Multidimensional pain assessment and opioid use after total knee arthroplasty: continuous vs single-injection regional vs systemic analgesia

Abstract

Introduction: Effective pain management after total knee arthroplasty (TKA) is essential for recovery. Continuous peripheral nerve blocks (PNBc) are often believed to provide superior pain relief compared with single-injection peripheral nerve blocks (PNBs). However, multidimensional pain-related patient-reported outcomes (PROs) have not been extensively studied.

Objective: Based on registry data, this study compared pain intensities summarized as a pain composite score (PCS) and postoperative opioid use between PNBc and PNBs nerve blocks in patients undergoing TKA, and evaluated additional PROs.

Methods: Data from 4,328 adults undergoing TKA enrolled in the PAIN OUT registry (ClinicalTrials.gov NCT02083835) were analyzed. Patients were categorized into general anesthesia (GA) or spinal anesthesia (SA), with subgroups general anesthesia only (GA-o) or spinal anesthesia only (SA-o), and combinations with single-injection PNB (GA&PNBs and SA&PNBs) or continuous PNB via catheter (GA&PNBc and SA&PNBc). The primary end point was PCS, summarizing pain intensities and time in severe pain during the first 24 hours. Secondary end points included opioid use and additional PROs.

Results: The use of GA&PNBc was associated with a higher PCS (+0.5 [0.0-0.9], P = 0.035) compared with GA&PNBs, while PCS was similar between SA&PNBs and SA&PNBc. Opioid use was more frequent in GA&PNBc (+20.3%) and SA&PNBc (+50.8%) compared with the respective PNBs groups (P < 0.001). Patient-reported outcomes were higher in PNBc groups (median score 3.2 vs 2.7-2.9 in other groups; P < 0.001).

Conclusion: Continuous PNBc showed no clear advantage over PNBs in pain relief, opioid use, or further PROs. Future research should incorporate comprehensive PROs to better evaluate analgesic techniques in TKA.

Keywords: Assessment of pain/pain assessment; Continuous peripheral nerve block; Multidimensional pain scores; Opioid use; Pain-related patient-reported outcomes; Peripheral nerve block; Postoperative pain management; Total knee arthroplasty.

Figure 1.

Adjusted estimated marginal means and adjusted pairwise contrasts for the PCS for groups and subgroups. (A) Adjusted estimated marginal means of the PCS for the anesthesia groups and the subgroups represented on the y-axis. The x-axis represents the NRS (0-10). Specific means and confidence intervals are provided on the right side. (B) Adjusted pairwise contrasts of PCS for all subgroups. Specific percentage differences, confidence intervals, and P values for each comparison are shown on the right. Statistical analyses were performed using a multivariable fixed-effects regression model with a beta-distributed outcome. A P value <0.05 was considered as statistically significant. GA, general anesthesia (all techniques); GA&PNBc, general anesthesia and PNB catheter used for postoperative analgesia; GA&PNBs, general anesthesia and PNB single-injection, with patients receiving a peripheral nerve block intraoperatively; GA-o, general anesthesia without an additional nerve block; NRS, numeric rating scale; PCS, pain composite score; SA&PNBc, spinal anesthesia and PNB catheter; SA&PNBs, spinal anesthesia and PNB single-injection; SA, spinal anesthesia (all techniques); SA-o, spinal anesthesia only.

Figure 2.

Adjusted estimated marginal means and adjusted pairwise contrasts for the PCS for patients without and with opioid or substance use before surgery. (A) Adjusted estimated marginal means of the PCS for the anesthesia groups and subgroups, represented on the y-axis and differentiated for patients with opioids and/or substances (red lines) and those without (blue lines). The x-axis represents the NRS (0-10). Specific means and confidence intervals are provided on the right. (B) Adjusted pairwise contrasts of the PCS with comparisons of all subgroups differentiated for patients with O&S (N = 412) and those without (N = 3916). Specific percentage differences, confidence intervals, and P values for each comparison are provided on the right. Statistical analyses were performed using a multivariable fixed-effect regression model with a beta-distributed outcome. A P value <0.05 was considered as statistically significant. GA, general anesthesia (all techniques); GA&PNBc, general anesthesia and PNB catheter; GA&PNBs, general anesthesia and PNB single-injection; GA-o, general anesthesia only, no additional nerve block; NRS, numeric rating scale; PCS, pain composite score; SA&PNBc, spinal anesthesia and PNB catheter; SA&PNBs, spinal anesthesia and PNB single-injection; SA, spinal anesthesia (all techniques); SA-o, spinal anesthesia only.

Figure 3.

Treatment contrasts for % of patients receiving postoperative opioids across anesthesia techniques for the entire population and for patients without and with opioid or substance use before surgery. (A) Treatment contrasts of anesthesia groups of the entire population. Each line represents a treatment contrast with the mean percentage difference and the associated 95% confidence intervals. The x-axis displays the difference in the percentage of patients with opioid use, with 0% indicating no difference among the groups. Specific percentage differences, confidence intervals, and P values for each comparison are provided on the right. (B) Treatment contrasts of anesthesia groups of the 2 subcohorts with (N = 412) or without (N = 3916) preoperative O&S. Statistical analyses were performed by using a multivariable logistic regression model. GA&PNBc, general anesthesia and PNB catheter; GA&PNBs, general anesthesia and PNB single-injection; GA-o, general anesthesia only, no additional nerve block; SA&PNBc, spinal anesthesia and PNB catheter; SA&PNBs, spinal anesthesia and PNB single-injection; SA-o, spinal anesthesia only.