Worst pain intensity and opioid intake during the early postoperative period were not associated with moderate-severe pain 12 months after total knee arthroplasty - a longitudinal study

Abstract

Objectives

There are several known predictors of pain after total knee arthroplasty (TKA). However, it is unclear whether acute postoperative pain intensity and postoperative opioid intake are associated with pain 12 months after TKA. Thus, the aim of this study was to assess whether worst pain intensity and opioid intake during the early postoperative period are associated with moderate-severe pain 12 months after TKA. Methods

A total of 202 patients undergoing primary TKA between October 2012 and September 2014 were prospectively enrolled. Age, sex, contralateral knee pain, BMI, physical status and opioid intake were collected preoperatively. Ketamine and daily opioid intake were collected on postoperative days (POD) 0–3. Using the Brief Pain Inventory, patients’ “worst pain intensity” was measured preoperatively, on POD 0–4, and 12 months after TKA. Two logistic regression models evaluated the independent association of early postoperative pain intensity (model 1) and postoperative opioid intake (model 2) with moderate-severe pain 12 months after TKA, adjusting for possible confounders. Results

In total, 187 patients with data at the 12 month postoperative follow-up were included in this analysis. Pain intensity on POD2 and POD3, as well as preoperative pain and BMI, were significantly associated with pain at 12 months in univariate models. However, in multivariable models adjusted for preoperative pain and BMI, neither pain intensity on POD 0–4 (model 1) nor opioid intake on POD 0–3 (model 2) were associated with pain at 12 months. Preoperative pain was still significant in both models, but BMI remained significant only in model 2. Conclusions

Worst pain intensity and opioid intake during the early postoperative period were not associated with moderate-severe pain 12 months after TKA when controlling for potential confounders. More research is needed to confirm these findings.