Background

The hallmark of laboratory test for the diagnosis of hereditary spherocytosis (HS) has been osmotic fragility (OF) test. However, OF test gives false negative results in 10% - 20% of HS patients and false positive results in autoimmune hemolytic anemia (AIHA) patients. Currently, the eosin 5-maleimide (EMA) binding test and hypertonic cryohemolysis (HCH) test have been proposed as screening tests in recently published guidelines for the diagnosis of HS (Br J Haematol 2012;156:37-49). And the flow cytometric OF (FC OF) test is recently developed as an assay that can replace the classical OF test. In this study, we evaluated the performance of 3 screening tests (EMA binding test, FC OF test, and HCH test) for the diagnosis of HS.

Methods

A total of 151 patients [32 for HS, 40 for AIHA, 40 for anemia of chronic disease (ACD), 39 for iron deficiency anemia (IDA)] and 140 normal controls (NCs) were enrolled in this study. EMA binding test, FC OF test, and HCH test were performed separately in each patient. Reference ranges (mean ± 1.96SD or 2.5th – 97.5th percentile) for each test were determined from 140 NCs. The cutoff value in EMA binding test (5th percentile) was estimated from 119 anemia patients (AIHA+ACD+IDA), and those in FC OF test (mean -1.65SD) and HCH test (95th percentile) were calculated from 140 NCs. Sensitivity, specificity, negative predictive values (NPV), positive predictive values (PPV), and accuracy of each test were calculated using these cutoff values and compared. Subsequently, receiver operating characteristics (ROC) analysis were performed to determine the best cutoff values of each test for the discrimination of HS from other anemia and area under curve (AUC) of three tests were compared.

Results

The reference ranges of EMA binding test, FC OF test, and HCH test were determined as 86.9% - 118.68%, 6.6% - 71.6%, and 1.5% - 11.8%, respectively, and estimated cutoff values for the diagnosis of HS were calculated to be< 89.4% (EMA binding test),< 8.0% (FC OF test), and > 9.8% (HCH test). In the EMA binding test, HS patients showed significantly decreased EMA binding to RBC (median 71.3%) than AIHA (median 102.0%, P< 0.001), ACD (median 103.7%, P< 0.001), and IDA (median 100.5%, P< 0.001) patients. In the FC OF test, HS patients also showed significantly decreased residual RBC% (median 3.0%) than AIHA (median 44.6%, P< 0.001), ACD (36.0%, P< 0.001), IDA (median 59.4%, P< 0.001) patients and NCs (median 25.2%, P< 0.001). In the HCH test, HS patients showed significantly increased hemolysis (median 10.2%) than AIHA (median 3.6%, P< 0.001) and ACD (median 4.2%, P< 0.001) patients and NC (median 4.7%, P< 0.001), but IDA patients (median 16.4%) showed significantly increased hemolysis than HS patients (P = 0.027). The sensitivity/specificity/NPV/PPV/accuracy of three tests for the discrimination of HS from other anemia using estimated cutoff values mentioned above were calculated as 96.9%/95.0%/99.1%/83.8%/95.4% for EMA binding test, 84.4%/96.6%/95.8%/87.1%/94.0% for FC OF test, and 50.0%/73.1%/84.5%/33.3%/68.2% for HCH test. When the best cutoff values (≤ 85.0% for EMA binding test, ≤ 10.8% for FC OF test, > 7.1% for HCH test) obtained from ROC analysis are applied, both EMA binding test (AUC 0.996) and FC OF test (AUC 0.986) showed satisfactory performance for the discrimination of HS from other anemia. However, HCH test (AUC 0.703) showed significantly poor performance compared than EMA binding test (P< 0.001) and FC OF test (P< 0.001).

Conclusion

Despite the current guideline recommends EMA binding test and HCH test as screening tests for the diagnosis of HS, our present results demonstrated that both EMA binding test and FC OF test possess satisfactory performance for the diagnosis of HS but HCH test has relatively poor performance compared to the EMA binding test or FC OF test and failed to discriminate HS patients from IDA patients. These results partly contradict the recommendations from HS guideline and therefore, both EMA binding test and FC OF test is recommended as screening tests for the diagnosis of HS, but HCH test is not recommended.

Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution