The Evaluation of Various Mathematical RBC Indices and Their Efficacy in Discriminating Between Thalassemic and Non-Thalassemic Microcytosis

Transcription

1 HEMATOPATHOLOGY The Evaluation of Various Mathematical RBC Indices and Their Efficacy in Discriminating Between Thalassemic and Non-Thalassemic Microcytosis JOHN D. LAFFERTY, ART, MARK A. CROWTHER, MD, 2 MAHMOUD A. ALI, MD, PHD, FRCP(C), AND MITCHELL LEVINE, MD, MSc, FRCP(C) 3 The differentiation between thalassemic and non-thalassemic microcytosis has important clinical implications in hematology and medicine. A simplified index, based on red cell parameters derived from automated blood cell analyzers, which could be used to discriminate between microcytic patients with a high probability of thalassemia minor and those with a low probability, would be an extremely useful tool. Five mathematical indices have been proposed as useful for this purpose. These are the: Bessman index, Shine and Lai index, England index, Mentzler index, and mean cell volume (MCV) alone. This study was designed to prospectively evaluate the efficacy of these indices. Patient samples were chosen every fourth day from all patient samples referred to the hematology laboratory at St. Joseph's Hospital over a 6-month period. All patient samples with an MCV < 80 fl and age > 8 years were considered eligible for the study. After enrollment and laboratory analysis were complete sensitivities and specificities were calculated for each of the indices using a variety of cut-off values and receiver operator characteristic (ROC) curves were constructed. Based on statistical analysis of the area under these curves, the authors conclude that MCV alone is as effective as the Mentzler and Shine and Lai indices in selecting microcytic patient samples with a high probability of thalassemia minor for thalassemia testing. They also conclude that the Bessman index and the England index are ineffective indices for this purpose. (Key words: Microcytic anemia; 0 thalassemia trait; a thalassemia trait; Iron deficiency anemia; Mean cell volume (MCV); Red cell distribution width RDVV ) Am J Clin Pathol 996; 06: Microcytosis is a common presenting manifestation of anemia. It is an important clinical clue as it narrows the differential diagnosis of the anemia to four well-defined conditions. These are iron deficiency, thalassemia minor, the anemias of chronic disease, and the sideroblastic anemias. The differentiation between thalassemic and non-thalassemic microcytosis has important clinical implications in hematology and medicine. The accepted protocol for the diagnosis of thalassemia minor in routine hematology laboratories is the performance of hemoglobin (Hb) electrophoresis, Hb A 2 quantification and investigation for the presence of Hb H inclusion bodies.' However, it has been proposed that the From the ^Regional Hemoglobinopathy Reference Laboratory, St. Joseph's Hospital. Hamilton. Ontario, Canada; 2 McMaster University Department of Medicine. Hamilton. Ontario, Canada; and * Father Sean O 'Sullivan Research Centre. St. Joseph's Hospital. Hamilton. Ontario, Canada. Manuscript received September 8, 995; revision accepted February Address reprint requests to John Lafferty: Regional Hemoglobinopathy Laboratory. St. Joseph's Hospital, 50 Charlton Avenue, E, Hamilton, Ontario L8N 4A6, Canada. routine electronic red blood cell (RBC) counts and indices derived from modern blood cell analyzers can provide valuable clues as to whether thalassemia minor may be present. 2 ' 3 Specifically, for a given mean cell volume (MCV), the Hb level and RBC count tend to be higher in thalassemia minor than in iron deficiency. A simplified index, based on the complete blood count (CBC), with a high level of specificity and sensitivity for thalassemia minor would be an extremely useful tool in the investigation of microcytic anemia. Various indices have been developed to differentiate between thalassemia minor and iron deficiency. Indices derived by England and colleagues 4 and Mentzler 5 exploit the above mentioned discrepancies between Hb, RBC count, and MCV. The usefulness of the England and Mentzler indices was evaluated by other researchers. 6 " 9 Shine and Lai 0 developed an index based on the MCV and MCH (mean cell hemoglobin), whereas Ghosh and associates" found that an MCV of <75 fl provided superior differentiation of ^-thalassemia trait from iron deficiency, in pregnant women of Chinese ancestry, than did the England and Mentzler indices. Bessman and 20

2 202 HEMATOPATHOLOGY Feinstein 2 relied on the observation that patients with iron deficiency had a greater range of red cell size than did patients with thalassemia minor. Their investigations led to the development of the red cell distribution width (RDW), which was widely publicized and added to the output of new cell analyzers. Bessman and colleagues' 3 subsequently proposed that a normal RDW (ie, <5.0 4 ), in combination with a decreased MCV was a strong indicator of thalassemia minor. There is significant controversy in the literature over the effectiveness of the RDW as an indicator of thalassemia minor with supporters 5 " 8 and detractors' 9 " 22 on both sides of the issue. To date, there has been no report prospectively assessing the efficacy of all the various formulas proposed and their ability to discriminate between patients with thalassemia minor and patients with various non-thalassemic causes of microcytosis. An effective way to critically assess the various formulas would be to perform a prospective study in which receiver operator characteristic (ROC) curves are constructed for each index. This would evaluate which formula has the best sensitivity and specificity profile and would be the most effective index for predicting the likelihood of thalassemia minor. The purpose of this communication is to report such a study performed at the Regional Hemoglobinopathy Reference Laboratory at St. Joseph's Hospital in Hamilton, Ontario, Canada. Entrance Criteria MATERIALS AND METHODS Samples were collected from the hematology laboratory at St. Joseph's Hospital in Hamilton, Ontario, Canada for a 6-month period. Samples were collected every fourth day to randomize the day of the week on which collection occurred. It is routine practice in this laboratory to screen all adult patients with an MCV of <75 fl for thalassemia syndromes. For the purpose of this study, all patients 8 years of age and older with an MCV < 80 fl were entered.' All patients were tested on their first entrance to the study and if encountered on subsequent days were not re-entered. Laboratory Tests All samples entered into the study were collected into potassium (K 3 ) EDTA Vacutainer tubes (Becton Dickinson Vacutainer Systems, Rutherford, NJ) and were delivered the same day to the hematology laboratory. Study testing was conducted on the following business day. Each sample had the following laboratory tests performed using standard published techniques: CBC (Coulter STKS, Coulter Electronics, Hialeah, FL.) Hb H inclusion body investigation 23 HbA 2 level 24 HbF level 25 Starch Gel Electrophoresis ph Free erythrocyte protoporphyrin (FEP) 27 Serum ferritin 28 High Performance Liquid Chromatography (HPLC) 29 was performed, when required, to confirm the identity of Hb variants detected on electrophoresis. Indices Used: Mentzler index: MCVRBC 5 Bessman index: RDW 3 Shine and Lai index: MCV 2 X (MCH00) 0 England index: MCV - RBC - (0.5 X Hb) Statistical Analysis Data were collected and sorted using a database in Lotus Version 2. (Lotus Development, Cambridge, MA). Data were later copied to Excel Version 5 (Microsoft, Toronto, ON). Receiver operator characteristic curves for these indices and the MCV were constructed in Excel using sensitivity and specificity calculations. 30 Cut-off values for the various indices were as follows: For the Mentzler index, <, 2, 3, 4, 5, 7, and > 7, for the Bessman index, < 3, 4, 5, 6, 7, 8, 9 and >9, for the Shine index, <900,,200,,400,,530,,560,,700,,900, and >,900, for the England index, <-4, -3, -, 0,,2, and >2 and for MCV, <57.5,60.0,64.0,68.0, 72.0, 76.0, 78.0, and >78.0. In addition, the estimate of area under each curve as well as comparisons of the areas under the curves were calculated. 3 ' 32 RESULTS Seven hundred eighty-nine patient samples were eligible for the study. Seventy-seven were excluded because of insufficient sample volume for full analysis. Of the 72 patient samples available, 40 were found to have thalassemia minor. Thirty-one of these were diagnosed as 3 thalassemia trait on the basis of a Hb A 2 level greater than Eight were diagnosed as a thalassemia trait based on the detection of the occasional RBC containing Hb H inclusion bodies (most likely a thalassemia cis deletion). 33 One was diagnosed as Hb E trait by detection of Hb E on starch gel electrophoresis, confirmed by HPLC. Six hundred seventy-two patients had no hemo- A.J.C.P.- August 996

3 LAFFERTY ET AL. 203 Mathematical RBC Indices for Thalassemic Microcytosis» _ 72 0 s 76.0 s 78.0! f m V *'ll 0 m J J -- MCV as index man Index a Men Izler Index - - Eng and index Shine and Lai index specificity Fie I. Receiver operator characteristic curves for the various published indices as well as MCV, derived from all the study data s 68 f s 60.0 r i > s »-MCV as index specificity FIG. 2. Receiver operator characteristic curve for MCV as an index derived from all the study data. globinopathy detected. Of this, 33 were categorized as nonanemic, but iron deplete (denned as ferritin < 20 ng L and Hb 2: 5 gl in females or > 35 gl in males), 9 were categorized as iron deficiency anemia (defined as ferritin < 20 ugl and Hb < 5 gl in females or < 35 gl in males). 34 Three hundred forty-eight patients did not fall into any of these categories and were not classifiable. This group would include patients with iron deficiency (with or without anemia and ferritin levels > 20 tgl), anemia of chronic disease, sideroblastic anemia as well as cases of a thalassemia trait, which did not yield positive Hb H inclusion body results (most likely a thalassemia 2 or a thalassemia trans deletion). 33 Receiver operator characteristic curves generated on all 72 patients for each of the indices as well as for MCV are displayed in Figure. Figure 2 displays the ROC curve for MCV alone with the various cut-off values. Table shows the sensitivity of an MCV < 72 as a predictor of the different thalassemia syndromes detected in the study using all patients. DISCUSSION Receiver operator characteristic curves readily demonstrate the utility of a test in the detection of a particular disease state. The greater the area under the curve the more useful, in general, the test will be. A straight diagonal line from the origin of the XY intercept to the upper right corner of the graph represents a useless test with no discriminating power. The MCV, Mentzler index, and the Shine and Lai index produced the best curves and were essentially identical (P <.7). In comparison, the England and Bessman indices performed poorly (P <.005). The advantage of the MCV over the Mentzler and Shine and Lai index is that it is routinely reported by most blood cell analyzers, requires no further calculations, and is already used by laboratory scientists and physicians on a routine basis. In order for a thalassemia index to be useful in selecting microcytic samples for thalassemia testing, it must detect the maximum number of thalassemic patients while at the same time eliminate as many nonthalassemic patients as possible. Therefore, the optimal cut-off point for a thalassemia index is the point on the ROC curve where the combination of sensitivity and specificity is maximal. In the case of MCV this is a cut-off of <72.0 fl (Fig. 2). At this point, a sensitivity of 0.88 and a specificity of 0.84 was achieved. TABLE. SENSITIVITY OF A MEAN CORPUSCULAR VOLUME <. 72 fl AS A PREDICTOR OF THE DIFFERENT THALASSEMIA SYNDROMES DETECTED IN THE STUDY Hemoglobin E trait a Thalassemia trait 3 Thalassemia trait Total n 3 40 Sensitivity Vol. 06 No. 2

4 204 HEMATOPATHOLOGY Based on the data from the study, had this laboratory used an MCV of <72.0 fl as an indication for thalassemia testing, 08 of the 672 non-thalassemic patients and 35 of the 40 thalassemic patients would have been tested, yielding a positivity rate of 25%. Table shows the sensitivity of an MCV < 72 fl in the detection of the different types of thalassemia trait encountered in the study. The sensitivity is essentially identical for a thalassemia trait and 3 thalassemia trait at 0.88 and 0.90, respectively. Insufficient numbers of Hb E trait were detected to yield significant data. The hemoglobinopathies detected in the study reflect the population of the Hamilton region where 3 thalassemia trait is the predominant thalassemia minor syndrome followed by a thalassemia trait. This reflects the multicultural nature of the region's population that contains large Italian and Southeast Asian populations. 35 Conclusion The England and Bessman indices are poor discriminators between thalassemic and non-thalassemic microcytosis and should not be used for this purpose. The Bessman index has a further disadvantage in that the RDW is known to be increased in 5? thalassemia trait 2 and Hb H disease. 3 The Bessman index would therefore fail to detect these conditions. The most advantageous index appears to be an MCV of <72 fl. However, it is not 00% accurate and its effectiveness in the detection of other forms of thalassemia minor (eg, a thalassemia 2, a thalassemia trans deletion, Hb Lepore trait, and Hb E trait) was not determined in this study. Thalassemia testing should be done on all microcytic patients who are not demonstrably iron deficient, who do not respond to iron therapy or are not suffering from chronic disease. Using an MCV < 72.0 fl as a laboratory-based criteria for the selection of samples for thalassemia testing should provide similar sensitivities and specificities regardless of the population. However, in certain centers, the population of individuals presenting with thalassemia trait may be so small that this would not be worthwhile. In many multicultural centres, similar results to the 25% positivity rate predicted by this study should be achievable. Acknowledgments. The authors thank the staff of the Regional Hemoglobinopathy Laboratory at St. Joseph's Hospital, Hamilton and the staff of the Radio-Immunoassay Laboratory at McMaster University Medical Centre for their contributions to this study. The authors also thank Mrs. Millie Danek, Mr. Andrew McFarlane, and Ms. Karen Lanigan for their assistance with this manuscript. REFERENCES. Weatherall DJ. The thalassemias. In: Williams WJ, Beutler E, Erslev AJ, Lichtman MA, eds. Hematology, ed 4. New York: McGraw Hill, 990, pp Shalev O, Yehezkel E, Rachmilewitz EA. Inadequate utilization of routine electronic RBC counts to identify beta thalassemia carriers. Am J Public Health 988; 78: Jandl JH. The thalassemias. In: Blood, Textbook of Hematology. Boston: Little Brown and Co., 987, pp England JM, Fraser PM. Differentiation of iron deficiency from thalassemia trait by routine blood count. Lancet 973; : Mentzler WC. Differentiation of iron deficiency from thalassemia trait. Lancet 973; : Klee GG, Fairbanks VF, Pierre RV, O'Sullivan MB. Routine erythrocyte measurements in diagnosis of iron-deficiency anemia and thalassemia minor. Am J Clin Pathol 976:66: Hedge UM, White JM, Hart GH, Marsh GW. Diagnosis of a thalassemia trait from Coulter counter "S" indices. J Clin Pathol 977;30: Berman BW, Ritchey AK, Jekel JF, et al. Hematology of3 thalassemia trait - age related developmental aspects and intrafamilial correlations. J Pecliatr 980;96: White JM, Richards R, Byrne M, et al. Thalassemia trait and pregnancy. J Clin Pathol 985;38: Shine I, Lai S. A strategy to detect (I thalassemia minor. Lancet 977;: Ghosh A, Woo JSK, Wan CW, ey al. Evaluation of a prenatal screening procedure for ft thalassemia carriers in a Chinese population based on the mean corpuscular volume (MCV). Prenat Diagn 985;5: Bessman JD, Feinstien DI. Quantitative anisocytosis as a discriminant between iron deficiency and thalassemia minor. Blood 979;53: Bessman JD, Gilmer PR, Gardner FH. Improved classification of anemias by MCV and RDW. Am J Clin Pathol 983:80: Bessman JD, Gilmer PR, Gardner FH. Too early to put down RDW for discriminating iron deficiency and thalassemia [Letter]. Am J Clin Pathol 986;86: Johnson CS, Tegos C, Beutler E. Thalassemia minor: Routine erythrocyte measurements and differentiation from iron deficiency. Am J Clin Pathol 983; 80: Roberts GT, El Badawi SB. Red blood cell distribution width index in some hematologic diseases. Am J Clin Pathol 985:83: Qurtom HA, Al-SalehQA, Lubani MM, etal. The value of red cell distribution width in the diagnosis of anemia in children. Eur J Pediatr] 989; 48: Bhambhani K, AronowR. Lead poisoning and thalassemia traitor iron deficiency: The value of the red blood cell distribution width. Am J Dis Child 990; 44: Flynn MM, Reppun TS, Bhagavan NV. Limitations of red blood cell distribution width (RDW) in evaluation of microcytosis. Am J Clin Pathol 986; 85: Helleman PW, Bartels PC. VanWaveren Hogervorst GD. Screening for thalassemia using the width of the Technicon H6000 H60 erythrocyte size histograms. Scand J Clin Lab Invest 988;48: Miguel A, Linares M, Miguel A, Miguel-Borja JM. Red cell distribution width analysis in differentiation between iron deficiency and thalassemia minor. Acta Haemal 988:80: Cesana BM, Maiolo AT, Gidiuli R, et al. Relevance of red cell distribution width (RDW) in the differential diagnosis of microcytic anemias. Clin Lab Haemal 99; 3: Dacie JV, Lewis SM. Red cell cytochemistry. In: Practical Hematology, ed 8. Edinburgh: Churchill Livingstone, 995. pp Efremov GD, Huisman TJH, Bowman K, Wrightstone RN. Schroeder WA. Microchromatography of hemoglobins. A rapid method for the determination of Hb A 2. J Lab Clin Med 975;83: A.J.C.P. August 996

5 LAFFERTY ET AL. 205 Mathematical RBC Indices for Thalassemic Microcytosis 25. Betke K. Marti HR, Schlicht I. Estimation of small percentages of fetal hemoglobin. Nature 959; 84: Huehns ER. Starch gel electrophoresis: Section 4 haemoglobins. In: Smith I. Chromatographic and Electrophoretic Techniques, vol 2. ed 2. Bath: Heinemann, 968, pp Piomelli S. A micromethod for free erythrocyte protoporphyrin: The FEP test. J Clin Med 973; 8: Luxton AW, Walker WHC. Ali MAM. A radioimmuno-assay for serum ferritin. Clin Chem 977;23: Wilson JB, Headlee ME, Huisman THJ. A new high-performance liquid chromatographic procedure for the separation and quantitation of various hemoglobin variants in adults and newborn babies. J Lab Clin Med 983; 02: Fletcher RH, Fletcher SW, Wagner EH. Clinical Epidemiology, the Essentials, ed 2. Baltimore: Williams and Wilkins, 988, pp Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radioloi*v 982;43: Hanley JA, McNeil BJ. A method of comparing areas under receiver operating characteristic curves derived from the same cases. Radiology 983; 48: Thompson CC, Ali M, Boyadjian S, Vacovsky M. Positional effect of cistrans alpha globin gene deletions on the formation of "H" bodies. AmJHemalol 988;3: Fairbanks VF, Beutler E. Erythrocyte disorders: Anemias related to disturbances of hemoglobin synthesis. In: Williams WJ, Beutler E, Erslev AJ, Lichtman MA, eds. Hematology, ed 4. New York: McGraw Hill, 990, pp Ali M, Lafferty J. The clinical significance of hemoglobinopathies in the Hamilton region: A 20-year review. Clin Invest Med 992;5: Vol. 06.No. 2