CE-HPLC Derived P2 and P3-Peaks in Health and in Hb D-Punjab and HbE States

Dear Sir,

Diagnostic interpretation of cation-exchange high-performance liquid chromatography (CE-HPLC) focusses on HbA0, HbF and HbA2 along with any variant peaks identified. However, additional chromatographic peaks are also always present, including the P2- and P3-peaks of post-translationally modified HbA0 [1]. The P2-peak represents glycated Hb and an increase helps detect unsuspected diabetes mellitus. It can also corroborate reticulocytosis if reduced. Rare variant hemoglobins like Hb Hope and Hb K-Woolwich also elute in the P2-window (retention time 1.28–1.50 min). The P3% increases on prolonged storage of blood samples and several hemoglobin variants including J-Meerut, Camden, Austin, Fukuyuma and N-Baltimore elute in the P3-window (retention time 1.50–1.90 min) [1–4]. Only limited information is available in literature on the ranges and variations of these two universal peaks in health and in common hemoglobinopathies.

In order to characterize these often-ignored Hb subsets, we compared the CE-HPLC findings from 103 healthy adult subjects who had normal CE-HPLC patterns with HPLCs of individuals with the HbD-Punjab variant (n = 53) and with HbE (n = 53). All 209 cases were run on the Variant II™ HPLC instrument (Bio-Rad Laboratories, Hercules, USA) using the beta-thal Short™ programme. D and E variants were verified by automated alkaline pH electrophoresis (Genio-S, Interlab, Rome, Italy) and further corroborated by hemogram findings and the ethnic backgrounds of the patients and family studies to distinguish homozygous from double heterozygous states.

We encountered 44 cases of HbD trait, 9 double heterozygous HbD + β thalassemia, 29 HbE trait, 21 double heterozygous HbE + β thalassemia and 3 HbE homozygotes. Mean P2% in controls was 3.8 ± 0.7 (range 2.8–7.8) while HbD-Punjab and HbE showed significantly lower P2% of 1.9 ± 1.0 (range 0.0–4.9) and 1.7 ± 1.4 (range 0–4.2), respectively (Table 1, p < 0.001 by 1-way ANOVA). P2% was inversely proportional to the HbD% and HbE% (Pearson correlation coefficient, r = − 0.854, p < 0.001). Double heterozygotes for HbD Punjab + β-thalassemia showed especially low P2% (mean 0.3%, range 0.0–1.2) as compared to HbD-traits (mean 2.2%, range 1.3–4.9). P2 was undetectable in all the 3 homozygous HbE cases and significantly lower in double heterozygote HbE + β-thalassemia vis-à-vis HbE-trait (p = 0.005, independent samples’ T test). Similarly, mean P3% in healthy persons was 4.3 ± 0.4 (range 3.4–5.8) while in HbD-Punjab and HbE cases it was reduced to 2.1 ± 0.9 (range 0.1–5.3) and 2.7 ± 0.9 (range 1.3–6.3) respectively (p < 0.001 by 1-way ANOVA).

Table 1.

Distributions of P2 and P3 percentages and unknown peaks in healthy individuals and in persons with Hb D-Punjab or HbE variants

Study group		N	HbD-Punjab%/HbE%	P2%	P3%
A.	Healthy individuals	103	–	3.8 ± 0.7 (2.8–7.8)	4.3 ± 0.5 (3.4–5.7)
B.	HbD-Punjab, overall	53	44.6 ± 17.6 (13–87.2)	1.8 ± 0.9 (0–2.9)	2.0 ± 0.9 (0.1–5.3)
	HbD-Punjab trait	44	37.2 ± 6.1 (13.0–40.9)	2.1 ± 0.4 (0.7–2.9)	2.5 ± 0.5 (1.2–5.3)
	HbD-Punjab + β thalassemia	09	80.5 ± 6.8 (54.7–87.2)	0.3 ± 0.4 (0–1.2)	0.5–0.4 (0.1–1.3)
C.	HbE, overall	53	41 ± 19.7 (18.8–81.2)	1.9 ± 1.3 (0.0–4.2)	2.9 ± 0.7 (1.3–6.3)
	HbE trait	29	27 ± 9.3 (18.8- 34.3)	2.6 ± 0.7 (0.4–3.7)	2.4 ± 0.8 (2.3–4.9)
	HbE + β thalassemia	21	55.2 ± 18.1 (38.3–77.6)	1.1 ± 1.5 (0–4.2)	2.8 ± 1 (1.8–4.7)
	HbE homozygous	03	75.5 ± 38 (72–81)	0	5.1 ± 2.7 (2.3–3.9)

All values are as mean ± SD (range)

The results of this small study are significant since HbD-Punjab and HbE are very frequently encountered hemoglobinopathies in several parts of India and south-east Asia [1]. The reduction in P2% in even heterozygous carriers for these variants reinforces that analytical laboratories need to be aware of their potential interferences with CE-HPLC-based assessment of HbA1c levels. Accurate HbA1c estimation forms the cornerstone of diabetes monitoring and levels over time are a major determinant of the risk of diabetic complications [5].

Previous papers too have shown HPLC-based tests to carry a risk of erroneous HbA1c results when HbD or HbE are present. One group studied 42 cases each of HbD-Punjab and HbE trait [6] and another had 27 cases of HbD-Punjab [7], hence our larger numbers build upon the prior data. The likely explanation for the lower P2/P3% in persons with HbD-Punjab or HbE trait is that the lower HbA% in these individuals (usually 45–60%) results in correspondingly lower quantities of modified HbA products/adducts (like glycated Hb). Additionally, in HbE + β-thalassemia patients, the elevated HbF peak tends to acquire a broad base, which may overlap (and therefore interfere) with the accurate quantitation of P2 peak [6, 7].

The normal ranges of P2% and P3% in healthy individuals are not easily available in indexed literature. Our data may aid diagnostic hemoglobinopathy laboratories using this specific instrument in interpreting their results. A major limitation of our study is that we evaluated only one instrument from one manufacturer. Even so, it highlights the fact that laboratories (especially those offering combined hematology and clinical chemistry testing) need to be aware of not just the various diagnostic connotations of the P2 and P3 peaks, but also the potential influences that the presence of common variant hemoglobins like HbD-Punjab and HbE can have on their percentages.

Compliance with Ethical Standards

Disclosure of Potential Conflict of interest

None of the authors have any conflict of interest.

Research Involving Human Participants and/or Animals

This retrospective study involved data collection from the paper records of routinely done investigations in an anonymized manner.

Informed Consent

Waiver of the requirement of informed consent for this paper is sought since the data is completely anonymized, no experimental procedures/tests were done, and the results of the study did not have any impact on the clinical management of the persons involved.