Interference of Hemoglobin N-Baltimore on Measurement of HbA1c Using the HA-8160 and HA-8180 HPLC Methods

Hemoglobin (Hb) variants may interfere with the HbA1c assay and can cause an incorrect diagnosis and/or inappropriate treatment. ¹

Here we report 2 type 2 diabetes cases who are known to carry the Hb N-Baltimore variant ² and whose HbA1c values were low. This structural variant (β95 Lys→Glu) does not cause any erythrocyte abnormality and therefore patients with this Hb variant are clinically asymptomatic.

HbA1c was measured using the ADAMS HA-8160 and HA-8180 high-performance liquid chromatography (HPLC) methods (A. Menarini Diagnostics, Florence, Italy), in diabetic mode and calibrated in accordance with the International Federation of Clinical Chemistry (IFCC) Reference Measurement Procedure. Both HPLC methods resulted in low HbA1c values, which were inconsistent with the Fasting Plasma Glucose (FPG) levels of the patients (Table 1). These findings concurred with our work published in 2013; the determination in that case was performed only using the HA-8160 HPLC method. ²

Table 1.

Blood Test Levels Corresponding to Patients (A and B) With the Hb N-Baltimore Variant.

Patient	Glucose (mg/dL)	HbA1c HA-8160 (mmol/mol) (%)	HbA1c HA-8180 (mmol/mol) (%)	HbA1c Cobas (mmol/mol) (%)
A	205; 267; 196	39 (5.7)	48 (6.5)	93 (10.7)
B	322; 297; 232	51 (5.9)	48 (6.5)	95 (10.8)

Abbreviations: Glucose, fasting blood glucose concentrations over the previous 6 months; HbA1c, hemoglobin A1c; HA-8160, HPLC ADAMS A1c HA-8160 of A.Menarini Diagnostic; HA-8180, HPLC ADAMS A1c HA-8180 of A.Menarini Diagnostic; Cobas, Tina-quant Hemoglobin A1c Gen 3 in a Cobas 8000 of Roche Diagnostics.

The third method for the HbA1c measurement used in this study was the Immunoturbidimetric Tina-quant^® HbA1c Gen. 3 assay, in a Cobas^® 8000 (Roche Diagnostics, Rotkreuz, Switzerland). The measurements, taken in accordance with IFCC standardization, revealed HbA1c values that were consistent with the high FPG concentrations obtained over the previous 6 months in these patients (Table 1).

The analysis of the HA-8160 HPLC chromatograms from the 2 patients revealed an abnormal pattern. It consisted of a slight elevation following the characteristic HbA1c peak that elutes at approximately 60 seconds; no warning signal was evident. Further analysis of the HA-8180 HPLC chromatograms from the 2 patients revealed that no anomalies were present. There was no extra peak, no delay in the appearance and width of the HbA1c peak, and no warning signal.

As described above, the determination of HbA1c using the HA-8160 HPLC in patients with Hb N-Baltimore produces an extra peak in the chromatogram, ² and it is known that this abnormal pattern is useful for detecting potential interferences in HbA1c and potential Hb variant patients. However, no unusual peak was observed in the chromatogram obtained by HA-8180. Nonetheless, the results of HbA1c by HA-8180 without an extra peak led to undervaluation. They are not consistent with glucose levels, as occurs when the determination is performed by HA-8160. The results with both HPLC methods were not consistent with the FPG concentration of the patient, although the immunoassay was.

Although it was initially thought that the routine inspection of chromatograms enables detecting the presence of Hb variants that are clinically silent, further studies should be conducted. However, our present results using the HA-8180 HPLC method show no discernible additional peak in the chromatogram. This finding concurs with other published works that support that not all HPLC methods for HbA1c determination have the resolution necessary to differentiate the >700 Hb variants described. ³

We believe that a turbidimetric immunoassay Tina-quant HbA1c Gen. 3 assay, which is less likely to be subject to interference from mutations occurring further from the N-terminus of the Hb β-chain, can be an alternative method in known Hb N-Baltimore variant carrier patients.