Validation of HATCH score to predict neurological outcome in patients with aneurysmal subarachnoid hemorrhage in a Latin American middle-income country.

To externally validate the HATCH score for predicting 6-month neurological outcomes in patients with aneurysmal subarachnoid hemorrhage in a multicenter Latin American cohort, and to compare its performance with the Hunt-Hess and WFNS scales. Retrospective study across three Argentine hospitals (2011-2024). Adults with aSAH admitted to the ICU were included. The primary outcome was 6-month functional status (mRS 0-3 vs. 4-6); the secondary outcome was 12-month mortality. Discrimination was assessed with AUROC (DeLong test), and calibration with calibration-in-the-large (CITL) and slope, using risks from original derivations. 177 patients were analyzed; 57% achieved a favorable 6-month outcome. Discrimination for 6-month mRS was comparable: AUROC 0.80 (95% CI 0.73-0.86) for HATCH, 0.79 (0.72-0.86) for Hunt-Hess, and 0.81 (0.74-0.87) for WFNS (p = 0.44). Calibration differed: HATCH had good agreement across probabilities (slope 0.94 [0.64-1.24]) with mild underprediction (CITL 0.61 [0.26-0.96]); Hunt-Hess overpredicted (CITL - 0.58 [- 0.94 to - 0.24]) with poor agreement (slope 0.47 [0.32-0.62]); WFNS showed good agreement (slope 1.19 [0.84-1.54]) but systematic underprediction (CITL 1.62 [1.28-1.96]). At 12 months, discrimination for mortality was moderate: AUROC 0.72 (0.64-0.79) for HATCH, 0.69 (0.61-0.77) for Hunt-Hess, and 0.73 (0.66-0.81) for WFNS (p = 0.12). Calibration favored WFNS (slope 0.87 [0.52-1.23]; CITL 0.74 [0.37-1.11]); HATCH showed slope 0.55 (0.28-0.83) and CITL - 0.36 (- 0.75 to 0.03); Hunt-Hess was substantially miscalibrated. In this multicenter Latin American external validation, HATCH provided the most balanced performance for 6-month functional outcome; for 12-month mortality, all scores showed only moderate discrimination, with WFNS calibrating most stably.