Analysis of covariance (ANCOVA)* of the difference between the rituximab versus control groups in the distributions of AUC mean values at 12 months of follow-up without and with transformations adjusted for age and sex and the baseline value.

*In each analysis, the like-transformed (or untransformed) baseline AUC mean value was employed as an adjusting covariate, e.g. of the baseline value in the analysis of the 12 month values. The means of the transformed (or untransformed) values within groups, and the difference, are presented along with the root mean square error (RMSE) of those values, the standardized difference, the F-value and one-sided p-value for the effect of treatment.†These analyses are based on the intention-to-treat cohort that includes 29 placebo-treated subjects who met the defined criteria. The results using the values are identical to those that appeared in the primary study manuscript [6].