BACKGROUND: Evidence suggests that pathophysiologic processes in bone are important determinants of outcome in osteoarthritis of the knee. Low intake and low serum levels of vitamin D may compromise favorable responses of bone to osteoarthritis, predisposing patients to progression.

OBJECTIVE: To determine whether dietary intake and serum levels of vitamin D would predict the incidence and progression of osteoarthritis of the knee in participants of the Framingham Study.

DESIGN: Prospective observational study.

SETTING: The Framingham Study.

PARTICIPANTS: Participants in the Framingham Heart Study who had knee radiography at examinations 18 (done between 1983 and 1985) and 22 (done between 1992 and 1993) and received interim assessments of vitamin D intake and serum levels.

MEASUREMENTS: Intake of vitamin D and serum levels of 25-hydroxyvitamin D, calculated on the basis of dietary habits and supplement use as reported on a questionnaire, were evaluated at examination 20 (1988 to 1989). Knee radiographs were given scores for global severity of osteoarthritis, using a modification of the scale of Kellgren and Lawrence (range, 0 to 4), and for the presence of osteophytes and joint-space narrowing (range, 0 to 3). Covariates measured at examinations 18 and 20 were age, sex, body mass index, weight change, injury, physical activity, health status, bone mineral density, and energy intake.

RESULTS: 556 participants (mean age at baseline +/- SD, 70.3 +/- 4.5 years) had complete assessments. Incident osteoarthritis occurred in 75 knees; progressive osteoarthritis occurred in 62 knees. Serum levels of vitamin D were modestly correlated with vitamin D intake (r = 0.24). Risk for progression increased threefold in participants in the middle and lower tertiles for both vitamin D intake (odds ratio for the lower compared with the upper tertile, 4.0 [95% Cl, 1.4 to 11.6]) and serum levels of vitamin D (odds ratio for the lower compared with the upper tertile, 2.9[Cl, 1.0 to 8.2]). Low serum levels of vitamin D also predicted loss of cartilage, as assessed by loss of joint space (odds ratio, 2.3 [Cl, 0.9 to 5.5]) and osteophyte growth (odds ratio, 3.1 [Cl, 1.3 to 7.5]). Incident osteoarthritis of the knee occurring after baseline was not consistently related to either intake or serum levels of vitamin D.

CONCLUSIONS: Low intake and low serum levels of vitamin D each appear to be associated with an increased risk for progression of osteoarthritis of the knee.