Revista Médica del Uruguay
Print version ISSN 0303-3295
Summary Neonatal sepsis is a frequent and severe disease in hospitalized neonates that requires a prompt diagnosis in order to chose an adequate treatment. The purpose of this paper is to assess clinical and paraclinical variables to establish diagnosis of late neonatal sepsis through likelihood ratio and probability post-test. A retrospective study that included 47 neonates histories with clinical diagnosis of suspected sepsis. The studied variables were as follow: surgical history, progressive metabolic acidosis in patients with proper ventilation and hepatomegalia; hyperglycemia, and white blood cell and platelets count. Sepsis was defined through the clinical information plus positive hemoculture. Likelihood ratio was calculated for both clinical information and positive hemoculture. Fagan’s nomogram was used to estimate accuracy of diagnosis when one of the variables was present. It was found that for a sepsis prevalence in a 34% unit the studied variables improved sepsis probability. Surgical history, acidosis and leucopenia reached a probability of 62.6%; hepatomegalia 57%; hyperglycemia, leukocytosis and plateletopenia 45%. A neonate with a suspected sepsis that presents any of the studied variables has a probability of at least 50% of having a sepsis, in which case an active response should be taken.
Keywords : SEPSIS [diagnóstico]; RECIÉN NACIDO.