Prospective study on the relationship between depressive symptoms and chronic musculoskeletal pain

Abstract

Chronic pain and depression often coexist, but there is still uncertainty about the nature of this relationship. Virtually all the available data are cross-sectional and therefore do not clarify the causal relationship between the two variables. In epidemiological studies, chronic pain has often been defined fairly liberally in terms of the actual duration. In this study, the definition of chronic pain was based upon self-reports of pain present for most of the days in at least 1 month of the 12 months preceding the interview. We tested the hypotheses that depression causes pain and that pain causes depression in a sample of 2324 subjects who were assessed for the presence of musculo-skeletal pain and the presence of depression, using for the latter a standardized published instrument called the Center for Epidemiologic Studies Depression scale (CES-D). The subjects were first examined using the National Health and Nutrition Survey (NHANES 1) of the United States National Center for Health Statistics from 1974 to 1975, and were followed-up from 1981 to 1984. Those with data on both occasions represent 76% of an initial population of 3059 persons. On logistic regression analysis depressive symptoms at year 1 significantly predicted the development of chronic musculo-skeletal pain at year 8 with an odds ratio of 2.14 for the depressed subjects compared with the non-depressed subjects. In patients in whom pain was present at baseline no socio-demographic variable alone predicted its persistence; however, male sex and white race together with 2 items of the CES-D did predict the persistence of existing pain. On regression analysis for the prediction of depression, chronic pain was the most powerful variable; other factors which were significant include low education, living in rural areas or areas containing up to 250,000 inhabitants, being unemployed and female sex. The odds ratio for the prediction of depression by chronic pain was 2.85. Although a necessary connection has not been established between the predictive and outcome variables nevertheless the evidence supports both the views that depression promotes pain and pain promotes depression. The latter effect was somewhat more powerful than the former, but both only accounted for a small proportion of the variance.