In a 10-year review of the utilization of continuous veno-venous hemofiltration (CVVH) for the treatment of neonatal hyperammonemia, 14 patients were identified with hyperammonemia due to either a urea cycle defect or an organic acidemia. Intensive care survival was 64%. The pretreatment level of serum ammonia and the rapidity of ammonia clearance did not differ between survivors and non-survivors (p = 0.16 and p = 0.93, respectively). Likewise, the duration of CVVH therapy did not differ between survivors and non-survivors (p = 0.1). Indicators of pretreatment physiological stress showed either a correlation with non-survival [Pediatric Risk of Mortality (PRISM) score, p = 0.006, cardioactive drug requirement, p = 0.003], or demonstrated a trend to such a correlation (serum lactate, p = 0.06). Complications associated with the CVVH technique were infrequent. Hypotension was seen in seven patients, but in only one patient did it arise de novo following the initiation of CVVH. In conclusion, neither the severity of the hyperammonemic state nor the efficacy of ammonia removal correlated with patient outcome. The pre-CVVH PRISM score and requirement for cardio-active medication were significantly greater in those patients who did not survive their acute illness. The pre-CVVH physiological condition of the neonates in this cohort was the main determinant of outcome.