Epidural and intravenous patient-controlled analgesia and the timing of epidural catheter removal after posterior lumbar interbody fusion: a retrospective cohort study.

purposeWe compared pain control and time to first ambulation between two successive institutional postoperative analgesic regimens-epidural patient-controlled analgesia (E-PCA) and intravenous PCA (IV-PCA)-in patients undergoing posterior lumbar interbody fusion (PLIF). We also evaluated whether early catheter removal (postoperative day [POD] 1-2) was associated with rebound pain or increased rescue analgesic requirements.

methodsThis retrospective cohort study included 90 patients who underwent PLIF between January 2019 and December 2025 (E-PCA, n = 49; IV-PCA, n = 41). The two groups represented two successive institutional analgesic regimens that differed not only in administration route but also in drug composition and PCA programming. To assess the effect of catheter removal timing, the E-PCA group was further subdivided into early (POD 1-2, n = 26) and late (POD 3-4, n = 20) removal groups, excluding three patients with POD 0 catheter removal for whom change measures could not be calculated. Outcome measures included visual analog scale (VAS) pain scores and time to first ambulation. In the E-PCA group, we assessed post-removal 24-hour VAS, changes in pain scores (ΔVAS), and changes in rescue analgesic consumption (ΔTramadol).

resultsRepeated-measures ANOVA revealed a significant time-by-group interaction (p < 0.001). The E-PCA group had lower VAS scores through POD 1 than the IV-PCA group, whereas no significant differences were observed from POD 2 to POD 4. The E-PCA group ambulated earlier (1.71 ± 1.12 vs. 2.39 ± 1.14 days; p = 0.006) and had a shorter PCA duration (2.22 ± 1.01 vs. 2.98 ± 1.15 days; p = 0.001). Although premature discontinuation was more frequent in the E-PCA group (69.4% vs. 24.4%; p < 0.001), with accidental catheter dislodgement accounting for 44.1% of premature discontinuations, there were no significant differences in adverse events. In the E-PCA group, no statistically significant differences were observed between the early and late removal groups in post-removal 24-hour VAS, ΔVAS, or ΔTramadol.

conclusionThe E-PCA regimen was associated with lower pain scores than the IV-PCA regimen during the first 24 h after surgery, although this difference cannot be attributed to the administration route alone because the two regimens also differed in drug composition and PCA programming. Among patients receiving E-PCA, earlier epidural catheter removal was not associated with statistically significant worsening of pain control or increased rescue analgesic requirements compared with later removal. However, this study was not designed to establish equivalence, and small to moderate differences cannot be excluded. These observational findings should therefore be interpreted cautiously and require confirmation in prospective studies.