[Evaluation of the Coulter MAXM. Differential leukocyte count and left-shift alarm]

To assess the reliability of the differential leucocyte count (DLC) and the left shift flagging (LSF) system provided by the Coulter MAXM (MAXM) haematology analyzer.

380 blood specimens (drawn with tri-K EDTA as anticoagulant) were studied.

By using the reference method (NCCLS H20-A), 50 out of the 380 blood specimens presented abnormal DLC (bands > 6%). Of from these, in 39 (80%) the MAXM displayed LSF of "bands 1 or 2". In 118 left shift flagged specimens (MAXM) with normal manual DLC, 87 (74%) had the "bands 1" alarm and 31 (26%) the "bands 2" alarm. Accordingly if the LSF "bands 1" is overlooked, the percentage of FP decreases from 36% to 10% but the percentage of false negatives (FN) increases from 22% to 58%. In order to improve the appreciation of LSF by decreasing the need of manual revisions, the visual examination of the leucocyte distribution scattergram (LDS), also provided by the MAXM, was conveniently evaluated. This study was performed on 190 blood specimens from which the MAXM displayed a normal DLC in 122 (64%), the LSF of "bands" in 44 (23%) and the LSF of "bands 2" in 24 (12.6%). Of from the 122 specimens with normal DLC, four were FN, of from the 44 specimens with "bands 1" LSF, 37 were FP and of from the 24 specimens with "bands 2" LSF, 16 were FP. The visual appreciation of the LDS showed in the majority of samples with "bands 1" and "bands 2" a definitely different shape consisting in a sharper image up to the top of the picture when compared to samples with normal DLC (without flags). According to this criteria, all the 122 specimens with normal DLC displayed a normal LDS and all the 24 specimens with "bands 2" flag displayed abnormal LDS. Of from the 44 specimens with "bands 1" flag, 26 (59%) showed an abnormal LDS and 18 (41%) a normal LDS. It is noteworthy that of from the 26 specimens with abnormal LDS only 7 were true positive (TP), whereas the 18 specimens with normal LDS all showed a normal DLC according to the reference method. These data allow us to conclude that manual revision was required in 26 out of 68 specimens with "bands 1" and abnormal LDS (13% of the total) and in all the 24 specimens with "bands 2" flag. Therefore by using the information provided by the LDS the need of manual revision decreases to 73% of the total sample with LSF.

Our results give further support to the idea that th VCS method used by the Coulter MAXM provides a high quality DLC with specific left shift detection.