Reduced Disability at Five Years with Early Treatment of Inflammatory Polyarthritis
Clinical trials in rheumatoid arthritis (RA) have demonstrated the efficacy of disease modifying antirheumatic drugs (DMARDs) in reducing disability in the short-term. Early aggressive treatment of RA is recommended based on short-term study results but there are only sparse data documenting longer term outcomes. The cost of long-term clinical intervention trials is prohibitive. Thus, long-term outcomes are increasingly being studied using prospective longitudinal observational cohorts which, by nature, are not “controlled” or randomized. A direct comparison of the outcomes of treated and untreated patients in this type of study could be misleading since treated patients may represent those with more severe disease. If a patient who is treated has a worse outcome than a patient who is not treated, it can be difficult to discern whether this is due to severe disease or a lack of efficacy of the treatment.
The “propensity score” is one method of adjusting for the bias in treatment assignment present in observational studies. The propensity score provides an estimate of the likelihood that the patient would receive treatment based only on the values of the individualss RA-related covariates (e.g., number of swollen or tender joints; rheumatoid factor; etc.)
Wiles et al (Arthritis and Rheumatism 44:1033-1042, 2001) utilized this methodology to investigate the effect of treatment with DMARDs and/or steroids on the 5-year disability outcome in patients with inflammatory polyarthritis. 384 patients registered by the Norfolk Arthritis Register were followed for 5 years. Treatment details and Health Assessment Questionnaire (HAQ) scores were recorded annually. Logistic regression was used to model differences in baseline factors associated with the start of DMARDs and/or steroids within 12 months of baseline. Based on this model, each subject was given a probability of starting treatment (“propensity score”). A second model compared the odds of disability (HAQ score > 1.00) in treated and untreated patients, adjusting for differences in disease severity using the propensity scores.
Results: Unadjusted analysis suggested that patients who received treatment had an increased odds of a worse outcome compared with those who did not receive treatment. When adjusted for differences in disease severity, using the propensity score, patients who received early treatment (within 6 months of symptom onset) had a similar odds of disability at 5 years as those not treated (odds ratio 0.71; 95% confidence interval 0.34, 1.44). In contrast, starting treatment later (> 6 months) was associated with a 2-fold increased odds of having a HAQ score of > 1.00 at 5 years.
Conclusion: These data support the notion that early aggressive treatment is associated with better functional outcomes at 5 years. The propensity score was used to adjust for “confounding by indication”.
Editorial Comment: As the authors point out, the propensity score methodology is becoming increasing popular in situations where randomized clinical trials are not feasible, most notably thus far in the field of cardiac medicine. In the current study, had an adjustment for disease severity not been made, the results would have indicated that treatment was detrimental rather than helpful. Further investigation of this methodology and application to arthritis cohorts is indicated.
