Presurgical anxiety and acute postsurgical pain predict worse chronic pain profiles after total knee/hip arthroplasty

Abstract

Introduction

Total joint arthroplasties generally achieve good outcomes, but chronic pain and disability are a significant burden after these interventions. Acknowledging relevant risk factors can inform preventive strategies. This study aimed to identify chronic pain profiles 6 months after arthroplasty using the ICD-11 (International Classification of Diseases) classification and to find pre and postsurgical predictors of these profiles.

Materials and methods

Patients undergoing total knee/hip arthroplasty (n = 209, female = 54.5%) were assessed before surgery, 48 h and 6 months postsurgery for sociodemographic, pain-related, disability and psychological characteristics. K-means-constrained cluster analysis identified chronic pain profiles based on 6-month pain intensity, pain interference and disability. Chi-square tests or one-way ANOVA explored between-cluster differences. Multinomial regression identified predictors of cluster membership. Separate models analyzed presurgical (model 1), postsurgical (model 2) and a combination of previously significant pre and postsurgical (model 3) variables.

Results

A three-cluster solution was selected, translating increasingly worse chronic pain severity: cluster 1 (C1, n = 129), cluster 2 (C2, n = 60) and cluster 3 (C3, n = 20). There were presurgical differences among clusters in the presence of other painful sites (p = 0.013, ϕc = 0.20), pain interference (p = 0.038, η² = 0.031), disability (p = 0.020, η² = 0.037), pain catastrophizing (p = 0.019, η² = 0.060), anxiety (p < 0.001, η² = 0.087), depression (p = 0.017; η² = 0.039), self-efficacy (p = 0.018, η² = 0.038) and satisfaction with life (p = 0.034, η² = 0.032), postsurgical pain frequency (p = 0.003, ϕc = 0.243) and intensity (p < 0.001, η² = 0.101). In model 1, disability predicted C2 (OR = 1.040) and anxiety predicted C3 (OR = 1.154) membership. In model 2, pain intensity predicted C3 (OR = 1.690) membership. In model 3, presurgical anxiety predicted C3 (OR = 1.181) and postsurgical pain intensity predicted C2 (OR = 1.234) and C3 (OR = 1.679) membership.

Conclusions

Most patients had low chronic pain severity at 6 months, but a relevant percentage exhibited poor outcomes. Membership to different outcome profiles was predicted by presurgical anxiety and acute postsurgical pain. These seem promising targets to prevent pain chronification that should be optimized for better surgical outcomes.