Acta Neurochirurgica (2019) 161:885–894
An increasing incidence of chronic subdural hematoma (CSDH) and an unchanging high recurrence rate of 10–20% call for individualized treatment. The aim of this study was to establish individualized prediction models for the risk of recurrence treating death as a competing risk.
Methods A retrospective national cohort of unilateral CSDH was included for analysis. Using competing risk survival analysis, we tested whether available covariates were associated with the risk of recurrence. We further established a preand a postoperative prediction model, where predictors were chosen using a LASSO approach. The models were visualized in nomograms. Predictive performance was evaluated by c index and calibrations plots.
Results A total of 763 patients with surgically evacuated unilateral CSDH were included for analysis. The recurrence rate was 14% while 12% of patients died during follow-up (1 year). In our association model, hematoma size, drain type, drainage time, presence of complications, and Glasgow Coma Score were significantly associated to recurrence. Subdural drain was associated with a lower recurrence risk than subgaleal drain. The preoperative model included hematoma size, hematoma density, and history of hypertension. The postoperative model included further drain type, drainage time, and surgical complications.
Conclusion The nomograms allow easy assessment of the recurrence risk for the individual patient, providing a better possibility for individual adjustment of treatment and follow-up. The predictive performance indicates that significant unaccounted or unknown factors still remain. The association test found passive subdural drain superior to passive subgaleal drain in minimizing the risk of CSDH recurrence