We describe the results of treatment of open tibial fractures in 92 children; 22 fractures were Gustilo type I, 51 type II and 19 type III. All children received tetanus prophylaxis, systemic antibiotics for 48 hours and thorough debridement and irrigation of the wound. Fifty-one wounds with minimal soft-tissue injury were closed primarily. The other 41 were initially left open; of these, 18 small wounds were allowed to heal secondarily and 23 larger wounds required split skin grafts or soft-tissue local or microvascular free flaps. Stable fractures were reduced and immobilised in an above-knee plaster cast (71%) and external fixation (28%) was used for unstable fractures, extensive soft-tissue injury and multiple injuries. Short-term complications included compartment syndrome (4%), superficial infection (8%), deep infection (3%), delayed union (16%), nonunion (7.5%) and malunion (6.5%): these incidences are similar to those reported in adults. Selective primary closure of wounds did not increase the incidence of infection. External fixation was associated with a greater occurrence of delayed and nonunion than plaster immobilisation, but this technique was used most often for the more severe injuries. Late review, at 1.5 to 9.8 years, showed a high incidence of continuing morbidity including pain at the healed fracture site (50%), restriction of sporting activity (23%), joint stiffness (23%), cosmetic defects (23%) and minor leg-length discrepancies (64%). Open tibial fractures in children are associated with a high incidence of early and late complications, which are more frequent in children with Gustilo type III injuries. The Gustilo classification was a useful guide for predicting the outcome and planning treatment.