Using the number of joints with erosion in a total of 68 joints throughout the body, we studied a population of patients with rheumatoid arthritis whose disease duration was 10-15 years. Three groups, each showing a Poisson distribution, were found: the subset with least erosive disease (LES), the subset with more erosive disease (MES), and the subset with mutilating disease (MUD). The mean number of joints with erosion was 10.9 in LES, 32.2 in MES, and 53.5 in MUD. In LES, erosive articular changes were primarily limited to the peripheral smaller joints. In MES, the larger axial joints were also involved. Almost all joints were extensively damaged in MUD. During the early period of disease, differences between the 3 groups were highly significant in the rapidity of carpal bone destruction, as assessed by the yearly reduction of carpal height ratio (P < 0.001), and in the serum Clq level ( P < 0.001).The average serum C I q level (mean ? SD 2 I 1.9 2 26.9 pg/ml) in patients with rheumatoid arthritis (RA) has been demonstrated to be elevated ( I ) and usually remains constant during the initial 7 years of disease (2). The level of serum CIq in RA has been found to be significantly higher than that of a healthy control group (mean ? SD 136.5 ? 22.6 pg/ml) (2,3).In a following investigation in which Clq levels and clinical findings over a >%year period were studied, it was found that the Clq level determined at the initial visit and during the first 5 years of RA seemed to have a prognostic value (2). Every patient whose Clq level was >250 pg/ml at the first clinic visit had extensive erosions and joint destruction by 10 years of disease, whereas in patients whose initial Clq level was <250 pg/ml, only a small number of joints had similar involvement. Thus, these data indicated the possible existence of at least 2 subsets of RA patients who had different natural histories based on the extent of joint damage.The present study was undertaken first to confirm that there were disease subsets of RA based on the extent of joint destruction, and then to determine the extent of the fluctuation of the serum Clq level in each disease subset over a long period of disease.
PATIENTS AND METHODSPatients. Patients in this study were seen at the Department of Orthopedics, Osaka University Hospital, Osaka, Japan; the Hoshigaoka Koseinenkin Hospital, Osaka; and the Shirahama Hot Spring Hospital, Wakayama, Japan. All met the American Rheumatism Association criteria for the diagnosis of classic or definite RA (4). We excluded patients who had received therapy with disease-modifying agents (9, such as gold or D-penicillamine, for more than 1 year.There were 198 women and 42 men, with a mean age of 54 years (range 28-76) at the time of this study. All had