Objective. To determine the risk and causes of death and to quantify mortality predictors in patients with rheumatoid arthritis (RA).Methods. RA patients (n = 3,501) from 4 centers (Saskatoon n = 905, Wichita n = 1,405, Stanford n = 886, and Santa Clara n = 305) were followed for up to 35 years; 922 patients died.Results. The overall standardized mortality ratio (SMR) was 2.26 (Saskatoon 2.24, Wichita 1.98, Stanford 3.08, Santa Clara 2.18) and increased with time. Mortality was strikingly increased for specific causes: infection, lymphoproliferative malignancy, gastroenterologic, and RA. In addition, as an effect of the SMR of 2.26, the expected number of deaths was increased nonspecifically across all causes (except cancer), with a large excess of deaths attributable to cardiovascular and cerebrovascular diseases. Independent predictors of mortality included age, education, male sex, function, rheumatoid factor, nodules, erythrocyte sedimentation rate, joint count, and prednisone use.

The objective of this study was to determine the relative risks of malignancy and of site-specific malignancies in patients with systemic lupus erythematosus (SLE). A cohort of 297 patients (91% Caucasian) with SLE were seen between 1975 and 1994 and followed for a mean of 12 years at the University of Saskatchewan Rheumatic Disease Unit. Expected cancer incidence rates were determined based on Province of Saskatchewan population statistics matched to each study patient for age, sex and calendar year of follow-up. Standardized incidence ratios (SIRs) of observed to expected cancers and 95% confidence intervals (95% CI) were calculated. A total of 27 cases of cancer were observed, whereas only 16.9 were expected (SIR 1.59 (95% CI 1.05-2.32)). For site-specific malignancies, an excess of cancer of the cervix (SIR 8.15 (95% CI 1.63-23.81)) as well as hemopoietic malignancy (SIR 4.9 (95% CI 1.57-11.43)) was found. The hemopoietic cancers were predominantly non-Hodgkin's lymphoma (SIR 7.01 (95% CI 1.88-17.96)). We did not find an association of malignancy with known risk factors, including use of cytotoxic agents. Increased risk of malignancy, notably non-Hodgkin's lymphoma and perhaps cervical cancer, should be regarded as a complication of SLE.

Objective. To determine the relative risks of malignancy and of site-specific malignancies in patients with rheumatoid arthritis (RA).Methods. In a prospective cohort study, 862 patients with RA (96% white) were enrolled from 1966 to 1974 and were followed up for up to 35 years (mean 17.4 years) at the University of Saskatchewan Rheumatic Disease Unit. All diagnoses of cancer were crossreferenced with the Provincial Cancer Registry. Expected cancer incidence rates were determined based on province of Saskatchewan population statistics matched to each study patient for age, sex, and calendar year. Standardized incidence ratios (SIRS) of the observedto-expected cancer incidence and 95% confidence intervals (95% CI) were then calculated.Results. A total of 136 cases of cancer were observed compared with 168 expected (SIR 0.80, P = 0.011 [95% CI 0.67-0.951). The relative risk of colorectal malignancy was significantly reduced in the RA study population (SIR 0.52, P = 0.037 [95% CI 0.25-0.961). A significant excess of leukemia was found (SIR 2.47, P = 0.026 [95% CI 1.12-4.69]), whereas the incidence rates for Hodgkin's disease and non-Hodgkin's lymphoma and all other site-specific malignancies were not found to be significantly different from general population rates.Conclusion. In our cohort of RA patients, colorectal cancer was detected in only half the expected number of patients. This risk reduction may be related to long-term nonsteroidal antiinflammatory drug (NSAID) use in RA, as has been suggested in several other studies of long-term NSAID use. An increased risk of leukemia was confirmed. This may be due to the persistent immune stimulation associated with RA it-

No abstract