Background: Scores to predict the occurrence of postoperative vomiting (PV) or nausea and vomiting (PONV) are well established in adult patients. The aim of this survey was to evaluate the applicability of risk scores developed and tested in adult patients in 983 paediatric patients (0-12 yr) undergoing various surgical procedures.
Method: The predictive properties of five models were compared with respect to discriminating power (measured by the area under a receiver operating characteristic curve) and calibration (comparison of the predicted and the actual incidences of the disease by weighed linear regression analysis).
Results: The cumulative incidence of PV was 33.2% within 24 h. The discriminating power was low and insufficient in all models tested (0.56-0.65). Furthermore, the predicted incidences of the scores correlated only vaguely with the actual incidences observed.
Conclusion: Specialized scores for children are required. These might use the history of PV, strabismus surgery, duration of anaesthesia > or =45 min, age > or =5 yr and administration of postoperative opioids as independent risk factors.