The application of next-generation sequencing technology in congenital haemolytic anaemia: a systematic review and meta-analysis

Abstract

Congenital haemolytic anaemia (CHA) poses diagnostic challenges due to genetic heterogeneity. This systematic review and meta-analysis evaluates the diagnostic efficacy of next-generation sequencing (NGS) in CHA. We systematically searched PubMed and Web of Science until April 2025. Inclusion criteria encompassed studies applying NGS (whole-exome sequencing [WES], whole-genome sequencing [WGS], clinical exome sequencing [CES], or targeted panels) in confirmed/suspected CHA patients. Pooled positive detection rates with 95% confidence intervals (CIs) were calculated using a random-effects model. Subgroup analyses stratified by family history and disease subtypes were performed. Study quality was assessed via modified STARD criteria. Ten studies involving 885 patients were included. The pooled positive detection rate of NGS was 44.3% (95% CI: 32.4–56.3%, p < 0.001). Subgroup analysis revealed significantly higher detection rates in patients with family history (51.0%; 95% CI: 32.8–69.2%) versus sporadic cases (16.9%; 95% CI: 8.4–27.2%, p < 0.001). Disease-specific yields varied: red cell membrane disorders showed the highest rate (45.3%; 95% CI: 35.2–55.7%), followed by enzymatic disorders (26.7%; 95% CI: 18.8–35.3%). Among all positive cases, pathogenic variants in five core genes accounted for 76.82% of detected mutations: SPTB (25.06%), PKLR (17.10%), ANK1 (11.94%), SLC4A1 (11.48%), and SPTA1 (11.24%). SPTB and ANK1 mutations were most frequently identified in red cell membrane disorders, while PKLR variants were exclusive to enzymatic disorders. NGS demonstrates substantial diagnostic utility in CHA, resolving nearly half of cases overall and over 50% of familial presentations. Its efficacy is particularly pronounced in red cell membrane disorders linked to cytoskeletal genes (SPTB, ANK1, SPTA1). These findings support integrating NGS into first-line CHA diagnostics, with prioritization of core gene panels for cost-effective implementation.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10238-025-01896-5.

Introduction

Congenital haemolytic anaemia (CHA) encompasses a spectrum of hematological disorders arising from genetic factors that cause anomalies in red blood cell membranes, enzymatic activities, or haemoglobin structure, ultimately resulting in the premature destruction of red blood cells [1, 2]. Epidemiological data demonstrate that, although the incidence of this condition exhibits geographical variability worldwide, it poses a significant threat to patients' quality of life and life expectancy, particularly among neonates and children. In these populations, CHA can lead to developmental delays, organ dysfunction, and potentially life-threatening complications [3, 4]. Traditional diagnostic approaches primarily depend on hematological evaluations, such as complete blood counts, blood smears, and osmotic fragility tests of red blood cells, alongside Sanger sequencing aimed at identifying specific gene mutations [2, 5]. However, these conventional methods are hampered by limitations, including low throughput, protracted testing durations, and challenges associated with the identification of unknown pathogenic genes [6]. Consequently, many patients encounter the difficulties of ambiguous or inaccurate diagnoses.

In recent years, next-generation sequencing (NGS) technology has transformed the diagnostic landscape for monogenic hereditary disorders due to its high throughput, exceptional sensitivity, and relatively low cost [7, 8]. NGS has revolutionized the detection of novel genes associated with CHA, addressing a critical limitation of traditional diagnostic approaches that rely on targeted Sanger sequencing—which is restricted to known pathogenic genes and thus fails to identify variants in uncharacterized or understudied genetic loci. By enabling high-throughput, unbiased sequencing of entire exomes, genomes, or expanded gene panels, NGS has facilitated the discovery of previously unrecognized CHA-associated genes and expanded the mutational spectrum of known genes. For instance, a study assessed 268 patients with haemolytic anaemia and identified pathogenic and likely pathogenic variants clearly responsible for the disease phenotype in 64/268 (24%) cases, discovering that half of the variants were novel mutations [9]. Additionally, multi-gene analysis modified the original clinical diagnosis in 45.8% of patients with congenital anaemia [10]. Such discoveries not only fill gaps in our understanding of CHA’s molecular etiology—including the genetic basis of atypical or idiopathic cases—but also improve diagnostic yields for patients who would otherwise remain undiagnosed, underscoring NGS’s unique value in advancing the characterization of CHA’s genetic landscape. Nonetheless, substantial variability exists in the application strategies, detection scope, and diagnostic efficacy of NGS across different studies, highlighting a notable deficiency in systematic reviews that summarize the clinical value of this technology [11–13].

In this context, the present study employs systematic reviews and meta-analyses to comprehensively synthesize existing literature evidence. The objectives are to elucidate the role of second-generation gene sequencing technology in the diagnosis of congenital haemolytic anaemia, assess its advantages and limitations relative to traditional diagnostic modalities, and furnish evidence-based medical insights that support rational decision-making regarding testing strategies in clinical practice, thereby promoting precision medicine within this domain.

Materials and methods

Literature search strategies

This systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. A multi-database search approach was employed to comprehensively collect relevant literature. The search databases included PubMed and Web of Science, with the time range set from the establishment of each database to April 2025. The search strategy employed the following Boolean operators: (“next-generation sequencing” OR “whole-exome sequencing” OR “whole-genome sequencing” OR “sequence analysis”) AND (“congenital haemolytic anaemia” OR “hereditary haemolytic anaemia”). A total of 2,961 results were retrieved from Pubmed, and 110 results were identified from Web of Science. No additional relevant studies were found in the reference lists of the included studies. The search and literature screening process is shown in Fig. 1.

Fig. 1.

The flow chart for the study

Literature selection criteria

Inclusion criteria: (1) study subjects were suspected or diagnosed with congenital haemolytic anaemia; (2) usage of gene sequencing methods and provision of related gene testing information; (3) clinical studies in English; (4) explicit reporting of both the total number of patients and the number of positive patients detected. Exclusion criteria: (1) non-human studies; (2) duplicated publications; (3) reviews, conference abstracts, letters, case reports, or other literature where original data could not be obtained; (4) studies with incomplete data that could not be supplemented by contacting the authors.

Data extraction and quality assessment

Two researchers independently conducted literature screening and data extraction. In case of disagreement, the issue was resolved through discussion or consultation with a third researcher. The extracted data included: first author, publication year, study region, sample size, study design type, next-generation gene sequencing technology platform, range of genes tested, diagnostic criteria, number of positive and negative results, follow-up time, etc. For continuous variables, the mean and standard deviation were extracted; for categorical variables, the frequency and percentage were extracted. Additionally, data related to the diagnostic efficacy of next-generation gene sequencing technology in the studies were extracted for subsequent meta-analysis. The modified Standards for Reporting of Diagnostic Accuracy (STARD) [14] for this project was used for the quality assessment of included studies.

Statistical analysis

Meta-analysis was conducted using Stata 17.0. A meta-analysis of the detection rate of CHA gene testing was performed using the single-group ratio method. Effect size (ES) and 95% confidence interval (CI) were calculated using a random effects model. Subgroup analysis was used to identify the detection rates of genetic testing in different populations and different diseases. The heterogeneity among the included studies was assessed using the chi-square test and the I² statistic to evaluate the magnitude of heterogeneity. If I² ≤ 50% and P ≥ 0.1, indicating low heterogeneity among studies, a fixed-effects model was used for the meta-analysis; if I² > 50% or P < 0.1, a random-effects model was employed to explore the sources of heterogeneity and conduct a pooled analysis.

Result

Characteristics of included studies

A total of 10 studies [11, 12, 15–22], encompassing 885 patients, were included in this meta-analysis. Among these, ten studies conducted familial analyses across various centres and countries. The studies employed diverse sequencing methodologies, including next-generation sequencing panel, whole-genome sequencing (WGS), targeted DNA sequencing, and clinical exome sequencing (CES). Detailed characteristics of the included studies are summarized in Table 1.

Table 1.

Characteristics of included studies

Study	Country	Included cases	Sequencing approach	Mutated genes
Chiguer et al. [11]	Moroccan	4	Exome sequencing, Sanger sequencing and qPCR to confirm	SPTB, SLC4A1, EPB42, ANK1, SPTA1
Songdej et al. [12]	Thai	21	Whole-exome sequencing	SPTB, ANK1, KLF1
Mansour-Hendili et al. [15]	France	40	Whole exome sequencing	ANK1, SLC4A1, SPTB, SPTA1, TRPV4, ADAR
Del Orbe Barreto et al. [16]	Spain	10	Personal genome machine (PGM) sequencing	SPTB, ANK1, SLC4A1, EPB41, GPI, HBB
Agarwal et al. [17]	USA	17	Next-generation sequencing panel	SPTB, SLC4A1, ANK1, SPTA1, α-LEPRA
Fermo et al. [18]	Italy	122	Targeted next-generation sequencing	SPTB, SLC4A1, ANK1, SPTA1, G6PD, PIEZO1
Choi et al. [19]	Korea	29	Targeted next-generation sequencing	ANK1, SPTB, G6PD, SPTA1, HBA1/HBA2, PKLR
Jamwal et al. [20]	India	43	Targeted sequencing using sequencing panel	G6PD, GPI, PKLR
Isik et al. [21]	Turkey	143	Clinical exome panels and whole-exome sequencing	SPTB, PKLR, ANK1, SLC4A1, HBA1/HBA2/HBB
Agarwal et al. [22]	USA	456	Targeted next-generation sequencing	SPTA1, SPTB, PKLR, ANK1, G6PD, SLC4A1, PIEZO1, EPB41, TP1

Primary outcome analysis

This study involved 885 cases, with 427 cases diagnosed as positive for NGS. Using a random effects model, the pooled positive detection rate was 44.3% (95% CI 0.324–0.563, p < 0.001) (Fig. 2). In addition, subgroup analyses based on family history and different disease types were performed in this study. In patients with a family history, the detection rate could be up to 51.0% (95% CI 0.328–0.692, p < 0.001), compared with 16.9% (95% CI 0.084–0.272, p < 0.001) in patients without a family history (Fig. 3). This study also chose as grouping criteria the types of diseases commonly mentioned in the included studies, including red cell membrane disorders, red cell enzymatic disorders, and haemoglobinopathies. The overall detection rate for these three types of diseases was 26.7% (95% CI 0.188–0.353, p < 0.001), with a detection rate of 45.3% (95% CI 0.352–0.557, p < 0.001) for red cell membrane disorders (Fig. 4). A total of 152 CHA-related genes were examined, and the most significant pathogenic mutations were mainly involving SPTB (25.06%), PKLR (17.10%), ANK1 (11.94%), SLC4A1 (11.48%), and SPTA1 (11.24%) (Fig. 5).

Fig. 2.

The overall detection rate of CHA

Fig. 3.

The detection rate of CHA in different populations

Fig. 4.

The detection rate of different types of CHA diseases

Fig. 5.

Gene types

Quality assessment of included studies

Quality assessment was conducted using the modified Standards for Reporting of Diagnostic Accuracy (STARD) for this project, and the results showed that the quality of the study was high (Supplementary Fig. 1). Since this meta-analysis was a single-group ratio analysis with descriptive results, publication bias assessment was considered unnecessary.

Discussion

This analysis consolidates evidence from ten studies encompassing 885 patients with CHA. The results demonstrate that NGS achieves a pooled diagnostic yield of 44.3% in identifying pathogenic variants underlying CHA. Notably, the detection rate varies significantly based on family history and disease subtype. The high-quality evidence and consistency across diverse sequencing methodologies underscore the robustness of these findings, positioning NGS as an effective tool in CHA diagnosis.

Molecular diagnostics for CHA were previously limited to targeted Sanger sequencing of genes most likely to undergo mutations or analysis of previously established familial mutations. The large number of genes involved in CHA and their composition significantly reduce the cost-effectiveness of Sanger sequencing. In contrast, the advantages of NGS include the ease of sequencing multiple genes, making this method a commonly used approach in diagnostic algorithms for these complex polygenic diseases [23, 24]. Recent studies have demonstrated its practicality in CHA diagnosis [12, 21, 22]. This study found that the combined positive detection rate of NGS for CHA was 44.3%, confirming its important value as a molecular diagnostic tool. Historically reliant on phenotypic assays (e.g., osmotic fragility tests, enzyme activity assays), traditional methods often fail to resolve genetically heterogeneous disorders [25]. This meta-analysis confirms that NGS resolves nearly half of clinically suspected CHA cases, enabling precise molecular diagnoses where previous methods yielded inconclusive results. The statistically significant heterogeneity reflects real-world variability in patient selection and sequencing platforms but does not diminish NGS’s clinical utility. Rather, it underscores the need for standardized implementation to maximize diagnostic efficiency. However, the wide confidence interval suggests significant heterogeneity among studies, which may stem from differences in sequencing strategies and the genetic complexity of the disease [26].

The subgroup analysis results showed that the detection rate for familial cases was 51.0%, while that for sporadic cases was only 16.9%. Patients with a family history are more likely to follow Mendelian inheritance patterns (such as autosomal dominant or recessive inheritance) [27]. Pathogenic gene mutations often exhibit familial clustering, and the mutation sites are relatively well-defined, making them easier to target and identify using NGS. This highlights the strong genetic nature of CHA and supports the use of NGS as the primary testing method for affected relatives. Additionally, a stratified analysis by disease type indicated that NGS detection rates were highest for red blood cell membrane diseases, which is consistent with the single-gene inheritance pattern and clearly identified pathogenic genes associated with this class of diseases [28]. The SPTB, ANK1, and SPTA1 genes, which were found to have high-frequency mutations in this study, are all key components of the red blood cell membrane skeleton [29]. Mutations in these genes directly lead to a decrease in membrane structural stability and trigger haemolytic anaemia [30]. NGS, through targeted or WES, can efficiently capture missense and deletion mutations in the coding regions of these genes, enabling precise diagnosis. In contrast, overall detection rates were lower for haemoglobinopathies, which may be related to the polygenic regulatory mechanisms of some haemoglobinopathies, such as promoter mutations. The pathogenic mechanisms of haemoglobinopathies involve both coding region mutations and abnormalities in non-coding regulatory elements [31]. NGS is sensitive in detecting coding region variants, but there are still technical bottlenecks in the coverage and interpretation of non-coding regulatory sequences, which may lead to the misdiagnosis of some cases. Combined with phenotypic analyses such as haemoglobin electrophoresis and high-performance liquid chromatography (HPLC), NGS can clearly identify mutation types, guiding genetic counselling and prenatal diagnosis. For example, in patients with sickle cell anaemia, NGS can accurately detect the HBB gene c.20A > T (p.Glu6Val) mutation, providing molecular evidence for family screening [32].

In the present study, pathogenic variants in five genes—SPTB, PKLR, ANK1, SLC4A1, and SPTA1—accounted for the majority of diagnoses. Previous study [15] has shown that all hereditary spherocytosis (HS) patients have mutations in known HS genes: SPTB, ANK1, SLC4A1, and SPTA1. All identified variants were classified as potentially pathogenic or pathogenic loss-of-function mutations. SPTB emerged as the most common mutated gene overall (25.06% of all positive variants) in this study, predominantly driving red cell membrane disorders—consistent with its role as a key component of the spectrin-ankyrin cytoskeleton complex, where missense or frameshift mutations disrupt membrane stability. ANK1 ranked third (11.94%), with variants concentrated in familial cases of HS. This was consistent with previous finding: 38 Chinese patients with suspected HS found that all patients had mutations in the ANK1 and SPTB genes [33]. Gallagher et al. [34] reported in 2019 on a group of 24 patients with recessive HS explored via WES and WGS, all of whom had mutations in SPTA1. Our findings, combined with previously published results, suggest that these genes may serve as marker genes for HS. Moreover, PKLR mutations (17.10%) were the second most common pathogenic variants after SPTB mutations, with a high frequency in sporadic cases, reflecting that most PKLR variants follow a recessive inheritance pattern. The PKLR gene, as the main causative gene for pyruvate kinase (PK) deficiency, was frequently detected in this study and became a key target in enzyme diseases [35]. PK is a key rate-limiting enzyme in the glycolysis pathway. Mammals have two PK genes: PK muscle (PKM) and PK liver/RBC (PKLR). PKLR gene controls the expression of both PK-L (liver), PR-R (RBC) by two tissue specific promoters. The mutations in PKLR can lead to insufficient ATP production in red blood cells, resulting in chronic haemolytic anaemia [36]. Predicting enzyme activity residual levels based on mutation types may provide a basis for developing treatment strategies for clinical conditions, such as splenectomy and blood transfusion. The mutation rates for SLC4A1 (11.48%) and SPTA1 (11.24%) were lower but consistent, both of which were key variants contributing to red blood cell membrane defects. Collectively, the concentrated prevalence of these five genes—covering 76.82% of all detected pathogenic variants—reinforces their utility as anchors for targeted NGS panels, as their prioritization can capture most clinically relevant mutations while reducing sequencing cost and interpretive complexity.

This study provides evidence-based support for the application of NGS in CHA, particularly for patients with a positive family history or suspected monogenic disorders. Emerging evidence from real-world clinical practice highlights that NGS not only improves diagnostic yield for CHA but also delivers tangible clinical benefits by streamlining the diagnostic pathway and optimizing treatment decision-making [37]. Regarding the time to diagnosis, traditional diagnostic workflows for CHA often result in prolonged diagnostic delays, with some patients experiencing months to years of uncertainty due to genetic heterogeneity and overlapping phenotypes. In contrast, NGS-based approaches (especially targeted panels or WES) consolidate multi-gene testing into a single assay, reducing the average time to molecular diagnosis from weeks to days [38]. In terms of treatment decision-making, NGS-derived molecular diagnoses directly inform personalized therapeutic strategies and avoid inappropriate interventions. For instance, identification of PKLR mutations confirms pyruvate kinase deficiency, guiding clinicians to prioritize supportive care (e.g., transfusion protocols tailored to avoid iron overload) or emerging targeted therapies instead of empirical treatments [39]. Similarly, detection of pathogenic variants in SPTB or ANK1 supports timely consideration of splenectomy in severe cases, while ruling out these variants in sporadic hemolysis prevents unnecessary surgical risk [40]. Additionally, NGS results facilitate precise genetic counseling for families, enabling informed reproductive decisions and proactive monitoring of at-risk relatives—an often underrecognized but critical component of CHA management [41]. In resource-limited regions, targeted panels focusing on genes associated with membrane diseases and enzyme disorders may represent a cost-effective strategy. Collectively, these clinical benefits underscore that NGS’s value extends beyond diagnostic accuracy, directly impacting patient outcomes by reducing diagnostic uncertainty, minimizing ineffective interventions, and aligning treatment with underlying molecular pathology.

Undeniably, the present study is subject to some limitations. First, it did not undertake separate analyses for rare subtypes. Given that the genetic makeup and clinical characteristics of rare subtypes may differ significantly from those of more common variants, the findings derived from the existing analysis may not adequately reflect the diagnostic efficacy of NGS within these specific populations, thereby limiting the comprehensiveness of the conclusions. Second, the clinical interpretation of sequencing results is contingent upon the completeness of relevant databases; however, this study did not evaluate strategies for addressing various genetic variants, which may ultimately impact diagnostic accuracy. Furthermore, due to the limited WGS-related data, reliable subgroup analyses cannot be conducted, nor can effective comparisons be made between WGS and WES regarding diagnostic yield, sensitivity, or specificity. Future research efforts should prioritize the establishment of specialized databases for rare CHA subtypes through multi-center collaboration, as well as conduct separate analyses to assess the diagnostic efficacy of NGS in these subtypes. Additionally, critical issues related to the interpretation of genetic variants should be meticulously explored, including evaluations of how the integrity of different databases, the criteria used for variant classification, and validation strategies influence diagnostic accuracy. Also, given the potential of WGS to detect structural variants, non-coding region mutations, and novel genes that may be missed by WES (particularly relevant for atypical or unexplained CHA cases), future studies should prioritize incorporating larger cohorts of WGS data through multi-center collaborations. Such efforts would enable a comprehensive evaluation of how WGS and WES perform relative to each other in CHA diagnosis, providing more targeted guidance for clinical sequencing strategy selection.

Conclusion

NGS demonstrates significant diagnostic utility in CHA, with robust performance in identifying pathogenic variants in membrane disorders and familial cases. While technical and methodological challenges persist, standardized NGS protocols and expanded genomic profiling hold promise for advancing CHA precision medicine. Clinicians should prioritize NGS for genetically suspected cases to enable timely molecular diagnosis and personalized care.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Yanling Zhong proposed the research approach and determined the meta-analysis theme and research questions. Shuning Xie conducted literature searches, screening, data extraction and verification, and wrote the first draft of the article.

Funding

None.

Data availability

No new data were generated during this study. All analyzed datasets are publicly available and cited appropriately.

Declarations

Conflict of interest

The authors declare no competing interests.

Ethical approval

All included original studies were approved by the relevant institutional ethics committees.

Informed consent

All included original studies obtained informed consent from patients or their legal representatives, in accordance with the requirements of the Declaration of Helsinki and relevant ethical guidelines.