RESUMEN
Leukocyte and neutrophil counts are two of the commonest laboratory tests used in clinical medicine. The usefulness of the tests is still controversial. It has been found that the tests were useful to a limited extent. We performed leukocyte and neutrophil counts in patients manifesting acute febrile illness, to identify the best cut-off point in differential diagnosis of underlying disorders causing such conditions. Patients were enrolled randomly on the basis of acute febrile illness. Those who suffered from diseases affecting myeloproliferative system were excluded. Leukocyte and neutrophil counts were performed by the standard method. Diagnosis of target organ disorder was obtained from the house staff's final diagnosis. Both laboratory and clinical performance were assessed in double blind fashion. Patients were divided into 2 groups. Group 1 consisted of 47 patients with acute bacterial infections and group 2 consisted of 53 cases with acute febrile conditions caused by diverse, non-bacterial disease. Various cut off points ranging from 5,000-25,000/mm3 and 10-100 per cent were used for leukocyte and neutrophil counts. Sensitivity, specificity and Youden index for each cut off point were determined. Receiver operating characteristic (ROC) curves were constructed. It was found that sensitivity and specificity for each cut off point for both leukocyte and neutrophil counts displayed perfect trade-offs and the Youden indices were similar. The ROC curves for both counts were depicted as a 45-degree lines of non-discrimination. Leukocyte and neutrophil counts were proved to be non-discriminate of acute febrile conditions.(ABSTRACT TRUNCATED AT 250 WORDS)