Bias-corrected diagnostic performance of the naked eye single tube red cell osmotic fragility test (NESTROFT): An effective screening tool for β-thalassemia

Abstract

It is being increasingly recognized that a majority of the countries in the thalassemia-belt need a cost-effective screening program as the first step towards control of thalassemia. Although the naked eye single tube red cell osmotic fragility test (NESTROFT) has been considered to be a very effective screening tool for beta-thalassemia trait, assessment of its diagnostic performance has been affected with the reference test- and verification-bias. Here, we set out to provide estimates of sensitivity and specificity of NESTROFT corrected for these potential biases. We conducted a cross-sectional diagnostic test evaluation study using data from 1563 subjects from Central India with a high prevalence of beta-thalassemia. We used latent class modelling after ensuring its validity to account for the reference test bias and global sensitivity analysis to control the verification bias. We also compared the results of latent class modelling with those of five discriminant indexes. We observed that across a range of cut-offs for the mean corpuscular volume (MCV) and the hemoglobin A2 (HbA2) concentration the average sensitivity and specificity of NESTROFT obtained from latent class modelling was 99.8 and 83.7%, respectively. These estimates were comparable to those characterizing the diagnostic performance of HbA2, which is considered by many as the reference test to detect beta-thalassemia. After correction for the verification bias these estimates were 93.4 and 97.2%, respectively. Combined with the inexpensive and quick disposition of NESTROFT, these results strongly support its candidature as a screening tool-especially in the resource-poor and high-prevalence settings.

Is NESTROFT sufficient for mass screening for β-thalassaemia trait?

“…Our research suggests that NESTROFT is sufficient for mass screening of b-thalassaemia trait, based on the following observations: (a) all the classification trees generated from the derivation subset invariably included NESTROFT as the first choice; (b) the classification tree provided a marginal improvement in false-negative error rate in the derivation subset and no improvement in the validation subset; (c) NESTROFT is indicated for a wide range of prevalence of the b-thalassaemia trait, and especially so when the prevalence is high; 11 (d) the screening performance of NESTROFT is further enhanced because HbA 2 concentration is only indicative of, but not a gold standard test for, b-thalassaemia trait 11 and (e) the cost of conducting NESTROFT is substantially less than the cost of automated cell counting. In a clinical setting, the information given by the haematological parameters is unquestionably useful; [32][33][34] however, for the purposes of large-scale population-level screening for b-thalassaemia trait, there appears little, if any, advantage in investing in automated cell counters in countries where the prevalence of b-thalassaemia is high and the resources for thalassaemia control are limited.…”

“…The sociodemographic characteristics of the subjects are described in detail elsewhere. 11,18,19 We estimated the RBC count, total haemoglobin concentration, haematocrit, MCV, MCH and MCH concentration (MCHC), using an automated cell counter, conducted NESTROFT and quantified haemoglobin A 2 (HbA 2 ) concentration as described previously. 11 All subjects in whom the band for HbA 2 was not detectable on electrophoresis were considered normal, and those with quantified HbA 2 concentration exceeding 3.5% of total haemoglobin concentration were considered to have a raised HbA 2 concentration.…”

“…[6][7][8][9] Recent studies have shown that the Naked Eye Single Tube Red cell Osmotic Fragility Test (NESTROFT) can be a very useful screening tool for b-thalassaemia trait. [10][11][12] However, Old et al 13 argue that the success of NESTROFT in large-scale screening programmes is lessened because of the associated false-negative error rate. Because of the high costs of treating b-thalassaemia major cases 14 that could result from the b-thalassaemia carriers missed by NESTROFT, it has been suggested that information from the haematological parameters output by automated cell counters should be combined with the results of NESTROFT to improve the yield from screening strategies for b-thalassaemia trait.…”

Hemoglobin F levels influence the results of NESTROFT: replication in two surveys

“…In a more recent study of 5045 blood donors reporting to Department of Transfusion Medicine, Bhopal, we could detect HbF levels in 378 subjects. The details of these two datasets have been described elsewhere (Mamtani et al. , 2006 and Chatterjee et al.…”

A Stopped Flow System with Hydrodynamic Injection for Red Blood Cells Osmotic Fragility Test: Possibility for Automatic Screening of Beta-thalassemia Trait

“…However, further study with a higher number of samples would be needed for determination of the exact cut off value to discriminate the two groups. Intensive statistical evaluations such as those conducted by Mamtani et al 25 compared to other gold standard techniques such as HbA2 quantification should be further studied.…”