Abstract
Reticulocyte count is a basic test in hematology. This study was done to compare manual and automated methods and to study the effect of sample storage on reticulocyte count. Analyses of samples (n = 86) were done at 2, 6, 24 and 48 h after blood collection. Manual counting was done from both freshly prepared slide and stored slide by microscopy on new methylene blue stained smears. Automated enumeration was on Sysmex XT-2000i analyser (Ret search II). The values of immature reticulocyte fraction (IRF) and low fluorescence reticulocytes (LFR) were also recorded. Comparison between two methods was done by Spearman’s correlation and Mann–Whitney test. Effect of storage was analysed by repeated measures ANOVA. There was strong positive correlation between both manual and automated methods at 2, 6, 24 and 48 h. The differences between the manual and automated methods were not significant at 2, 6 and 24 h (p 0.975, 0.967 and 0.227). The difference between the freshly prepared slide and stored slide were significant at 6, 24 and 48 h (p 0.015, 0.004 and 0.001). The change in reticulocyte count with time, decrease in IRF and increase in LFR were not significant up to 6 h but were significant at 24 and 48 h after blood collection. Both the methods were accurate and correlated well with each other. Freshly prepared smears for manual counting were better than counting on stored slide. Up to 6 h after blood collection results obtained by both methods are acceptable.
Keywords: Automated reticulocyte enumeration, Immature reticulocyte fraction, Manual method, Reticulocyte count, Storage effect
Introduction
Reticulocyte counting is one of the basic tests in hematology for evaluating bone marrow erythropoietic activity. Laboratory methods for detection of reticulocytes are based on the presence of ribonucleic acid in them which can be stained by various dyes. Till 1980′s reticulocyte quantitation was carried out by microscopic examination of blood smears stained by supravital stains like New methylene blue [1]. With the advent of automation, the precision of reticulocyte counts has remarkably improved [2]. Now in most of the laboratories manual counting is replaced by automated methods which use fluorescent dyes for staining cells and a flow cytometer for cell enumeration [3].
Reticulocyte count is considered as less stable at room temperature because of in vitro maturation of reticulocytes to red blood cells and the count begins to fall within 6–8 h of blood collection [4]. The Clinical and Laboratory Standards Institute (CLSI) recommends that samples stored at room temperature should be analyzed for reticulocytes within six hours of collection. For storage greater than six hours, refrigerated temperatures (4 °C) should be preferred to prevent deterioration of the sample [5].
The objective of this study was to compare manual and automated methods in counting reticulocytes and to study the effect of sample storage on reticulocyte count. The purpose of the study was to assess comparability between manual and automated methods of reticulocyte counting and to understand the variation of reticulocyte count over time.
Materials and Methods
This was a cross-sectional study conducted in the Department of Pathology of a tertiary care hospital in Southern India over a period of one year. EDTA anticoagulated blood samples of patients sent to hematology laboratory for routine complete hemogram in the first morning batch where reticulocyte count was indicated were included. Insufficient or grossly hemolysed samples were excluded.
Sample size was calculated using OpenEpi Software version 3 with estimated mean difference between reticulocyte count measured using manual and automated method to be 0.2 [manual- 1.81(0.416) vs automated- 1.61(0.236)] from previously published data [6]. The sample size was calculated as 86 with 99% confidence interval and 90% power.
Manual microscopic counting was done on new methylene blue stained smears. Standard technique of mixing equal volumes of whole blood and of dye solution (100 µl) followed by incubation at 37 °C for 15 min and blood film preparation on glass slide using wedge method was used. Number of reticulocytes was counted in a maximum of 10 fields to a total of 1000 RBC and reticulocyte percentage was calculated as number of reticulocytes counted/1000 × 100.
The same sample was analysed by automated hematology analyser Sysmex XT-2000i (Sysmex corporation, Kobe, Japan). CBC + RETIC mode was used for analysis of sample using RET search II reagent. In the analyser after a predefined response time the stained sample is introduced into the detector, where forward light scatter and side fluorescence emission are measured. The parameters noted were reticulocyte percentage, absolute reticulocyte count, immature reticulocyte fraction (IRF), low fluorescence ratio (LFR), medium fluorescence ratio (MFR) and high fluorescence ratio (HFR).
The samples were analysed repeatedly by both manual and automated methods at different time intervals (2, 6, 24 and 48 h). The storage temperature was at room temperature (RT) which is usually maintained around 25 °C for 2 and 6 h and at 2–8 °C for 24 and 48 h. Repeat manual counting was done both on freshly prepared slide as well as stored initial slide (prepared within 2 h). Manual count mentioned as such in the text by default indicates counts done on freshly prepared slide.
Data was analysed in SPSS version 20. Outcome continuous variables like reticulocyte count based on manual and automated methods were first summarized as mean with SD and compared using independent t-test. As the distribution of data was found to be non-normally distributed by Shapiro–Wilk test, so non-parametric equivalent Mann–Whitney test was used to compare medians which is presented. Association between manual and automated methods of counting reticulocytes was done by Spearman’s correlation. For normally distributed data of reticulocyte parameters, effect of storage was done using repeated measures ANOVA.
Results
Comparison of methods Association between the two methods (automated vs manual) revealed a strong positive correlation at 2, 6, 24 and 48 h with rs values 0.960, 0.973, 0.974 and 0.954 respectively (Fig. 1). Median (IQR) of reticulocyte count percentage by automated and manual methods at various storage times is given in the Table 1. The difference between automated and manual method was not statistically significant at 2, 6 and 24 h (p values 0.975, 0.967 and 0.227 respectively) but was significant at 48 h (p value 0.013). When manual method (freshly prepared) was compared with manual method using old slide, the difference was statistically significant at 6, 24 and 48 h (p values 0.015, 0.004 and 0.001 respectively).
Fig. 1.
Scatterplot showing strong positive correlation between manual and automated methods of reticulocyte count at 2 h with Spearman’s correlation coefficient of 0.960
Table 1.
Median (IQR) values of reticulocyte count percentage by automated and manual methods using freshly prepared slide and old slide at various time points
| Time | Automated method | Manual method (freshly prepared) | p value (manual vs automated) | Manual method (old slide) | p value (freshly prepared vs. old slide) |
|---|---|---|---|---|---|
| 2 h | 1.73 (0.73–3.19) | 1.6 (0.80–3.40) | 0.975 | ||
| 6 h | 1.75 (0.80–3.21) | 1.6 (0.87–3.05) | 0.967 | 1.0 (0.60–2.00) | 0.015 |
| 24 h | 1.60 (0.69–2.84) | 1.0 (0.60–2.57) | 0.227 | 0.60 (0.20–1.60) | 0.004 |
| 48 h | 1.57 (0.64–2.64) | 0.75 (0.20–2.05) | 0.013 | 0.20 (0.00–1.52) | 0.001 |
Bold indicates significant p value (< 0.05)
Effect of storage Mean (SD) of reticulocyte count by manual and automated methods at various storage times is given in Table 2. Comparison of the reticulocyte count obtained by automated method at various time points showed that the difference between 2 and 6 h was not significant (p value 1.00) but was significant between 2 and 24 h and 2 and 48 h. (p value < 0.001). Similarly, in the manual method, the difference between 2 and 6 h was not statistically significant (p value 1.00) but difference between 2 and 24 h and 2 and 48 h was statistically significant (p value < 0.001).
Table 2.
Mean (SD) values of reticulocyte count percentage by automated and manual methods, IRF and LFR by automated method compared at various time points
| Time(h) | Reticulocyte count percentage by automated method | Reticulocyte count percentage by manual method | IRF by automated method | LFR by automated method | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Mean (SD) | Compared to (h) | p value | Mean (SD) | Compared to (h) | p value | Mean (SD) | Compared to (h) | p value | Mean (SD) | Compared to (h) | p value | |
| 2 | 2.66 (3.07) | 6 | 1.00 | 2.89 (3.58) | 6 | 0.365 | 18.10 (10.62) | 6 | 0.34 | 81.89 (10.62) | 6 | 0.365 |
| 24 | < 0.001 | 24 | 0.001 | 24 | 0.001 | 24 | 0.001 | |||||
| 48 | < 0.001 | 48 | < 0.001 | 48 | < 0.001 | 48 | < 0.001 | |||||
| 6 | 2.68 (3.07) | 2 | 1.00 | 2.92 (3.56) | 2 | 0.365 | 16.38 (9.09) | 2 | 0.34 | 83.60 (9.07) | 2 | 0.365 |
| 24 | < 0.001 | 24 | 0.041 | 24 | 0.04 | 24 | 0.041 | |||||
| 48 | < 0.001 | 48 | < 0.001 | 48 | < 0.001 | 48 | < 0.001 | |||||
| 24 | 2.48 (2.96) | 2 | < 0.001 | 2.36 (3.18) | 2 | 0.001 | 14.76 (8.88) | 2 | 0.001 | 85.21 (8.89) | 2 | 0.001 |
| 6 | < 0.001 | 6 | 0.041 | 6 | 0.04 | 6 | 0.041 | |||||
| 48 | 0.002 | 48 | < 0.001 | 48 | 0.003 | 48 | < 0.001 | |||||
| 48 | 2.38 (2.94) | 2 | < 0.001 | 1.85 (2.91) | 2 | < 0.001 | 13.13 (8.88) | 2 | < 0.001 | 87.13 (8.58) | 2 | < 0.001 |
| 6 | < 0.001 | 6 | < 0.001 | 6 | < 0.001 | 6 | < 0.001 | |||||
| 24 | 0.002 | 24 | < 0.001 | 24 | 0.003 | 24 | < 0.001 | |||||
Bold indicates significant p value (< 0.05)
Immature reticulocyte fraction (IRF) and Low fluorescence ratio (LFR) by automated method Mean (SD) of IRF and LFR by automated method at various storage times is given in Table 2. The IRF reflects the proportion of immature reticulocytes and is calculated from the sum of MFR plus HFR. The effect of storage on IRF was not significant when comparing values obtained at 2 h to 6 h (p value 0.34) but was significant on comparing 2 h to 24 h and 48 h (p value 0.001 and < 0.001).
Similarly, for LFR, comparison of values obtained at 2 h and 6 h was not statistically significant (p value 0.365) but was significant on comparing 2 h to 24 h and 48 h (p value 0.001 and < 0.001).
Discussion
Both manual and automated methods showed a strong positive correlation in the detection of reticulocytes at different time intervals. The value which was found to be high by one method was found to be high by other method also. This data was similar with the data from previous studies which also showed an excellent correlation between the methods [7, 8]. The medians at 2, 6 and 24 h by both the methods were superimposable which implies that both the methods are comparable up to 24 h after blood collection and this data is in agreement with a previous study which also showed that there is no statistical difference between the automated and manual methods [9]. The difference between the two methods at 48 h may be due to observer’s difficulty in identifying the reticulocytes due to their maturation from immature reticulocytes (contains clumps of RNA) to mature reticulocytes (contain few granules of RNA). Automation was also found to be superior in detecting low reticulocyte count. In most of the cases with low reticulocyte count, in which automated method had given a value, corresponding manual method value was found to be zero. This is because of the number of cells surveyed in automated method will be more than 10,000 cells whereas in manual method it was only 1000 cells.
The values obtained in our study showed statistically significant results when comparing freshly prepared slide and old slide. These results imply that instead of counting from the stored slide, it is better to use a newly prepared slide from the sample. We also observed some dye crystal-like artefacts in stored slides.
We studied the effect of storage of sample on reticulocyte count and found that irrespective of the method employed, values obtained on samples processed within 6 h after blood collection were not significantly different from baseline values but were so at 24 and 48 h. CLSI also recommends, samples should be processed within 6 h for reticulocyte analysis if they are kept at room temperature. Studies in the literature about the effect of storage of sample at room temperature have found reticulocyte count to be stable up to 12 h,[10] and unstable with significant reduction over 24 h [11] Some other authors observed a stability of four days for reticulocyte percentage and absolute values when analysis was done in an automated hematology analyser [12].
In our study, we observed a reduction in IRF and an increase in LFR on storage of sample. It indicates that there was a decrease in immature or young reticulocytes and increase in mature reticulocytes which represents the maturation of reticulocytes in vitro. So, if the IRF and LFR have to be studied, samples should be processed within 6 h of blood collection. Similar results were obtained in another study. They analysed the samples by keeping at both room temperature and at 4 °C. At both the storage conditions, they observed maturation of reticulocytes and it was more for samples stored at RT. A reduction in IRF and increase in LFR was evident from 6 h after blood collection [7].
In conclusion, both manual and automated methods can be used for the enumeration of reticulocytes. If manual method is employed, it is preferable to make a freshly prepared slide from sample instead of counting from the old slide. Effect of storage is apparent at 24 h. For more accurate and reliable results, reticulocyte analysis should be performed within 6 h of blood collection irrespective of the method employed. If automated reticulocyte parameters like IRF and LFR have to be studied, it is advisable to perform the reticulocyte analysis within 6 h of blood collection.
Authors Contribution
LG performed the study, analyzed the results, and prepared the manuscript; DB co-guided the study and reviewed the manuscript; RK conceptualized and guided the study and edited the manuscript.
Funding
This study was supported by JIPMER Intramural Grant for MSc-MLT project (JIP/Res/Intramural/Phs 1/2018- 19 dated 16–11-2018).
Declaration
Conflict of interest
Authors declare that they have no conflict of interest.
Ethical Approval
This study was approved by the Institute Ethics Committee (JIP/IEC/2017/0378). As no direct patient interaction or additional sampling was needed, hence waiver of consent was obtained. All were in accordance with the national ethical guidelines for biomedical and health research involving human participants as per Indian Council of Medical Research 2017 guidelines.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Riley RS, Ben-Ezra JM, Tidwell A. Reticulocyte enumeration: Past & present. Lab Med. 2001;32(10):599–608. doi: 10.1309/U0PB-1XCK-FLQ7-V36V. [DOI] [Google Scholar]
- 2.Koepke JA. Update on reticulocyte counting. Lab Med. 1999;30:339–343. doi: 10.1093/labmed/30.5.339. [DOI] [Google Scholar]
- 3.Siekmeier R, Bierlich A, Jaross W. Determination of reticulocytes: three methods compared. ClinChem Lab Med. 2000;38(3):245–249. doi: 10.1515/CCLM.2000.036. [DOI] [PubMed] [Google Scholar]
- 4.Briggs C, Bain BJ. Basic haematological techniques. In: Bain BJ, Bates I, Laffan MA, Lewis SM, editors. Dacie and Lewis practical haematology. 11. Edinburgh: Churchill Livingstone; 2011. p. 35. [Google Scholar]
- 5.CLSI (2004) Methods for reticulocyte counting (Automated Blood Cell Counters, Flow Cytometry, and Supravital Dyes); Approved Guideline—Second Edition. CLSI document H44-A2, Wayne, PA: Clinical and Laboratory Standards Institute
- 6.Buttarello M, Bulian P, Prà MD, Barbera P, Rizzotti P. Reticulocyte quantification by Coulter MAXM VCS (volume, conductivity, light scatter) technology. ClinChem. 1996;42(12):1930–1937. [PubMed] [Google Scholar]
- 7.Lacombe F, Lacoste L, Vial JP, Briais A, Reiffers J, Boisseau MR, Bernard P. Automated reticulocyte counting and immature reticulocyte fraction measurement comparison of ABX PENTRA 120 Retic, Sysmex R-2000, flow cytometry and manual counts. Am J ClinPathol. 1999;112(5):677–86. doi: 10.1093/ajcp/112.5.677. [DOI] [PubMed] [Google Scholar]
- 8.Maconi M, Danise P, Cavalca L, Formisano D. Flow cytometric reticulocyte counting: a comparison between two methods. J Clin Lab Anal. 2010;24(4):252–255. doi: 10.1002/jcla.20394. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Simionatto M, de Paula JP, Chaves MA, Bortoloso M, Cicchetti D, Leonart MS, do Nascimento AJ. Manual and automated reticulocyte counts. Hematology. 2010;15(6):406–409. doi: 10.1179/102453310X12647083621128. [DOI] [PubMed] [Google Scholar]
- 10.Schapkaitz E, Pillay D. Prolonged storage-induced changes in haematology parameters referred for testing. Afr J Lab Med [Internet] 2015;4(1):a208. doi: 10.4102/ajlm.v4i1.208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Pintér E, László K, Schüszler I, Konderák J. The stability of quantitative blood count parameters using the ADVIA 2120i hematologyanalyzer. Pract Lab Med. 2015;4:16–21. doi: 10.1016/j.plabm.2015.12.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.de Baca ME, Gulati G, Kocher W, Schwarting R. Effects of Storage of Blood at Room temperature on Hematologic parameters measured on Sysmex XE-2100. Lab Med. 2006;37(1):28–36. doi: 10.1309/1EERK1M02QFJRX6P. [DOI] [Google Scholar]