• rheumatoid arthritis;
  • mortality;
  • radiograph;
  • cohort;
  • methotrexate



There is an excess of mortality in patients with rheumatoid arthritis (RA) but no long-term Australian cohort data.


To determine median life years lost, all-cause standardised mortality ratio (SMR) and cause-specific SMR, their predictors and secular change in Australian patients with RA.


Study population was all patients seen by a rheumatologist between 1990 and 1994. Record linkage with Australian National Death Index was performed to determine fact and cause of death up to 2004. All-cause and cause-specific SMR, and median life years lost were determined.


There were 35 (31%) deaths in the early 1990s cohort (n = 113), SMR 1.31 (95% 0.93, 1.80). There were 216 (44%) deaths in the pre-1990s established cohort (n = 495), SMR 1.73 (1.49, 1.95). Median life years lost in the early cohort was 6 years for males and 7 years for females compared with 8 and 10 years, respectively, in the established cohort. Patients with low disease activity score at baseline (DAS < 3.2), SMR was 0.8 (0.3, 2.2) and 1.5 (1.1, 2.2) for the early and established cohorts, and if DAS ≥3.2, SMR was 1.4 (1.02, 1.98) and 1.8 (1.5, 2.1) respectively. Primary cause of death was cardiovascular disease (SMR 1.43 (1.17, 1.74). Patients at most risk were those age 45–54 years. RA was listed as a comorbid condition on the death certificate in only 16% of patients.


Within a period of 14 years, median life expectancy of patients with RA with disease onset in the early 1990s is reduced by 6–7 years. However, our results also suggest a secular reduction in excess mortality.