The most award winning
healthcare information source.
TRUSTED FOR FOUR DECADES.
And the Bands Played On
Abstract & Commentary
By Stan Deresinski, MD, FACP, Clinical Professor of Medicine, Stanford; Associate Chief of Infectious Diseases, Santa Clara Valley Medical Center. This article originally appeared in the December 2010 issue of Infectious Disease Alert. It was peer reviewed by Timothy Jenkins, MD. Dr. Jenkins is Assistant Professor of Medicine, University of Colorado, Denver Health Medical Center. Dr. Deresinski does research for the National Institute of Health, and is an advisory board member and consultant for Merck, and Dr. Jenkins reports no financial relationships relevant to this field of study.
Synopsis: Physicians may not be able to rely upon their clinical laboratory to accurately report the neutrophil "band count."
Source: Geissal ED, Coffey T, Gilbert DN. Clinical importance of the failure to detect immature neutrophils by an automated hematology analyzer. Infect Dis Clin Pract. 2010;18:374-378.
Knowledge of the presence of an increased proportion of band neutrophils (neutrophils with non-segmented nuclei) is believed by many clinicians to assist them in the diagnosis and management of some patients with suspected or known infection. Automated hematology analyzers, however, are not capable of providing a "band count." The identification of band neutrophils instead depends upon manual review of a blood smear. The need for such a review may be indicated as a result of the machine having detected a predetermined degree of abnormality, such as in the cell volume, cell number, or light scatter. This may be followed by a rapid visual scan of the smear and, then, if abnormalities are suggested by that review, by a more extensive examination with reporting, among other things, of the proportion of white blood cells (WBC) made up of bands and other earlier immature forms. Greissal and colleagues determined the overall sensitivity of this process in the detection of "bandemia" by comparing the proportion of smears with an increased proportion of band forms when processed with this screening procedure, or by routine visual examination of smears of all blood samples (i.e., the degree of sensitivity of flagging a possible abnormality by the automated analyzer).
In addition to other triggers, specimens were flagged if, in addition to other findings, the machine (Beckman Coulter LH 750) detected a total WBC < 3,000 cells/μL or > 30,000 cells/μL, or an absolute neutrophil count < 3,000 cells/μL or > 20,000 cells/μL. Blood from 101 consecutive patients with positive blood cultures was tested. The need for a quick visual examination of a smear was indicated for 87 of the patients, and complete visual examination confirmed bandemia in 42 of the 46 (91%) whose quick scan had indicated a need for full examination. The rapid exam was deemed to be negative and, therefore, not indicative of a need for a full examination in 41. Performance of a complete examination of these specimens found, however, that 20 had > 13% band forms on their peripheral blood smears. Thus, the rapid screen missed 20 of 41 (49%) patients with bandemia despite automated flagging. Complete smear examination of blood from the 14 patients whose specimens were not flagged found that 4 (29%) had > 13% band forms. Overall, using a conservative threshold for bandemia of > 13%, the standard process missed its detection in 24 of 101 (24%) of these selected patients.
The detection of bandemia is often of particular interest to the clinician evaluating patients who do not have leukocytosis, as was the case in 11 of these bacteremic patients with WBC 2,400-9,800 cells/ μL. Four of the 11 were not flagged, and three of these had elevated band counts. In the other seven, rapid scanning of the smear was felt to not indicate the need for a complete examination but, in fact, four had bandemia. Thus, the normal process failed to detect bandemia in seven of 11 bacteremic patients who did not have leukocytosis.
While pathologists seem to often disagree, clinicians commonly believe that measurement of the percentage of band neutrophils is often useful in patient management. While some published evidence suggests that knowledge of bandemia is not useful, others suggest that knowledge of this may provide useful clinical information. One setting in which this may be so is in emergency departments. A very recent evaluation of 289 bacteremic patients seen in an emergency department found that one-third had a normal temperature (36°C-38°C) and 52% had a normal WBC.1 Of the 210 patients who had a "full differential" performed, 172 (82%) had > 5% bands. The band count was elevated by this criterion in 79% of those with a normal total WBC and 80% with a normal temperature. Fifty-two patients had both normal temperature and WBC; 28 of these had a "full differential" and 21 (75%) had bandemia. Thus, knowledge of bandemia may be helpful as a clue to the presence of sepsis in patients in whom other suggestive findings are absent.
Careful examination of blood smears in order to determine the proportion of neutrophil bands is, unfortunately, time consuming and, therefore, adds significantly to the labor costs of a laboratory procedure that is otherwise totally automated. It would be desirable if automated systems could accurately detect band neutrophils.
The machine used by Greissal et al, the Coulter LH 750, examines 8,000 leukocytes per sample and determines cell volume for each cell type by measurement of direct current impedance, the internal cell composition by measurement of conductivity by radio frequency opacity, and cytoplasmic granularity by measuring light scatter with a laser. Neutrophils of septic patients have increased mean volume, as well as a greater distribution of volumes and decreased light scatter. Measurement of mean neutrophil volume and neutrophil volume distribution width has been reported to be more sensitive and specific as indicators of the presence of sepsis than manual band count, total neutrophil count, and CRP.2 That, however, cannot be true for the patients in whom bandemia is not detected.
Greissal et al have performed a careful analysis of the sensitivity of a standard laboratory algorithm, starting with an automated analyzer in the detection of bandemia and found it wanting in patients with bacteremia. This finding was perhaps of greatest importance to the clinician in patients without other common markers suggestive of sepsis, even with the use of a much higher threshold for bandemia (> 13%) than in the study by Seigel and colleagues (> 5%).
Thus, I believe that an accurate band count is of clinical value in the patient with suspected infection with a normal total WBC.3 Another frequently encountered circumstance in which I, correctly or incorrectly, utilize the band count is in patients receiving corticosteroid therapy, which routinely causes leukocytosis. The detection of bandemia may provide a mechanism for judging whether the elevated WBC observed in a patient receiving corticosteroids is due to the medication alone or whether it is a reflection, at least in part, of inflammation, including inflammation resulting from infection.4 The leukocytosis associated with prednisone administration consists of mature neutrophils, and one investigation found that the presence of > 6% band forms is suggestive of the presence of infection.4
Thus, while our pathology colleagues may not like it, I believe there is good reason to request full visual examination of blood smears to determine the presence of bandemia in selected circumstances.
1. Seigel TA, et al. Inadequacy of temperature and white blood cell count in predicting bacteremia in patients with suspected infection. J Emerg Med. 2010 Jul 29. [Epub ahead of print]
2. Bagdasaryan R, et al. Neutrophil VCS parameters are superior indicators for acute infection. Lab Hematol. 2007;13:12-16.
3. Wile MJ, et al. Manual differential cell counts help predict bacterial infection. A multivariate analysis. Am J Clin Pathol. 2001;115:644-649.
4. Schoenfeld Y, et al. Prednisone-induced leukocytosis. Influence of dosage, method and duration of administration on the degree of leukocytosis. Am J Med. 2001;71:773-778.