| Abstract|| |
Introduction: Beta thalassemia trait (BTT) must be differentiated from iron deficiency anemia to avoid unnecessary iron therapy and for the prevention of thalassemia major by genetic counseling. In a tertiary care hospital, it is vital that the screening tool is not only sensitive but also specific so as to be cost effective and save time. Aim: The aim of this study was to evaluate the new Sehgal index and compare it to existing complete blood count-based indices for the best combination of sensitivity and specificity to predict BTT. Materials and Methods: Study was done in 2 phases - Phase 1: A retrospective analysis of 1022 consecutive high-performance liquid chromatography (HPLC) cases from July 2008 to June 2011. Phase 2: A prospective analysis of 973 consecutive HPLC cases from July 1, 2011 to June 10, 2013 was done to confirm the results of Phase 1 and the applicability of the new Sehgal index. Results: Prevalence of BTT was 28.8% (294/1022) and 25.39% (247/973) in Phase 1 and Phase 2, respectively. Receiver operating characteristic-area under the curve and Youden index was highest for new Sehgal index, followed by Mentzers index <14. The prospective study shows results similar to those in Phase 1 confirming the superiority of the above two indices. Conclusion: Sehgal index and Mentzers index <14 showed the best combination of sensitivity and specificity in predicting BTT. The best indices or combination can be used as a "validated flag rule" in the analyzer middleware program in a hospital for identifying suspected cases of BTT.
Keywords: Beta thalassemia trait, complete blood count, iron deficiency anemia, mentzers index, Sehgal index
|How to cite this article:|
Sehgal K, Mansukhani P, Dadu T, Irani M, Khodaiji S. Sehgal index: A new index and its comparison with other complete blood count-based indices for screening of beta thalassemia trait in a tertiary care hospital. Indian J Pathol Microbiol 2015;58:310-5
|How to cite this URL:|
Sehgal K, Mansukhani P, Dadu T, Irani M, Khodaiji S. Sehgal index: A new index and its comparison with other complete blood count-based indices for screening of beta thalassemia trait in a tertiary care hospital. Indian J Pathol Microbiol [serial online] 2015 [cited 2019 Aug 25];58:310-5. Available from: http://www.ijpmonline.org/text.asp?2015/58/3/310/162862
| Introduction|| |
Iron deficiency anemia (IDA) and beta thalassemia trait (BTT) are two of the most common causes of microcytic anemia.  It is essential to differentiate between the two, so as to avoid unnecessary iron therapy which is contraindicated in beta thalassemia and secondly, for the prevention of beta thalassemia major by genetic counseling. Through genetic counseling birth rate of thalassemia major can be reduced by as much as 90%. 
Beta thalassemia minor is the most common form of thalassemia.  Most homozygotes suffer from a severe syndrome and require regular blood transfusions to survive. There are over 250 million carriers of thalassemia in the world with higher prevalence in some parts of the world (Mediterranean regions up to 8%, countries in the Middle East up to 10%, India 3-15% and South East Asia up to 9%) which can potentially be a major public health problem. 
Routine testing to differentiate between IDA and BTT include: Complete blood count (CBC), serum iron, serum ferritin, total iron binding capacity (TIBC), bone marrow iron stores, levels of HbA2, free erythrocyte protoporphyrin, and zinc protoporphyrin levels. Despite their usefulness, these tests are often expensive and time-consuming.  Routine electronic red blood cell (RBC) counts and other CBC-based indices derived from modern blood cell analyzers can be rapidly obtained and are inexpensive. Various CBC-based indices (discriminant factors [DF]) like Mentzers Index, Green and King (G and K) index, Shine and Lal (S and L) index etc., have been published and evaluated for screening of BTT. ,, These indices can be used as preliminary screening tools which can help distinguish between IDA and BTT and to allow "reflex" HbA2 analysis when a proper cut-off is chosen.  Red cell indices have often been used for initial screening of couples before marriage in high-risk areas. 
In India, screening camps for Thalassemia are conducted by nongovernment organizations and health organizations at schools, colleges, and community centers. The aim of screening camps is to identify individuals with suspected BTT based on CBC based indices followed by confirmatory analysis. For the same, a screening index should have highest level of sensitivity so as not to miss any BTT individuals as far as possible. However, many of the highly sensitive indices have a poor specificity especially in our country where IDA is extremely common leading to additional confirmatory analysis and higher costs.
In a tertiary level health care center, it is important to identify suspect BTT individuals from the routine CBC workload. As this involves additional high-performance liquid chromatography (HPLC) testing for confirmation of BTT, it is important that a screening index is not only sensitive but also very specific to make it cost effective. The ideal scenario would be if the CBC analyzer itself would flag suspected BTT individuals based on the CBC indices and middleware formulas. The aim of this study was to compare the Sehgal index (new index) with the existing CBC based indices (DF's) to identify the index having the best combination of sensitivity and specificity to predict BTT, using HPLC as a gold standard.
| Materials and Methods|| |
All cases included in this study were patient samples sent to hematology lab for suspected hemoglobinopathies. Patients were diagnosed to have BTT on HPLC (BIORAD-D10) based on high HbA2 levels (>3.6). CBC was run on Sysmex XE 2100 analyzer for all HPLC samples. Both the instruments were maintained and run as per manufacturer instructions.
The study was done in two phases. Phase 1 was a retrospective analysis of 1022 cases whereas Phase 2 was a prospective analysis of 973 patients as shown below.
The exclusion criteria for the study were:
- All patients <1-year of age
- All data with mean cell volume (MCV) >120fl.
A retrospective analysis of 1022 consecutive cases was done from July 2008 to June 2011. Iron studies were available for 316 patients of whom 192 patients had Iron Deficiency based on low serum iron levels, increased TIBC, low transferrin saturation and low ferritin. Based on this information, the patients were divided into 3 categories as shown below. All hemoglobinopathies other than BTT were noted for prevalence but excluded from further analysis in the categories shown below.
- Category A: Cases with normal iron status (124 cases out of 316 cases).
- Category B: Cases with Iron deficiency (192 cases out of 316 cases).
- Category C: All cases (irrespective of iron status) (1022 cases).
Data entries and statistics were performed using Microsoft Excel (Microsoft Office, Redmond, Washington state, USA, version 14) and MedCalc (Ostend, Belgium, version 12). The data were adapted and calculated to fulfill the formulas of the DF's (CBC based indices) shown in [Table 1].
|Table 1: List of various discriminant factors (CBC based indices) used in the study|
Click here to view
All the DF's were evaluated to calculate sensitivity, specificity, receiver operating characteristic-area under the curve (ROC-AUC) and Youden Index. For the Sehgal index, the cut-off of 972 was calculated using the ROC curves and then using the same cut-off, above calculations were done. The idea was to identify the CBC-based index (DF) with the best combination of sensitivity and specificity. As the above are essentially screening indices, sensitivity >85% was desirable.
It was a prospective analysis of 973 consecutive cases from July 2011 to June 2013. Inclusion and exclusion criteria were similar to Phase 1. Patents iron status was not sought, and all cases were taken for evaluation. The aim of the Phase 2 study was to verify the DF (CBC based-index) considered as best in Phase 1 and to evaluate the indices prospectively.
Sehgal index is a new index used in this study, and it was calculated as follows - Sehgal index = MCV × MCV/RBC.
| Results|| |
We evaluated 1022 cases and 973 cases of suspected hemoglobinopathies in Phase 1 and Phase 2, respectively. The age, male to female ratios and incidence of BTT and other hemoglobinopathies in the Phase 1 and 2 are shown in [Table 2]. The incidence of BTT in both phases of the study was similar being 28.76% and 25.39%, respectively.
In Phase 1 of the study, there were three categories of patients. Category A was patients with normal iron status. This category had 124 cases of BTT, which were pure BTT without IDA. Category B was patients with IDA, and there were 192 cases of BTT with concomitant IDA. Category C was all cases of clinically suspected BTT irrespective of iron status. The various CBC indices/DFs were evaluated for all three categories separately, and the same is shown in [Table 3], [Table 4], [Table 5], respectively.
|Table 3: Evaluation of all DF's for category A patients – pure BTT (124/316)|
Click here to view
|Table 4: Evaluation of all DF's for category B patients – BTT with concomitant IDA (192/316)|
Click here to view
|Table 5: Evaluation of all DF's for category C patients – all 1022 cases|
Click here to view
As seen in [Table 3], in Category A patients, Sehgal index and Mentzer's index <14 had the highest combination of sensitivity and specificity with ROC-AUC values of 0.88 and 0.85, respectively. The RDW-CV value and S and L index had very high sensitivity but very poor specificities.
[Table 4] shows a similar analysis for patients with BTT and concomitant IDA (Category B). In this scenario, the RDW Index had the highest ROC-AUC value of 0.84. Mentzers index <13 had better ROC values as compared to Mentzer's <14 in this group. Though Sehgal index had a lower ROC value compared to RDW index, it still had a high sensitivity of 89.4%. The RDW value of 14 as a cut-off and S and L index had 100% sensitivity but had an extremely poor specificity of 8% and 16%, respectively.
[Table 5] shows analysis of all the indices for all patients (category C-1022 cases) irrespective of iron status. This is an important group to evaluate as this is how most patients will come to a laboratory with no definite knowledge of the iron status. Sehgal index, Mentzer's <14 and RDW index were the top three performing indices with ROC values of 0.81, 0.80, and 0.79, respectively. As seen in earlier categories, RDW value of 14 and S and L Index had highest sensitivities but poor specificity.
Phase 2 of the study was a prospective analysis of 973 patients suspected to have BTT. They were evaluated as a single group irrespective of iron status. This group was similar to category C patients of Phase 1. The Phase 2 prospective study shows results [Table 6] similar to those in the retrospective study. The Sehgal index was noted to have the highest combination of sensitivity and specificity with ROC value of 0.82 and Youden index of 64.57%. Second to this was the Mentzer's index <14 that had a ROC value of 0.80 and Youden index of 61.63%. RDW value of 14 and S and L Index again showed highest sensitivity with poor specificity.
|Table 6: Evaluation of all the DF's for Phase 2 of the study – 973 cases|
Click here to view
| Discussion|| |
In the Indian population, IDA and BTT are common causes of microcytic hypochromic blood picture. , The real danger of nondiagnosis and misdiagnosis of carriers of thalassemia trait is potential homozygous offspring. Appropriate screening, detection of patients, and counseling of couples at risk are the most important measures for the reduction of morbidity and mortality.  A discriminant formula or CBC-based index with a combination of high sensitivity and specificity for detecting BTT would be a very useful tool in the investigation of microcytic hypochromic anemia in countries and areas where nutritional deficiencies and thalassemia are present with higher prevalence.
In Phase 1 (retrospective), we found BTT in 294 of 1022 patients (i.e., 28.76%). In Phase 2 (prospective), we found a similar incidence of BTT with 247 BTT patients out of 973 patients (25.39%). Madan et al. found the prevalence of BTT in school children as 4.05%.  The incidence in this study is much higher than the prevalence rate of BTT because of the referral bias as it was a hospital-based study done on patients who were referred for HPLC studies, whereas the comparative study was community based with an expected lower incidence rate.
Hemoglobinopathies other than BTT were found in 50 patients (4.89%) in Phase 1 and 59 patients (6.06%) in Phase 2. Both phases showed a similar pattern with Hb S being the most common abnormality followed by HbE, HbD-Punjab, HbQ-India, and HPFH.
Rathod et al.  did a similar study on 200 suspected cases of BTT. They also evaluated pure BTT and patients of BTT with IDA and found S and L as the best index with high sensitivity of 99% and 95% in the two categories, respectively. However, even in their study in the group of patients having BTT with IDA they showed poor specificity. S and L index showed high sensitivity in index study (Phase 1-95%, Phase 2-97%) also but had very poor specificity, thus not being a cost effective test in a hospital set up.
Zahid et al.  did a study on 50 cases of diagnosed BTT and found that S and L index could correctly identify BTT in 92% cases. However, they could not study the specificity of the indices as they did not evaluate all cases of the microcytic hypochromic picture.
Madan et al.  studied 658 (126: BTT with IDA, 337: Pure BTT, 195: IDA, 40: Normal) cases of HPLC. They found Meltzer's index <14 to have a BTT detection rate of 88.7% which is comparable to the current study (Phase 1-83%, Phase 2-85%).
AlFadhli et al.  studied 153 cases of Microcytic hypochromic anemia (IDA: 56, BTT-47, alpha thalassemia trait: 50). They found England and Fraser (E and F) index to be the best in identifying BTT. In contrast, we found E and F to be very low on sensitivity (46.80%) but had high specificity (96.20%) with low Youden index (42.90). The difference between these studies and present study could possibly due to the high incidence of concomitant IDA in our study subset. In addition Suad et al. also had a big number of alpha thalassemia patients in their study group which was not evaluated in the index study.
Urrechaga et al.  studied 90 healthy adults (45 males and 45 females) and 270 diagnosed cases of beta thalassemia. They found G and K to have the best AUC of 0.99%. In contrast in pure BTT patients (category A of Phase 1) of the index study G and K index had a ROC AUC of only 0.73 compared to the highest ROC for Sehgal index of 0.88.
Yeo et al.  studied antenatal samples and used an MCV cut-off <80 as the first screening tool followed using the S and L index as the next screening tool as it had a high sensitivity rate and very high negative predictive value. When S and L index was not predictive of BTT, HPLC was not done to make screening procedure cost effective.
As seen above, none of the formulas and parameters can be 100% sensitive and specific, and various studies have shown different results showing one formula to be better than the other. There are many reasons for the discrepancies with the main being the difference in the population subsets evaluated, the prevalence of nutritional deficiency anemia in the population studied and different yardsticks used for evaluation of formulas. In addition, most studies have used HbA 2 levels for identification of BTT and its well-known that there are silent BTT cases with normal HbA 2 levels. Furthermore, the prevalence of concomitant alpha thalassemia and other Hb disorders may not have been extensively evaluated in most studies.
In this study, Sehgal index and Meltzer's Index <14 had the highest Youden index in both the phases of the study (Phase 1-62 and 59.9, Phase 2-64.5 and 61.6, respectively). A high Youden index indicates a good combination of sensitivity and specificity implying its utility for screening of BTT in a tertiary hospital setup. The Sehgal index (MCV × MCV/RBC) by squaring the MCV value essentially utilizes the power of MCV to separate out BTT cases more effectively. Instead of using a two-step screening process of first identifying a case with low MCV and then applying the Mentzers index (MCV/RBC); the Sehgal index converts it into a one-step screening process.
If screening of BTT using high sensitivity was the only goal, S and L index or simply using a MCV cut-off , or RDW cut-off would serve the purpose for advising further tests for BTT screening camps, even if it was at rate of comprising on specificity and over testing by HPLC so as to not to miss any BTT. However, these indices have low specificity especially in areas with high prevalence of IDA and are not ideal to meet the needs of a tertiary care hospital which would aim to pick up BTT and at the same time not to over test and economize their resources.
This study found that the new index "Sehgal index" and MI <14 met the requirements of both high sensitivity and high specificity to predict BTT irrespective of iron levels in a large number of patients studied both retrospectively and prospectively. As the next step, we have implemented the Sehgal index as a "validated flag rule" in our new CBC analyzer (XN-2000, Sysmex) middleware program for identifying suspected cases of BTT. This has drastically improved the detection rate for BTT cases in our routine practice. Using one of the above-mentioned indices as a "validated flag" in the current generation of CBC analyzers will be definitely helpful for tertiary care centers with a high incidence of BTT.
| Acknowledgments|| |
KS designed the study, analyzed data and wrote the manuscript. PM helped in study design, collected the study data, analyzed the data, and wrote the manuscript. TD helped in Study design, data analysis and edited the manuscript. MI collected the study data and analyzed the data. SK helped in data analysis and edited the manuscript.
| References|| |
Urrechaga E, Borque L, Escanero JF. The role of automated measurement of RBC subpopulations in differential diagnosis of microcytic anemia and ß-thalassemia screening. Am J Clin Pathol 2011;135:374-9.
Zahid H, Naumaan M, Chughtai AS. Diagnostic significance of red cell indices in beta-thalassemia trait. Biomedica 2005;21:129-31.
AlFadhli SM, Al-Awadhi AM, AlKhaldi D. Validity assessment of nine discriminant functions used for the differentiation between Iron deficiency anemia and thalassemia minor. J Trop Pediatr 2007;53:93-7.
Rathod DA, Kaur A, Patel V, Patel K, Kabrawala R, Patel V, et al.
Usefulness of cell counter-based parameters and formulas in detection of beta-thalassemia trait in areas of high prevalence. Am J Clin Pathol 2007;128:585-9.
Trivedi DP, Shah HA. Discriminant functions in distinguishing beta thalassemia trait and Iron deficiency anemia: The value of the RDW-SD. Internet J Hematol 2010;7. [Last accessed on 2015 May 22].
Madan N, Sikka M, Sharma S, Rusia U, Kela K. Red cell indices and discriminant functions in the detection of beta-thalassaemia trait in a population with high prevalence of Iron deficiency anaemia. Indian J Pathol Microbiol 1999;42:55-61.
Yeo GS, Tan KH, Liu TC. The role of discriminant functions in screening for beta-thalassaemia traits during pregnancy. Singapore Med J 1995;36:615-8.
Dr. Kunal Sehgal
Department of Laboratory Medicine, P.D. Hinduja National Hospital and Medical Research Centre, Hematology Laboratory, Mumbai, Maharashtra
Source of Support: None, Conflict of Interest: None
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6]