Prognostic Model to Predict Outcome of Adult Patients with Tuberculous Meningitis Undergoing Ventriculoperitoneal Shunt Surgery: A Prospective Observational Study

Background: The prognosis of adult patients with tuberculous meningitis undergoing ventriculoperitoneal (VP) surgery is not well known. Prognostic models developed to predict the prognosis might help clinicians immensely.

Methods: This was a prospective study. Adult patients (≥18 years) with tuberculous meningitis undergoing VP shunt surgery were included in the study. Patients were followed for 6 months. The primary outcome was death and the secondary outcome was a composite outcome of death plus disability. Prognostic models were developed using binary logistic regression. The model performance was assessed using discrimination and calibration. Internal validation of the model was performed using bootstrap. The models were plotted on a nomogram and e-calculator for bedside use.

Results: Of 92 patients, 28 (30.4%) died and 36 (39.1%) experienced a composite poor outcome. The final model showed Medical Research Council grade III, papilledema, seizures, and size of the temporal horn to be predictors of poor outcome. The models showed good discrimination: area under the curve of 83.7% (95% confidence interval, 75.1%-92.3%; P < 0.001) for death and area under the curve of 84.7% (95% confidence interval, 75.9%-93.5%; P < 0.001) for composite poor outcome. Both the models showed good calibration; the Hosmer-Lemeshow test P value of 0.448 for death and 0.093 for composite outcome. The model remained valid after performing a bootstrap procedure, thus signifying internal validation.

Conclusions: About 39% of adult patients with tuberculous meningitis die or remain disabled after VP shunt surgery. Medical Research Council grade III illness, papilledema, seizures, and size of hydrocephalus are the most important baseline predictors that can help in prognostication. The nomograms developed may help the treating physician with prognostication.