Use of an Investigation Algorithm for the Identification of Hemoglobin Variants (HbV) and Hemoglobinopathies (HbP) by High Performance Liquid Chromatography (HPLC) Significantly Reduces the Need for Confirmatory Testing.

Abstract

Alkaline hemoglobin (Hb) electrophoresis, the most widely used investigative tool for the laboratory diagnosis of HbV and HbP, is capable of separating common HbV, such as Hbs A, F, S and C, but Hbs S, D, G and Lepore are unresolved from each other as are Hbs C, A2, O-Arab and E. There are yet other HbV with identical/similar electrophoretic mobilities. Consequently, acid electrophoresis is needed for the identification of these HbV. The identity of HbV, nevertheless, is generally inferred from their electrophoretic mobility, their quantity and the patient’s ethnic background. Hb fraction analysis by HPLC in recent years has been shown to be an attractive alternative. It has the advantage of quantifying Hb F and Hb A2 along with HbV screening and quantification in a single, highly reproducible system. It has been reported by us and others that the retention time (RT), i.e the time in minutes from sample injection to the maximum point of the elution peak, on HPLC of Hbs A, F, A2, S, D, G, Lepore, C, O-Arab and E are all distinct. This laboratory has recently reported its experience with 60,293 samples analyzed for quantification of Hb fractions and screening for HbV using the Bio-Rad Variant II HPLC system. The mean RT, mean percent Hb fraction (%Hb), chromatogram characteristics and electrophoretic mobilities for 39 normal and abnormal Hb fractions were determined. An investigation algorithm was proposed. Here we report the results of a 12-month prospective test of this algorithm in which 18,655 additional specimens were analyzed by HPLC from which an interpretation was made. Confirmatory testing was then performed on detected HbV by appropriate laboratory tests. Comparison of the interpretations based solely on HPLC and after confirmatory testing was analyzed. Abnormal HbV were found in 1588 (8.5%) samples. Hb S (5.8%), Hb C (1.6%), Hb A2′ (0.8%) and Hb E (0.2%) were the most common HbV encountered. Nineteen additional HbV were encountered in varying incidences in 34 samples. Use of the algorithm allowed the correct identification of 18 HbV in 1581 (99.62%) samples prior to confirmatory testing (88.1% by the RT alone and 11.5% samples by the RT and %Hb with peak characteristics providing supportive evidence for the identification of 4). For 5 HbV seen in 7 (0.44%) samples confirmatory testing was necessary for final identification. Three were α-variants in the Hb J family with indistinguishable RTs (1.72–1.74 minutes) and %Hbs (22.6–27.1%) and have no hematological or clinical significance. Two were β-variants with RTs and %Hbs not statistically different (p<0.1, respectively) and no hematological or clinical significance. These data demonstrate HPLC to be a powerful diagnostic tool for the direct identification of HbV suitable for routine laboratory investigation. The integration of a proper algorithm involving RT, %Hb and peak characteristics in HPLC allowed the clinical laboratory to correctly identify ~80% of the HbV encountered in >99% of the samples without the need for expensive and labor-intensive confirmatory testing, a significant financial saving to the laboratory. More importantly, the identification of the common variants (i.e. Hbs C, D-Punjab, E, Hope, Lepore, O-Arab and S) that in combination with Hb S result in a clinically significant sickling disorder was accurately accomplished in 100% of the cases.