Accounting for artifactually elevated HbA2 in cases of Hb hope when measured by capillary electrophoresis.

To the Editor Panyasai et al1 point out an important problem with the measurement of hemoglobin (Hb) A2. By comparing 11 cases of Hb Hope analyzed by capillary electrophoresis (CE) and high-performance liquid chromatography (HPLC; Bio-Rad β-Thalassemia Short Program, Bio-Rad, Hercules, CA), they document a much greater increase in the HbA2 value measured by CE than by HPLC. They also point out that this increase could lead to an errant interpretation of β-thalassemia trait. The authors suggest that the increase in HbA2 could result from a preferential synthesis of HbA2 because the normal δ chains compete better than the variant Hb Hope β chain for the available α chains. This phenomenon has been reported with other variants.2-4 However, if this were the explanation for the increased HbA2 in the present case, the HbA2 should be increased to a similar extent by HPLC and CE. To explain the considerable difference in HbA2 results between HPLC and CE, the authors hypothesize a coelution of several Hb Hope adducts with HbA2. However, in reviewing the electrophoretogram in their article, we noted that the Hb Hope and HbA overlap in the CE pattern. Because of this, the software shifts the baseline up to approximate the percentage of HbA and Hb Hope. However, while this provides an estimate of the relative amounts of HbA and Hb Hope, it does not account for the overlapping hemoglobin below it in the pattern. Because the percentage of HbA2 is calculated as a percentage of the total hemoglobin, by not accounting for the overlap area, the percentage of HbA2 is falsely elevated. To look at this possibility, we went back to 2 cases of Hb Hope that were diagnosed in our laboratory by CE (Sebia Capillarys2, Sebia, Norcross, GA) and confirmed by HPLC (Ultra2 method, Trinity Biotech, Kansas City, MO). The first was a case of Hb Hope similar to the cases reported by Panyasai et al1 that had an elevated HbA2 of 6.0% ❚Figure 1A❚, whereas by HPLC, the HbA2 was 3.1%. However, when we constructed the baseline to include the area under the 2 peaks, the HbA2 gave a value of 3.4% ❚Figure 1B❚, much more comparable to the HPLC because it was now being calculated as a percentage of the total hemoglobin. To see if HbA2 was elevated in a condition in which there was no overlap with the other β chain product, we examined a case of doubly heterozygous HbS/Hb Hope. As shown in ❚Figure 1C❚, there was no overlap between the products of the 2 β variants. The HbA2 was 3.4%. This consideration complements the important observation that when examined by CE, HbA2 is artifactually elevated in the presence of Hb Hope. We predict that similar artifactual elevations of HbA2 could occur in other variant situations in which a large portion of the major hemoglobins involved are excluded by the baseline. By adjusting the baseline in such cases, a false impression of possible β-thalassemia can be avoided.