Longitudinal analysis of the relationship between serum insulin-like growth factor-I and radiographic knee osteoarthritis.

OBJECTIVE To examine the relation between serum insulin-like growth factor I (IGF-I) levels and both incident and progressive radiographic knee osteoarthritis (OA) in the Framingham Osteoarthritis Study. DESIGN Subjects had bilateral weight-bearing, anterior-posterior knee radiographs performed in 1983-1985 and again in 1992-1993. IGF-I levels were measured from blood specimens obtained in 1988-1989 by a competitive binding radio-immunoassay (RIA) after separation with octadecasilyl-silica cartridges of serum IGF-I from binding proteins. Participants without baseline radiographic OA [Kellgren and Lawrence grades (K&L) = 0-1] were classified as having incident disease if they had K&L > or = 2 grades at follow-up. Progressive OA was defined as an increase in K&L score of > or = 1 in knees with baseline OA (K&L > or = 2). All analyses were knee-based and sex-specific. We examined IGF-I tertiles in relation to the risk of incident and progressive radiographic OA separately, adjusting for age, body mass index (BMI), and baseline K&L score, and used generalized estimating equations to adjust for the correlation between fellow knees. RESULTS Four hundred and forty-one participants had knee radiographs and serum IGF-I levels measured. No associations were found for serum IGF-I levels and incident [women: OR = 0.9 (0.6-1.7), men OR = 1.2 (0.6-2.6)] or progressive [women OR = 0.9 (0.6-1.6), men OR = 0.9 (0.3-3.0)] radiographic knee OA in either sex. Neither did we observe any association between IGF-I and worsening of individual radiographic features of OA (i.e., osteophyte growth and joint space loss). CONCLUSION In summary, this longitudinal study did not demonstrate any association of serum IGF-I and incident or progressive radiographic knee OA. Further studies are needed to clarify the role of IGF-I in OA.