Over a 5-year period (1984-1988) intra- and periventricular hemorrhage (IVH/PVH) was observed in 299 preterm infants. Sixty-eight infants developed posthemorrhagic hydrocephalus (PH); of these, 23 infants died and 40 infants could be followed up for assessment of neurological development (5 patients were lost to follow-up). At 1 year of corrected age 15% (25% at 5 year follow-up) of the infants were determined to have developed normally, 35% (25% at 5-year follow-up) showed mild neurological symptoms and/or slight developmental delay, 32.5% (28% at 5-year follow-up) had handicaps and/or moderate mental retardation, and 17.5% (22% at 5-year follow-up) had severe handicaps and/or severe mental retardation. There was a significantly worse outcome in infants with grade 4 IVH/PVH (P < 0.05) and a significantly worse outcome in the group requiring ventriculoperitoneal (VP) shunt (P < 0.05). The results at 1 year of corrected age proved to be a quite realistic predictor of neurological functioning at 5 years of age (80% predicted correctly in the non-shunted-group--one patient lost to follow-up; 95% predicted correctly in the shunted group--four patients lost to follow-up). Cystic periventricular leukomalacia had been diagnosed in 7 (10%) patients and was associated with poor neurodevelopmental outcome. Gestational age, birth weight, time of shunt placement, and peripartum asphyxia had no significant influence on neurodevelopmental outcome. Infants with shunt infections and a high number of shunt revisions were found to have a significantly worse neurodevelopmental outcome (P < 0.01).