Trojano and colleagues,1 based on their observational study, suggest that interferon-β slows progression in relapsing-remitting multiple sclerosis patients. They have used propensity scores to correct significant differences between interferon-β–treated and untreated control groups, and then conducted a sensitivity analysis to detect how the magnitude of an unmeasured binary confounder might affect the propensity score adjusted hazard ratios. In fact, several potential confounders, which are significant predictors of disease progression,2–4 have been excluded from the propensity score such as symptoms at onset, completeness of recovery from initial attack, interval between first and second attack, number of relapses during the first 2 or 5 years, T2 lesion load on first magnetic resonance image, and so on. Taking account of these confounders may change the result of their study.


  1. Top of page

Olivier Gout MD*, * Department of Neurology, Fondation Opthalmologique Adolphe de Rothschild, Paris, France