OBJECTIVES--To collect and evaluate all trials on clinical efficacy of topical treatments for head lice. DESIGN--Systematic review of randomised trials identified from following data sources: Medline, International Pharmaceutical Abstracts, Science Citation Index, letters to key authors and companies, and hand search of journals. SETTING--Trials in schools or communities. SUBJECTS--Patients infested with lice. MAIN OUTCOME MEASURE--Cure rate (absence of live lice and viable nits) on day 14 after treatment. RESULTS--Total of 28 trials were identified and evaluated according to eight general and 18 lice specific criteria. Of the 14 trials rated as having low to moderate risk of bias, seven were selected as they used the main outcome measure. These seven trials described 21 evaluations of eight different compounds and placebo (all but two evaluations were of single applications). Only permethrin 1% creme rinse showed efficacy in more than two studies with the lower 95% confidence limit of cure rate above 90%. CONCLUSIONS--Only for permethrin has sufficient evidence been published to show efficacy. Less expensive treatments such as malathion and carbaryl need more evidence of efficacy. Lindane and the natural pyrethrines are not sufficiently effective to justify their use.