Acute Pain | Chronic Pain | General

Intrathecal opioid use during caesarean section with spinal anaesthesia

Clinical bottom line:

Morphine 0.1 to 0.2 mg significantly decreased postoperative pain and decreased the need for postoperative analgesia compared with spinal anaesthesia given alone for women undergoing caesarean section. These benefits were of clinical relevance. However, patients experience more adverse effects, including pruritus, nausea and vomiting. Less information was available for other opioids, making direct comparisons difficult. However, it is likely that fentanyl and sufentanil are less effective than morphine, but are also associated with less harm. It is unclear whether benefits of fentanyl and sufentanil are of clinical relevance. Concerning the efficacy of intrathecal opioid use for intraoperative analgesia, evidence suggests that this is of benefit. However, based on the observation that only 24% required additional analgesia intraoperatively, routine use of intrathecal opioids to reduce intraoperative analgesia is inappropriate.

Based on the evidence within the review, reviewers recommend the drug of choice as 0.1 mg intrathecal morphine. For every 100 women with 0.1 mg intrathecal morphine added to spinal anaesthetic, 43 will experience pruritus, 10 will experience nausea and 12 will experience vomiting postoperatively, all of whom would not have experienced these adverse events without intrathecal morphine.

Spinal anaesthesia is commonly used for caesarean section, and opioids are often added to spinal solutions to enhance and prolong intraoperative and postoperative analgesia. Morphine and fentanyl are the most frequently used opioids in this practice.

Systematic review

Dahl JB, Jeppesen IS, Jørgensen H, et al. Intraoperative and postoperative analgesic efficacy and adverse effects of intrathecal opioids in patients undergoing cesarean section with spinal anesthesia. Anesthesiology. 1999; 91:1919-1927.

Inclusion criteria were randomised, double-blind, controlled trials of opioid added to spinal anaesthesia in women undergoing caesarean section; single dose of opioid added to a spinal anaesthetic; placebo or no treatment control; full journal publication.

Where possible, reviewers pooled information and calculated number-needed-to-treat and number-needed-to-harm with 95% confidence intervals, and assumed that if the confidence intervals included infinity, then there was no benefit of treatment. Reviewers assumed that if the NNT confidence intervals between two treatments did not overlap, then treatments were significantly different. Where doses were weight normalised (one trial), these were recalculated to total dose based on demographic data in the report.


Four different opioids were studied: morphine, fentanyl, sufentanil and buprenorphine. All trials used bupivacaine (13 trials), lidocaine (one trial) or tetracaine (one trial). There was not always sufficient information to assess efficacy of all drugs and at all doses studied. Information could not be pooled for efficacy analysis, and establishing dose-response relationships was therefore hindered.

Postoperative pain

Six comparisons looked at postoperative pain scores. Four of four comparisons showed benefit with morphine (two prolonged postoperative pain relief with 0.1 mg and 0.25 mg, and two decreased pain scores for 24 hrs postoperatively with 0.1 mg).

Time to first administration of supplemental analgesics

Twelve trials looked at time to first administration of supplemental analgesics postoperatively (mainly upon request or pain reached certain criteria). Morphine 0.1 mg and 0.2 mg significantly increased median time to analgesia in five of five comparisons, whereas 0.05 mg had no effect (one comparison). Just looking at the effective doses of morphine, median time to administration was 27 hours (range 11 to 29) compared with controls of 2 hours (range 1 to 4) for bupivacaine alone, 1 hour for lidocaine alone and 8 hours for tetracaine alone.

Fentanyl doses between 2.5 g and 50 g were examined in 15 comparisons. Doses between 6.25 g and 50 g significantly increased time to analgesia in most comparisons. These doses had a median time to analgesia of 4 hours (range 2 to 13). There was evidence of a dose-response relationship. Buprenorphine 0.03 mg and 0.045 mg was evaluated in one trial and sufentanil 10 g, 15 g and 20 g in two trials. Both drugs significantly increased time to first administration of analgesic.

24 hour postoperative analgesic consumption.

Seven of seven comparisons showed that morphine decreased analgesic consumption, with a dose range of 0.05 mg to 0.2 mg. Fentanyl decreased analgesic consumption in only two of 14 comparisons (at doses of 40 g and 60 g), and sufentanil was effective at 0 to 6 hours only (two comparisons).

Intraoperative need for supplemental analgesia.

Ten of 28 comparisons showed significant reduction in intraoperative analgesia with intrathecal opioid compared with control. Pooled data from all of these trials showed that 464/485(96%) patients did not need summplemental analgesia compared with 192/254 (76%) with control. Reviewers calculated NNTs for different doses and drugs, and report that the median NNT for not needing supplemental analgesics intraoperatively was 4.9 (3.9 to 6.9), and that there were no significant differences between the various opioids.

Adverse effects

Numbers-needed-to-harm for pruritus, nausea and vomiting were calculated from pooled adverse effects information collected up to 48 hours. Logistic regression analysis was used to determine whether adverse effects increased with dose.

All drugs were associated with increased incidence of adverse effects. In general, the higher the dose, the more harm was caused, although not always to a significant degree. When all information was pooled, NNTs suggested that intrathecal opioids caused significantly more pruritus, nausea and vomiting. There was not always sufficient information to establish which drugs and which doses were specifically associated with significant increases in adverse effects, but based on the 95% confidence interval not including infinity, morphine (for which the most information was available) was associated with increases in all adverse effects at doses of 0.05 to 0.25 mg combined. The number of patients needed to harm with 0.05 to 0.25 mg morphine compared with no intrathecal opioid was 2.6 (2.1 to 3.3) for pruritus, 6.3 (4.2 to 13) for nausea and 10 (5.7 to 41) for vomiting. The NNHs with 0.1 mg morphine were 2.3 (1.8 to 3.3) for pruritus and 8.3 (4.4 to 65) for vomiting. Confidence intervals for nausea included infinity suggesting the point estimate (9.9) is unreliable. The risk of these adverse effects increased with increasing dose.

Respiratory depression, defined as respiratory rate less than 10 breaths per minute, was monitored in 12 trials of 485 opioid patients and 250 controls. Only one case was noted in a patient receiving 0.1 mg morphine.

Related topics