Aims To externally validate the Endoscopic Third Ventriculostomy Success Score (ETVSS) recently found by Kulkarni et al1 to predict successful ETV for hydrocephalus, that is, child not requiring shunt or repeat ETV, on the basis of a child's individual characteristics. The ETVSS is based upon patient age, aetiology and presence of a previous shunt. To date there has been no external validation of this model. We consider both short term and long term outcome, using a detailed large single centre series of paediatric patients with a long period of follow-up.
Methods We retrospectively identified consecutive children undergoing ETV at a single regional neurosurgery centre. We compared actual success at both 6 and 36 months with mean predicted probabilities for low, moderate and high chance of success strata based on the ETVSS. Long-term success was calculated using Kaplan Meier methods and comparisons made by unpaired t tests.
Results In total 166 primary ETV were performed at a median age of 39 (range 0.03–230) months. There was a greater number of patients in younger age groups in this series; 49 patients were under 6 months of age (29.5%) compared to 129 (20.9%) in the model dataset (difference 8.6%, CI 1.0 to 16.3; p=0.07). Overall, ETV success was 72.9% at 6 and 64.5% at 36 months. Although derived to predict outcome at 6 months, the model predicted outcome better on long-term follow-up than at 6 months. At long-term follow-up, the mean predicted probability was significantly higher in those with a successful ETV (n=99) than in those who failed (n=67) (p=0.001). The ETVSS accurately predicted long term success rates; the low, medium and high groups had mean predicted probabilities of 82%, 63% and 36% and overall success at 36 months of 76%, 66%, 42% respectively. The overall complication rate was 6%.
Conclusion The ETVSS closely predicted the overall long term success rates in high, moderate and low risk groups. Our study suggests the ETVSS will be useful in the clinical decision-making in predicting long term outcome of ETV but further refinement of the model, validation and comparison with shunt treatment is required.