Acknowledgements: we would like to thank Jacques Dugailliez for the critical review of the manuscript.
Haematologica 2002; 87:(10)ELT40
[Medline] [prev] [index] [next]Underestimation of haemoglobin concentration in blood samples with high hyperleukocytosis: case report and alternative method for determination on blood cells automated analyzers
Anne Fohlen-Walter, Jean-Jacques Goupil, Thomas Lecompte, Jean-François Lesesve
Service d'Hématologie Biologique, CHU de Nancy, 54511 Vandoeuvre, France
Correspondence: Mrs Anne WALTER, Service d'Hématologie Biologique, CHU Nancy-Brabois, 54511 Vandoeuvre-Les-Nancy Cedex, France. Tel : 03 83 15 37 66. Fax : 03 83 15 37 89. Email: a.walter@chu-nancy.fr
Complete blood cells counts (CBC) are now performed on automated instruments. Some CBC may be inaccurate due to unusual specimen characteristics or instrument malfunction/ interference. Erroneous measurements of haemoglobin (Hb) concentration results most often from abnormal turbidity in case of high white blood cell (WBC) count or lipid in the plasma.1 In these case, attending data is a falsely high Hb estimation. Severe hyperleukocytosis may lead to an overestimation of Hb when measured by spectrophotometry, consequent on leukocyte-induced turbidity in the hemolysate.2,3 As far as we know, we report here for the first time 3 cases of unexpected low Hb in association with elevated WBC count.
Diagnosis were: T-cell prolymphocytic leukaemia (patient #1), chronic myelogenous leukaemia (patient #2), and acute monocytic leukaemia (patient #3). The blood cells counts were performed using a Technicon H*2 automated counter (Bayer, Tarrytown, NY). Relevant haematologic findings are summarized in Table 1. WBC counts were obtained after diluting the blood in order that measurements fall within the predetermined limits. The Hb measurement was associated with a turbidity alarm in both cases, plus a flatness alarm (showing flow abnormality) in cases #1 and #3. The discrepancy between the calculated and the measured mean cell haemoglobin concentration (respectively MCHC and CHCM), did not allow to validate Hb (alarm "Count Excess"). According to the H*2 technology, the MCHC was computed as follows: Hb/haematocrit, whereas the CHCM, like the mean cell volume (MCV), directly derived from the amount of light scattered at 2 angles (5 to 15° and 2 to 3°) by individual isovolumetrically sphered red cells flowing through a narrow sensing aperture. We corrected Hb (cHb) from CHCM, MCV and the corrected RBC count (cRBC = automated RBC count &endash; automated WBC count) as follows: cHb = CHCM x cRBC x MCV. This way of correction was justified by the following points: 1) The RBC count included the WBC count. This negligible error in normal subjects (approximately 0.1%) might become significant in the presence of markedly high leukocytosis and lead to spuriously elevated RBC count and haematocrit. 2) The CHCM and the MCV were determined only from the RBC population, i.e. cells that had a volume and a refractive index compatible with a RBC as measured by the scattered-light (RBC map). They are not affected neither by the medium turbidity neither by the elevated WBC count. Nevertheless, as a precaution, the assessment of the MCV and the CHCM should be performed with a 1:5 or 1:10 dilution to eliminate the possible interference of high WBC count (particulary monocytic and granulocytic series) on the RBC coincidence count, causing insufficient trimming of the RBC map and consequently an overestimation of the MCV and the CHCM (Table 1). The falsely low Hb could result from an unstable colour reaction over the time as depicted by the graphs of the light transmitance at 546 nm after the lysis of RBC and the cyanmethemoglobin reaction (Figure 1). The large amount of leukocytes and proteins in the reaction cuvette could have induced optical aberration such as bubbles. These 3 cases showed that, despite the "turbidity" alarm, Hb could be underestimated in case of high hyperleukocytosis probably depending on impaired colour reaction.
Several techniques have been described to remove the interference of WBC in Hb measurement, such as filtration of the blood to get leucocyte-poor blood,3 determination of Hb in the supernatant after centrifugation of the hemolysate.4 The method that we propose, using CHCM, MCV and the substraction of the number of leukocytes from the number of erythrocytes measured by the analyzer allows an accurate determination of Hb. This very simple and quick method fits for cases of underestimation as well as for overestimation of Hb. It is also convenient for new generation analyzers (Advia serie).References
Acknowledgements: we would like to thank Jacques Dugailliez for the critical review of the manuscript.