Primary efficacy measures yielded pain scores, which were analysed with repeated measures analysis of variance (ANOVA), and patient preference was analysed with a goodness-of-fit chi 2 test.