Intraventricular fibrinolytic enhancement with plasminogen activators is an experimental treatment for posthaemorrhagic hydrocephalus, but some infants do not respond. The objectives of this study were to investigate whether plasminogen activator inhibitor-1 is detectable in normal or posthaemorrhagic neonatal cerebrospinal fluid and whether higher neonatal cerebrospinal fluid concentrations of plasminogen activator inhibitor-1 are associated with failure of fibrinolytic therapy. Cerebrospinal fluid samples from 7 controls and 16 infants with posthaemorrhagic hydrocephalus (15 treated with exogenous fibrinolytic agents) were analysed for plasminogen activator inhibitor-1. Plasminogen activator inhibitor-1 was not detectable in any of the control samples but was detectable in all but one of the posthaemorrhagic samples, and at significantly higher levels in the treatment failures (median 94 ng ml(-1)) than in the treatment successes (median 25 ng ml(-1)). High levels of plasminogen activator inhibitor-1 in the cerebrospinal fluid are predictive of, and provide a plausible biological explanation for, failure of intraventricular fibrinolytic therapy.