Data are presented as counts or either as medians with full or interquartile ranges or as means with SDs. storehealthmall.eu
The effect of MMF on lung function was examined by using longitudinal data analytic methods. Longitudinal models were utilized to examine each of three continuous pulmonary physiology variables (FVC, total lung capacity [TLC], and diffusing capacity of the lung for carbon monoxide [Dlco]). Each model modeled the mean of the given physiology variable at each of three time points: (1) before MMF initiation; median, 166 days; (2) at the time of MMF initiation ±33 days; and (3) after MMF initiation; median, 371 days. Thus, a model yielded least-squared means estimates for its physiology variable at each of the three time points, and the null hypothesis of equality among the three mean values for a given variable was tested.
Compound symmetry was confirmed as the appropriate method for handling the within-subject correlation of the repeated measures in these models by examining fit statistics (eg, —2 residual log likelihood and the Akaike information criterion) in the output from SAS PROC MIXED (SAS Institute; Cary, NC). By using PROC MIXED, we were able to take advantage of using all available data points (ie, patients who had missing data were not excluded from the longitudinal analyses). All p values from the longitudinal analysis were adjusted for multiple comparisons using the Tukey-Kramer method. Prednisone doses before and after initiation of MMF were compared by using the Wilcoxon signed-rank test. Mean differences in pulmonary physiology variables that occurred over the two study time intervals— the first interval was from before MMF initiation to the time of MMF initiation, and the second interval was from the time of MMF initiation to the time of the most recent study—were compared using paired t tests. All statistical analyses were performed using statistical software (version 9.1; SAS Institute; Cary, NC). We considered p values < 0.05 to be statistically significant.