ISLH

Laboratory Hematology 6:73-78 2000 Carden Jennings Publishing Co., Ltd.

Official Publication

Reticulocyte Counting in Thalassemia Using Different Automated Technologies

PORNVAREE LAMCHIAGDHASE,1 KOVIT PATTANAPANYASAT,2 WANNA MUANGSUP3
Department of Clinical Microscopy, Faculty of Medical Technology; 2Center of Excellence for Flow Cytometry, Office for Research and Development, Faculty of Medicine, Siriraj Hospital; 3Department of Medicine, Faculty of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand
1

ABSTRACT
The reticulocyte count represents an important test in the study of marrow erythropoietic activity. Automated techniques have led to a signicant improvement in the accuracy and precision of reticulocyte enumeration compared with manual counting. This study compared automated reticulocyte counting by 2 technologies (Coulter Gen S and Sysmex R-2000) in samples from 45 healthy subjects and 66 thalassemic patients (21 hemoglobin [Hb] H, 8 Hb H/Constant Spring [CS], 20 -thalassemia [-thal]/Hb E nonsplenectomy, and 17 -thal/Hb E splenectomy). There was no statistically signicant difference between the 2 methods, although the Sysmex R-2000 showed a slightly higher mean reticulocyte count (RC). The overall correlation between the 2 instruments was r = 0.877. RCs in all thalassemia groups were signicantly higher than in control subjects (P < .0001). There was no significant difference among -thalassemia syndromes, although absolute RCs of Hb H/CS patients had the highest values, indicating better compensatory erythropoiesis. Compared with nonsplenectomized -thal/Hb E patients, splenectomized -thal/Hb E patients had a higher percentage of reticulocytes, a higher absolute RC, a higher percentage of high-light-scatter reticulocytes, and greater mean reticulocyte volume, mean sphered cell volume, and high-fluorescence-intensity ratio. Patients with -thal/Hb E showed a more severe degree of ineffective erythropoiesis than patients with -thalassemic syndromes. In conclusion, we have shown good agreement between the 2 automated reticulocyte counting technologies. Lab Hematol. 2000;6:73-78.

INTRODUCTION
The reticulocyte count (RC) is the fundamental routine measurement used to evaluate the erythropoietic activity of bone marrow. Although the manual counting method is inexpensive and relatively simple to perform, it has certain major limitations. The technique is time-consuming, is subject to statistical error because of the small number of cells counted, and has significant interobserver variation [1-4]. Development of automated counting by ow cytometry has tremendously increased the accuracy and precision of counting reticulocytes and determining their maturity [5-8]. Different maturity indexes, such as the reticulocyte maturation index (RMI) or the different maturational fractions obtained by the Sysmex R series automated counter (Toa Medical Electronics, Kobe, Japan), have been designed based on the relative amount of cellular RNA. The Sysmex counter provides information about reticulocyte maturation based on the different reecting intensities relative to RNA content, divided into low-fluorescence-intensity ratio (LFR), middle-uorescence-intensity ratio (MFR), and highfluorescence-intensity ratio (HFR). The HFR contains the most immature reticulocytes, MFR contains immature reticulocytes, and LFR contains mature reticulocytes. In the Sysmex R-3500, the immature reticulocyte fraction (IRF), which is equal to MFR plus HFR, is also reported. Recently, Coulter (Hialeah, FL) developed an automated method using the new methylene blue (NMB) procedure for use with the Gen S hematology analyzer. In addition to providing normal parameters, the Coulter Gen S also provides parameters such as mean reticulocyte volume (MRV), IRF, highlight-scatter reticulocytes (HLR) (percentage and absolute value), and mean sphered cell volume (MSCV), which is the average of the total red blood cells (RBCs) in the reticulocyte analysis. These new parameters have generated clinical interest, and some studies have evaluated their possible role in the study of anemia [9] and in monitoring post-therapeutic hematologic recovery [10]. The thalassemias are a heterogeneous group of genetic hemoglobin disorders resulting from reduced synthesis of - and -globin chains [11]. Clinically, they are divided into 3 groups: thalassemia minor, which is the symptomless thalassemia heterozygote; thalassemia

KEY WORDS:

Reticulocyte count Thalassemia Coulter Gen S Sysmex R-2000

Correspondence and reprint requests: Pornvaree Lamchiagdhase, Department of Clinical Microscopy, Faculty of Medical Technology, Mahidol University, Siriraj Hospital Campus, Bangkoknoi, Bangkok 10700, Thailand (e-mail: [email protected]) (Received November 19, 1999; accepted April 4, 2000)

73

74

P. Lamchiagdhase et al
RET%: The number of reticulocytes per 100 RBCs, directly measured and reported as a percentage of RBCs. RET#: The absolute number of reticulocytes, calculated from RET% and RBC number and expressed as 106 cells/L or 109 cells/L. IRF: Immature reticulocyte fraction, calculated from the ratio of HLR and total reticulocytes. HLR%: High-light-scatter reticulocytes, directly measured and reported as a percentage of RBCs. HLR#: The absolute number of HLR, calculated from HLR% and RBC number and expressed as 106 cells/L or 109 cells/L. MRV: Mean reticulocyte volume, the average volume of individual reticulocytes, expressed in fL. MSCV: Mean sphered cell volume, the average of total RBCs in the reticulocyte analysis, expressed in fL. Automated Reticulocyte Counting by Sysmex R-2000 The Sysmex R-2000 is a fully automated reticulocyte counter. The measuring principle of the system is based on ow cytometry combined with hydrodynamic focusing. EDTA blood (100 L) is aspirated, and intracellular RNA is stained by auramine O, which is uorescent under argon laser light [17]. The reticulocyte population is further subdivided into LFR, MFR, and HFR. The percentage of reticulocytes is given as the sum of LFR, MFR, and HFR. Approximately 3.2 104 RBCs are usually analyzed for each sample. The IRF is manually calculated by MFR plus HFR. Statistical Analysis For statistical analysis, SPSS/PC+ version 4.0 (Microsoft, Redmond, WA) was used to calculate descriptive statistics (mean, SD, coefcient of variation, and range). Comparisons of statistical difference between reticulocyte counting methods were performed with Friedman 2-way analysis of variance and Wilcoxon signed rank test. The simple linear regression and correlation coefcient (r) were also determined. Among the groups of patients, KruskalWallis 1-way analysis of variance and the Mann-Whitney U test were used. The threshold for statistical signicance for all comparisons was chosen as P = .05.

intermedia, characterized by moderate anemia; and transfusiondependent thalassemia major. Although the molecular biology of the genetic defect has been studied extensively, knowledge of RBC pathology is still insufficient for understanding the premature removal of abnormal RBCs in the marrow and peripheral circulation. Patients suffer from anemia because of untimely RBC destruction in the bone marrow and spleen. Compensatory erythropoiesis is still exacerbated by premature RBC destructionso-called ineffective erythropoiesis in the marrow. Evaluation of erythropoiesis is a very interesting approach to investigating pathogenesis of anemia. In Thailand, -thalassemia, -thalassemia, abnormal hemoglobin (Hb) E, and Hb Constant Spring (CS) are common [11-14]. In the present study, we compared manual reticulocyte counts and reticulocyte counts performed on the Coulter Gen S and the Sysmex R-2000 in various genotypes of thalassemia.

MATERIALS AND METHODS

Samples and Subjects Forty-ve healthy adults (21 men, 24 women) aged 26 8 years who were not on medication and had normal hematologic data by automated complete blood count (CBC) were used as the control group in this study. We also examined 66 thalassemic patients aged 34 12 years, all with clinical symptoms of thalassemia intermedia. Criteria for the diagnosis was based on previous reports [12,13]. The thalassemic group included 21 Hb H patients (-thalassemia [-thal] 1/-thal 2; 4 men, 17 women), 8 Hb H/CS patients (-thal 1/Hb CS; 4 men, 4 women), and 20 -thal/Hb E patients (6 men, 14 women), all of whom were nonsplenectomized and had no blood transfusions for at least 3 months before this study. Seventeen splenectomized -thal/Hb E patients (7 men, 10 women) were also included. Blood samples were collected and placed in K2EDTA anticoagulant and analyzed for CBC, blood smear, and RC within 4 hours. Manual Reticulocyte Counting Supravital staining of unxed RBCs with NMB was performed by mixing 100 L whole blood with 100 L of 1% NMB solution in a 12 75-mm tube [15]. After incubation at room temperature for 15 minutes, the dilution was remixed, and a wedge smear was performed for each sample. The number of reticulocytes per 1000 RBCs was determined microscopically. A reticulocyte was dened as a RBC containing at least 2 granules of reticulum. The absolute RC was calculated from the RBC count using Coulter Gen S, and the reticulocyte production index (RPI) was calculated manually. The RPI (also known as shift correction) is a general indicator of the rate of RBC production above normal in anemia. Automated Reticulocyte Counting by Coulter Gen S The performance characteristics are described with the Coulter Gen S reticulocyte measurement in RETIC mode. Briey, 34 L whole blood is automatically aspirated and incubated with 166 L NMB to precipitate the basophilic RNA in reticulocytes [16]. Hemoglobin and unbound stain are removed by adding a clearing reagent, which leaves clear, spherical, mature RBCs and darkly stained reticulocytes. Stained reticulocytes are differentiated from mature RBCs and other cell populations by direct current measurements for volume, conductivity, and light scatter (VCS). Cells (3.2 104) are counted to measure the following parameters:

RESULTS
The RCs of healthy subjects and thalassemic patients are summarized in Table 1. The percentage reticulocytes and absolute RCs of the patients were signicantly higher than those of normal subjects. In healthy subjects, the percentage reticulocytes did not differ between men and women, but the absolute RC in men was higher (P < .0001). The comparison of the manual RC and the 2 automated RCs showed no difference in the thalassemic group. Among the thalassemic patients, however, percentage reticulocytes and absolute RCs of Hb H/CS were high by the manual method, and splenectomized -thal/Hb E patients showed the highest values when using the Sysmex R-2000. In addition, the correlation between the manual RC and the 2 automated RCs was good (P = .001), with r = 0.874 and r = 0.816 for the manual RC versus the 2 automated methods and r = 0.877 between the 2 automated methods (Figure). The percentage reticulocytes and absolute RCs of Hb H/CS were signicantly higher than those of the other group of nonsplenectomized patients. However, when results were compared between the 2 types of Hb H diseases, there was no statistically signicant

Automated Reticulocyte Count in Thalassemia TABLE 1. Reticulocyte Count of Healthy Subjects Using the Manual Method, the Coulter Gen S, and the Sysmex R-2000*
n Reticulocyte count, % Healthy subjects Men Women Total Hb H Hb H/CS -Thal/Hb E (NS) -Thal/Hb E (S) Total Absolute reticulocyte count, 109/L Healthy subjects Men Women Total Hb H Hb H/CS -Thal/Hb E (NS) -Thal/Hb E (S) Total Manual Coulter Gen S

75

Sysmex R-2000

21 24 45 21 8 20 17 111

1.0 0.3 (0.5-1.6) 1.0 0.3 (0.4-1.7) 1.0 0.3 (0.4 -1.7) 7.5 2.6 (3.4-13.4) 9.5 2.3 (6.3-12.3) 3.3 1.1 (1.4-5.9) 9.2 4.6 (4.4-19.5) 4.5 4.1 (0.4-19.5)

1.1 0.3 (0.6-1.7) 1.2 0.4 (0.7-1.7) 1.1 0.4 (0.6-1.7) 7.1 2.2 (3.6-11.2) 7.4 1.4 (5.0-10.0) 3.9 1.0 (2.3-6.7) 8.5 4.6 (3.4-21.1) 4.3 3.6 (0.5-21.1)

1.2 0.3 (0.7-2.3) 1.1 0.4 (0.7-2.0) 1.2 0.4 (0.7-2.3) 7.0 2.7 (2.7-12.9) 8.6 3.5 (5.8-16.2) 2.5 0.9 (1.3-4.8) 11.4 4.5 (4.0-21.5) 4.5 4.4 (0.5-21.6)

21 24 45 21 8 20 17 111

53 16 (25-87 ) 44 17 (26-83) 48 17 (25-87) 297 94 (116-552) 407 156 (263-649) 144 43 (38-194) 278 135 (126-529) 167 146 (18-649)

58 15 (35-92) 53 16 (24-91) 55 16 (24-92) 269 91 (37-407) 310 67 (218-388) 144 43 (67-221) 221 99 (64-404) 154 111 (24-407)

65 18 (44-78) 49 20 (23-99) 57 20 (23-99) 306 119 (105-651) 389 219 (208-853) 105 36 (49-163) 342 123 (163-633) 170 156 (21-853)

*Data are means SD (range). Hb indicates hemoglobin; CS, Constant Spring; -thal, -thalassemia; NS, nonsplenectomy; S, splenectomy.

difference in any parameter or method used. Splenectomized -thal/Hb E reticulocyte percentages were significantly higher (P = .001) than the 2 types of Hb H on the Sysmex R-2000. Table 2 shows the mean values and SDs of IRF, the percentage and absolute HLR obtained by the Coulter Gen S, and the percentage of LFR, MFR, and HFR obtained by the Sysmex R-2000 in healthy subjects and thalassemic patients. In 2 types of Hb H diseases, none of the parameters obtained by the 2 automated methods showed a signicant difference. However, nonsplenectomized -thal/Hb E showed signicantly lower IRF, HLR, MRV, MSCV, MFR, and HFR compared with -thalassemia and splenectomized -thal/Hb E; only the LFR percentage was comparable. For the immature fraction using the Sysmex R-2000 (Table 2), only LFR (P = .01) and MFR (P = .001) of Hb H patients were significantly different from those of nonsplenectomized and splenectomized -thal/Hb E patients. However, when comparing nonsplenectomized and splenectomized -thal/Hb E, the LFR and MFR did not show any signicant difference (P > .05). The RPI, MRV, HLR#, MFR, HFR, and IRF of Hb H/CS patients were signicantly higher than those of Hb H and nonsplenectomized -thal/Hb E patients. However, when results were compared between the 2 types of Hb H diseases, there was no statistically signicant difference in any parameter or method used. The mean percentage reticulocytes, absolute RC, HLR#, and MRV were slightly higher than those of Hb H patients, but in contrast, all reticulocyte values except IRF, LFR, and MFR of nonsplenectomized -thal/Hb E patients were signicantly lower than those of other patients. Table 3 demonstrates the average of RBC size and reticulocytes. MRV was consistently higher than mean corpuscular volume (MCV), with a mean difference in healthy subjects of 27.9 7.5 fL. In splenectomized -thal/Hb E patients, the highest MRV was

found, with a mean difference of 60.7 14.2 fL. The ratio of MCV to MRV in every group was >1. MRV of splenectomized -thal/Hb E patients was signicantly higher than that of other patients, with the mean volume ratio (MRV:MCV) of those cell populations being approximately 1.8 0.2. Correlation between percentage reticulocytes and absolute RC with IRF, HLR, MRV, and MSCV is shown in Table 4. Poor correlation was found with IRF when using the 2 instruments, but

Correlation of reticulocyte count (RC) as determined by the Coulter Gen S and Sysmex R-2000.

76

P. Lamchiagdhase et al

TABLE 2. Reticulocyte Parameters of Healthy Subjects and Thalassemic Patients Using the Coulter Gen S and Sysmex R-2000*
Healthy Subjects (n = 45)
RPI Coulter Gen S IRF HLR, % HLR#, 109/L MRV, fL MSCV, fL Sysmex R-2000 LFR, % MFR, % HFR, % IRF, % ()

Hb H (n = 21)
2.6 1.0 ()

Hb H/CS (n = 8)
3.6 1.4 () 0.375 0.122 (0.080-0.440) 2.80 1.14 (0.590-4.38) 0.114 0.042 (0.031-0.162) 110.5 5.0 (104.3-117.2) 85.7 7.6 (76.0-102.8) 77.5 9.0 (67.8-96.0) 23.4 3.3 (19.5-28.4) 2.8 1.9 (1.5-6.9) 26.2 4.5 (21.5-32.0)

-Thal/Hb E (NS) (n = 20)

0.8 0.3 () 0.333 0.056 (0.230-0.450) 1.23 0.58 (0.080-2.84) 0.047 0.018 (0.020-0.088) 96.6 13.2 (75.0-130.2) 68.1 11.2 (55.2-99.3) 83.9 4.8 (73.8-93.7) 14.3 4.0 (5.8-21.9) 1.8 1.1 (0.5-4.3) 16.1 4.8 (6.3-26.2)

-Thal/Hb E (S) (n = 17)

2.0 0.8 () 0.355 0.206 (0.020-0.640) 3.03 2.40 (0.150-7.74) 0.091 0.068 (0.004-0.210) 136.0 17.6 (98.2-165.9) 106.8 14.8 (62.7-124.0) 82.6 2.9 (75.7-86.8) 14.7 1.9 (11.6-18.6) 2.7 1.3 (1.4-5.7) 17.3 2.9 (13.2-24.3)

0.308 0.055 (0.243-0.438) 0.388 0.079 (0.200-0.490) 0.355 0.111 (0.168-0.513) 2.86 1.22 (0.750-5.27) 0.017 0.006 (0.008-0.026) 0.112 0.043 (0.035-0.186) 114.1 7.9 (98.5-125.5) 106.7 11.9 (82.1-127.3) 92.1 4.3 (82.8-97.8) 76.5 6.6 (61.6-88.9) 92.9 2.8 (86.7-97.4) 6.7 2.6 (2.6-13.0) 0.5 0.4 (0.0-1.2) 7.1 2.9 (2.4-13.3) 77.1 9.2 (65.4-97.1) 20.6 6.9 (2.7-28.5) 2.3 1.6 (0.1-6.1) 22.9 8.2 (2.9-34.6)

*Data are means SD (range). Hb indicates hemoglobin; CS, Constant Spring; -thal, -thalassemia; NS, nonsplenectomy; S, splenectomy; RPI, reticulocyte production index (corrected reticulocyte count + maturation time in peripheral blood); IRF, immature reticulocyte fraction; HLR%, high-light-scatter reticulocyte percentage; HLR#, high-light-scatter reticulocyte number; MRV, mean reticulocyte volume; MSCV, mean sphered cell volume; LFR, low-uorescenceintensity ratio; MFR, middle-uorescence-intensity ratio; HFR, high-uorescence-intensity ratio.

percentage and absolute HLR on the Coulter Gen S were signicantly correlated with MFR and IRF on the Sysmex R-2000.

DISCUSSION
The peripheral blood RC is commonly used as an indicator of the erythropoietic activity of bone marrow. Manual counting with the microscope usually gives results with a high degree of inaccuracy and imprecision and a large inter- and intraobserver variability. Automated counting has overcome those limitations, and the availability of reticulocyte maturation indexes, based on the measurement of RNA content, extends the clinical utility of reticulocyte counting. Using 2 different principles of automated reticulocyte counting, we compared automated reticulocyte counting with the manual method in thalassemic and healthy subjects. Correlation analysis of manual RC versus the Coulter Gen S and Sysmex R-2000 produced r values of 0.874 and 0.816, respectively. Correlation analysis between the 2 automated instruments gave an r value of 0.877. Kessler et al [18] reported good correlation between Coulter Gen S and manual RC (r = 0.800), Coulter MAXM (r = 0.936), and R-1000 (r = 0.890). Kojima et al [19] and Tatsumi et al [20] also found good correlations

between manual RCs and Sysmex R-1000, with r values of 0.945 and 0.893, respectively. The percentage reticulocytes and absolute RCs of thalassemic patients were significantly higher than those of normal subjects. High RCs in Hb H and Hb H/CS patients have been reported using the manual microscope procedure [13,21]. In this study, no methods showed statistically signicant differences between Hb H and Hb H/CS patients. However, RPI of Hb H/CS patients (3.6) was signicantly higher than that of Hb H patients (2.0). This nding is similar to ndings in our previous report (Hb H = 1.1; Hb H/CS = 2.3) [22] and to the ndings of Khuhapinant et al (Hb H = 1.97; Hb H/CS = 3.62) [23]. RPI is an indicator of the rate of RBC production and increases in response to anemia. An RPI value >3 generally indicates an adequate bone marrow response to anemia, whereas an RPI value <2 represents an inadequate response [24] and is interpreted as evidence of severe RBC destruction in Hb H/CS versus Hb H patients. In our study, the nonsplenectomized -thal/Hb E patients had a low RPI (0.8), which reects ineffective erythropoiesis and is more pronounced in -thalassemic patients than in -thalassemic patients [23]. There was also no signicant difference in the IRF, HLR%, HLR#, and MFR between the 2 types of Hb H diseases in our data, but there was a signicant

TABLE 3. Mean Corpuscular Volume (MCV) and Mean Reticulocyte Volume (MRV) of Healthy Subjects and Thalassemic Patients Using the
Coulter Gen S*
Healthy Adults (n = 45) MCV, fL MRV, fL MRV-MCV, fL MRV:MCV ratio 86.8 3.2 114.1 7.9 27.9 7.5 1.3 0.1 Hb H (n = 21) 69.3 5.4 106.7 11.9 37.4 8.6 1.5 0.1 Hb H/CS (n = 8) 76.8 4.5 110.5 5.0 33.6 5.1 1.4 0.1 -Thal/Hb E (NS) (n = 20) 60.8 6.6 96.6 13.2 35.8 12.3 1.6 0.2 -Thal/Hb E (S) (n = 17) 75.2 8.2 136.0 17.6 60.7 14.2 1.8 0.2

*Data are means SD. Hb indicates hemoglobin; CS, Constant Spring; -thal, -thalassemia; NS, nonsplenectomy; S, splenectomy.

Automated Reticulocyte Count in Thalassemia TABLE 4. Correlation Between Reticulocyte Count and Other Parameters of 111 Samples Using the Coulter Gen S and the Sysmex R-2000*
Coulter Gen S RET# Coulter Gen S RET% RET# IRF HLR% HLR# MRV MSCV Sysmex R-2000 RET% RET# LFR MFR HFR IRF HLR % HLR# MRV MSCV RET% RET# Sysmex R-2000 LFR MFR HFR

77

IRF

0.7998

0.5167 0.3564

0.8823 0.8034 0.7125

0.8398 0.8764 0.6791 0.9577

0.2360 0.0477 0.2330 0.2044 0.0698

0.0328 0.2203 0.1604 0.1157 0.2229 0.8236

0.8772 0.7077 0.2844 0.7133 0.6554 0.4178 0.2593

0.8136 0.7897 0.2475 0.6727 0.6884 0.2671 0.1394 0.9439

0.7112 0.6839 0.4675 0.7648 0.7547 0.0281 0.2131 0.5234 0.4736

0.7250 0.7518 0.3889 0.7393 0.7408 0.0167 0.1964 0.5931 0.6064 0.9378

0.6215 0.5362 0.2556 0.5663 0.5122 0.1509 0.0275 0.5480 0.4683 0.8036 0.7527

0.7316 0.7410 0.3755 0.7334 0.7263 0.0379 0.1789 0.6034 0.6026 0.9479 0.9938 0.8177

*RET# indicates number of reticulocytes; IRF, immature reticulocyte fraction; HLR%, high-light-scatter reticulocyte percentage; MRV, mean reticulocyte volume; MSCV, mean sphered cell volume; RET%, percentage reticulocytes; LFR, low-uorescence-intensity ratio; MFR, middle-uorescence-intensity ratio; HFR, high-uorescence-intensity ratio. P = .001. P = .01. Note. Italicized numbers represent nonsignicant gures.

difference when we compared these measurements in the nonsplenectomized -thal/Hb E patients. However, when we compared nonsplenectomized and splenectomized -thal/Hb E patients, only the HLR% and HFR showed statistically significant differences. This result suggests that the spleen may play an important role in pooling reticulocytes before releasing them into the circulation [25]. A marked increase in reticulocytes was found in splenectomized -thal/Hb E patients, especially on the Sysmex R-2000. However, there was no statistically signicant difference between the manual RC (9.2 4.6) and that of the Coulter Gen S (8.5 4.6), and RCs with both methods were signicantly lower than those with the Sysmex R-2000 (11.4 4.5) (P = .0012 and .0037, respectively). One explanation is that the supravital stain used for the Coulter Gen S is similar to that used for manual RC but different from that for the Sysmex R-2000, which uses a different uorochrome. Villamor et al [26] presented evidence that when using flow cytometrybased methods, white blood cells could erroneously be identied as reticulocytes, which may affect the reticulocyte differential and RMI. This result occurs because the uorochrome can stain both RNA and DNA. The presence of Howell-Jolly bodies and nucleated red blood cells, as seen in splenectomized patients, can also interfere with automated RC [27], resulting in high numbers. Our data show no statistically significant difference among methods in Hb H and Hb H/CS patients, but there was a signicant difference in nonsplenectomized and splenectomized -thal/Hb E patient values. For MRV, we found no difference in the 2 types of Hb H diseases. However, in splenectomized -thal/Hb E patients, the MRV and MRV:MCV ratio were signicantly higher. This nding suggests that the spleen plays an important role in modifying the size of immature RBCs and supports previous ndings that maturation in abnormal erythropoiesis of thalassemia patients is similar to that of normal subjects, except after splenectomy [28,29].

In summary, the present investigation demonstrated a linear relationship between manual RC and 2 automated instruments with no statistically signicant difference between these methods. RCs by automated methods are useful not only for providing precise counts, but also for determining RBC maturation. This is especially true of the Coulter Gen S, which uses the same principle in staining reticulocytes as the manual method and has less interference from inclusion-containing RBCs. By using these 2 automated technologies to evaluate erythropoietic activity in thalassemia, we found that both thalassemia intermedia and nonsplenectomized -thal/Hb E had severe ineffective erythropoiesis when compared with 2 types of Hb H diseases. -Thal/Hb E splenectomized patients showed a marked increase in reticulocyte number with increasing MRV and immature reticulocytes.

ACKNOWLEDGMENTS
The authors wish to thank PCL Holding Co. Ltd., Thailand, for support with the reagents and all the cost of publication. We also thank Janice M. Darden for reading the manuscript and Amphun Chaiphan and Sirintorn Mahem for technical assistance.

REFERENCES
1. Peebles DA, Hochberg A, Clarke TD. Analysis of manual reticulocyte counting. Am J Clin Pathol. 1981;76:713-717. 2. Koepke JF, Koepke JA. Reticulocytes [review]. Clin Lab Haematol. 1986;8:169-179. 3. Cavill I. The rejected reticulocyte [review]. Br J Haematol. 1993;84: 563-565. 4. Lohmann RC, Crawford LN, Wood DE. Prociency testing in reticulocyte counting. Clin Lab Haematol. 1994;16:57-64.

78

P. Lamchiagdhase et al
18. Kessler C, Machin SJ, Pollard Y, et al. Reticulocyte performance on the Coulter Gen S System. Lab Hematol. 1997;3:41-47. 19. Kojima K, Niri M, Setoguchi K, Tsuda I, Tatsumi N. An automated optoelectronic reticulocyte counter. Am J Clin Pathol. 1989;92:57-61. 20. Tatsumi N, Tsuda I, Kojima K, Nari M, Setoguchi K. An automated reticulocyte counting method: preliminary observations. Med Lab Sci. 1989;46:157-160. 21. Winichagoon P, Thonglairuam V, Fucharoen S, Tanhpaichitr VS, Wasi P. -Thalassemia in Thailand. Hemoglobin. 1988;12:485-498. 22. Lamchiagdhase P, Pattanapanyasat K, Lebnark T. Reticulocyte count in two types of hemoglobin H disease. Thai J Hematol Transfus Med. 1995;5:195-201. 23. Khuhapinant A, Bunyaratvej A, Tatsumi N, Pribwia M, Fucharoen S. Number and maturation of reticulocytes in various genotypes of thalassemia as assessed by flow cytometry. Acta Haematol. 1994;91:119125. 24. Hillman RS, Finch CA. Red Cell Manual. 6th ed. Philadelphia, PA: FA Davis; 1992. 25. Bowdler AJ. Splenomegaly and hypersplenism [review]. Clin Haematol. 1983;12:467-488. 26. Villamor N, Kirsch A, Huhn D, Vives-corrons JL, Serke S. Interference of blood leukocytes in the measurements of immature red cells (reticulocytes) by two different (semi-) automated ow-cytometry technologies. Clin Lab Haematol. 1996;18:89-94. 27. Lofsness KG, Kohnke ML, Geier NA. Evaluation of automated reticulo cyte counts and their reliability in the presence of Howell-Jolly bodies. Am J Clin Pathol. 1994;101:85-90. 28. Lowenstein LM. The mammalian reticulocyte. Int Rev Cytol. 1959;8:135-147. 29. dOnofrio G, Chirillo R, Zini G, et al. Simultaneous measurement of reticulocyte and red blood cell indices in healthy subjects and patients with microcytic and macrocytic anemia. Blood. 1995;85:818-823.

5. Davis BH, Bigelow NC, Koepke JA, et al. Flow cytometric reticulocyte analysis: multiinstitutional interlaboratory correlation study. Am J Clin Pathol.1994;102:468-477. 6. Buttarello M, Bulian P, De Pre M, Barbera P, Rizzotti P. Reticulocyte quantification by Coulter MAXM VCS (volume, conductivity, light scatter) technology. Clin Chem. 1996;42:1930-1937. 7. Davis BH, DiCorato M, Bigelow NC, Langweiler MH. Proposal for standardization of ow cytometric reticulocyte maturity index (RMI) measurements. Cytometry. 1993;14:318-326. 8. Tichelli A, Gratwohl A, Driessen A, et al. Evaluation of the Sysmex R-1000: an automated reticulocytes analyzer. Am J Clin Pathol. 1990;93:70-78. 9. Tsuda I, Tatsumi N. Maturity of reticulocyte in various haematological disorders. Eur J Haematol. 1989;43:251-254. 10. Tanke HJ, von Vianen PH, Emiliani FMF. Changes in erythropoiesis due to radiation or chemotherapy as studied by ow cytometric determination of peripheral blood reticulocytes. Histochemistry. 1986;84:544-548. 11. Weatherall DJ, Cleg JB. The Thalassemia Syndromes. 3rd ed. Oxford, England: Blackwell; 1981. 12. Wasi P, Na-nakorn S, Pootrakul S, et al. Alpha- and beta-thalassemia in Thailand. Ann N Y Acad Sci. 1969;165:60-82. 13. Wasi P, Na-nakorn S, Pootrakul S. The -thalassemias. Clin Haematol. 1974;3:383-410. 14. Wasi P. Hemoglobinopathies in Southeast Asia. In: Bowman JE, ed. Distribution and Evaluation and Evolution of the Hemoglobin Gene Loci. New York, NY: Elsevier; 1983:176-208. 15. Evatt BL, Gibbs WN, Lewis SM, McArthur JR. Fundamental Diagnostic Hematology: Anemia . Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control; 1991:78-79. 16. Coulter Corporation. Coulter Gen S System On Line Reference Manual. Coulter Corp. PN 4237221; Miami, FL, December 1996. 17. TOA Medical Electronics. TOA Medical Electronics Operation Manual for R-2000. Kobe, Japan: TOA Medical Electronics; 1993.