Key points are not available for this paper at this time.
The use of neutrophil left-shift parameters in the diagnosis of inflammatory and infective disease (ID) was evaluated. The level of C-reactive protein (CRP), currently the best quantitative parameter of inflammation, was used as the gold standard. Of 292 patients, 230 (79%) had a level of CRP of 1.0 mg/dL or greater and were classified as having inflammation, whereas 62 (21%) had normal levels. The neutrophil band count in each patient was determined by microscopic examination of 200 WBCs. The diagnostic value of the band count as an indicator for ID was evaluated in comparison to the WBC count, the neutrophil count, and the left-shift indicators of two automated hematologic analyzers, H*1 Technicon (Bayer Technicon Instruments, Tarrytown, NY) and Coulter MAX M (Coulter Electronics, Hialeah, Fla). When receiver operating characteristics were used, the band count was superior to the immature to total neutrophil count (I/T) ratio, the total WBC count, and the neutrophil count. The sensitivity and specificity in identifying ID at designated cutoff points were as follows: band count of 20% or greater of total WBC count (53% and 79%, respectively), I/T ratio of 0.25 or greater (59% and 63%), total WBC count of 9.6 x 10(6)/mL or greater (68% and 56%), and neutrophil count of 8.0 x 10(6)/mL or greater (60% and 58%). The performance of the H*1 Technicon left-shift flag was similar but slightly inferior to the band count (sensitivity, 44%; specificity, 79%), whereas the Coulter MAX M flags had a clearly higher sensitivity (79%) and lower specificity (53%). In addition, microscopic evaluation to determine the presence of reactive morphologic changes in neutrophils, such as toxic granules, Döhle bodies, and cytoplasmic vacuoles, had a high sensitivity (80%) but a low specificity (58%) in predicting ID. The diagnostic value of both microscopic and automated neutrophil left-shift parameters as indicators for ID is limited. Morphologic changes in neutrophils, however, either have a high specificity (band count) or a high sensitivity (toxic signs) in predicting ID and therefore may be a clinically useful tool.
Seebach et al. (Thu,) studied this question.