| Abstract|| |
Since the advent of automation in the field of hematological cell counters there has been a constant refinement of the technology and increase in the number of newer parameters available on CBC analysers. Many novel parameters are being put into routine clinical use and both clinical evaluation and monitoring critically depend on knowledge of laboratory reference ranges. Here, we present reference interval for the Sysmex XE-2100, with emphasis on the novel or newer research parameters. Blood samples from a total of 122 clinically asymptomatic and apparently healthy subjects were evaluated and a final of 100 subjects (54-M, 46-F) were included in the study. A broad spectrum of parameters available with the analyser was assessed and reference ranges for the same evaluated.
Keywords: Complete blood count, reference ranges, research parameters, reticulocyte, Sysmex XE-2100
|How to cite this article:|
Sehgal KK, Tina D, Choksey U, Dalal RJ, Shanaz KJ. Reference range evaluation of complete blood count parameters with emphasis on newer research parameters on the complete blood count analyzer Sysmex XE-2100. Indian J Pathol Microbiol 2013;56:120-4
|How to cite this URL:|
Sehgal KK, Tina D, Choksey U, Dalal RJ, Shanaz KJ. Reference range evaluation of complete blood count parameters with emphasis on newer research parameters on the complete blood count analyzer Sysmex XE-2100. Indian J Pathol Microbiol [serial online] 2013 [cited 2019 Jan 23];56:120-4. Available from: http://www.ijpmonline.org/text.asp?2013/56/2/120/118698
| Introduction|| |
Since the advent of automation in the field of hematological cell counters, there has been a constant refinement of the technology and the number of different parameters available on complete blood count (CBC) analyzers. Most new parameters get introduced first as research parameters and after various validation studies and approvals they are put in routine practice.
A measured or observed laboratory test result from a person (usually a patient) is compared with a reference interval for the purpose of making a medical diagnosis, therapeutic management decision or other physiological assessment. The interpretation of clinical laboratory data is, therefore, a comparative decision-making process. For this decision-making process to occur, reference values are needed for all tests in the clinical laboratory. 
The Sysmex XE-2100 is a CBC analyzer, which uses laser light to measure on a cell-by-cell basis the side scatter, forward scatter and side fluorescence light. In addition, impedance based counts are performed for red blood cell (RBC) and platelet counts. Further counts and measurements use radio frequency and direct current methods and hemoglobin colorimetric method. In addition, by using specialized reagents for nucleated red blood cells (nRBC) mode, immature myeloid information (IMI) mode and reticulocyte mode, the XE-2100 is able to quantify a number of novel parameters many of which are food and drug administration approved The novel parameters are as follows-immature reticulocyte fraction (IRF), low fluorescence reticulocytes, medium fluorescence reticulocytes, high fluorescence reticulocytes, immature platelet fraction (IPF), reticulocyte hemoglobin (Hb)-Ret He and red blood cell hemoglobin RBC He, nRBC, immature granulocyte fractions (Ig%) and hemopoietic progenitor cell% (HPC%). ,,,
The novel parameters reported on the Sysmex XE-2100 have been found to have significant clinical potential. Enumeration of nRBC is not only important in identifying disease, but also because their presence can have a direct effect on the accuracy of the white blood cell count.  IRF has been found to be extremely useful in monitoring of bone marrow regeneration post-transplant or chemotherapy and for classification of anemia and monitoring of treatment.  IPF has been used for differential diagnosis of thrombocytopenia and for prediction of platelet recovery post-transplant or chemotherapy.  These studies have brought such novel parameters into routine clinical use and hence deriving reference ranges for these parameters are the first and essential step before the patient results are shared with clinicians.
It was the purpose of this study to determine the exact reference intervals for the CBC parameters that are measured by the XE-2100 analyzer, with emphasis on novel/research parameters enumerated above.
| Materials and Methods|| |
All patients coming into the out-patient department of our tertiary care hospital for routine evaluation under the health check category were included for the study. All patients selected were clinically asymptomatic and apparently healthy. The following blood examinations were performed for all study samples: CBC, reticulocyte count, erythrocyte sedimentation rate, iron studies, vitamin B12 levels and folic acid levels.
Only patients who showed a completely normal CBC with no flags and normal values for all above tests were further evaluated. A total of 122 samples were evaluated and a final of 100 patients (54-M, 46-F) were included in the study after excluding patients with high erythrocyte sedimentation rate levels (n = 6), iron deficiency (n = 10), vitamin B12/ folate-deficiency (n = 6).
The Sysmex XE-2100 CBC analyzer was calibrated as per manufacturer guidelines and daily controls were used as per College of American Pathology (CAP) guidelines. The coefficient of variation were calculated and observed for each parameters evaluated in the study.
SPSS V20.0 Statistical Software was used for the statistical analysis. While evaluating reference range data it is essential to understand, which statistical method to use and why. The philosophy behind the various statistical analyses used in the study is explained in [Figure 1].
All parameters in this study were evaluated for their distribution to check if the population is normally distributed or skewed as shown in [Figure 1]. This was performed subjectively by looking at histograms and objectively by the Shapiro-Wilk test. When the population was normally distributed mean ± 2 standard deviation (SD) was used as representative of its range; however if the population data was skewed median, interquartile range (IQR) and 2.5-97.5 percentile were used as representative of the range. 
After evaluation of the mean/median and ranges for various parameters, the next logical step was to evaluate the difference if any between males and females for all parameters. As explained in [Figure 1], as the reference values of many analytes do not follow the Gaussian form, the non-parametric Mann-Whitney-U method was used in this study. 
| Results|| |
A total of 100 healthy volunteers (56-males, 44-females) were evaluated for calculating the normal ranges for various CBC parameters. They were all adult patients with an age range of 22-72 years.
As shown in [Table 1] below, the normality of distribution was studied for all parameters by using the Shapiro-Wilk test and a significant P value of less than 0.05 was applied. Wherever the P value was below the significant level [Table 1] that parameter was considered to have a skewed distribution. For normally distributed parameters population data were described using mean, SD and range as mean ± 2SD, whereas for skewed distribution parameters population was defined by median, IQR and range as 2.5-97.5 percentile. As shown in [Table 2], the non-parametric Mann-Whitney U test was applied to all parameters to check for any significant difference between male and female subgroups. The null hypothesis was that the distribution of data is same for males and females. If P value was found to be less than 0.05 the null hypothesis was rejected indicating a significant difference between males and females. The separate ranges for male and female subgroups for the relevant parameters are shown in [Table 2].
|Table 2: Ranges for parameters with difference in male and female subpopulations|
Click here to view
The value of certain parameters such as nRBC, Ig% and HPC % was zero, which is expected in the normal population and hence is not detailed further in [Table 1] and [Table 2].
| Discussion|| |
Various studies for biological reference ranges have been published including those on Sysmex-X class analyzers; , however, most of them have focused on routine CBC parameters. This study is the first of its kind from India focusing on evaluating reference ranges for research/novel CBC parameters.
Number of study samples
Clinical and Laboratory Standards Institute (CLSI) recommends a minimum of 120 observations for each parameter. This has the advantage of allowing 90% confidence limits to be computed non-parametrically for each reference limit.  Moreover, if separate intervals were needed for different subclasses (by sex or age-class, for example), each such interval should be based on the recommended number (at least 120) of reference observations. In the present study, it was difficult to obtain appropriate age-related reference subjects in sufficient numbers. However, CLSI recommends that whatever number of values are obtained, the data should still be analyzed by the non-parametric method and reported by percentiles appropriate to the number of values obtained. 
Using the non-parametric method, it is impossible to distinguish between two percentiles of a distribution that are P 0% apart unless at least n = (100/ P )–1 observations have been obtained.  The reason for this is that the non-parametric method is based solely on the ranks of the observations (in order of magnitude) and ignores their measured values. For example, if a sample of nine observations is taken at random from some population, only nine estimates of percentiles can be obtained from the nine rankings when these have been ranked in order of magnitude. The smallest observation is the nonparametric estimate of the 10 th percentile of the population; the largest observation is the non-parametric estimate of the 90 th percentile of the population. Thus as the formula says, a sample of nine observations (9 = (100/P )–1, where P = 10.0) represents the minimum sample size necessary to obtain distinct nonparametric estimates of the ordered population deciles, which are, by definition, percentiles of the population exactly 10% apart from each other.  Similarly, to estimate the 2.5 th percentile distinct from the 5 th percentile or the 95 th percentile distinct from the 97.5 th (i.e., P = 2.5), a minimum of 39 measurements are required. The smallest observation in the sample would be the non-parametric estimate of the 2.5 th percentile of the population while the largest observation would estimate the 97.5 th percentile.  In the present study, a final total of 100 samples were evaluated and hence using the above formula it was possible to distinguish population subsets, which are one percentile apart.
The ranges of various parameters in published literature established by other workers show mild variation from each other. This may be due to pre-analytical variables or more importantly the different populations selected by different institutes for such studies. For instance, some calculate the reference intervals from a large number of patients who visit the clinics. Others use a double selection by using for instance, only samples from the eye disease clinics, reasoning that few patients with hematological diseases will be found in this category. Others use samples of blood bank donors or volunteers from the hospital or laboratory staff while some refine this strategy by paying attention to the age of the volunteer. In short, these variables create different reference intervals; although, the differences are sometimes quite small. 
The reference intervals we report here for the various parameters match quite well those found by others [Table 3]. ,,,,, Minor variations of reference ranges were seen in few parameters in our study as compared with published literature. Such probably population-based shifts stress the importance of determining reference limits in every laboratory , or verifying the applicability of a published reference interval for a given hematology analyzer before its use as suggested by CLSI. 
|Table 3: Comparison of reference ranges of novel parameters vis a vis published literature|
Click here to view
We confirmed the statistically significant differences of reference intervals for men versus women for Hb, RBC, hematocrit (HCT), mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration and platelet counts as reported, in other studies. ,, However, we observed such differences for Ret He, RBC He, RBC Y and Ret Y amongst the research/novel parameters. Nonetheless, the differences in some parameters like platelet counts, while statistically significant, were not clinically relevant.
While we have established reference intervals for PLT-I- impedance platelet count, PLT-O- fluorescence-optical count, it should be noted that the analyzer uses a switching algorithm deciding, which count is more reliable (PLT-I or PLT-O) on a case-by-case basis and reports the more reliable one as PLT. The accuracy of the XE-2100 platelet counting in thrombocytopenic samples is best when the switching algorithm is used (unpublished data) and overriding of the algorithm is not a good practice. ,
As reported by Pekelharing et al., for the use of the reference interval for reticulocyte production index (RPI), some limitations do apply. Whereas, it is usually assumed that a RPI of "1" is normal, our data also shows that the reference interval can actually reach as low as 0.3. This suggests that the assumptions made in calculating RPI have a rather monocausal regulation of reticulocyte maturation time in the bone marrow tied directly to the HCT. It needs to be re-emphasized that RPI should only be used for adult anemic patients. While an RPI higher than 2 can indeed suggest a significantly increased hematopoiesis, our data also suggests that the clinical value of a low RPI should not be overestimated. Erythropoietic bone marrow activity can alternatively be assessed by the IRF and the RetHe content, in addition to, of course, the reticulocyte counts itself.
The data obtained in this study definitely matches with that of western literature and other published articles, but there is not any significant Indian published data on the novel parameters. Though it is believed by some that range of Hb, RBC and such parameters may be lower in Indian population the same has not been concretely published. A major thing that was kept in mind during this study was to exclude patients with latent iron or vitamin B12 deficiency. As nutritional deficiency anemia are quite common in India, unless one rules out the latent deficiency patients we might be recognizing borderline low values of Hb, RBC, mean corpuscular volume, etc., as a normal distribution. There was a restriction in doing a larger number of samples in view of the costs involved for doing iron studies, vitamin B12 and folate levels for all study patients. As the normal ranges of routine parameters matches to that in literature, we strongly believe that the novel parameter ranges are also representative of our population. The normal ranges derived in this study do need to be validated across larger sample numbers and in a multicentric fashion for determining a thorough countrywide normal reference range.
Automated blood cell counters are becoming more sophisticated and the range of reportable parameters available is ever increasing. Both laboratory scientists and clinicians need to keep up to date with new parameters and methods in hematology. Often it is the laboratory scientists that need to introduce the new parameters and their clinical utility to the medical staff. Standardization of the tests and establishing biological reference ranges is the first step toward good laboratory practice that ensures reliable results are reported to the clinician. The availability of novel parameters such as nRBC, Ig%, IPF, Ret He as part of external quality assurance program agencies such as CAP and UK NEQAS is a welcome prospect and a step in the right direction.
| Acknowledgments|| |
We would like to acknowledge National Health & Education Society for funding the project and Hinduja Hospital, where the entire work has been conducted. We would like to acknowledge the support of Ms. Monisha Sethi for coordinating with the data collection and data analysis during the project.
| References|| |
|1.||Edward AS, Basil TD, Miller WG, D'Orazio P, Eckfeldt JH, Evans SA, et al. How to Define and Determine Reference Intervals in the Clinical Laboratory; Approved Guideline. 2 nd ed. Vol. 20. Wayne, Pennsylvania: NCCLS; C28-A2; 2000. p. 1-59. |
|2.||Briggs C. Quality counts: New parameters in blood cell counting. Int J Lab Hematol 2009;31:277-97. |
|3.||Briggs C, Harrison P, Machin SJ. Continuing developments with the automated platelet count. Int J Lab Hematol 2007;29:77-91. |
|4.||Pfaeffli J. Reference limits for the automated haematology analyser sysmex XE-2100. Sysmex J Int 2002;12:18-23. |
|5.||Pekelharing JM, Hauss O, de Jonge R, Lokhoff J, Sodikromo J, Spaans M, et al. Haematology reference intervals for established and novel parameters in healthy adults. Sysmex J Int 2010;1:1-9. |
|6.||Wakeman L, Al-Ismail S, Benton A, Beddall A, Gibbs A, Hartnell S, et al. Robust, routine haematology reference ranges for healthy adults. Int J Lab Hematol 2007;29:279-83. |
|7.||Briggs C, Kunka S, Machin SJ. The most accurate platelet count on the Sysmex XE-2100. Optical or impedance? Clin Lab Haematol 2004;26:157-8. |
|8.||Haeckel R, Wosniok W, Arzideh F. A plea for intra-laboratory reference limits. Part 1. General considerations and concepts for determination. Clin Chem Lab Med 2007;45:1033-42. |
|9.||Briggs C, Kunka S, Hart D, Oguni S, Machin SJ. Assessment of an immature platelet fraction (IPF) in peripheral thrombocytopenia. Br J Haematol 2004;126:93-9. |
|10.||Cho YG, Lee JH, Kim DS, Lee HS, Choi SI. Clinical usefulness of the simple technique to diagnose thrombocytopenia using immature platelet fraction. Korean J Lab Med 2007;27:1-6. |
|11.||d'Onofrio G, Kuse R, Foures C, Jou JM, Pradella M, Zini G. Reticulocytes in haematological disorders. Clin Lab Haematol 1996;18 Suppl 1:29-34. |
|12.||Brugnara C, Schiller B, Moran J. Reticulocyte hemoglobin equivalent (Ret He) and assessment of iron-deficient states. Clin Lab Haematol 2006;28:303-8. |
Kunal K Sehgal
Department of Hematology, P.D. Hinduja Hospital & Medical Research Centre, Veer Savarkar Marg, Mahim, Mumbai - 400 016, Maharashtra
Source of Support: National Health & Education Society (work
done at P.D.Hinduja Hospital & Medical Research Centre, Mumbai), Conflict of Interest: None
[Table 1], [Table 2], [Table 3]