Serum COMP was measured by inhibition enzyme- linked immunosorbent assay (ELISA) using a polyclonal antiserum w4x. Serum CRP was quantified by a sensitive sandwich ELISA

SIR, Traditionally osteoarthritis (OA) has been considered a process involving a disturbance of the normal balance between degradation and repair in the articular cartilage and subchondral bone. In contrast to rheumatoid arthritis where inflammation, both local and systemic, is a key feature, OA is considered a primarily non-inflammatory condition. The clinical signs of lowgrade inflammation that are seen in some patients, representing later stages of the process, have been thought to reflect secondary events in the joint. This view has been questioned and in a recent review arguments in favour of inflammation being of major importance in OA pathophysiology were presented w1x. In support, in established OA, low-level increases in serum C-reactive protein (CRP) have been reported w2, 3x. To examine the role of inflammation in early knee OA, we determined serum CRP in individuals with knee pain at baseline and after 3 yr and correlated the levels to the presence of radiographic OA as defined below after 3 yr. We also correlated serum CRP to serum cartilage oligomeric matrix protein (COMP) to investigate further the role of COMP as a cartilage marker and examine the relationship between CRP and COMP as markers for different pathophysiological features w4x. The examined individuals have been described including the COMP results w5x. Thirty-eight individuals with chronic knee pain (>3 months at inclusion) were monitored over a 3-yr period and divided into two groups, those who after 3 yr had normal radiographs (n=15) and those who showed joint-space narrowing in the tibiofemoral (joint space <3 mm) anduor the patellofemoral joint (joint space <5 mm) (n=23) w6, 7x. For comparison, sera from 20 healthy blood donors were examined. There was no significant age difference between the three groups. The median (range) age of the individuals with positive radiographs was 49 yr (39–54), for the individuals with knee pain with negative radiographs 44 yr (37–54) and for the blood donors 44 yr (37–55). There were 10 male and 10 female blood donors. Of the persons developing positive radiographs, 14 were female and nine male and of the persons with only knee pain, three were female and 12 were male. Serum COMP was measured by inhibition enzymelinked immunosorbent assay (ELISA) using a polyclonal antiserum w4x. Serum CRP was quantified by a sensitive sandwich ELISA (detection limit <0.05 mgul) developed using commercially available reagents. Immunoplates C96 Maxisorp (NUNC, Copenhagen, Denmark) were coated by a dilution of rabbit anti-human CRP antibodies (DAKO A-073, Copenhagen, Denmark) and dilutions of human serum CRP-calibrator (DAKO X-0923, Copenhagen, Denmark) were used for constructing a standard curve. Sheep anti-human CRP Letters to the Editor 903