidural hydromorphone to this ropivacaine regimen was associated with phase-dependent pain differences and fewer intraoperative nausea or vomiting episodes. They called for prospective confirmation of exploratory physiological and post hoc findings.

Clinician Questions

Does lower nausea with epidural hydromorphone during cesarean delivery demonstrate a direct antiemetic effect?

No. Among women receiving epidural ropivacaine for elective cesarean delivery, the lower frequency of intraoperative nausea or vomiting with added hydromorphone does not establish an intrinsic antiemetic effect. Investigators noted that nausea has multiple possible causes and hypotension occurred at similar frequencies; reduced visceral traction input is one possible contributor, but the comparison did not isolate that mechanism.

How was visceral traction response graded in the epidural hydromorphone cesarean trial, and why did that affect its win-ratio analysis?

During elective epidural cesarean delivery, the attending anesthesiologist graded the overall visceral traction episode from no response to severe response. Moderate or severe grades incorporated nausea or vomiting, while a severe grade could include abdominal distension requiring intervention. The investigators’ post hoc win-ratio hierarchy included nausea or vomiting as a separate tier, partly overlapping the grade definition; patient-reported VAS pain at T2 measured a different aspect of the episode.