Determinants of pain severity in knee osteoarthritis: effect of demographic and psychosocial variables using 3 pain measures.

OBJECTIVE

Why some patients with knee osteoarthritis (OA) report greater pain severity than others is unclear. We examined the demographic variables, psychosocial variables, and physical findings that predict severity of pain in patients with symptomatic knee OA comparing 3 different pain scales.

METHODS

Pain severity was measured in 68 outpatients with knee OA using the WOMAC OA Index, the McGill Pain Questionnaire (MPQ), and a 0-100 visual analog scale (VAS). Depression, anxiety, fatigue, helplessness, self-efficacy, and quality of life were measured using standard instruments. Pain threshold was measured by dolorimetry and a standard knee examination performed. Radiographs were viewed when available.

RESULTS

Significant correlations (r = 0.39-0.61) were found between pain measures. In unadjusted analysis, BMI and helplessness correlated with all 3 scales; race, education, female sex, and osteophyte score also correlated with at least one instrument. Depression, anxiety, and fatigue correlated only with the MPQ. Age, duration, and quality of life were not related to pain severity. After adjusted analysis the following variables remained: education, helplessness, and osteophyte score (WOMAC); BMI and helplessness (MPQ); duration, education, helplessness, and osteophyte score (VAS). "Sitting pain" and "night pain" had different associations from pain on walking, standing, or using stairs.

CONCLUSIONS

Different pain scales measure different facets of the pain experience in knee OA and cannot be used interchangeably. The WOMAC pain scale has advantages over other instruments. Helplessness, education, and BMI appear to be important, potentially treatable, factors in determining self-reported pain severity in knee OA: other associations vary with both the pain scale used and the situation in which pain occurs, supporting the hypothesis that pain in knee OA is heterogeneous.