AIM: To determine the agreement rates between clinical and autopsy diagnoses in a neonatal intensive care unit (NICU), distinguishing between the main diagnosis and cause of death. METHODS: Clinical and autopsy records of 75 infants who died in two consecutive years in the NICU (autopsy rate 42.6%) of a pediatric hospital in Mexico City were reviewed. RESULTS: Ninety-two percent of main clinical diagnoses were confirmed by autopsy. Four conditions (congenital cardiopathy, prematurity, specific congenital syndromes and hyaline membrane disease) accounted for more than two-thirds of diagnoses. However, for cause of death, the global agreement was only 50%. The most common conditions considered by clinicians (77%) and pathologists (56%) to be the causes of death were cardiogenic, septic or mixed shocks. Additionally, clinicians omitted 34 relevant conditions in 30 (40.0%) patients, and 21 of these conditions possibly played a role in the deaths of 17 (22.7%) patients. The most frequently omitted diagnosis was pneumonia, in 9 (26.5%) patients. Omissions were not related to gestational age, age at death, days as an inpatient, or gender. CONCLUSION: Despite a high agreement rate in the main diagnoses, notable imprecisions were present regarding cause of death and antemortem overlooking of potentially fatal conditions, confirming the useful role of autopsy to verify clinical diagnoses and suggesting that differentiation between the main diagnosis and cause of death should be carried out in future studies.
AIM: To determine the agreement rates between clinical and autopsy diagnoses in a neonatal intensive care unit (NICU), distinguishing between the main diagnosis and cause of death. METHODS: Clinical and autopsy records of 75 infants who died in two consecutive years in the NICU (autopsy rate 42.6%) of a pediatric hospital in Mexico City were reviewed. RESULTS: Ninety-two percent of main clinical diagnoses were confirmed by autopsy. Four conditions (congenital cardiopathy, prematurity, specific congenital syndromes and hyaline membrane disease) accounted for more than two-thirds of diagnoses. However, for cause of death, the global agreement was only 50%. The most common conditions considered by clinicians (77%) and pathologists (56%) to be the causes of death were cardiogenic, septic or mixed shocks. Additionally, clinicians omitted 34 relevant conditions in 30 (40.0%) patients, and 21 of these conditions possibly played a role in the deaths of 17 (22.7%) patients. The most frequently omitted diagnosis was pneumonia, in 9 (26.5%) patients. Omissions were not related to gestational age, age at death, days as an inpatient, or gender. CONCLUSION: Despite a high agreement rate in the main diagnoses, notable imprecisions were present regarding cause of death and antemortem overlooking of potentially fatal conditions, confirming the useful role of autopsy to verify clinical diagnoses and suggesting that differentiation between the main diagnosis and cause of death should be carried out in future studies.