Evolución y modelo pronóstico de los pacientes con el diagnóstico de hemorragia subaracnoidea espontánea que ingresan en la unidad de cuidados intensivos

Resumo

]Objectives: • To determine the prognosis of patients admitted by spontaneous subarachnoid haemorrhage (SAH) and associated factors. • To develop predictive models at admission of in-hospital mortality and at long-term survival in these patients. Material and methods: Retrospective follow-up study of patients admitted in 2003-2013 to the Intensive Care Unit of the Complexo Hospitalario Universitario de A Coruña with SAH. We analyzed on demographic characteristics, comorbidity, clinical debut, prognostic seores at admission, diagnostic-therapeutic management, complications, inhospital mortality and survival at 10 years after diagnosis. Results: We included 536 patients (Age: 56.9±14.1 years, 59.3% women, Charlson index 0.7±1.1), the most frequent manifestations being headache (76.5%), dizziness (54.1 %) and coma (34.0%). The presence of aneurysms (SAH-A) was found in 78.1 % of cases, idiopathic etiologic (SAH-I) was found in 17.9%, and 3.2% presented perimesencephalic bleeding (SAH-PM). Neurological complications occurred in 53.7% of cases during admission and 48.9% of cases presented systemic complication, which were more frequent in SAHA. In-hospital mortality was 27.8% and probability of survival since admission was 62.8% at 5 years and 59,1 % at 10 years. The predictive variables ofin-hospital mortalitywere debut coma (OR=l.87), a score of 3-4 on the Fisher scale (OR=2.26), the APACHE II scale in the first 24 hours (OR=l.05) and total SOFA day O (OR=l.19), the same variables together with age, being associated with long-term survival. Two predictive models of in-hospital mortality and survival were constructed, with adequate calibration and discriminating capacity. 37 Resumen del estudio Conclusions: Patients with SAH are young, with low comorbidity and significant prognostic impact depending on the clinic of debut, severity on admission and complications during the stay. Depending on the subtype, patients have different characteristics and prognosis. The proposed prediction models improve the predictive capacity of other published scales.