Background/Purpose

Rheumatoid arthritis (RA) is associated with a 20 to 30% increased risk of mortality from all-causes compared to the general population. The aim of the present study is to examine whether, as observed in the general population, mortality rates in RA have decreased over the past 25 years. Furthermore, to assess whether changes are due to a potentially milder disease at presentation or change in treatment practise.

Methods

Data from 32 centres in the UK that recruited 2763 patients to two inception cohorts between 1986 and 2012 were combined for the analysis: Early RA Study and Early RA Network. Both recruited DMARD naïve patients at presentation to the rheumatology clinic. Death certificates were provided by the NHS central register. All-cause mortality was standardised against population rates (stratified by age, sex and calendar year) to examine whether excess mortality risk had changed with time. Pooled logistic survival models estimated the relative yearly reduction in mortality hazard. Marginal structural modelling was used to examine the effect on methotrexate on survival, adjusting for confounding by indication of the treatment effect.

Results

The excess mortality risk in RA compared to the general population reduced over time. Restricting the analysis to deaths within 10 years of onset, for ERAS (recruitment 1986 to 2001) the all-cause standardised mortality ratio (SMR) was significantly increased (1.21; 95%CI 1.10 to 1.34), whereas for ERAN (recruitment 2002 to 2012) it was non-significant (1.04; 95%CI 0.88 to 1.22). The difference between SMR’s for the two cohorts was non-significant. Combining the two cohorts, year of symptom onset was significantly associated with all-cause mortality risk (HR = 0.96, p<.001; 95%CI .95 to .98). Controlling for demographic and clinical features at baseline did not reduce the magnitude of the effect for year of onset (HR = 0.95, p<.001; 95%CI .93 to .97). Extending the model to control for treatment using a marginal structural modelling approach, the use of methotrexate (use of which increased dramatically over the period of recruitment) was associated with a 60% reduction in the risk of death (HR = 0.40, p<.001; 95%CI .25 to .64). After controlling for methotrexate use the effect of year of onset was reduced to non-significant (Hazard ratio = 0.98, p<.001; 95%CI .96 to 1.01).

Conclusion

Substantial gains in life expectancy have been observed for people with RA in the UK over the last 25 years. The excess mortality risk appears to have been greatly diminished. This is probably due to changes in treatment practise, rather than RA becoming milder at presentation.

ACR Meeting Abstracts - https://acrabstracts.org/abstract/reduced-mortality-risk-in-rheumatoid-arthritis-findings-from-two-uk-inception-cohorts/