Randomized controlled trial of intrathecal oxytocin on speed of recovery after hip arthroplasty

Abstract

Recovery from surgery is quicker in the postpartum period, and this may reflect oxytocin action in the spinal cord. We hypothesized that intrathecal injection of oxytocin would speed recovery from pain and disability after major surgery. Ninety-eight individuals undergoing elective total hip arthroplasty were randomized to receive either intrathecal oxytocin (100 μg) or saline. Participants completed diaries assessing pain and opioid use daily and disability weekly, and they wore an accelerometer beginning 2 weeks before surgery until 8 weeks after. Groups were compared using modelled, adjusted trajectories of these measures. The study was stopped early due to the lack of funding. Ninety patients received intrathecal oxytocin (n = 44) or saline (n = 46) and were included in the analysis. There were no study drug-related adverse effects. Modelled pain trajectory, the primary analysis, did not differ between the groups, either in pain on day of hospital discharge (intercept: -0.1 [95% CI: -0.8 to 0.6], P = 0.746) or in reductions over time (slope: 0.1 pain units per log of time [95% CI: 0-0.2], P = 0.057). In planned secondary analyses, postoperative opioid use ended earlier in the oxytocin group and oxytocin-treated patients walked nearly 1000 more steps daily at 8 weeks ( P < 0.001) and exhibited a clinically meaningful reduction in disability for the first 21 postoperative days ( P = 0.007) compared with saline placebo. Intrathecal oxytocin before hip replacement surgery does not speed recovery from worst daily pain. Secondary analyses suggest that further study of intrathecal oxytocin to speed functional recovery without worsening pain after surgery is warranted.

Trial registration: ClinicalTrials.gov NCT00301130.

Figure 1.. Missing data

Proportion of data points missing from daily electronic diaries of pain and opioid use (teal) and of daily steps measured by actigraphy (red). Bars represent the proportion on each day and lines represent the cumulative proportion of subjects with that number of missing data points or fewer.

Figure 2.. Primary outcome: Recovery from pain after surgery

Worst daily pain for the first 8 weeks after surgery. Lines with shaded limits represent the modelled worst daily pain score after adjustment for pre-defined prognostic covariates for subjects randomized to oxytocin (red) or placebo (black) with their 95% confidence limits shown in pink and grey, respectively. The background is a heat map of the raw, unmodeled, and unadjusted data showing the decreasing pain scores over time in the study population as a whole. Groups do not differ (P=0.057).

Figure 3.. Secondary outcome: Opioid use

Probability of opioid medication use after surgery. Lines with shaded limits represent the modelled probability for a subject with median values for covariates after adjustment for pre-defined prognostic covariates for subjects randomized to oxytocin (red) or placebo (black) with their 95% confidence limits shown in pink and grey, respectively. Groups differ by t-test (P=0.023). Insert: Exploratory analysis of modeled change in opioid medication use over time in subjects taking opioids preoperatively (m=23) who were randomized to oxytocin (red; N=11) or placebo (blue; N=12). Shaded areas represent 95% confidence intervals.

Figure 4.. Secondary outcomes: Steps and disability after surgery

A) Daily steps for the first 8 weeks after surgery. Lines with shaded limits represent the modelled daily steps after adjustment for pre-defined prognostic covariates for subjects randomized to oxytocin (red) or placebo (black) with their 95% confidence limits shown in pink and grey, respectively. The background is a heat map of the raw, unmodeled, and unadjusted data showing the increasing number of steps over time in the study population as a whole. Groups differ by t-test (P<0.001). B) Disability as measured on the World Health Organization Assessment of Disability 2.0 (WHODAS 2.0) for the first 8 weeks after surgery. WHODAS 2.0 has a range of 12 (no disability) to 60 (severe disability). Lines with shaded limits represent the modelled change in WHODAS 2.0 score in an individual with median values of covariates after adjustment for pre-defined prognostic covariates for subjects randomized to oxytocin (red) or placebo (black) with their 95% confidence limits shown in pink and grey, respectively. The small circles show the raw, unmodeled, and unadjusted data showing the decreasing disability score over time in the study population. Raw data and summary statistics (mean ± SD) shown in small circles centered at time 0 and the large circles to their left, respectively. Groups differ by t-test (P=0.003). C) Difference between groups in modeled WHODAS 2.0 scores over time, expressed as score in oxytocin group minus placebo group as a percentage of the placebo group value. Negative values represent less disability, and the minimum clinically meaningful threshold of 5% [34] shown in the dotted line.