Evaluation of long-term outcome of patients with urea cycle diseases (UCD) is needed for medical decisions and counselling. Own data comparing outcome of UCD patients with the old treatment limited to protein restriction (i.e. close to the natural history) with that of patients on the modern conservative treatment have shown that gains in survival occur at the cost of more mentally retarded surviving patients. We discuss the possible bias in long-term outcome studies of those rare inheritable disorders where non-predictable environmental factors leading to catabolic crises have a crucial impact on prognosis. A combination of peak or initial ammonia value combined with the duration of coma is discussed as a criterion for prognosis of handicap. The neglect of dietary compensation of branched chain amino acid deficiency worsened by phenylbutyrate treatment in some published protocols could well be an additional cause of the non satisfactory long-term results of conservative treatment which - in our view - mainly aim at bridging optimally the period of late neonatal presentation until liver transplantation in patients with CPS and OTC deficiency (except for mild forms).