[Acute pulmonary infections at autopsy. A study of clinical and macroscopic diagnoses at autopsy compared with microscopic autopsy findings].

The clinical diagnoses and gross diagnoses of acute pulmonary infections were investigated for 100 consecutively performed hospital autopsies, and the diagnoses were compared with the histological findings. We found 34 cases of pneumonia and one case of tuberculosis. Of these, 29 infections represented principal diagnoses, i.e. as causative of or contributive towards death. The predictive values, the sensitivities and specificities were estimated. The predictive value for gross diagnostics was low with a value of 42.9% against 58.8% for clinical diagnostics. The sensitivities for both clinical diagnostics and gross diagnostics were about 30% and the specificities about 85%. The study shows that sampling for histology from all pulmonary lobes is essential for correct autopsy diagnoses, either from areas that appear to be infected on gross examination or from the peripheral parts. Furthermore autopsy performance is still of great value for clinical diagnostics and for medical statistics.