Estudio de la correlacion diagnostica clinica y necropsica en una Unidad de Cuidados Intensivos Pediatrica

Objective: To investigate the relationship between clinical and autopsy diagnoses in a Pediatric ICU of a university hospital in order to assess the incidence of class 1,2 correlations. Methods: Retrospective study of clinical records of patients who died between January 1991 and June 1994. Diagnoses were made by two pediatricians and a pathologist was consulted in cases of controversial diagnosis. Results: We reviewed 110 patients who died in PICU. Autopsy was performed in 44 patients (40%). Five of them were legal procedures and results were not available. There were 39 children: 23 were male and 16 were female. Their mean age was 30.4 months (range 1 month to 17 years). The mean PICU lenght-of-stay was 3.9 days (range 1 hour to 30 days). Clinical diagnoses were cathegorized as follows: septic shock 8, coma 8, cardiogenic shock 5, acute respiratory failure 4, hypovolemic shock 2, and several etiologies 8. There were two class 1 (a diagnosis that if it have been known before it could have indicated a different treatment) correlations (5%), eleven class 2 (autopsy findings that would not have indicated a different treatment: untreatable disease, correct treatment despite ignoring the real diagnosis) correlations (28%), two class 3 (secondary findings associated to the primary cause of death) correlations (5%) and three class 4 (secondary findings not related to the primary cause of death that could have only infuenced the outcome in terminal patients) correlations (7%). Clinical-autopsy correlation was complete in 21 cases (54%). No relationship was observed between PICU short lengths-of-stay and the incidence of class 1,2 correlations. Conclusions: Autopsy frequently shows new findings and/or unsuspected diagnoses. For this reason in all deceased patients authorization for autopsy should systematically be asked for. Short lengths of PICU-stay before death do not increase the incidence of class 1,2 errors.