Prenatal Screening and Diagnosis of ß-Thalassemia in India: Is ARMS-PCR Enough?

Abstract

Accurate and timely prenatal diagnosis of thalassemia is cornerstone to the success of thalassemia control; currently parents are screened for ß-thalassemia mutations by ARMS-PCR and subsequently chorionic villus sampling is done. We did an audit to ascertain whether the present design is adequate and determined the role of sequencing for pre-natal diagnosis of beta-thalassemia. This was a retrospective analysis of prenatal testing data collected over 10 years, (2010–2019). ARMS-PCR was done to identify the beta-globin mutation followed by CVS wherever indicated. Data was classified into 3 groups:—5 most commonly occurring mutations (group 1), less common mutations (group 2) and mutations not detected (group 3). Total number of cases studied were 2128. Mean age of the cohort was 29.30 years (range 18–48 years). Approximately 90% individuals had one of the 5 common mutations in decreasing order of frequency: IVS 1-5 G>C (1297/2128); Codon 26G>A/HbE (451/2128); codon 30G>C (69/2128); codon 15G>A (61/2128); FS 41-42–CTTT (48/2128). Undetected mutations amounted to 7.3% (156/2128). Mean haemoglobin was highest in the group 2 (12.46 g/dl) followed by the group 1 (11.20 g/dl) and least in group 3 (10.99 g/dl). MCV, MCH and MCHC showed similar trends. ANOVA on all these parameters, except RDW, within groups and for individual mutations, were statistically significant (p < 0.001). The hemogram-HPLC-ARMS-PCR-CVS approach is a cost-effective and established method but tends to miss out a considerable number of thalassemia mutations (~7%), emphasizing the role of sequencing in difficult cases. This needs to be addressed while formulating guidelines for thalassemia screening in future.

Introduction

Population screening for carrier status and timely prenatal diagnosis of thalassemia is cornerstone to the success of the thalassemia control programme in any country. ß-Mutation detection or ß-genotyping in couples at-risk is an essential pre-requisite to prenatal diagnosis. For pre-natal diagnosis, mutations should ideally be detected before conception, but in India most couples seek help in the post-conception period. Current practice is to screen the parents for ß-thalassemia mutations by amplification refractory mutation system—polymerase chain reaction (ARMS-PCR) using a panel of 5 or 6 common mutations prevalent in the Indian sub-continent, [1, 2] followed by extended panel of 30 mutations, if the former doesn’t yield results. A chorionic villus sampling (CVS) is done subsequently for foetal ß-genotyping, usually at 10–12 weeks of gestation. For couples arriving late for testing, amniocentesis for foetal ß-genotyping is an alternative method [3]. However, all these techniques have shortcomings. Very few centres in India have the facility for pre-natal diagnosis of thalassemia. Institute of Hematology and Transfusion Medicine, Kolkata, is one of these centres, doing pre-natal diagnosis of thalassemia for the last 10 years. We did an audit of the data to ascertain whether the present design is adequate for pre-natal diagnosis of thalassemia.

Patients and Methods

This was a retrospective analysis of data collected over a period of 10 years, (January 2010–December 2019), from 2128 individuals (1064 couples) who attended our institute for prenatal diagnosis of thalassemia. As ours is a referral cum nodal centre for thalassemia in the region, most couples were referred to us from elsewhere, with either history of a prior birth of a thalassemic child, or where both partners were detected to be thalassemia carriers after marriage. Some couples however, attended our centre on their own volition. Carrier status was reconfirmed by high performance liquid chromatography (HPLC). Subsequently, ARMS-PCR was done to identify the mutation, initially, with 6 primers and after that with the extended panel if required. Subsequent to the identification of ß-mutations in the parents, CVS was done to identify whether the foetus was affected or not, after duly ruling out maternal contamination by variable number tandem repeat/short tandem repeat (VNTR/STR) analysis [3, 4]. Informed consent, explaining the risks of the procedure was taken from every participant.

Data Collection, Mutation Testing and Statistical Analysis

A detailed history and clinical examination findings were recorded in the proforma as per institutional protocol. Mutation detection was done for the 5 common mutations in the ß-globin gene prevalent in the geographical region, viz. IVS 1-5G>C [HGVS: HBB:c.92+5G>C], Codon 30G>C [HGVS: HBB:c.92G>C], codon 15G>A [HGVS: HBB:c.48G>A], frameshift 41–42(-TTCT) [HGVS: HBB:c.126_129delCTTT], frameshift 8–9(+G) [HGVS: HBB:c.27dupG]. Using clues from HPLC, mutation testing for variant haemoglobins like HbE, [codon 26G>A; HGVS: HBB:c.79G>A] or HbS [codon 6A>T; HGVS: HBB:c.20A>T] were also done wherever indicated. An extended panel ARMS-PCR containing codon 121 G>C/HbD Punjab [HGVS: HBB:c.364G>C], IVS 1-1G>A [HGVS: HBB:c.92+1G>A], 619 base pair deletion [HGVS: NG_000007.3:g.71609_72227del619], Asian-Indian inversion deletion/delta beta [HGVS–not available], to name a few were done, when the first panel didn’t identify the mutation.

After determining the frequencies of the mutations in our cohort, we re-classified the data into 3 groups for subsequent analysis:—5 most commonly occurring mutations in our cohort (group 1), less common mutations (group 2) and mutations not detected (group 3). Variations in the RBC indices were analysed across the groups.

Statistical analysis was done using SPSS ver. 23 [IBM Inc, Armonk, New York, USA]. Continuous variables were displayed as frequency and percentages. For categorical variables, Chi square test or Fisher exact test were applied as required. Analysis of Variance (ANOVA) was used to distinguish between the groups. A p value less than 0.05 was considered significant and all analysis were two-tailed.

Results

Demographics

Total number of cases studied were 2128. The males and females, 50% each, were uniformly distributed (1064 couples). Mean age of the cohort was 29.30 years (range 18–48 years). Out of the total, approximately 5% (53 couples) attended the hospital for pre-conceptional testing. The rest came for post-conceptional testing at a median gestational age of 12 weeks (range 8–17 weeks). Twenty five couples (2.34%), however approached for testing quite late, when the pregnancy was already at 15–17 weeks gestational age. These cases underwent amniocentesis for prenatal screening and diagnosis.

ARMS-PCR Testing Data

In our cohort, 90.5% individuals had one of the 5 common mutations and they comprised in decreasing order of frequency: IVS 1-5 G>C (1297/2128); Codon 26(G>A)/HbE (451/2128); codon 30(G>C) (69/2128); codon 15(G>A) (61/2128); FS 41-42 [–CTTT] (48/2128). Other less common mutations (46/2128; 2.1%) were found in the remaining cases (Table 1). On the other hand, un-detected mutations accounted for a considerable (156/2128; 7.3%) proportion of the dataset. We did not find any IVS 1-1 (G>T) mutations or 619 base deletion.

Table 1.

Frequency of mutations, RBC indices and HPLC data

Type of ß-mutation	Cases n = 2128 (%)	Hb (mean ± 2SD)	MCV (mean ± 2SD)	MCH (mean ± 2SD)	MCHC (mean ± 2SD)	RDW (mean ± 2SD)	HbA0 (mean ± 2SD)	HbA2 (mean ± 2SD)	HbF (mean ± 2SD)
IVS 1- 5 (G > C)	1297 (60.9)	11.08 ± 1.77	66.42 ± 4.57	20.67 ± 1.83	31.17 ± 1.57	16.8 ± 2.62	84.03 ± 3.79	5.97 ± 1.35	1.27 ± 2.34
Codon 26/HbE (G > A)	451 (21.2)	11.64 ± 2.35	75.6 ± 6.17	24.61 ± 2.57	32.86 ± 1.8	15.49 ± 3.64	55.74 ± 18.22	38.96 ± 19.53	3.17 ± 7.88
Codon 30 (G > C)	69 (3.2)	11.07 ± 1.66	64.94 ± 3.46	20.08 ± 1.49	31.38 ± 0.98	16.95 ± 2.88	83.62 ± 1.51	5.53 ± 0.48	1.14 ± 1.14
Codon 15 (G > A)	61 (2.9)	10.81 ± 1.54	62.82 ± 4.35	19.6 ± 1.34	31.17 ± 1.83	17.2 ± 1.62	83.14 ± 0.99	5.63 ± 0.39	1.46 ± 1.11
FS 41–42 (-TTCT)	48 (2.3)	11.03 ± 1.42	64.85 ± 5.25	19.9 ± 1.46	30.93 ± 0.69	16.73 ± 1.53	83.8 ± 1.25	5.8 ± 0.42	0.97 ± 0.93
Codon 6/HbS (A > T)	19 (0.9)	11.89 ± 2.22	78.69 ± 8.41	26.37 ± 3.68	33.83 ± 1.44	15.93 ± 2.77	54.78 ± 14.81	3.44 ± 0.55	1.92 ± 3.32
Asian-Indian Inv deletion/Delta-beta	14 (0.7)	13.14 ± 1.88	75.49 ± 2.74	24.25 ± 1.07	32.8 ± 0.94	16.79 ± 1.06	70.37 ± 3.7	2.58 ± 0.21	20.73 ± 4.5
FS 8–9 (-G)	5 (0.2)	12.1 ± 2.51	66.2 ± 5.01	21.42 ± 1.76	31.15 ± 0.35	15.65 ± 0.64	83.75 ± 0.17	5.93 ± 0.48	0.98 ± 0.822
Codon 121/HbD Punjab (G > C)	5 (0.2)	13.5 ± 1.64	83.18 ± 2.7	28.54 ± 1.45	34.3 ± 0.77	14.2 ± 0.99	52.66 ± 2.81	3.26 ± 0.45	0.74 ± 0.42
Codon 126/Hb Dhonburi	3 (0.1)	12.9 ± 1.64	74.72 ± 11.26	25.6 ± 5.02	34.10 ± 2.25	16.2 ± 2.13	19.27 ± 26.01	4.8 ± 2.1	1.1 ± 0.7
IVS 1 – 1 (G > T)	0 (0)	–	–	–	–	–	–	–	–
619 base pair deletion	0 (0)	–	–	–	–	–	–	–	–
Mutation undetected	156 (7.3)	10.99 ± 1.22	70.38 ± 6.58	22.2 ± 3.26	31.47 ± 2.25	16.65 ± 1.15	82.55 ± 6.35	4.9 ± 1.6	3.67 ± 1.12
P value	All RBC indices and HPLC data amongst individual mutations p < 0.001

Type of ß-mutation		Hb (mean ± 2SD)	MCV (mean ± 2SD)	MCH (mean ± 2SD)	MCHC (mean ± 2SD)	RDW (mean ± 2SD)
IVS 1- 5 (G > C)	5 common mutations (group 1) N = 1927	11.20 ± 1.92	68.35 ± 6.41	21.51 ± 2.64	31.53 ± 1.77	16.52 ± 2.89
Codon 26/HbE (G > A)
Codon 30 (G > C)
Codon 15 (G > A)
FS 41–42 (-TTCT)
Codon 6/HbS (A > T)	Less common mutations (group 2) N = 46	12.46 ± 2.03	76.23 ± 7.98	25.22 ± 3.52	33.29 ± 1.57	15.94 ± 2.04
Asian-Indian Inv deletion/Delta-beta
FS 8–9 (-G)
Codon 121/HbD Punjab (G > C)
Codon 126/Hb Dhonburi
IVS 1 – 1 (G  > T)
619 base pair deletion
Mutation undetected	Undetected mutations (group 3) N = 156	10.99 ± 2.39	70.39 ± 9.39	22.22 ± 3.77	31.47 ± 1.77	16.65 ± 2.6
P values		< 0.001	< 0.001	< 0.001	< 0.001	0.363

CVS Analysis

A total of 657 CVS samplings were done out of which 163 (24.8%) foetuses were affected (homozygous or compound heterozygous for ß-thalassemia), 166 (25.2%) foetuses were ß-thalassemia/hemoglobinopathy carriers and 328 were unaffected (49.9%). CVS was done only for those couples where a beta-mutation was identified by ARMS-PCR.

Haemogram Parameters

The means (±2SD) of Hb, MCV, MCH, MCHC, and RDW were 11.21 g/dl (±1.97 g/dl), 68.67 fl (±6.8 fl), 21.65 pg (±2.8 pg), 31.56 g/dl (±1.78 g/dl) and 16.52 (±2.8) respectively. Re-classifying the data into 3 groups, the mean Hb was highest in the group with less common mutations (12.46 g/dl) followed by the group with common mutations (11.20 g/dl) and least in the undetected group (10.99 g/dl). Similarly, MCV was highest in group 2 (76.23 fl) possibly because this group contained cases with HbD/HbS, followed by the group with undetected mutations (70.39 fl) and group containing the five common mutations (68.35 fl) (Table 1). MCH and MCHC showed similar trends. ANOVA on all these parameters, except RDW, within groups and for individual mutations, were statistically significant (p < 0.001) (Table 1).

Discussion

Thalassemia is a major public health problem in India and it imposes a financial burden also. This country has the largest number of thalassemia major children worldwide: –150,000 and every year 10,000–15,000 thalassemic children are born here. In addition, there are around 42 million ß-thalassemia carriers in the country with an average prevalence rate of 3–4% [5, 6]. Universal prenatal screening for hemoglobinopathies has been advocated by many, [3, 7] and is under consideration by the policy makers of the country [5]. Success of thalassemia control depends on effective and adequate pre- marital and pre-conceptional screening. This program cannot run successfully without a facility for pre-natal diagnosis. Ante-natal screening is a time-bound process as a conclusive diagnosis from CVS is needed to decide whether to terminate the pregnancy or not. The platform for screening, viz. ARMS-PCR vs sequencing, becomes an important consideration for such a time-bound process.

There are few important revelations from our data analysis. The most important was that, mutations in 7.3% carriers could not be elucidated by ARMS-PCR. ARMS-PCR is a step-wise approach and requires considerable amount of time. In these cases, where a definite mutation is not identified, crucial time may be lost and the pregnancy may advance to a stage where termination is no longer allowed. Another important revelation was that couples, especially in India, made it to a healthcare facility quite late in the pregnancy, thus emphasizing the value of timely diagnosis even further. Colah et al. in 2018 published their experience of prenatal diagnosis of E-beta thalassemia in Western India and resorted to sequencing to identify rare ß-mutations in their couples at-risk [8]. In addition, sequencing can pick up additional mutations in the foetus, viz. co-inheritance of α-thalassemia or presence of XMN-1 polymorphisms as reported in a previous publication from our Institute [9]. These features would confer a thalassemia intermedia or a non-transfusion dependent phenotype to the foetus and would possibly save the couple from the trauma of termination of pregnancy and save the life of the unborn child [9]. Considering all these findings, it underscores the need for sequencing platforms atleast at the Nodal centres for thalassemia control and should be seen as a necessary investment by the thalassemia control project.

Setting up a sequencing lab at a nodal centre may cost upwards of one crore rupees, but the running cost of sequencing of a sample is only around Rs. 1000. Of course there are manpower costs, logistics, quality control and training issues, but this is money well spent, when we compare it to the financial burden of transfusion programmes, chelation drugs, monitoring and management of complications of thalassemic children, let alone offering them transplants as a curative measure for this debilitating disease. Therefore it is imperative to do a cost–benefit analysis in this regard.

One may ask whether there are easier, rapid and less expensive methods of prenatal diagnosis of thalassemia. Work has been done to standardize HPLC on foetal RBCs to accurately diagnose a thalassemic fetus [3, 8]. Other prenatal screening techniques for thalassemia have been used and involve testing of foetal lymphocytes, trophoblasts, erythroblasts extracted from maternal blood, but haven’t found much popularity due to various constraints. Newer non-invasive techniques involve extraction of foetal cell-free DNA from maternal blood [4, 10] for genotyping. However, all these approaches require sequencing platforms. More recently, matrix assisted laser desorption ionization- time of flight (MALDI-TOF) and APEX or arrayed primer extension have been used for non-invasive prenatal detection of ß-thalassemia but most of these methods neither are cheap nor have been validated on large populations and therefore are far away from being adopted by national programmes [4, 11].

When compared to national data [12], we demonstrated that Eastern India has a different genotypic signature of the thalassemia and hemoglobinopathy syndromes. The common mutations (IVS 1-1 and 619 base-pair deletion), prevalent in northern India, aren’t found here frequently. Instead we found a considerable number of codon 30, codon 15 and FS 41-42–TTCT mutations in our cohort. These 5 mutations along with codon 26 mutation/HbE, accounted for almost 90% of all the mutations and could be detected by ARMS-PCR. Therefore, we suggest that each thalassemia prevalent region in the country, develop their own panel of ß-mutations for screening purposes based on local genotypic variations as already mentioned in previous guidelines and resource documents [3, 5, 6, 12–14].

Another interesting finding, not yet reported elsewhere in literature is that individuals with the less common mutations (group 2) and those in whom mutations could not be detected (group 3) had significantly different RBC indices and HPLC values (p < 0.001). A scoring system or a software, based on RBC and HPLC indices could be developed, to help identify and segregate these individuals. ARMS-PCR could might as well be avoided, and sequencing done for these individuals straightaway, thus reducing costs further.

Conclusion

It suffices to say that prenatal diagnosis is of paramount importance to reduce the burden of thalassemia in the country. The haemogram-HPLC-ARMS PCR-CVS approach is a cost effective and established method for prenatal screening but tends to miss out a considerable number of less common and rare mutations, necessitating the need for sequencing in these difficult cases. Regional variations in mutations need to be taken into account before designing panels for mutation testing. With the existing evidence and experience collected from ß-globin gene sequencing at research laboratories and apex institutes, guidelines need to be revised to incorporate this essential tool for the future of thalassemia screening in the country.

Author Contributions

Shouriyo Ghosh: Concept, data collection and processing, statistical analysis, manuscript drafting. Sila Chakrabarti: Technical processing, molecular work, concept. Maitreyee Bhattacharyya: Concept, supervision, manuscript editing.

Funding

The authors declare no source of funding for purpose of preparing or publishing this paper.

Availability of Data and Material

Confidential.

Code Availability

Not applicable.

Compliance with Ethical Standards

Conflict of interest

The authors declare that they have no conflict of interest.

Ethics Approval

Retrospective data analysis.

Consent to Participate

Informed consent was taken from all participants.

Consent for Publication

The authors give full consent for publication.