Abstract
While high-performance liquid chromatography (HPLC) is well-established for β-thalassaemia and haemoglobinopathies, phenotypic screening for α0-thalassaemia has been limited. To address this limitation, we aimed to translate the discovery of the α-thalassemia early eluting peak (αEEP) in HPLC into clinical practice by comparing its diagnostic performance with other existing methods (haemoglobin H inclusion test [HbHi] and immunochromatographic strip test [ICT]) in a multicentre setting, and elucidating the nature of the αEEP by liquid chromatography-tandem mass spectrometry (LC-MS/MS). With a cohort of 820 genotyped patients, the αEEP showed superior diagnostic performance in detecting --SEA (sensitivity 99.6%, specificity 100%) compared with HbHi (sensitivity 95.8%, P = 0.006; specificity 97.3%, P < 0.001) and ICT (sensitivity 95.8%, P = 0.006; specificity 75.4%, P < 0.001). Both HbHi and ICT showed reduced sensitivity in β-thalassaemia carriers versus non-carriers. ICT showed reduced specificity when Hb F ≥ 1% compared with < 1%. The αEEP remained robust across all subgroups. LC-MS/MS revealed a strong association between the αEEP and embryonic ζ-globin chains (P < 0.001). The αEEP offered cost reductions of 98.6% over HbHi and 97.3% over ICT. Collectively, the αEEP is a highly reliable and cost-effective marker for detecting --SEA carriers, enabling a novel “all-in-one” HPLC screening strategy for --SEA, β-thalassaemia and haemoglobinopathies. Trial registration number: not applicable.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1007/s00277-026-07102-0.
Keywords: Alpha-thalassaemia, Immunochromatographic strip test, Haemoglobin H inclusion test, Alpha-thalassaemia early eluting peak
Introduction
Thalassaemia is one of the most common inherited disorders globally, causing quantitative imbalance of globin chains due to mutations in the α-, β- and δ-globin genes and consequently ineffective erythropoiesis [1]. The α- and β-thalassaemia are particularly prevalent in regions such as Southeast Asia, resulting in significant public health burden [2]. The global prevalence of α- and β-thalassaemia is approximately 5% and 1.5%, respectively [3, 4]. Clinically significant thalassaemia can develop in offspring if both parents are carriers, including β-thalassaemia intermedia or major (both functional β-globin genes lost), haemoglobin (Hb) H disease (three α-globin genes lost), and the fatal Hb Bart’s hydrops foetalis (all four α-globin genes lost).
The α0-thalassaemia carrier screening is critical especially in high-prevalence areas. The Southeast Asian type deletion (--SEA) is the most prevalent mutation in Southeast Asia and the West Pacific region [5–8]. While β-thalassaemia and other haemoglobinopathies are readily detected by high-performance liquid chromatography (HPLC) or capillary electrophoresis, only red cell indices have been recommended for triaging patients for α0-thalassaemia genotyping [9, 10]. Although genotyping remains the gold standard, a reliable and cost-effective phenotypic screening for high-prevalence areas is essential, as red cell indices in iron deficiency can mimic those seen in thalassaemia [11, 12].
Microscopy-based Hb H inclusion test (HbHi) detecting Hb H inclusions (β4 tetramer) is still widely used but is labour-intensive, not standardised and has limited sensitivity [13, 14]. In the last decade, an immunochromatographic strip test (ICT) based on the detection of Hb Bart’s (γ4 tetramer) has been developed, with a reported sensitivity of 97.4% [15, 16].
More recently, the α-thalassaemia early eluting peak (αEEP) has emerged as a potential screening method for --SEA carriers by utilising the mostly-unanalysed early eluting peak pattern in HPLC (Fig. 1A) [17, 18]. We hypothesised that the αEEP had better diagnostic performance compared with existing methods (i.e. ICT and HbHi), and was a biochemically sound marker for diagnostic use. This study endeavours to translate the discovery of the αEEP into clinical practice by conducting a multicentre, assessor blinded, head-to-head diagnostic comparison of the αEEP with HbHi and ICT, and to characterise the nature of the αEEP. Ultimately, we aimed to enable a novel “all-in-one” HPLC screening strategy for --SEA carriers, β-thalassaemia and other haemoglobinopathies (Fig. 1B).
Fig. 1.
(A) The α-thalassaemia early eluting peak (αEEP) assessment involves magnifying the area of interest (red box) in the chromatogram of high-performance liquid chromatography (HPLC). The αEEP appears at a retention time of about 0.24 min, and is interpreted as either ‘positive’ (arrowhead, upper panel) or ‘negative’ (lower panel). (B) The novel “all-in-one” strategy in thalassaemia and haemoglobinopathy screening combines standard haemoglobin analysis (Hb F, Hb A2, Hb variants and Hb H peak) by HPLC with an additional early eluting peak assessment for the αEEP. Results guide further genetic counselling and confirmatory DNA testing as indicated. (C) 820 patients were included in this multicentre diagnostic test comparison for blinded assessment of the αEEP, Hb H inclusion test and the immunochromatographic strip test. The results were compared against genotyping as the gold standard. (D) Early eluates from twenty patients (ten αEEP-positive and ten αEEP-negative) were analysed by liquid chromatography-tandem mass spectrometry to characterise the substance underlying the αEEP
Materials and methods
This multicentre, retrospective diagnostic test evaluation consisted of two major parts (Fig. 1C, D). The first was a multicentre diagnostic comparison of the αEEP, HbHi and ICT conducted at six tertiary referral hospitals in Hong Kong (Tuen Mun Hospital, Queen Mary Hospital, Prince of Wales Hospital, United Christian Hospital, Queen Elizabeth Hospital and Princess Margaret Hospital) providing regional haematology laboratory service with a catchment population of about 6.8 million (90.5% of the population) [19]. The Hb pattern study requests were received for investigation of anaemia, clinically suspected thalassaemia or haemoglobinopathies in daily practice. The second was a liquid chromatography-tandem mass spectrometry (LC-MS/MS) study to investigate the biochemical nature of the αEEP, performed by Tuen Mun Hospital and the Proteomics and Metabolomics Core Facility, Centre for PanorOmic Sciences, Li Ka Shing Faculty of Medicine, the University of Hong Kong. The study was performed in accordance with the Declaration of Helsinki, with ethics approval obtained for all participating centres after review by the Hospital Authority Central Institutional Review Board (Ref. CIRB-2023-174-1), Institutional Review Board of the University of Hong Kong / Hospital Authority Hong Kong West Cluster (Ref. UW 24–098), and The Joint Chinese University of Hong Kong (CUHK) - New Territories East Cluster Clinical Research Ethics Committee (Ref. 2024.113).
The diagnostic comparison was conducted across the six participating hospitals from March to December 2024. Inclusion criteria were: (1) ethylenediaminetetraacetic acid (EDTA)-anticoagulated peripheral blood sample received for Hb pattern study; (2) age ≥ 1 year; and (3) sufficient to perform all the tests: αEEP, ICT, HbHi and α-globin genotyping for common deletional and non-deletional mutations in Chinese (--SEA, 3.7 kb deletion [-α3.7], 4.2 kb deletion [-α4.2], Hb Constant Spring, Hb Quong Sze and codon 30 mutation), as described previously [17, 18]. Patients with Hb H disease or a history of blood transfusion within three months were excluded. Extended genotyping for Filipino (--FIL), Thai (--THAI) and Mediterranean (--MED) deletions modified from Tan et al. [20] was performed for non-Chinese patients with a negative initial α-globin genotyping result. The αEEP (from Variant II HPLC, Bio-Rad Laboratories, California, USA), a commercial ICT kit (i + Med Laboratories, Thailand), and in-house HbHi were assessed by trained, blinded observers (Fig. 1C). Detailed methods including materials, sample preparation, detection and interpretation are provided in the Supplementary Information.
For LC-MS/MS analysis, 10 mL of first-minute HPLC eluates from each of the twenty peripheral EDTA blood samples (ten αEEP-positive, ten αEEP-negative) were collected (Fig. 1D). Unique peptide sequences for haemoglobin subunits were identified via database search during software analysis. Further details are provided in the Supplementary Information.
Statistical analysis
Sample size was determined assuming a 5% discordance rate between ICT or HbHi and αEEP results. A sample size of 700 provided 85% power to detect differences at a two-sided type I error rate of 5%.
Statistical analyses were performed using GraphPad Prism version 10.4.2 (GraphPad Software, California, USA) and SPSS Statistics version 30.0 (IBM Corporation, New York, USA). Equivocal αEEP results were considered negative in the calculation of diagnostic performance due to potential inter-observer variability. The test performance was primarily assessed for --SEA detection based on sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) with 95% confidence intervals (95% CI) derived by the Wilson-Brown method. Differences in sensitivity and specificity between αEEP and HbHi or ICT for detecting --SEA were assessed by McNemar’s test. Subgroup analyses compared test performance in detecting --SEA according to the β-thalassaemia status and Hb F levels (< 1% versus ≥ 1%). Fisher’s exact test was used to evaluate associations between subgroups and test performance, and between the αEEP result and globin chain peptides detected by LC-MS/MS. All hypothesis tests were two-sided, and P values of < 0.05 were considered statistically significant. Costs for αEEP, HbHi and ICT were compared based on total manpower and reagent/consumable costs per 100 tests.
Results
Multicentre diagnostic comparison
Patient characteristics
The characteristics of the 820 patients included are summarised in Table 1. The cohort consisted predominantly of adults aged 18 years or older (691 patients, 84.3%), and most were Chinese (776 patients, 94.6%). 362 (44.1%) carried one or more α-thalassaemia mutations: α0-thalassaemia trait (263, 32.1%); heterozygous α+-thalassaemia (97, 11.8%); and homozygous or compound heterozygous α+-thalassaemia (2, 0.2%) (Fig. 2A). Among the 263 α0-thalassaemia mutations detected, most (259, 98.5%) were --SEA and 4 (1.5%) were --FIL. Furthermore, 186 (22.7%) patients were β-thalassaemia carriers, 3 (0.4%) were δβ-thalassaemia carriers, and 11 (1.3%) had heterozygous or homozygous Hb E. 287 (35.0%) patients had no thalassaemia or haemoglobinopathy detected.
Table 1.
Clinical characteristics of the patients included in the multicentre evaluation
| Total no. of cases | 820 |
|---|---|
| Age (year) - median (range) | 43 (30–63) |
| -1 to < 18 | 129 (15.7%) |
| -18 or above | 691 (84.3%) |
| Female sex - no. (%) | 472 (57.6%) |
| Chinese ethnicity - no. (%) | 776 (94.6%) |
| RBC (×1012/L) - median (IQR) | 5.06 (4.54–5.61) |
| HGB (g/dL) - median (IQR) | 11.5 (10.3–12.7) |
| MCV (fL) - median (IQR) | 71 (65–78) |
| MCH (pg) - median (IQR) | 22.2 (20.5–25.3) |
| MCHC (pg) - median (IQR) | 31.9 (31.1–32.7) |
| HCT (L/L) - median (IQR) | 0.36 (0.324–0.397) |
| RDW (%) - median (IQR) | 15.6 (14.5–17) |
| α-thalassaemia - no. (%) | |
| -Heterozygous α0-thalassaemia --SEA mutation | 259 (31.6%) |
| -Heterozygous α0-thalassaemia --FIL mutation | 4 (0.5%)* |
| -Heterozygous or compound heterozygous α+-thalassaemia mutations | 2 (0.2%) |
| -Heterozygous α+-thalassaemia mutations | 97 (11.8%) |
| β-thalassaemia, δβ-thalassaemia or Hb E - no. (%) | |
| -β-thalassaemia trait | 186 (22.7%) |
| -δβ-thalassaemia trait | 3 (0.4%) |
| -Hb E trait | 9 (1.1%) |
| -Homozygous Hb E | 1 (0.1%) |
| -Hb E trait + Hb Kaohsiung | 1 (0.1%) |
| Other non-thalassaemic Hb variants - no. (%) | |
| -Hb S trait | 1 (0.1%) |
| -Hb G-Taipei | 1 (0.1%) |
| -Hb J-Kaohsiung | 1 (0.1%) |
| -Hb Hekinan II | 1 (0.1%) |
| Hb F level - no. (%) | |
| -<1% | 633 (77.2%) |
| -≥1% | 187 (22.8%) |
Abbreviations: --FIL Filipino deletion, Hb haemoglobin, HGB haemoglobin level, HCT haematocrit, MCH mean cell haemoglobin, MCHC mean cell haemoglobin concentration, MCV mean cell volume, RBC red blood cell count, RDW red cell distribution width, --SEA Southeast Asian deletion
*Extended genotyping performed only in non-Chinese patients
Fig. 2.
(A) The α-globin genotypes of patients included in the multicentre diagnostic test comparison. The Southeast Asian type deletion was the predominant mutation in this cohort (31.6% of patients). (B) The α-thalassaemia early eluting peak (αEEP) showed significantly higher sensitivity and specificity compared with haemoglobin H inclusion test (HbHi) and immunochromatographic strip test (ICT). (C) Liquid chromatography-tandem mass spectrometry (LC-MS/MS) results from an αEEP-positive patient. The unique peptide sequence of the ζ-globin chain (ISTQADTIGTETLER) was detected in the tandem mass spectrometry (upper panel), and the corresponding peak (red) appeared at 33.03 min in liquid chromatography (lower panel). (D) LC-MS/MS showed that free ζ-globin chains were detectable only in αEEP-positive patients and absent in αEEP-negative patients, with no difference observed in other free globin chains. (E) Cost analysis showed that the αEEP method was 98.6% and 97.3% less expensive than HbHi and ICT, respectively
Test performance in detecting different α-globin genotypes
The αEEP, HbHi and ICT results are compared against the α-globin genotype in Table 2. The αEEP was 99.6% sensitive (258/259) and 100% specific (561/561) for --SEA carriers. The αEEP was negative in all four non-Chinese --FIL carriers and was equivocal (analysed as negative in the statistical analysis) in one --SEA carrier (Supplementary Fig. S1). The αEEP showed superior diagnostic performance compared with both HbHi (sensitivity 95.8% [P = 0.006], specificity 97.3% [P < 0.001]) and ICT (sensitivity 95.8% [P = 0.006], specificity 75.4% [P < 0.001]) for detecting --SEA (Fig. 2B, Supplementary Table S1). All the αEEP, HbHi and ICT were insensitive for α+-thalassaemias (0%, 4.0% and 40.4%, respectively) (Supplementary Table S2).
Table 2.
Summary of the αEEP, HbHi and ICT for α-thalassaemia according to the genotype
| α-β Genotypes | αEEP | HbHi | ICT | |||||
|---|---|---|---|---|---|---|---|---|
| Positive | Equivocal | Negative | Occasional | Negative | Positive | Weak Positive | Negative | |
| α0-thalassaemia trait (N = 263) | ||||||||
| --SEA/αHekinan IIα; βAβA | 1 | 0 | 0 | 1 | 0 | 1 | 0 | 0 |
| --SEA/αα; βAβA | 245 | 1 | 0 | 240 | 6 | 162 | 78 | 6 |
| --SEA/αα; β0βA | 11 | 0 | 0 | 6 | 5 | 1 | 5 | 5 |
| --SEA/αα; δβ0βA | 1 | 0 | 0 | 1 | 0 | 1 | 0 | 0 |
| --FIL/αα; βAβA | 0 | 0 | 2 | 2 | 0 | 1 | 1 | 0 |
| --FIL/αα; β0βA | 0 | 0 | 1 | 0 | 1 | 0 | 1 | 0 |
| --FIL/αα; βJ−KaohsiungβA | 0 | 0 | 1 | 1 | 0 | 1 | 0 | 0 |
| Homozygous or compound heterozygous α+-thalassaemia (N = 2) | ||||||||
| αCSα/αQSα; βAβA | 0 | 0 | 1 | 0 | 1 | 1 | 0 | 0 |
| -α3.7/-α3.7; β0βA | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 1 |
| Heterozygous α+-thalassaemia (N = 97) | ||||||||
| -α3.7/αα; βAβA | 0 | 0 | 48 | 0 | 48 | 1 | 14 | 33 |
| -α4.2/αα; βAβA | 0 | 0 | 15 | 1 | 14 | 1 | 7 | 7 |
| -α3.7/αα; βSβA | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 1 |
| -α3.7/αα; β0βA | 0 | 0 | 8 | 0 | 8 | 1 | 1 | 6 |
| -α4.2/αα; β0βA | 0 | 0 | 2 | 0 | 2 | 0 | 0 | 2 |
| -α3.7/αα; βEβKaohsiung | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 1 |
| -α4.2/αα; βEβA | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 1 |
| -α3.7/αα; δβ0βA | 0 | 0 | 1 | 0 | 1 | 0 | 1 | 0 |
| αCodon30α/αα; βAβA | 0 | 0 | 2 | 1 | 1 | 1 | 1 | 0 |
| αCSα/αα; βAβA | 0 | 0 | 3 | 0 | 3 | 0 | 2 | 1 |
| αQSα/αα; βAβA | 0 | 0 | 12 | 2 | 10 | 1 | 7 | 4 |
| αCSα/αα; β0βA | 0 | 0 | 1 | 0 | 1 | 1 | 0 | 0 |
| αQSα/αα; β0βA | 0 | 0 | 2 | 0 | 2 | 0 | 0 | 2 |
| Normal α-globin genotype (N = 458) | ||||||||
| αα/αα; βAβA | 0 | 0 | 287 | 6 | 281 | 4 | 49 | 234 |
| αα/αα; βG−TaipeiβA | 0 | 0 | 1 | 0 | 1 | 0 | 1 | 0 |
| αα/αα; β0βA | 0 | 0 | 160 | 2 | 158 | 4 | 34 | 122 |
| αα/αα; βEβA | 0 | 0 | 8 | 0 | 8 | 0 | 1 | 7 |
| αα/αα; βEβE | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 1 |
| αα/αα; δβ0βA | 0 | 0 | 1 | 0 | 1 | 0 | 1 | 0 |
| Subtotal | 258 | 1 | 561 | 263 | 557 | 182 | 204 | 434 |
| Total | 820 | 820 | 820 | |||||
Abbreviations: αEEP α-thalassaemia early eluting peak, CS Constant Spring, --FIL Filipino deletion, HbHi haemoglobin H inclusion test, ICT immunochromatographic strip test, QS Quong Sze, --SEA Southeast Asian deletion
Subgroup analyses for --SEA detection according to β-thalassaemia status and Hb F levels
Subgroup analyses for the test performance in detecting --SEA stratified by β-thalassaemia status and Hb F levels are summarised in Supplementary Tables S3 and S4.
Notably, sensitivity was significantly lower in β-thalassaemia carriers than in non-carriers for both HbHi (54.6% vs. 97.6%, P < 0.001) and ICT (54.6% vs. 97.6%, P < 0.001). Among the eleven patients with concomitant --SEA and β-thalassaemia trait, four were negative for both HbHi and ICT, one was HbHi-positive/ICT-negative, and one was HbHi-negative/ICT-weak positive. All were αEEP-positive. The PPV for ICT was also significantly lower in the β-thalassaemia subgroup (12.5% vs. 71.6%, P < 0.001). In contrast, αEEP performance was unaffected by β-thalassaemia status (β-thalassaemia carriers vs. non-carriers: sensitivity 100% vs. 99.6% [P = 1]; specificity 100% vs. 100% [P = 1]; PPV 100% vs. 100% [P = 1]; NPV 100% vs. 99.7% [P = 1]) (Supplementary Table S4).
In patients with Hb F ≥ 1%, ICT showed significantly lower specificity (68.6% vs. 78.0%, P = 0.02) and PPV (38.8% vs. 70.9%, P < 0.001) compared with patients with Hb F < 1%. On the other hand, HbHi had reduced PPV when Hb F ≥ 1% compared with < 1% (83.3% vs. 96.0%, P = 0.009). The αEEP remained unaffected by the Hb F levels (Hb F ≥ 1% vs. Hb F < 1%: sensitivity 100% vs. 99.6% [P = 1]; specificity 100% vs. 100% [P = 1]; PPV 100% vs. 100% [P = 1]; NPV 100% vs. 99.8% [P = 1]) (Supplementary Table S4).
LC-MS/MS study results
Upon database search for globin chain peptides detected in LC-MS/MS, unique peptide sequences for the ζ-globin chain were found in all ten αEEP-positive patients (Fig. 2C). In contrast, unique ζ-globin peptide sequences were absent in all ten αEEP-negative cases (P < 0.001; Fig. 2D, Supplementary Table S5). No significant differences were observed in other free globin chains. These findings suggest that the αEEP likely corresponds to ζ-globin chains.
Cost analysis
Among the three tests, HbHi incurred the highest cost, while the αEEP had the lowest cost. The αEEP required only 6 s per case for interpretation and recording of the results, with no additional reagent/consumable cost. ICT had the highest reagent/consumable costs. Overall, the αEEP offered a cost reduction of 98.6% and 97.3% compared with HbHi and ICT, respectively (Fig. 2E, Supplementary Table S6).
Discussion
Thalassaemia screening is important in prevalent regions for identifying potential carriers. Testing is typically requested during investigations for hypochromic microcytic anaemia or as part of the antenatal screening. While carriers are generally asymptomatic, reliable screening is crucial for identifying pregnancies at risk of having offspring with lifelong transfusion dependence or fatal hydrops foetalis [3, 21].
Current antenatal thalassaemia screening guidelines indicate that existing α-thalassaemia screening tests are unreliable, necessitating genotyping for many patients with low mean cell haemoglobin or mean cell volume [9, 10]. This study addresses this challenge by enabling a novel “all-in-one” HPLC screening strategy for --SEA carriers, β-thalassaemia and haemoglobinopathies through the additional analysis of the previously unassessed early eluting peaks.
The sensitivity of the αEEP for detecting --SEA in this study was significantly higher than initially reported [17]. This discrepancy is attributed to limited, non-random sampling for molecular confirmation in the previous study, which overrepresented discordant cases. In contrast, our study evaluated diagnostic performance more accurately, with full genotyping results available in all included cases.
Interestingly, one (0.1%) equivocal αEEP result was observed (Table 1; Supplementary Fig. S1). We adopted a conservative approach due to the potential inter-observer variability, as previously reported [18], resulting in the only false-negative αEEP result. However, an equivocal result should prompt further investigation with genotyping, as it may indicate a hidden αEEP (Supplementary Fig. S1). Further optimisation of the HPLC protocol to better separate the αEEP from other early eluting peaks may also improve the diagnostic performance.
This study revealed a finding of reduced ICT sensitivity in β-thalassaemia carriers, which has not been reported previously. The reduced HbHi sensitivity in β-thalassaemia is well established, as the ameliorated α-/β-globin chain imbalance leads to reduced Hb H (β4) formation [22]. Similarly, β-thalassaemia could result in fewer free γ-globin chains available for Hb Bart’s (γ4) formation, thereby reducing ICT sensitivity. This phenomenon was unrecognised in previous studies due to a lack of concomitant α- and β-thalassaemia cases [16, 23]. Additionally, ICT demonstrated significantly reduced specificity when Hb F was ≥ 1% compared with < 1% (68.6% vs. 78.0%, P = 0.02). Low ICT specificity (45.0%) has been reported in clinical samples with Hb F ≥ 2%, although no comparison arm was available [18]. Our findings also support that ICT is susceptible to interference by elevated Hb F levels, attributable to increased free γ-globin chains and consequently Hb Bart’s formation [18].
In contrast, the αEEP remained accurate in the presence of β-thalassaemia and elevated Hb F levels, likely due to its nature being the embryonic ζ-globin chain confirmed by LC-MS/MS. Although the ζ-globin gene is normally silenced during the developmental globin switch, it is de-repressed in --SEA carriers due to α-globin gene deletion and potential regulatory sequences in cis [24, 25]. As ζ-globin is more stable as monomer than as dimer (ζβ) or tetramer (Hb Portland II, ζ2β2), it is detectable within the early phase of HPLC (containing free globin chains) as an additional early eluting peak (i.e. the αEEP) [26, 27]. Therefore, the αEEP detects --SEA independent of the α-/β-globin imbalance, leading to robustness in cases with β-thalassaemia and elevated Hb F levels. Furthermore, four non-Chinese --FIL carriers were identified, all of whom showed negative αEEP results. This confirms our LC-MS/MS results, as the --FIL mutation involves ζ-globin gene deletion and lacks detectable ζ-globin chains [13]. Collectively, the αEEP is more accurately described as the “free ζ-globin (fζ) peak” to reflect its underlying principle of α-thalassaemia detection.
Our results have significant implications for the αEEP applicability in other ethnicities. The αEEP will likely be highly effective in most Southeast Asia and West Pacific countries where --SEA is the single prevalent double gene deletion, including China, Thailand, Vietnam and Malaysia [5–8, 28]. The αEEP will also be beneficial in countries where other double gene deletions are prevalent in addition to --SEA, such as the Philippines [29]. The αEEP is likely to have limited utility in regions where --SEA is uncommon. As illustrated in this study, the importance of ethnicity information cannot be overemphasised to provide appropriate screening and genetic counselling [9].
The characteristics of the αEEP, HbHi and ICT are compared in Table 3. All three methods are based on distinct mechanisms of α-thalassaemia detection: αEEP detects free ζ-globin chains, HbHi detects Hb H (β4), and ICT detects Hb Bart’s (γ4). Overall, αEEP offers the best diagnostic performance in detecting --SEA, demonstrating robustness in the presence of β-thalassaemia and elevated Hb F levels, alongside significant cost reduction. The receiver operating characteristic curve analysis also showed that the αEEP has the greatest area under curve, and both ‘equivocal’ and ‘positive’ results were found to be predictive for --SEA mutation (Supplementary Fig. S2; Supplementary Table S7). Nevertheless, any equivocal αEEP result should be further investigated with additional confirmatory test due to potential inter-observer variation and limited number of cases encountered in this study. Furthermore, αEEP was shown to be suitable for clinical laboratory use, as it is stable, robust against other clinical interferences, and can be easily trained with high inter-observer agreement [18]. The perfect inter-observer agreement of the αEEP was also confirmed in this study (κ = 1.000) (Supplementary Table S8). This method is also highly scalable, potentially also to resource-limited settings, as HPLC equipment is widely used for thalassaemia and haemoglobinopathy screening. Based on our findings, a comprehensive thalassaemia screening flowchart incorporating the αEEP was proposed (Supplementary Fig. S3).
Table 3.
Summary for the comparison of the αEEP, HbHi and ICT tests
| Test | αEEP (or fζ peak) | HbHi | ICT |
|---|---|---|---|
| Testing principle |
Free embryonic ζ-globin chain |
Hb H (β4 tetramers) | Hb Bart’s (γ4 tetramers) |
| --SEA detection |
99.6% sensitivity 100% specificity |
95.8% sensitivity 97.3% specificity |
95.8% sensitivity 75.4% specificity |
| Effect of β-thalassaemia | Not affected | Reduced sensitivity (reduced α-/β-globin chain imbalance) | Reduced sensitivity (reduced α-/β-globin chain imbalance) |
| Effect of elevated Hb F level | Not affected | Not affected | Reduced specificity (increased γ-globin production) |
| Applicability to other double gene deletions | Limited | Applicable | Applicable |
| Manpower | 6 s per case | Up to 15 min per case | 3 min per case |
| Consumables | No additional cost (if HPLC is used already) |
USD 12.8 per 100 tests |
USD 455.1 per 100 tests |
| Total cost | Minimal | High manpower cost | High consumable cost |
Abbreviations: αEEP α-thalassaemia early eluting peak, fζ free ζ-globin, Hb haemoglobin, HbHi haemoglobin H inclusion test, HPLC high-performance liquid chromatography, ICT immunochromatographic strip test, --SEA Southeast Asian deletion
There are several key strengths of our study. First, it is the largest study to date in which all cases underwent confirmatory genotyping, enabling a head-to-head comparison of newer α-thalassaemia screening tests with impartial assessment by blinded observers. Second, the unbiased inclusion of cases from service laboratories ensured real-world applicability. Notably, Hb H disease cases were excluded from this study as they show a characteristic Hb H peak on HPLC and are easily detected by HbHi. This was to avoid overestimating the tests’ ability to detect the asymptomatic α0-thalassaemia carriers, which was the primary outcome of interest. Third, we elucidated the previously unknown nature of the αEEP, providing a scientific basis for its use as a reliable diagnostic marker. Given its perfect specificity in our cohort, a positive αEEP result is virtually diagnostic of --SEA carrier status. This finding suggests predictable applicability in other ethnic populations. Lastly, our study offers a practical one-stop solution for screening --SEA, β-thalassaemia and haemoglobinopathy carriers by repurposing existing HPLC equipment, thereby streamlining the process. A clear negative αEEP result makes --SEA highly unlikely, overcoming the historical lack of a reliable screening test.
The primary limitation is that this study was conducted in a predominantly Chinese population. Although the applicability of the αEEP is predictable given the well-documented prevalence of α-thalassaemia mutations globally, it is prudent to validate its diagnostic performance in specific regional settings. Since the genotyping study was a targeted panel on common mutations, rare mutations could be missed by the genotyping. Some mutations prevalent in other regions (e.g. --MED and --THAI mutations) were not encountered in our study. While --THAI mutation involves ζ-globin gene deletion and is mechanistically less likely to be αEEP-positive, --MED mutation has intact ζ-globin gene in cis. However, the diagnostic performance of the αEEP in these situations should be confirmed in real patients. In addition, while subgroup analyses suggested that the diagnostic performance of αEEP remained robust, ICT and HbHi were significantly affected by concomitant β-thalassaemia or elevated Hb F levels. However, the number of cases with concomitant --SEA mutations was limited (Supplementary Table S3), leading to wider 95% confidence intervals in sensitivity and PPV calculations (Supplementary Table S4). A larger cohort of patients with these conditions may be needed to better confirm diagnostic performance in these settings. Finally, the applicability of the αEEP in neonates and infants are yet to be validated.
Conclusion
Overall, the αEEP, or free ζ-globin (fζ) peak, is a highly reliable diagnostic marker for --SEA carriers showing superior diagnostic performance compared with existing methods including ICT and microscopy. It enables an “all-in-one” HPLC screening strategy for --SEA, β-thalassaemia and other haemoglobinopathies. This method can be readily evaluated and implemented in clinical laboratories, as the required HPLC instrument is widely available.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
We would like to acknowledge the contributions of the Proteomics and Metabolomics Core Facility, Centre for PanorOmic Sciences, Li Ka Shing Faculty of Medicine, the University of Hong Kong for performing the liquid chromatography-tandem mass spectrometry analysis.
Author contributions
Wing Kit Lam was responsible for conceptualisation, data curation, formal analysis, investigation, methodology, project administration, software, validation, visualisation, writing the original draft, and reviewing & editing the manuscript.Carmen Michelle Yuen, Lawrence Lap Chi Tsui, Ting Hon Stanford Li, Vivian Ka Pik Yeung and Albert Chun Fung Sin were responsible for data curation, investigation, project administration, and reviewing & editing the manuscript.Tsz Fung Wong, Winnie Yim Fong Law, Christina Pui Ying Fan, Lok Nga Ko, Vivian Hoi Kei Woo, Kit Yu Chan, Tsz Ning Chan, Tze Wing Fan, Lok Han Too, Chi Keung Cheng, Man Ling Wong, Aves Hui Hsuan Wu, Benny Man Wai Lit, Yu Fong Wong, Man Wai Chan, Chun Him Ip, Julia Cheuk Ting Leung, Po Chun Wong, Kei Ching Yuen, Wang Ho Yuen, Hoi Ching Wong and Jamilla Wai Yan Li were responsible for investigation, and reviewing & editing the manuscript.Anskar Yu Hung Leung and Sze Fai Yip were responsible for methodology, resources, supervision, and reviewing & editing the manuscript.Joyce Sin Cheung, Natalie Pui Ha Chan, Margaret Heung Ling Ng, Joyce Hoi Yi Kwong, Eudora Yu De Chow, Wai Shan Wong, Kate Fung Shan Leung were responsible for resources, supervision, and reviewing & editing the manuscript.
Funding
None declared.
Data availability
The authors confirm that the data supporting the findings of this study are available within the article and/or its supplemental data.
Declarations
Consent to participate
A waiver of written informed consent was obtained from the local institutional review boards.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Wing Kit Lam, Email: lwk936@ha.org.hk.
Sze Fai Yip, Email: yipsf1@ha.org.hk.
References
- 1.Kattamis A, Kwiatkowski JL, Aydinok Y, Thalassaemia (2022) Lancet 399(10343):2310–2324 [DOI] [PubMed] [Google Scholar]
- 2.Tuo Y, Li Y, Li Y, Ma J, Yang X, Wu S et al (2024) Global, regional, and national burden of thalassemia, 1990–2021: a systematic analysis for the global burden of disease study 2021. EClinicalMedicine 72:102619 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Piel FB, Weatherall DJ (2014) The α-thalassemias. N Engl J Med 371:1908–1916 [DOI] [PubMed] [Google Scholar]
- 4.Kattamis A, Forni GL, Aydinok Y, Viprakasit V (2020) Changing patterns in the epidemiology of β-thalassemia. Eur J Haematol 105:692–703 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Lau YL, Chan LC, Chan YY, Ha SY, Yeung CY, Waye JS et al (1997) Prevalence and genotypes of alpha- and beta-thalassemia carriers in Hong Kong -- implications for population screening. N Engl J Med 336:1298–1301 [DOI] [PubMed] [Google Scholar]
- 6.Chaibunruang A, Prommetta S, Yamsri S, Fucharoen G, Sae-Ung N, Sanchaisuriya K et al (2013) Molecular and hematological studies in a large cohort of α(0)-thalassemia in northeast Thailand: data from a single referral center. Blood Cells Mol Dis 51:89–93 [DOI] [PubMed] [Google Scholar]
- 7.O’Riordan S, Hien TT, Miles K, Allen A, Quyen NN, Hung NQ et al (2010) Large scale screening for haemoglobin disorders in southern Vietnam: implications for avoidance and management. Br J Haematol 150:359–364 [DOI] [PubMed] [Google Scholar]
- 8.Tritipsombut J, Sanchaisuriya K, Phollarp P, Bouakhasith D, Sanchaisuriya P, Fucharoen G et al (2012) Micromapping of thalassemia and hemoglobinopathies in diferent regions of northeast Thailand and Vientiane, Laos People’s Democratic Republic. Hemoglobin 36:47–56 [DOI] [PubMed] [Google Scholar]
- 9.Bain BJ, Daniel Y, Henthorn J, de la Salle B, Hogan A, Roy NBA et al (2023) BSH Committee. Significant haemoglobinopathies: A guideline for screening and diagnosis: A British Society for Haematology Guideline. Br J Haematol 201:1047–1065 [DOI] [PubMed] [Google Scholar]
- 10.HKCOG (2003) Guidelines of antenatal thalassaemia screening. Hong Kong College of Obstetricians and Gynaecologists. Accessed on 27 July 2025. Available on: https://www.hkcog.org.hk/hkcog/Download/Guidelines_of_Antenatal_Thalassaemia_Screening_2003.pdf
- 11.Cao A, Rosatelli MC, Monni G, Galanello R (2002) Screening for thalassemia: a model of success. Obstet Gynecol Clin North Am 29:305–vii [DOI] [PubMed] [Google Scholar]
- 12.Zhou Y, Zhang J, Wang C, Zhou L, Zhou L, Ou D et al (2020) Application of HbA2 levels and red cell indices-based new model in the differentiation of thalassemia traits from iron deficiency in hypochromic microcytic anemia Cases. Int J Lab Hematol 42:526–532 [DOI] [PubMed] [Google Scholar]
- 13.Lafferty JD, Barth DS, Sheridan BL, McFarlane AG, Halchuk LM, Raby A et al (2008) A multicenter trial of the effectiveness of zeta-globin enzyme-linked immunosorbent assay and hemoglobin H inclusion body screening for the detection of alpha0-thalassemia trait. Am J Clin Pathol 129:309–315 [DOI] [PubMed] [Google Scholar]
- 14.Lee SY, Yap ES, Lee EY, Goh JH, Liu TC, Yip C (2019) Evaluation of thalassaemia screening tests in the antenatal and non-antenatal populations in Singapore. Ann Acad Med Singap 48:5–15 [PubMed] [Google Scholar]
- 15.Prayalaw P, Fucharoen G, Fucharoen S (2014) Routine screening for α-thalassaemia using an immunochromatographic strip assay for haemoglobin Bart’s. J Med Screen 21:120–125 [DOI] [PubMed] [Google Scholar]
- 16.Winichagoon P, Kumpan P, Holmes P, Finlayson J, Newbound C, Kabral A et al (2015) Validation of the immunochromatographic strip for α-thalassemia screening: a multicenter study. Transl Res 165:689–695 [DOI] [PubMed] [Google Scholar]
- 17.Lam WK, Law WYF, Wong TF, Woo VHK, Wu AHH, Chow EY et al (2023) A specific early eluting peak in Bio-Rad Variant II cation-exchange high-performance liquid chromatography (CE-HPLC) for the screening of alpha-thalassaemia in a Chinese population. Pathology 55:835–842 [DOI] [PubMed] [Google Scholar]
- 18.Lam WK, Fan CPY, Law WYF, Wong TF, Too DLH, Ko LN et al (2025) Evaluation of stability and potential interference on the α-thalassaemia early eluting peak and immunochromatographic strip test for α-thalassaemia --SEA carrier screening. Clin Chem Lab Med 63:2549–2557 [DOI] [PubMed] [Google Scholar]
- 19.Public healthcare services in Hong Kong East (2025) The Government of the Hong Kong Special Administrative Region. https://www.info.gov.hk/gia/general/202505/07/P2025050700660.htm
- 20.Tan AS, Quah TC, Low PS, Chong SS (2001) A rapid and reliable 7-deletion multiplex polymerase chain reaction assay for alpha-thalassemia. Blood 98:250–251 [DOI] [PubMed] [Google Scholar]
- 21.Taher AT, Musallam KM, Cappellini MD (2021) β-Thalassemias. N Engl J Med 384:727–743 [DOI] [PubMed] [Google Scholar]
- 22.Li D, Liao C, Li J, Xie X, Huang Y, Zhong H (2006) Detection of alpha-thalassemia in beta-thalassemia carriers and prevention of Hb Bart’s hydrops fetalis through prenatal screening. Haematologica 91:649–651 [PubMed] [Google Scholar]
- 23.Bunkall C, Ghallyan N, Elliott C, Van de Water N, Chan G (2018) Evaluation of an immunochromatographic strip test for alpha-thalassaemia screening. Int J Lab Hematol 40:691–696 [DOI] [PubMed] [Google Scholar]
- 24.Tang W, Luo HY, Albitar M, Patterson M, Eng B, Waye JS et al (1992) Human embryonic zeta-globin chain expression in deletional alpha-thalassemias. Blood 80:517–522 [PubMed] [Google Scholar]
- 25.Luo HY, Deisseroth AB, Chui DH (1995) Human embryonic zeta-globin gene expression in mouse-human hybrid erythroid cell lines. Blood 86:1212–1217 [PubMed] [Google Scholar]
- 26.Manning LR, Russell JE, Padovan JC, Chait BT, Popowicz A, Manning RS et al (2007) Human embryonic, fetal, and adult hemoglobins have different subunit interface strengths. Correlation with lifespan in the red cell. Protein Sci 16:1641–1658 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Abdullah UYH, Ibrahim HM, Jassim HM, Salleh MZ, Kek TL, Fakhruzzaman Bin Noorizhab MN et al (2019) Relative proteome quantification of alpha, beta, gamma and delta globin chains in early eluting peaks of Bio-Rad variant II® CE-HPLC of hemoglobin from healthy and beta-thalassemia subjects in Malaysia. Biochem Biophys Rep 18:100635 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Ahmad R, Saleem M, Aloysious NS, Yelumalai P, Mohamed N, Hassan S (2013) Distribution of alpha thalassaemia gene variants in diverse ethnic populations in malaysia: data from the institute for medical research. Int J Mol Sci 14:18599–18614 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Ko TM, Hwa HL, Liu CW, Li SF, Chu JY, Cheung YP (1999) Prevalence study and molecular characterization of alpha-thalassemia in Filipinos. Ann Hematol 78:355–357 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The authors confirm that the data supporting the findings of this study are available within the article and/or its supplemental data.