Neonatal screening programs for congenital hypothyroidism (CH) were initiated in the 1970s to ensure early treatment in order to prevent mental retardation. With screening, developmental prognosis is considerably improved, but follow-up studies still report developmental delay compared to controls. To explain the delay, most studies have focused on effects of CH severity, supposedly caused by prenatal hypothyroidism. Nonoptimal treatment could also be an explanatory factor. Treatment guidelines have changed; now, a higher levothyroxine starting dose is recommended. In this review, we first summarize outcome studies in CH versus controls and in mild versus severe CH. Second, we report results on the association between levothyroxine treatment variables and developmental outcome. Six research groups have reported results on the association between levothyroxine starting dose and intelligence, and eight groups on the association between circulating thyroid hormones and intelligence. Most studies found a positive association between a high level of treatment, primarily treatment during the first year and later intelligence. However, negative associations between high-dose treatment and outcome have been reported, and the question of optimal treatment in relation to developmental outcome has not been answered. Effects of high levothyroxine dosage should be documented in samples that have been treated according to recent recommendations.