Postoperative Density of Residual Hematoma Is a Predictor of Chronic Subdural Hematoma Recurrence—A Single-Center Retrospective Study

Background/Objectives: Chronic subdural hematoma (cSDH) is one of the most common neurosurgical disorders in older adults and remains associated with a relevant recurrence rate despite standardized surgical treatment. Identifying postoperative factors that predict recurrence may improve postoperative surveillance and surgical strategy. We therefore investigated whether the density of the residual hematoma on early postoperative computed tomography (CT) is associated with recurrence after burr-hole trepanation. Methods: We retrospectively analyzed all consecutive patients who underwent burr-hole trepanation for cSDH at our institution between 2016 and 2021. Residual hematoma density was measured in Hounsfield units (HU) on postoperative CT scans. Recurrence was defined as symptomatic hematoma reaccumulation requiring reoperation. Results: A total of 223 patients met the inclusion criteria. The overall recurrence rate was 20.2%. Mean residual hematoma density was significantly higher in the recurrence group than in the non-recurrence group (23.2 ± 13.1 HU vs. 19.0 ± 9.1 HU; p = 0.005). ROC analysis identified an optimal cutoff value of 19 HU (AUC 0.578, 95% CI 0.511–0.645; p = 0.022) with sensitivity of 57.8%, and a specificity of 57.9%. Patients with residual hematoma density ≥19 HU had a significantly higher recurrence rate than those with lower density (25.7% vs. 15.6%; p = 0.043). In multivariable logistic regression, higher residual density remained an independent predictor of cSDH recurrence (p = 0.008, OR 1.9; 95% CI 1.2–3.0). Conclusions: Higher postoperative residual hematoma density is an independent predictor of recurrence after burr-hole trepanation for cSDH. Residual density on early postoperative CT may serve as a simple and clinically accessible imaging marker of incomplete hematoma clearance and may help refine postoperative risk assessment. Nevertheless, the low AUC limits interpretation of predictive performance.