The optimal dose of oral midazolam with or without intranasal S-ketamine for premedication in children: a randomised, double blinded, sequential dose-finding trial