|Year : 2016 | Volume
| Issue : 2 | Page : 65-70
Reproducibility of hematological parameters: Manual versus automated method
Aliyu A Babadoko1, Ismaila N Ibrahim1, Abubakar U Musa2, Nasiru Usman1
1 Department of Haematology and Blood Transfusion, Ahmadu Bello University Teaching Hospital, Zaria, Nigeria
2 Department of Haematology and Blood Transfusion, Usmanu Danfodiyo University Teaching Hospital, Sokoto, Nigeria
|Date of Submission||04-Mar-2015|
|Date of Acceptance||30-May-2016|
|Date of Web Publication||21-Jun-2016|
Dr. Aliyu A Babadoko
Department of Haematology and Blood Transfusion, Ahmadu Bello University Teaching Hospital, Zaria
Background: The automated hematology analyzer has replaced the traditional manual assay methods and eye count for determination of hematological parameters. Recently, they are widely used in laboratories and hospitals for counting of blood cells necessary for diagnosis and monitoring treatment of various disorders. Objective: To compare the blood count results of automated hematology analyzer with the traditional manual method in the determination of some hematological parameters. Methodology: A total of 100 samples were randomly selected from samples received at the reception of hematology laboratory in June 2014. The anticoagulated blood was collected and subjected to analysis of hematological parameters by automated hematology analyzer (Swelab Alfa) and standard manual methods. Results: The mean hematocrit, total white cell and platelets count, neutrophils, and lymphocytes percentages by manual method were 37.5 ± 7.2%, 7.2 ± 3.7 × 10 9 /L, 244.8 ± 171.8 × 10 9 /L, 53.8 ± 16.0%, and 41.8 ± 28.2%, respectively while that by automation were 37.2 ± 7.3%, 7.9 ± 6.1 × 10 9 /L, 278.1 ± 162.0 × 10 9 /L, 52.6 ± 16.0%, and 41.0 ± 14.3%, respectively. Whereas the mean platelets count was significantly (P < 0.05) higher in the automated method, there was no significant statistical difference between the mean hematocrit, total white cell and platelets count, neutrophils, and lymphocytes percentages of all the study samples (P > 0.05) and this remained so in male gender. The Pearson correlation test showed a positive significant (P < 0.05) correlation between both methods even after gender stratification. Conclusion: Automated analyzers can be used in all laboratories to provide quick and accurate results for patient care. However, it should be accompanied by microscopic blood film examination to provide confirmatory and additional useful diagnostic information.
Keywords: Blood film examination, hematology analyzers, manual methods
|How to cite this article:|
Babadoko AA, Ibrahim IN, Musa AU, Usman N. Reproducibility of hematological parameters: Manual versus automated method. Sub-Saharan Afr J Med 2016;3:65-70
|How to cite this URL:|
Babadoko AA, Ibrahim IN, Musa AU, Usman N. Reproducibility of hematological parameters: Manual versus automated method. Sub-Saharan Afr J Med [serial online] 2016 [cited 2021 Jun 16];3:65-70. Available from: https://www.ssajm.org/text.asp?2016/3/2/65/184352
| Introduction|| |
Diseases are characterized by changes in individual blood parameters which are more or less typical for the underlying disease. Therefore, measurement of these parameters at any time with high precision, reproducibility, and high accuracy allows a precise diagnosis.
Quantitative determination of the individual cellular components was first enabled in 1852 by the work of Karl Vierordt (1818-1884), a physiologist.  In 1924, Neubauer published his net structure, and this led to the manual cell counts, which is considered as "gold standard" in most developing countries.  The precision and accuracy were highly dependent on the number of counted cells and at a reasonable level of effort, were subjected to fluctuations of up to 10%. 
The first step toward automation was made in 1934 with the Moldavan capillary method.  However, the actual breakthrough in the development of hematological instruments suitable for routine work was achieved by Wallace Coulter in 1956 with his patent "High-speed automatic blood cell counter."  Hematology analyzer is an automatic instrument programed to give an idea about the number of the blood cells through aspiration of a blood sample flow through an electric field.  This method has proved its value when used clinically in hospitals instead of the traditional manual method that depends on the visual counts of the blood cells, which takes time and effort. 
The reasons for improving blood cells counting, particularly in severely anemic, leukopenia and thrombocytopenic patients, stem from current hematooncology practice in relation to bone marrow failure syndromes (myelodysplastic syndrome, aplastic anemia, and paroxysmal nocturnal hemoglobinuria), leukemias, lymphomas, chemotherapy induction, postchemotherapy bleeding, prophylactic antibiotic administration, platelet transfusions, parasitic infestations, viral (HIV), and fungal infections. In daily practice, an increasing request for blood cell counts particularly platelet count in Ahmadu Bello University Teaching Hospital (ABUTH), Zaria was observed in recent time.
In our hospital, automated analyzers are recently being introduced for routine work, some results commonly hematocrit, leukocyte differential counts, and platelets counts were not accepted by physicians and request the repeat of these counts manually once the result does not correlate with the clinical condition of the patient. Thus, the decision to conduct this study to obtain objective yardstick necessary to give a conclusive advice to the clinicians that will help in patient evaluation, treatment, and monitoring.
To compare the results of some hematological parameters obtained by automated counts with that obtained by manual methods using the same sample at the same time.
| Methodology|| |
This is a cross-sectional study, involving randomly selected blood samples of patients attending ABUTH in Zaria, in the month of June through August 2014. A total number of 100 anticoagulated blood samples (nonvacuum bottles) of patients for complete blood counts at hematology department reception with various presumptive diagnoses were selected during the study period. Selection was by the simple random sampling, blood samples were taken each day by the odd numbers, i.e., samples number 1, 3, 5, etc. Data of age and sex of patients were taken from the laboratory request forms. Collection of blood sample was based on the two-step sample appraisal procedure operated by the department.  Samples were processed manually (microhematocrit estimation and hemocytometer cell counting) examined (thin blood film prepared, stained with Leishman's stains, examined under the light microscope) by Medical Laboratory Hematology Scientists in the department using standard methods adopted from Dacie and Lewis Practical Haematology.  Thereafter, the same sample was processed by the automated analyzer (Swelab Alpha, Sweden) immediately. The procedure of the two methods was subsequently reviewed by the Hematologist. Although the department does not participate in any proficiency testing programs, an internal quality control measures is conducted weekly using three commercially prepared control samples for low, normal, and high counts. Data processing was performed by the Statistical Program (SPSS 20, Statistical Package for Social sciences IBM Corp. Released 2011 version 20. Armonk, NY). Data were tested by the Student's t-test for two means and the hypothesis test for two proportions. Coefficient of correlation (r) was determined between the two methods used. All tests were applied at a level of significance (α =0.05). P ≤ 0.05 was considered as statistically significant.
| Results|| |
Of a total number of 100 samples randomly selected 56% were males while 44% were females. The mean age of the male was 30.70 ± 11.48 years and that of the females was 33.14 ± 14.06 years. There was no statistically significant difference in age between the gender P > 0.05. The hematological parameters of all the study samples and the correlations are shown in [Table 1] and [Figure 1]a-e respectively while those of the males is shown in [Table 2] and [Figure 2]a-e and the females in [Table 3] and [Figure 3]a-e.
|Figure 1: (a) Correlation of hematocrit of all the study samples (r .736, P < 0.001), (b) correlation of total white blood cell count × 109/L, in all the study samples (r .432, P < 0.001), (c) correlation of Platelet count × 109/l, in all the study samples (r - 0.531, P < 0.001), (d) correlation of neutrophil differential count in all the study samples (r .790, P < 0.001), (e) correlation of lymphocyte differential count in all the study samples (r .429, P < 0.001)|
Click here to view
|Figure 2: (a) Correlation of hematocrit of in male samples (r .643, P < 0.001), (b) correlation of total white blood cell count × 109/L, in male samples (r .607, P < 0.001), (c) correlation of platelet count × 109/L, in male samples (r - 0.471, P < 0.001), (d) correlation of neutrophil differential count in male samples (r .0.733, P < 0.001), (e) correlation of lymphocyte differential count in male samples (r .874, P < 0.001)|
Click here to view
|Figure 3: (a) Correlation of hematocrit in female samples (r .841, P < 0.001), (b) correlation of total white blood cell count × 109/L, in female samples (r .386, P = 0.01), (c) correlation of platelet count × 109/L, in female samples (r .569, P < 0.001), (d) correlation of neutrophil differential count in female samples (r .859, P < 0.001), (e) correlation of Lymphocyte differential count in female participants (r .329, P = 0.029)|
Click here to view
| Discussion|| |
The manual method (manual phase-contrast microscopy), although has significant limitations in terms of performance, particularly in the area of imprecision, up to date it remains the only "Gold Standard" in cell counting available to assess any degree of accuracy of the automated count. The International Council for Standardization in Haematology recently recommended a new immunologically-based reference method.  This immunological method is not available in our laboratories and so we still rely on the manual method to confirm our results once the accuracy of such method is in doubt.
The studied samples represented patients who are complaining of various diseases, which may or may not affect the cell count. In this study, the mean hematocrit and all the cell counts estimated by the manual and the automated methods for all the studied samples (n = 100) did not show significant statistical difference [Table 1], except the mean platelets count which was significantly different [P = 0.043, [Table 1]. This remained so even when the sample was separated according to the gender of the patients, [Table 3] and [Table 4], except the platelets count that was significantly different in the male gender [P = 0.01, [Table 3]. Although this is similar to the finding in previous studies , but differs in that gender difference in the mean platelets counts between the manual and automated counts was not observed. These findings may be attributed to the presence of high number of samples with normal count, 73 samples were normal by both methods [Table 2]. This large number dominates in the correlation of all sample and yields positive correlation. Meanwhile, samples of low count (6 samples) or those of high count (12 samples) may not be enough to show such positive correlation, and so we cannot rely on either method alone once there is suspicion of low or high count and the blood film examination of platelet estimate is mandatory in such situations. This is similar to the findings by Bakhubaira  and Charie et al.  in the evaluation of 88 specimens with platelet counts of ≤20 × 10 9 /L tested in three studies, the automated analyzer counts matched the platelet reference count method in 87 cases, representing an agreement of 98.9%.  That is in low platelet counts there should be another reference method which is reliable like the method used by Charie et al. (immunologically-based method). 
When all the samples were analyzed by the Pearson correlation test, we observed significant positive (P < 0.05) correlation between the result of both methods [Figure 1]a-e and this remains so despite gender stratification [Figure 2]a-e and [Figure 3]a-e. This means that both methods can be used for cell count without producing a significant difference in results except for the platelet count which should be used with caution and preferably platelet estimate conducted on blood film which is an arbitrary method of assurance in the absence of new immunological methods. Our finding is in line with previous reports by Atilola and Kamentsky and McCarthy et al. ,
The automated leukocyte differential count (precisely neutrophil and lymphocyte percentage distributions) was also not significantly different with the manual count [P > 0.05, [Table 1] and demonstrated a significant positive correlation [Figure 1]d and e and these remain so even after gender stratification [Table 3], [Table 4] and [Figure 2]d and e, [Figure 3]d and e. Although earlier reports by Lewis and Bentley in 1977 and previous report by Takubo and Tatsumi documented the inability of the automated machine to identify or differentiate the leukocytes, more especially the immature cells, which was attributed to poor differentiation of segmented neutrophils and band neutrophils. In this study, we did not identify such discrepancies; this may be due to a large proportion of our samples being within normal limits. ,,
We advise physicians to rely on the hematology analyzer results, however, medical decisions should not be delayed awaiting for repeat or confirmation as Steele et al.  reported in their study, where they found that analyzer results can be used for medical decision making without long delays for repeat and confirmatory testing.
In our laboratory, experienced laboratory scientists and technicians are responsible for the daily work in addition to a periodic quality control measures applied and hematologists consultation is only sought for when there is a low or very high counts for confirmation. However, once there is an additional confirmatory request particularly of platelets count, our physicians should rely on the results obtained by the manual counts and blood film estimate by the hematologist.
| Conclusion|| |
There is a significant positive correlation between the manual and the automated methods. Hematology auto-analyzers can be used routinely in our laboratories to provide quick and accurate results, however, blood film examination and platelet estimate is recommended for every sample of platelet count, especially when the count is lower or higher than normal by both methods.
We acknowledge the efforts of all laboratory scientists and technicians in the Department of Haematology and Blood Transfusion in the ABUTH, Zaria for reception, collection, processing, and sample analysis.
Financial Support and Sponsorship
Conflicts of Interest
There are no conflicts of interest.
| References|| |
Graham M. Principles of automated blood cell counters. In: David B, John C, editors. The Science of Laboratory. 2 nd
ed. England: John Wiley and Sons Ltd.; 2005. p. 289-96.
Joachim L, Burkhard G, Uwe C. Automation in haematology. Transfus Med Hemother 2007;34:328-39.
Soulaf JK. The evaluation of traditional and automatic coulter method in estimation of haematological parameters in adult rats. Beni Suef Univ Basic Appl Sci 2013;1:31-5.
Mamman AI, Muktar HM, Aminu SM, Babadoko AA, Suleiman AI, Hassan A, et al
. Laboratory User Handbook for Department of Hematology. 1 st
ed. Zaria: Nigeria. ABU Press ltd; 2015. p. 4-5.
Barbara JB, Imelda B. Basic haematologic techniques. In: Lewis SM, Bain BJ, Bates I, editors. Dacie and Lewis Practical Haematology. 11 th
ed. London: Churchill Livingston; 2001. p. 19-46.
International Council for Standardization in Haematology Expert Panel on Cytometry; International Society of Laboratory Hematology Task Force on Platelet Counting. Platelet counting by the RBC/platelet ratio method. A reference method. Am J Clin Pathol 2001;115:460-4.
Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307-10.
O′Connell SM, Impeduglia T, Hessler K, Wang XJ, Carroll RJ, Dardik H. Autologous platelet-rich fibrin matrix as cell therapy in the healing of chronic lower-extremity ulcers. Wound Repair Regen 2008;16:749-56.
Bakhubaira S. Automated versus manual platelet count in Aden. J Clin Exp Pathol 2013;3:149.
Charie LA, Harrison P, Smith CU, Cobb JR, Briggs C, Machin S. Accuracy in the low platelet count range: A comparison of automated platelet counts on Beckman coulter high-volume hematology analyzers with ISLH/ICSH platelet reference method. Lab Hematol 2001;7:236-44.
Atilola LR, Kamentsky LA. Routine differential leucocyte count. Clin Lab Med 1996;15:289-91.
McCarthy JM, Capullari T, Spellacy WN. The correlation between automated hematology and manually read smears for the determination of nucleated red blood cells in umbilical cord blood. J Matern Fetal Neonatal Med 2005;17:199-201.
Lewis SM, Bentley SA. Haemocytometry by laser-beam optics: Evaluation of the Hemac 630L. J Clin Pathol 1977;30:54-64.
Takubo T, Tatsumi N. Quality control in a manual and an automated leukocyte differential count. Southeast Asian J Trop Med Public Health 1999;30 Suppl 3:66-74.
Ike SO, Nubila T, Ukaejiofo EO, Nubila IN, Shu EN, Ezema I. Comparison of haematological parameters determined by the Sysmex KX-2IN automated haematology analyzer and the manual counts. BMC Clin Pathol 2010;10:3.
Steele BW, Wu NC, Whitcomb C. White blood cell and platelet counting performance by hematology analyzers: A critical evaluation. Lab Hematol 2001;7:255-66.
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3], [Table 4]