Psychological predictors of persistent pain after total knee arthroplasty: A retrospective cohort study

Objective Although total knee arthroplasty is a successful procedure, approximately 20% of patients still complain of postoperative pain of unknown cause. This single-center retrospective study aimed to investigate the association between preoperative psychological factors and persistent pain following primary total knee arthroplasty, with functional outcomes examined as secondary descriptive comparisons. Methods A review of 3674 medical records from our joint registry database (2004.01–2021.01) identified 2576 patients who met the inclusion criteria (unilateral primary total knee arthroplasty and ≥3-year follow-up). Based on visual analog scale pain scores at 1 year post-total knee arthroplasty, patients were stratified into a pain cohort (visual analog scale > 2, n = 377) and a pain-free cohort (visual analog scale ≤ 2, n = 2199) as the primary analytic approach, with a sensitivity analysis treating visual analog scale as a continuous variable to assess the robustness of our findings. Clinical outcomes and preoperative psychological assessments were evaluated to examine the association between psychological factors and postoperative pain. Results The pain cohort exhibited significantly worse postoperative outcomes, including higher visual analog scale pain scores, poorer functional outcomes, and reduced quality of life (all p = 0.001). In multivariable logistic regression, pain catastrophizing scale (adjusted odds ratios = 1.06, p = 0.003), depressive symptoms (Beck Depression Inventory) (adjusted odds ratios = 1.12, p = 0.002), and anxiety (Hospital Anxiety and Depression Scale-Anxiety) (adjusted odds ratios = 1.18, p = 0.002) remained independently associated with persistent pain after Bonferroni correction, whereas optimism (Life Orientation Test-Revised) and Hospital Anxiety and Depression Scale-Depression were attenuated and no longer significant. Patients with persistent pain also reported poorer functional outcomes (Knee Society Score, Western Ontario and McMaster Universities Arthritis Index, and European Quality of Life-5 Dimensions) at all time points (all p = 0.001), representing descriptive between-group differences. Conclusion Preoperative psychological factors, such as pain catastrophizing scale, depressive symptoms, and anxiety, were independently associated with persistent pain at 1 year after total knee arthroplasty, whereas optimism was not independently associated after adjustment for other psychological factors. These findings support the potential role of psychological factors in prognostic assessment and expectation setting. Patients with persistent pain also had inferior functional outcomes as secondary descriptive findings. However, the clinical value of routine psychological assessments and their impact on outcomes require further investigation in broader prospective studies.