The effect of smoking on clinical, laboratory, and radiographic status in rheumatoid arthritis

(1) To determine the degree to which rheumatoid factor (RF) positivity is associated with smoking; (2) to determine the quantitative effect of smoking and smoking length on the concentration of RF in all patients, and, in the seropositive patients separately, determining if a "dose-response" effect exists; (3) to investigate these relationships in men and women to clarify whether the effect of smoking is similar in both sexes; (4) to determine the effect of smoking, controlling for RF, on a variety of measures of disease status, severity, and activity.

Six hundred ten consecutive patients with rheumatoid arthritis seen for routine clinical care provided information on their smoking history. All underwent a complete joint examination, completed a series of health status questionnaires, provided information concerning pulmonary illnesses, underwent determinations for RF and erythrocyte sedimentation rate (ESR), and had hand radiographs.

RF concentration was linearly related to the number of years smoked. This association was present in both sexes, but was stronger in men. Smoking was similarly related to rheumatoid nodule formation. A nonlinear relationship was found between smoking and radiographic abnormalities as determined by the Larsen method, even controlling for RF. Similarly, pulmonary illness was independently related to smoking and RF. No effect of smoking was seen on disease process variables such as ESR, pain, joint count, global severity, or functional ability.

Quantitative relationships exist between smoking extent and RF positivity, RF concentration, nodule formation, radiographic progression, and pulmonary disease. These 3 latter effects are independent of RF positivity or concentration. Smoking does not contribute to alterations in disease activity measures, but appears to play a role in overall severity of disease.