Abstract
With the advent of high performance liquid chromatography (HPLC) accurate and precise measurement of hemoglobin A2 has been available for over 20 years. When microcytosis is present an elevated A2 is used as a surrogate marker for identification of beta thalassemia carriers. This combination is a very useful screening test. In contrast, A2 measurement in the low range is used to a lesser extent in screening for alpha thalassemia, delta chain variants and delta thalassemia. In the absence of structural variants A2 measurement by HPLC is accurate in the normal, high, and low ranges. This accuracy however does not hold true when samples harbor structural variants, particularly in the presence of Hb S as there are glycated and degraded fractions of S that co-elute with A2, thus falsely elevating the A2 value. In addition, with E and “D” type variants there is no A2 result or a falsely decreased value because A2 partially or completely co-elutes with the variant. In contrast to HPLC, capillary zone electrophoresis (CZE) separates the hemoglobin fractions according to their electrophoretic mobility with an alkaline buffer. We compared HbA2 measurement by high resolution (HR) HPLC (Primus Diagnostics) and CZE (Sebia CapillaryS) in samples without structural variants in the normal, high, and low ranges and in samples containing common structural variants. The 112 normals were from healthy, hematologically normal adult volunteers. The 20 samples in the high A2 range did not include structural variants and all had microcytosis. The 42 low A2 samples were from individuals without microcytosis in whom isoelectricfocusing (IEF) was also performed to identify delta chain variants. The heterozygotes for S, C, E and “D” had adult phenotypes and based on the percentages of the variant did not appear to have alpha thalassemia. In the “D” group there were 6 D-Punjab, 1 Osu-Christiansburg, and 1 Korle-Bu. G-Philadelphia was excluded as it occurs in association with alpha thalassemia. The results of our comparison study are shown in Table I. Besides what is indicated in the table we identified 9 individuals in the low A2 group that had delta chain variants identified on CZE but silent on HR-HPLC. These delta variants were confirmed by IEF. In summary, in the absence of structural variants these data indicate excellent correlation between HR-HPLC and CZE in the normal, high and low A2 ranges. In the low A2 range CZE has an added advantage over HR-HPLC in detecting delta variants. In the C-Trait group both methodologies appear equivalent but as expected in the S-Trait, E-Trait, and “D” Trait groups the CZE value appears to be more accurate. Compared to the normal group, however, there is a slight positive bias in S-Trait and E-Trait. In conclusion both methodologies appear complementary and can be used in combination for greater accuracy in hemoglobin identification and quantification of fractions.
Parameter . | N . | HR-HPLC Mean . | HR-HPLC 1 SD . | CZE Mean . | CZE 1 SD . | y . | R value . |
---|---|---|---|---|---|---|---|
Normal range A2 (1.8–3.5) | 112 | 2.6 | 0.17 | 2.6 | 0.15 | 1.02 | 0.90 |
High range A2 > 3.6 | 20 | 5.1 | 0.59 | 5.0 | 0.5 | 0.98 | 0.94 |
Low range A <1.8 | 42 | 1.4 | 0.25 | 1.5 | 0.29 | 1.08 | 0.91 |
S Trait | 32 | 4.0 | 0.25 | 3.1 | 0.18 | 0.76 | 0.66 |
C trait | 15 | 3.1 | 0.28 | 3.1 | 0.3 | 0.98 | 0.66 |
E Trait | 12 | 2.0 | 0.55 | 3.4 | 0.41 | 1.52 | 0.38 |
“D” Trait | 8 | N/A | N/A | 2.9 | 0.18 | N/A | N/A |
Parameter . | N . | HR-HPLC Mean . | HR-HPLC 1 SD . | CZE Mean . | CZE 1 SD . | y . | R value . |
---|---|---|---|---|---|---|---|
Normal range A2 (1.8–3.5) | 112 | 2.6 | 0.17 | 2.6 | 0.15 | 1.02 | 0.90 |
High range A2 > 3.6 | 20 | 5.1 | 0.59 | 5.0 | 0.5 | 0.98 | 0.94 |
Low range A <1.8 | 42 | 1.4 | 0.25 | 1.5 | 0.29 | 1.08 | 0.91 |
S Trait | 32 | 4.0 | 0.25 | 3.1 | 0.18 | 0.76 | 0.66 |
C trait | 15 | 3.1 | 0.28 | 3.1 | 0.3 | 0.98 | 0.66 |
E Trait | 12 | 2.0 | 0.55 | 3.4 | 0.41 | 1.52 | 0.38 |
“D” Trait | 8 | N/A | N/A | 2.9 | 0.18 | N/A | N/A |
Author notes
Disclosure: No relevant conflicts of interest to declare.