Changes in Outcome Prediction During the First Week After Subarachnoid Hemorrhage

Abstract

BACKGROUND: Prediction of long-term outcome based on initial neurological condition after aneurysmal subarachnoid hemorrhage varies with time. To date, studies have been limited to early time points and have reported that prognostication is best after resuscitation. OBJECTIVE: To describe how prediction of outcome varies from ictus through the first week of admission. METHODS: A retrospective analysis of patients with a diagnosis of aneurysmal subarachnoid hemorrhage recruited to a prospective database. Neurological condition was recorded on each day of the inpatient stay, up to day 7, using World Federation of Neurological Societies score (WFNS). Poor outcome was defined by modified Rankin scale of 3-6 at 3 months. Outcome prediction was assessed using area under the curve (AUC) after binary logistic regression. RESULTS: Of 645 patients, 55(14%) patients with WFNS 1&2 and 77(45%) patients with WFNS 4&5 on day 0 had a poor outcome. 30(8%) patients with WFNS 1&2 and 54(81%) patients with WFNS 4&5 on day 7 had a poor outcome. Prognostication using WFNS improved from day 0 to day 7 (AUC = 70.1%, CI 65.0%–75.1% vs AUC = 81.9%, CI 77.4%–86.0%) with an incremental improvement with each day in between, and the largest increases early around the time of resuscitation. CONCLUSION: Prediction of outcome improves beyond the initial resuscitation, up to day 7 of admission, with no evidence of any deterioration around the time of treatment or delayed complications like delayed cerebral ischemia. This is important when prognosticating for clinical purposes and emphasizes the importance of standardization of timing of WFNS in research.