70 CHAPTER 3 Drooling severity was evaluated objectively using the validated, semi-quantitative drooling quotient, a method where the presence of new saliva on the lip or chin is directly observed every 15 seconds and recorded by a specialized speech and language therapist for 5 minutes.17 Participants were evaluated at least 1 hour after a meal, while awake and sitting up straight. To personalize the evaluation of treatment success, this was defined as a ≥50% reduction in VAS and/or drooling quotient from baseline.16 Other research groups have used similar definitions of treatment success or relied on a ≥2SD reduction in either VAS and/or drooling quotient. Statistical analysis Data were analysed with SPSS v22.0 for Windows (IBM Corp., Armonk, NY, USA). Descriptive statistics were used to summarize patient characteristics. Continuous variables were analysed with paired and unpaired t-tests. If the assumption of normallydistributeddatawas violatedaccording to theShapiro–Wilk test, continuous variables were analysed with the Wilcoxon signed-rank and Mann–Whitney U tests. Evaluation of treatment response over time was performed using a linear mixed model with change in the VAS as the dependent variable and treatment and visits as the fixed factors. The predictive value of BoNT-A treatment success after 8 and 32 weeks on treatment success 32 weeks after bilateral submandibular duct ligation was tested with a c2 test. To evaluate a potential selection bias for the inclusion of only poor responders to BoNT-A or differences in patient characteristics, we compared the participants in our study to a cohort of 232 children who received a total of 407 BoNT-A injections at our centre between January 2002 and May 2013.1 Differences in mean VAS and drooling quotient and baseline patient characteristics (i.e. age, epilepsy, intellectual disability, CP diagnosis, Gross Motor Function Classification System level) between the two groups were analysed with Mann–Whitney U and c2 tests respectively. Statistical significance was defined as p≤0.05 for all statistical analyses; 95% confidence intervals (CIs) were provided for the interpretation of point estimates.
RkJQdWJsaXNoZXIy MjY0ODMw