Key PointsQuestion
Is susceptibility to rheumatic heart disease (RHD) heritable in African individuals, and if so, what are the common genetic variants associated with RHD risk?
Findings
In this genome-wide association study of 4809 African individuals, 1 genetic risk locus at 11q24.1 (rs1219406) was associated with RHD at genome-wide significance in Black African individuals but not in other groups, although 1 previously described association was replicated at nominal significance. Polygenic heritability of RHD is estimated at 0.49 in African individuals.
Meaning
This study suggests that there is an important polygenic component to RHD risk in African individuals, highlighting genetic features exclusive to African individuals as well as genetic similarities with non-African individuals.
Importance
Rheumatic heart disease (RHD), a sequela of rheumatic fever characterized by permanent heart valve damage, is the leading cause of cardiac surgery in Africa. However, its pathophysiologic characteristics and genetics are poorly understood. Understanding genetic susceptibility may aid in prevention, control, and interventions to eliminate RHD.
Objective
To identify common genetic loci associated with RHD susceptibility in Black African individuals.
Design, Setting, and Participants
This multicenter case-control genome-wide association study (GWAS), the Genetics of Rheumatic Heart Disease, examined more than 7 million genotyped and imputed single-nucleotide variations. The 4809 GWAS participants and 116 independent trio families were enrolled from 8 African countries between December 31, 2012, and March 31, 2018. All GWAS participants and trio probands were screened by use of echocardiography. Data analyses took place from May 15, 2017, until March 14, 2021.
Main Outcomes and Measures
Genetic associations with RHD.
Results
This study included 4809 African participants (2548 RHD cases and 2261 controls; 3301 women [69%]; mean [SD] age, 36.5 [16.3] years). The GWAS identified a single RHD risk locus, 11q24.1 (rs1219406 [odds ratio, 1.65; 95% CI, 1.48-1.82; P = 4.36 × 10−8]), which reached genome-wide significance in Black African individuals. Our meta-analysis of Black (n = 3179) and admixed (n = 1055) African individuals revealed several suggestive loci. The study also replicated a previously reported association in Pacific Islander individuals (rs11846409) at the immunoglobulin heavy chain locus, in the meta-analysis of Black and admixed African individuals (odds ratio, 1.16; 95% CI, 1.06-1.27; P = 1.19 × 10−3). The HLA (rs9272622) associations reported in Aboriginal Australian individuals could not be replicated. In support of the known polygenic architecture for RHD, overtransmission of a polygenic risk score from unaffected parents to affected probands was observed (polygenic transmission disequilibrium testing mean [SE], 0.27 [0.16] SDs; P = .04996), and the chip-based heritability was estimated to be high at 0.49 (SE = 0.12; P = 3.28 × 10−5) in Black African individuals.
Conclusions and Relevance
This study revealed a novel candidate susceptibility locus exclusive to Black African individuals and an important heritable component to RHD susceptibility in African individuals.
In 2018, the World Health Organization made rheumatic heart disease (RHD) a global health priority, given that 40 million individuals are affected worldwide.1-4 Annually, RHD is associated with an estimated 10.5 million disability-adjusted life-years and 300 000 premature deaths.5-8 The prevalence of RHD has decreased significantly in high-income countries in the last 75 years owing to the advent of penicillin and major improvements in indices of social and economic development.9 During the last 2 decades, there has been progress in documenting RHD’s burden in low- and middle-income countries. However, RHD remains a significant cause of morbidity and mortality in poorer communities of low- and middle-income countries.9,10 Despite Africa being home to 17% of the world’s total population,11 in 2017, sub-Saharan Africa comprised 23% of the global RHD caseload,5 with a high prevalence of 864 per 100 000, compared with only 9.8 per 100 000 in North America and 7.7 per 100 000 in Western Europe.12Quiz Ref ID Currently, in sub-Saharan Africa, RHD remains both the most common form of acquired cardiovascular disease in women, children, and young adults and the leading cause of cardiac surgery,13-17 with prevalence peaking between 25 and 45 years of age.18
Rheumatic heart disease is a consequence of an autoimmune response to untreated, or inadequately treated, Streptococcus pyogenes (Lancefield Group A β-hemolytic Streptococcus [GAS]) infection, often pharyngitis, in a susceptible host.19 Although not all patients with RHD have a history of rheumatic fever, the precursor of RHD, close to 60% of rheumatic fever cases progress to RHD.20 Important gaps remain in our understanding of individual host susceptibility to RHD after GAS infection.21,22 Although RHD development can be prevented by treating GAS infections with penicillin, this strategy has not been successful in poverty-stricken, overcrowded regions in Africa. Quiz Ref IDIn addition to prevailing social and economic factors (poverty and overcrowding), differences in genetic susceptibility and GAS strain virulence have also been hypothesized to be associated with interindividual differences in RHD susceptibility.10,23 Understanding genetic susceptibility will help devise better prevention, control, and interventions.
Quiz Ref IDThe precursor of RHD, rheumatic fever, is widely considered to have a strong genetic component,23 with heritability estimated to be 60% from twin studies.23,24 Although there are no heritability estimates for RHD, reports of familial cases provide support for genetic susceptibility.25,26 Several candidate gene studies tested for RHD susceptibility; however, these studies had small sample sizes, and the results remain equivocal.27 Three previous genome-wide association studies (GWASs) for RHD susceptibility in 4 different populations reported (1) a genome-wide significant (GWS) locus (rs11846409; P = 4.1 × 10−9) in the immunoglobulin heavy chain (IGH) locus in Pacific Islander individuals28; (2) a suggestive locus (rs9272622; P = 1.86 × 10−7) in the HLA class II gene’s HLA-DQA1 (OMIM 146880) region in Australian Aboriginal individuals, which failed to reach GWS29; and (3) a GWS locus (rs201026476; P = 7.45 × 10−9) in the HLA class III gene’s locus situated in the 3′ untranslated region of the PBX2 (pre–B-cell leukemia transcription factor 2; OMIM 176311) gene in Indian and European individuals.30 However, these aforementioned RHD GWASs did not include African individuals. In the present study, to better understand the RHD pathophysiologic characteristics and genetic determinants associated with RHD susceptibility, we performed a GWAS for RHD susceptibility in African individuals.
The Genetics of Rheumatic Heart Disease (RHDGen) Network is a project within Human Hereditary and Health Africa (H3Africa).31 The University of Cape Town Human Research Ethics Committee approved the present study as a substudy of RHDGen. This study followed the Strengthening the Reporting of Genetic Association Studies (STREGA) reporting guideline.32 Written informed consent was obtained from all adult study participants. In the case-control study group, 4809 participants (2548 cases and 2261 controls) passed GWAS quality control, and in the family-based study group, 116 independent trio families (348 participants) passed GWAS quality control.
Participants were from 8 African countries (ie, Kenya, Mozambique, Namibia, Nigeria, South Africa, Sudan, Uganda, and Zambia; eFigure 1 in the Supplement). Any Black African ethnic group from any of the 8 African countries (mostly of Bantu descent) was included in the Black African cohort. The South African ethnic group composed primarily of persons of mixed race, comprising any admixture combination of individuals of European, Southeast Asian, South Asian, Bantu-speaking African, and/or indigenous Southern African hunter-gatherer ancestries (Khoikhoi, San, or Bushmen),33 was renamed admixed African individuals. The race/ethnicity of an individual was self-reported and verified through visualization of top genetic principal components.
Cases were existing, and new patients who consented to participate in the study were recruited from cardiology units with echocardiography facilities. All study participants for the GWAS group and the probands for the trio study were assessed according to the evidence-based guideline from the 2012 World Heart Federation criteria for echocardiographic diagnosis of RHD.14,34,35 Rheumatic heart disease was diagnosed in adults on the basis of features of definite RHD, with central verification of echocardiographic images by an experienced cardiologist (C.C.) at the Project Coordinating Office through a web-based portal.
Genotyping and Imputation
We extracted DNA at the cardiovascular genetics laboratory at the Hatter Institute for Cardiovascular Research in Africa and the Cape Heart Institute, University of Cape Town, South Africa. We genotyped DNA samples using the Infinium Human Omni 2.5-8 (Illumina Inc) bead chips, version 1.1 and version 1.3 according to the manufacturer’s protocol at the Genetic and Molecular Epidemiology Laboratory located in the Population Health Research Institute, Hamilton, Ontario, Canada. We adopted the 2011 recommendations of Turner et al36 for per-sample and per-marker quality control. After quality control, genotyped variants were prephased on the Sanger Imputation Server37 using EAGLE2, version 2.0.5.38 We then imputed genetic variants with the African Genome Resources reference panel (from the African Genome Variation Project [AGVP]) using the Positional Burrows-Wheeler Transform imputation method.39,40
GWAS Testing and Meta-analysis
We used Genome-wide Complex Trait Analysis, version 1.90.1 beta41 to conduct a mixed linear model association42 to retain the highest number of individuals by including 336 related individuals (7%; non–trio study participants). We assumed an additive model and adjusted for covariates, including sex, 10 genetic principal components, and Infinium Human Omni 2.5-8 bead chip version (1.1 or 1.3). We performed stratified analyses for Black African individuals and admixed African individuals separately. Effect sizes were transformed from linear to logistic odds ratios (ORs).43 A meta-analysis of the GWAS results from Black African individual and admixed African individual was subsequently performed using an inverse-variance fixed-effects model in META, version 1.7.44,45 The GWS P value threshold was set at less than 5 × 10−8.
Regional association plots were generated using LocusZoom.46 We investigated lead associations for novelty using the GWAS Catalog,47,48 GeneAtlas Phenome-Wide Association Studies (PheWAS),49,50 GWAS Atlas PheWAS,51,52 and the Open Targets Genetics portal.53,54 We adjusted the GeneAtlas PheWAS P values for false-discovery rates using the Benjamini-Hochberg procedure and fine-mapped GWS single-nucleotide variations (SNVs) (eTable 1 in the Supplement). We also interrogated the Genotype-Tissue Expression Project (GTEx) portal,55,56 version 8 database for expression quantitative trait loci.
Pathway analyses were conducted using 2 different methods. First, data-driven expression-prioritized integration for complex traits (DEPICT), an integrative tool based on estimated gene functions, systematically prioritizes the most likely causal genes at associated loci, highlights enriched pathways, and identifies tissue or cell types where genes from associated loci are highly expressed. DEPICT was run using SNVs with P ≤ 1 × 10−5,57 as recommended. Multi-marker Analysis of GenoMic Annotation (MAGMA) was also used for gene and gene-set analyses while correcting for type 1 error rate and linkage disequilibrium (LD), using all the GWAS SNVs.58,59
We estimated narrow-sense heritability (h2) on the liability scale using Genome-wide Complex Trait Analysis–REML (restricted maximum likelihood estimation).42,60 To calculate the h2, we adjusted for sex, chip type, and 10 principal components, with the prevalence of disease set to 0.0047 in each ethnic group.12
TaqMan Validation of rs1219406
We directly genotyped the lead imputed SNV, rs1219406, in a random subset of 1218 RHDGen study participants from the Black African GWAS using a Custom TaqMan SNP Assay,61 which uses real-time quantitative polymerase chain reaction technology.
Family-Based Validation With Polygenic Transmission Disequilibrium Testing
The Black African GWAS (training and validation sample) was used to develop RHD polygenic risk scores (PRS) for polygenic transmission disequilibrium testing (pTDT) within the Black African trio families (testing sample). The optimal P value threshold (P0) for inclusion of genetic variants into the PRS for pTDT was established through the clumping and thresholding procedure of PRSice-2,62-64 within training and validation data sets. Subsequently, PRS (P0) were derived in affected probands and unaffected parents, and pTDT was performed to test whether the proband PRS deviated from the mean parental PRS62,65 (detailed methods in eAppendix 2 in Supplement).
GWAS of RHD Susceptibility
We performed a GWAS of 7 605 010 autosomal variants in Black African individuals (1687 cases and 1492 controls) and admixed African individuals (601 cases and 454 controls) separately and a meta-analysis of both groups. Study participants’ characteristics are provided in Table 1. In Black African individuals, we found a GWS association at 11q24.1 (rs1219406; minor allelic frequency (MAF) = 0.092; OR, 1.65; 95% CI, 1.48-1.82; P = 4.36 × 10−8) (Figure 1 and Table 243). No other association was observed at GWS. Bayesian fine-mapping66 ±1 Mb of rs1219406 (eTable 1 in the Supplement) produced 27 candidate causal variants in the 99% credible set. The lead SNV (rs1219406) accounted for 43% of the posterior probability out of the total 27 SNVs in the 99% credible set; rs1219406 tags a 100-Kb LD block within which there are no genes identified (Figure 2). The nearest gene (152 276 base pairs [bp] to the canonical transcription start site) is AP001924.1 (a novel transcript; GenBank AP001924.5), and the nearest coding gene (242 554 bp to the canonical transcription start site) is the BH3-Like Motif Containing, Cell Death Inducer (BLID [OMIM 608853]).53
Because rs1219406 was imputed in our study (with high imputation quality [ie, information score = 0.98]), we also sought to strengthen the association findings by direct genotyping (TaqMan) in a subset of 1218 randomly selected samples. We observed a high concordance (99%) between direct genotypes and imputed genotypes (eFigure 3 in the Supplement). Nonetheless, no association was observed in admixed African individuals (MAF = 0.41; OR, 1.02; 95% CI, 0.85-1.22; P = .84). We also tested rs1219406 for association with RHD in external data sets, including populations of Melanesian, Polynesian, Northern Indian, Fijian Indian, and European descent.28,30 None of the aforementioned samples confirmed the association (P > .05), neither individually nor with a meta-analysis of external replication sets (pooled OR, 0.96; 95% CI, 0.86-1.06; P = .45) (eTable 2 in the Supplement).
To our knowledge, no associations or expression quantitative trait loci have previously been reported with rs1219406 in GWAS Catalog47 and GTEx.55 We also performed a PheWAS portal49 search for rs1219406 in the predominantly European GeneAtlas. We observed 9 associations with a false-discovery rate of 0.05 or lower (eTable 1 in the Supplement), including immune system–associated traits (eg, lymphocyte count and M05 seropositive rheumatoid arthritis), skeletal system–associated traits (eg, height), and metabolism-associated traits (eg, body mass index). We also performed an Open Targets Genetics and GWAS atlas search. The GWAS atlas PheWAS identified 14 phenotypes significantly associated (P < .05/231 tests) with rs1219406 (eTable 1 in the Supplement) height at GWS,51,53 as well as other anthropometric and cardiovascular traits.
In admixed African individuals, the highest nominally significant association was on chromosome 2p16.3 (rs11125426; OR, 0.41; 95% CI, 0.09-0.72; P = 2.10 × 10−7) (Figure 1 and Table 243). The nearest gene is the long intergenic nonprotein coding RNA-1867 (LINC01867 [HGNC 52687]), and the nearest coding gene is the neurexin-1 (NRXN1 [OMIM 600565]).53 In the meta-analysis of Black and admixed African individuals, the highest nominally significant association was at chromosome 2p11.2 (rs2386325, OR, 0.80; 95% CI, 0.71-0.88; P = 3.17 × 10−7) (Figure 1 and Table 243). The nearest gene is the immunoglobulin κ variable 1D-33 (IGKV1D-33 [HGNC 5753]), and the nearest coding gene is the ribose 5-phosphate isomerase A (RPIA [HGNC 10297]).53 Neither of these loci reached GWS and were not replicated (eTable 2 in the Supplement).
Replication of Previously Reported RHD Loci
We systematically evaluated 75 candidate genes and loci of interest21 but found no significant enrichment for genetic associations (Figure 321,28,67; eTable 1 in the Supplement). A prior GWAS of 2852 Pacific Islander individuals implicated a variant at the IGH locus on chromosome 14 (rs11846409). In RHDGen, this locus replicated (P < .05/2 independent tests) in admixed African individuals (OR, 1.37; 95% CI, 1.17-1.56; P = 2.38 × 10−3) and in the combined GWAS meta-analysis (OR, 1.16; 95% CI, 1.06-1.27; P = 1.19 × 10−3),28 with a consistent direction of effect to earlier analyses. Pooling these association statistics with those from the Pacific, European, and Indian populations, each risk allele was associated with a 1.2-fold increased risk of disease (OR, 1.23, 95% CI, 1.15-1.31; P = 9.1 × 10−10; Figure 321,28,67).
The HLA DQA1-DQB1 locus reported to be nominally associated with RHD in Aboriginal Australian individuals (rs9272622; OR, 0.90; P = 1.86 × 10−7)29 was not present in the unimputed RHDGen GWAS for replication. Because rs9272622 was not available in the African Genome Resources (from the AGVP) imputation reference panel, we imputed from the 1000 Genomes 1000G Phase 3 African imputation panel and reran associations for comparison. Using this approach, rs9272622 did not show evidence of association with RHD in either Black (OR, 0.96; 95% CI, 0.66-1.26; P = .44) or admixed African individuals (OR, 0.85; 95% CI, 0.65-1.05; P = .16). The closest imputed HLA region SNV in RHDGen, rs28746888, was near HLA-DQB1 and had suggestive significance (OR, 0.70; 95% CI, 0.57-0.84; P = 3.61 × 10−7; Figure 2).
We also tried to replicate a second HLA locus (rs201026476) associated with RHD in North Indian and European individuals.30 However, rs201026476 was not present in either of our available imputed data sets of African descent (AGVP or 1000G phase 3 African individuals), and hence replication was not possible.
In Silico Pathway Analyses
DEPICT analyses57 were conducted using suggestively significant SNVs (P < 1 × 10−5). Gene-set enrichment (eTable 3 in the Supplement) and tissue enrichment (eTable 4 in the Supplement) analyses showed some biologically relevant tissues among the top hits, including heart valve tissue; however, these results were not significant after adjustment for multiple hypothesis testing (false-discovery rate P < .05).68
We also performed MAGMA gene analysis and generalized gene-set analysis of GWAS data.58 We observed nominally significant associations (P < .05/number of total genes) with genes and gene sets of potential RHD relevance, using the competitive approach. These include the loci for IGH (chr14q32) and IGK (chr2p12), gene sets encoding components of the complement system, and genes for the regulation of cell proliferation involved in heart morphogenesis and CC chemokine receptor activity. The top MAGMA results per ethnic group are shown in eTables 5, 6, and 7 in the Supplement. No gene set was significant after adjustment for 18 000 tests.
Polygenic Association With RHD Susceptibility
The heritability of RHD explained by common (MAF ≥0.05) genetic variants was estimated at h2 = 0.49 (SE = 0.12; P = 3.28 × 10−5) in Black African individuals and h2 = 0.35 (SE = 0.24; P = 6.56 × 10−2) in admixed African individuals. The presence of significant heritability for RHD suggests a highly polygenic architecture. To further investigate this polygenicity, we conducted pTDT in trio families, hypothesizing that polygenic risk for RHD would be overtransmitted from unaffected parents to affected probands.65
A PRS comprising 195 putative risk variants (Black African GWAS P < 8.01 × 10−5) was found to be overtransmitted to probands. Probands with RHD had a PRS that was on average 0.27 SDs higher than that of their unaffected parents (pTDT mean, 0.27 SD; pTDT SE, 0.16 SD; and pTDT P value, .04996).
Our study provides insights into genetic loci and pathways associated with RHD susceptibility. We identified a GWS (P < 5 × 10−8) association at a novel locus on chromosome 11q24.1 that appears to be specific to Black African individuals, given that we did not replicate the finding in admixed African individuals or those of other ancestries; rs1219406 is intergenic and located in a region harboring only long noncoding transcripts, making it challenging to pinpoint potential causal mechanisms. Owing to the paucity of functional data on Black African individuals, it remains difficult to assess the biological implications of the observed association.55,69-71Quiz Ref ID PheWAS analyses identified associations of rs1219406 with traits relevant to RHD, such as autoimmune traits, including lymphocyte count, atrial fibrillation, and M05 seropositive rheumatoid arthritis (a type of inflammatory polyarthropathy); rs1219406 is also associated with height at GWS,72 increasing confidence that it is either a functional locus or it is in LD with 1 or more functional variant(s). Nonetheless, our results require further replication, ideally in Black African individuals.
To our knowledge, we provided the first chip-based heritability estimates for RHD in African samples. Our estimates of 0.35 in admixed African individuals and 0.49 in Black African individuals support an important association of host genetics with RHD susceptibility.23 The high heritability observed, combined with a lack of strong associations, suggests a polygenic architecture for RHD. The weakly significant overtransmission of a PRS comprising 195 putative risk alleles in 116 trios supports this hypothesis. More important, the pTDT analysis is robust to population stratification.
Several loci have been previously reported to be associated with RHD, and our study provided the opportunity to test these in African individuals. Quiz Ref IDThe IGH locus SNV rs11846409, discovered in an RHD GWAS in Pacific Islander individuals of Oceanian and South Asian ancestry, was replicated in our data. These results establish IGH as a key risk locus in RHD development across multiple distinct populations.28 By contrast, we did not replicate the Aboriginal Australian HLA SNV, rs9272622.29 We could not test another HLA locus in major histocompatibility complex class III (rs201026476) reported in an RHD GWAS of North Indian individuals and European individuals because it was not genotyped and could not be imputed in RHDGen.30 It is possible that the complex LD structure in the highly polymorphic HLA region, its geographic variability in populations73 (eg, from 8 different African countries vs Aboriginal Australian individuals), and our limited GWAS sample size may have contributed to the lack of replication.
Pathway analysis and GWAS searches to assess evidence for other functional associations implicate putative biological candidates associated with immunologic and cardiovascular development. These align with some of the current theories regarding RHD pathophysiologic characteristics.74 However, the results were nominally significant, and further replication is needed. Furthermore, source data for pathway analysis are derived from predominantly European populations and tissues and might not be fully reflective of biological processes in African individuals. Similarly, our competitive MAGMA analyses failed to meet stringent Bonferroni correction for multiple hypothesis testing.57,59 Hence, more functional work is needed to elucidate the appropriate pathways and how they connect to RHD pathophysiologic characteristics in African individuals.
Our study had several limitations. First, despite being the largest study of RHD genetics, the overall sample size was small compared with contemporary GWASs for other diseases. Considerable challenges remain for the collection and analysis of biological samples in Africa.75,76 Second, the novel GWS rs1219406 locus could not be replicated in Black African individuals because no independent cohort was available. Third, the functional relevance of the associations identified is difficult to assess because of the lack of functional genetics data specific to African individuals.60,75,76 Fourth, regional environmental factors, such as the prevalence of HIV, could potentially impact the results. Individuals’ HIV status remains difficult to obtain because of the considerable social stigma and discrimination attached to it.77 A study in 2016 showed that patients with RHD have more severe outcomes in the setting of HIV.78 These findings were published after we designed our study; hence, HIV status was not elicited in the case report forms nor accounted for via RHDGen’s informed consent process.
Our study establishes the polygenic nature of RHD in African individuals and provides a comprehensive comparison of key genetic associations with other populations. Our study mandates further research on the 11q24.1 locus as a novel candidate for RHD in Black African individuals. Our results also support a high heritability for RHD in African individuals, providing a basis for the familial aggregation of RHD cases beyond a shared environment. As with other polygenic traits, large sample sizes will be necessary to uncover the numerous variants with smaller effect sizes underlying RHD risk.
Accepted for Publication: March 25, 2021.
Published Online: June 9, 2021. doi:10.1001/jamacardio.2021.1627
Corresponding Author: Guillaume Paré, MD, MSc, Population Health Research Institute, David Braley Cardiac, Vascular and Stroke Research Institute, 237 Barton St E, Hamilton, ON L8L 2X2, Canada (pareg@mcmaster.ca).
Author Contributions: Ms Machipisa and Dr Paré had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Engel and Paré contributed equally to this work.
Concept and design: Machipisa, Shaboodien, de Vries, Lwabi, Okello, Musuku, ElSayed, Elhassan, Zühlke, Mulder, Ramesar, Lesosky, Parks, Cordell, Engel, Paré.
Acquisition, analysis, or interpretation of data: Machipisa, Chong, Muhamed, Chishala, Pandie, Laing, Joachim, Daniels, Ntsekhe, Hugo-Hamman, Gitura, Ogendo, Okello, Damasceno, Novela, Mocumbi, Madeira, Musuku, Mtaja, ElSayed, Elhassan, Bode-Thomas, Okeahialam, Zühlke, Mulder, Lesosky, Parks, Cordell, Keavney, Engel, Paré.
Drafting of the manuscript: Machipisa, Chong, Shaboodien, Pandie, Ogendo, Okello, Madeira, Zühlke, Engel, Paré.
Critical revision of the manuscript for important intellectual content: Machipisa, Chong, Muhamed, Chishala, de Vries, Laing, Joachim, Daniels, Ntsekhe, Hugo-Hamman, Gitura, Lwabi, Okello, Damasceno, Novela, Mocumbi, Musuku, Mtaja, ElSayed, Elhassan, Bode-Thomas, Okeahialam, Zühlke, Mulder, Ramesar, Lesosky, Parks, Cordell, Keavney, Engel, Paré.
Statistical analysis: Machipisa, Chong, Pandie, Daniels, Lesosky, Parks, Paré.
Obtained funding: Machipisa, Hugo-Hamman, Okello, Zühlke, Cordell, Engel, Paré.
Administrative, technical, or material support: Machipisa, Muhamed, Chishala, Shaboodien, Laing, Daniels, Hugo-Hamman, Lwabi, Okello, Damasceno, Madeira, Musuku, ElSayed, Elhassan, Zühlke, Ramesar, Engel, Paré.
Supervision: Chong, Hugo-Hamman, Gitura, Lwabi, Damasceno, Musuku, ElSayed, Elhassan, Bode-Thomas, Zühlke, Mulder, Cordell, Keavney, Engel, Paré.
Conflict of Interest Disclosures: Ms Machipisa reported receiving grants from Wellcome Trust during the conduct of the study; nonfinancial support from H3Africa; attending a sponsored training course at Wellcome Genome Campus; was funded by the University of Cape Town (under the Mayosi Research Group RHDGen Fellowship funded by the Wellcome Trust) the Crasnow Travel Scholarship, and via Population Health Research Institute (PHRI) and McMaster University through the inaugural Bongani Mayosi UCT (University of Cape Town)-PHRI Scholarship 2019/2020 to complete the submitted work. Mr Chong reported receiving a Canadian Institute of Health Research doctoral award and consulting fees from Bayer. Ms Pandie reported receiving grants from Wellcome Trust during the conduct of the study. Dr Damasceno reported receiving grants from University of Cape Town during the conduct of the study. Dr Musuku reported receiving grants from Wellcome Trust during the conduct of the study. Dr Mtaja reported receiving nonfinancial support from Wellcome Trust during the conduct of the study. Dr Mulder reported receiving grants from the National Institutes of Health during the conduct of the study. Dr Parks reported receiving grants from National Institute for Health Research during the conduct of the study and the British Medical Association. Dr Cordell reported receiving support from Newcastle University and the Wellcome Trust. Dr Keavney reported receiving support from the British Heart Foundation, the Wellcome Trust, and Manchester University. Dr Engel reported receiving support from the RHDGen Wellcome Trust grant, the South African National Research Foundation (NRF) No. 116287, grant NW17SFRN33630027 from the American Heart Association, and UCT. Dr Paré reported receiving grants from Bayer and Esperion; personal fees from Bayer, Bristol Myers Squibb, Lexicomp, Amgen, Illumina, and Sanofi; and the Canada Research Chair in Genetic and Molecular Epidemiology, and CISCO Professorship in Integrated Health Systems outside the submitted work. No other disclosures were reported.
Funding/Support: RHDGen (https://h3africa.org/index.php/consortium/the-rhdgen-network-genetics-of-rheumatic-heart-disease-and-molecular-epidemiology-of-streptococcus-pyogenes-pharyngitis) was supported by grants awarded from the Wellcome Trust under the H3Africa; grant099313/B/12/A (https://app.dimensions.ai/details/grant/grant.3640606 and https://europepmc.org/grantfinder/grantdetails?query=pi%3A%22Mayosi%2BBM%22%2Bgid%3A%22099313%22%2Bga%3A%22Wellcome%20Trust%22).
Role of the Funder/Sponsor: Wellcome Trust under the H3Africa arm, provided the grant and received project progress reports at H3Africa meetings and other internal reviews. These meetings provided the funder with all details and the opportunity to provide input to the principal investigator’s plans regarding the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We would like to thank all participants for being a part of this study, as well as the members of the Mayosi Research Group Coordinating Office team and collaborating sites’ staff for the study coordination, recruitment, data entry, and cleaning. We also thank the Genetics of Rheumatic Heart Disease (RHDGen) Network Consortium members listed in eAppendix 1 in the Supplement. We acknowledge the Hatter Institute of Cardiovascular Research in Africa (HICRA), directed by Karen Sliwa, MD PhD, and its cardiovascular genetics laboratory for assisting with all of the preliminary wet laboratory work (primarily, students and staff funded by the Mayosi Research Group or RHDGen, namely: Maryam Fish, PhD, Gaurang Deshpande, PhD, Stephen Kamuli, MSc, Timothy Spracklen, PhD, Lameez Pearce, BSc, Janine Saaiman, and Zukiswa Jiki, BSc). We would also like to thank staff like Reina Ditta, MSc, of the Genetic and Molecular Epidemiology Laboratory (GMEL) for managing the laboratory and notably, Amanda Hodge, Gianluca Situm, BSc, Gillian Lampkin, BSc and Taylor MacIsaac, BSc, for assisting with the GMEL wet laboratory work. Also, we acknowledge the GMEL dry laboratory and PHRI students, affiliates, and members for providing technical consultations (especially Jenny Sjaarda, PhD, Pedrum Mohammadi-Shemirani, MSc, Marie Pigeyre, MD, PhD, Ricky Lali, MSc, Shihong Mao, PhD, Loubna Akhbir, PhD, Amel Lamri, PhD, Ann Le, MSc, Godsent Isiguzo, MD, PhD, Tinashe Chikowore, PhD, Rob Morton, PhD, Viwe Mtwise, MD, MMed, and Alexander P. Benz, MD MSc). We thank the Pacific Islands Rheumatic Heart Disease Genetics Network, South Asia Rheumatic Heart Disease Genetics Network and the UK Biobank Resource (application 11537) on which pooled analyses were based. Finally, we would like to thank everyone who took the time to provide valuable input throughout this study.
Additional Information: This article is dedicated to the memory of our mentor, friend, and colleague Bongani M Mayosi, MD, DPhil (1967-2018), who inspired and established the RHDGen Network Consortium through his great vision, leadership, guidance, and mentorship. This article is also dedicated to the memory of Lungile Pepeta, MD, recruitment site cardiologist, and Veronica Francis, RN, study coordinator, Department of Medicine, University of Cape Town and Groote Schuur Hospital, Cape Town, South Africa. Genotype and phenotype data used in this article will be deposited in the Pan African Bioinformatics Network for the Human Heredity and Health in Africa (H3ABioNet) Data Archive, as per the H3Africa guidelines.79 Some additional restrictions on access and use apply (eg, focus on RHD research). Access to certain components of the dataset requires regulatory approval from the country where the samples were obtained. The H3A BioNet Data Archive repository will provide further information about access to the dataset (link: will be provided by H3AbioNet & European Genome-phenome Archive). Please cite this article whenever any data or method provided in the resources mentioned above is used. A complete list of individuals who worked in The Genetics of Rheumatic Heart Disease (RHDGen) Network Consortium is provided in eAppendix 1 in the Supplement.
2.GBD 2013 Mortality and Causes of Death Collaborators. Global, regional, and national age-sex specific all-cause and cause-specific mortality for 240 causes of death, 1990-2013: a systematic analysis for the Global Burden of Disease Study 2013.
Lancet. 2015;385(9963):117-171. doi:
10.1016/S0140-6736(14)61682-2
PubMedGoogle ScholarCrossref 4.Beaton
A, Kamalembo FB, Dale J,
et al. The American Heart Association’s call to action for reducing the global burden of rheumatic heart disease: a policy statement from the American Heart Association. Circulation. 2020;142(20):e358-e368.
5.Yuyun
MF, Sliwa
K, Kengne
AP, Mocumbi
AO, Bukhman
G. Cardiovascular diseases in sub-Saharan Africa compared to high-income countries: an epidemiological perspective.
Glob Heart. 2020;15(1):15. doi:
10.5334/gh.403
PubMedGoogle ScholarCrossref 6.GBD 2017 DALYs and HALE Collaborators. Global, regional, and national disability-adjusted life-years (DALYs) for 359 diseases and injuries and healthy life expectancy (HALE) for 195 countries and territories, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017.
Lancet. 2018;392(10159):1859-1922. doi:
10.1016/S0140-6736(18)32335-3
PubMedGoogle ScholarCrossref 7.Zühlke
L, Engel
ME, Karthikeyan
G,
et al. Characteristics, complications, and gaps in evidence-based interventions in rheumatic heart disease: the Global Rheumatic Heart Disease Registry (the REMEDY study).
Eur Heart J. 2015;36(18):1115-22a. doi:
10.1093/eurheartj/ehu449
PubMedGoogle ScholarCrossref 13.Zühlke
L, Karthikeyan
G, Engel
ME,
et al. Clinical outcomes in 3343 children and adults with rheumatic heart disease from 14 low- and middle-income countries: two-year follow-up of the Global Rheumatic Heart Disease Registry (the REMEDY Study).
Circulation. 2016;134(19):1456-1466. doi:
10.1161/CIRCULATIONAHA.116.024769
PubMedGoogle ScholarCrossref 14.Remenyi
B, Carapetis
J, Wyber
R, Taubert
K, Mayosi
BM; World Heart Federation. Position statement of the World Heart Federation on the prevention and control of rheumatic heart disease.
Nat Rev Cardiol. 2013;10(5):284-292. doi:
10.1038/nrcardio.2013.34
PubMedGoogle ScholarCrossref 15.Huntley
GD, Thaden
JJ, Nkomo
VT. Epidemiology of heart valve disease. In: Kheradvar A, ed.
Principles of Heart Valve Engineering. Elsevier; 2019:41-62. doi:
10.1016/B978-0-12-814661-3.00003-4
16.Yangni-Angate
KH, Meneas
C, Diby
F, Diomande
M, Adoubi
A, Tanauh
Y. Cardiac surgery in Africa: a thirty-five year experience on open heart surgery in Cote d’Ivoire.
Cardiovasc Diagn Ther. 2016;6(suppl 1):S44-S63. doi:
10.21037/cdt.2016.10.06
PubMedGoogle ScholarCrossref 17.Karthikeyan
G, Connolly
SJ, Ntsekhe
M,
et al; INVICTUS Investigators. The INVICTUS rheumatic heart disease research program: rationale, design and baseline characteristics of a randomized trial of rivaroxaban compared to vitamin K antagonists in rheumatic valvular disease and atrial fibrillation.
Am Heart J. 2020;225:69-77. doi:
10.1016/j.ahj.2020.03.018
PubMedGoogle ScholarCrossref 25.Poomarimuthu
M, Elango
S, Soundrapandian
S, Mariakuttikan
J. “
HLA-G 3'UTR gene polymorphisms and rheumatic heart disease: a familial study among South Indian population”.
Pediatr Rheumatol Online J. 2017;15(1):10. doi:
10.1186/s12969-017-0140-x
PubMedGoogle ScholarCrossref 26.Aliku
T, Sable
C, Scheel
A,
et al. Targeted echocardiographic screening for latent rheumatic heart disease in northern Uganda: evaluating familial risk following identification of an index case.
PLoS Negl Trop Dis. 2016;10(6):e0004727. doi:
10.1371/journal.pntd.0004727
PubMedGoogle Scholar 28.Parks
T, Mirabel
MM, Kado
J,
et al; Pacific Islands Rheumatic Heart Disease Genetics Network. Association between a common immunoglobulin heavy chain allele and rheumatic heart disease risk in Oceania.
Nat Commun. 2017;8:14946. doi:
10.1038/ncomms14946
PubMedGoogle ScholarCrossref 29.Gray
L-A, D’Antoine
HA, Tong
SYC,
et al. Genome-wide analysis of genetic risk factors for rheumatic heart disease in Aboriginal Australians provides support for pathogenic molecular mimicry.
J Infect Dis. 2017;216(11):1460-1470. doi:
10.1093/infdis/jix497
PubMedGoogle ScholarCrossref 32.Little
J, Higgins
JP, Ioannidis
JP,
et al. STrengthening the REporting of Genetic Association Studies (STREGA)—an extension of the STROBE statement.
Genet Epidemiol. 2009;33(7):581-598. doi:
10.1002/gepi.20410
PubMedGoogle ScholarCrossref 35.Nunes
MCP, Sable
C, Nascimento
BR,
et al. Simplified echocardiography screening criteria for diagnosing and predicting progression of latent rheumatic heart disease.
Circ Cardiovasc Imaging. 2019;12(2):e007928. doi:
10.1161/CIRCIMAGING.118.007928
PubMedGoogle Scholar 38.Loh PR, Danecek P, Palamara PF, et al. Reference-based phasing using the Haplotype Reference Consortium panel. Nat Genet. 2016;48(11):1443-1448.
40.McCarthy
S, Das
S, Kretzschmar
W,
et al; Haplotype Reference Consortium. A reference panel of 64,976 haplotypes for genotype imputation.
Nat Genet. 2016;48(10):1279-1283. doi:
10.1038/ng.3643
PubMedGoogle Scholar 41.Yang J, Zaitlen NA, Goddard ME, Visscher MP, Price AL. Advantages and pitfalls in the application of mixed-model association methods.
Nat Genet. 2014;46(2):100-106. doi:
10.1038/ng.2876 43.Pirinen
M.,, Donnelly
P., and Spencer
C.C., Efficient computation with a linear mixed model on large-scale data sets with applications to genetic studies.
Ann Appl Statistics. 2013;7(1):369-390. doi:
10.1214/12-AOAS586
Google ScholarCrossref 53.Ghoussaini
M, Mountjoy
E, Carmona
M,
et al. Open Targets Genetics: systematic identification of trait-associated genes using large-scale genetics and functional genomics.
Nucleic Acids Res. 2021;49(D1):D1311-D1320. doi:
10.1093/nar/gkaa840
PubMedGoogle ScholarCrossref 57.Pers
TH, Karjalainen
JM, Chan
Y,
et al; Genetic Investigation of ANthropometric Traits (GIANT) Consortium. Biological interpretation of genome-wide association studies using predicted gene functions.
Nat Commun. 2015;6(1):5890. doi:
10.1038/ncomms6890
PubMedGoogle ScholarCrossref 61.Andreson
R, Puurand
T, Remm
M. SNPmasker: automatic masking of SNPs and repeats across eukaryotic genomes.
Nucleic Acids Res. 2006;34(Web Server issue):W651-5. doi:
10.1093/nar/gkl125PubMedGoogle Scholar 65.Weiner
DJ, Wigdor
EM, Ripke
S,
et al; iPSYCH-Broad Autism Group; Psychiatric Genomics Consortium Autism Group. Polygenic transmission disequilibrium confirms that common and rare variation act additively to create risk for autism spectrum disorders.
Nat Genet. 2017;49(7):978-985. doi:
10.1038/ng.3863
PubMedGoogle ScholarCrossref 66.Maller
JB, McVean
G, Byrnes
J,
et al; Wellcome Trust Case Control Consortium. Bayesian refinement of association signals for 14 loci in 3 common diseases.
Nat Genet. 2012;44(12):1294-1301. doi:
10.1038/ng.2435
PubMedGoogle ScholarCrossref 67.Auckland
K, Mittal B, Cairns BJ,
et al. The human leukocyte antigen locus and susceptibility to rheumatic heart disease in South Asians and Europeans. Preprint Posted July 26, 2019. medRxiv 19003160. doi:
10.1101/19003160
70.Sallah
N, Carstensen
T, Wakeham
K,
et al. Whole-genome association study of antibody response to Epstein-Barr virus in an African population: a pilot.
Glob Health Epidemiol Genom. 2017;2:e18-e18. doi:
10.1017/gheg.2017.16
PubMedGoogle ScholarCrossref 72.Yengo
L, Sidorenko
J, Kemper
KE,
et al; GIANT Consortium. Meta-analysis of genome-wide association studies for height and body mass index in ∼700000 individuals of European ancestry.
Hum Mol Genet. 2018;27(20):3641-3649. doi:
10.1093/hmg/ddy271
PubMedGoogle ScholarCrossref 74.Allen
HD,
et al. Moss & Adams' Heart Disease in Infants, Children, and Adolescents: Including the Fetus and Young Adult. Lippincott Williams & Wilkins; 2013.