The aim of this study was to define factors that influence therapy outcome of submandibular botulinum toxin injections for drooling in children with cerebral palsy or mental disability. We postulated that differences in response may be explained by the variation of dysfunctions in the various cerebral palsy subtypes. Prospectively collected data were evaluated of 80 spastic and 48 dyskinetic children, of whom 70% had an IQ of <70. In addition, the data of 23 fully ambulant children with mental disability only were examined. Flow and Drooling Quotient were assessed at baseline and at 8 weeks after injection. After treatment, both the Drooling Quotient and submandibular flow decreased in all children. Morbidity associated with the procedure was limited. Ninety-three children responded to botulinum. Decrease of submandibular flow in these children was associated with reduction of parotid flow. In those who did not respond to therapy, spread across all 3 diagnostic classifications, parotid flow increased after injection. Response failure is characterized by increased parotid flow after injection; however, the precise role of parotid flow in therapy failure remains unclear. We cannot predict who will respond to botulinum toxin to treat drooling.