Monogenic and Polygenic Contributions to QTc Prolongation in the Population

Study Design and Patient Population

United Kingdom Biobank Cohort

The UKBB is a prospective cohort of ∼500,000 individuals from the UK with deep phenotyping and multiple genomic data types.⁷ Among participants with ECG data in the UKBB, we defined independent derivation and validation sub-cohorts for the heart-rate corrected QT interval (QTc) polygenic risk score (PRS). Two types of ECGs used in this analysis were performed on partially overlapping subsets of the UKBB: supine resting 12-lead ECGs (N=42,635) and 3-lead resting ECGs prior to a bicycle exercise protocol (N=60,799). Individuals who underwent both 12-lead and 3-lead ECGs (N=7,844) were only included in the validation sub-cohort and their QT interval was extracted from the 12-lead ECGs for this study. Heart rate corrected QT intervals were calculated using the Bazett formula, defined as (Supplemental Methods).

The derivation sub-cohort included participants from the UKBB with both imputed genomic data and the pre-exercise ECG data only. We excluded individuals with poor quality genetic data and with characteristics that would affect their QT interval such as Wolff-Parkinson-White Syndrome, a paced rhythm, QRS interval >120 ms, digoxin use or class I or III antiarrhythmic drug use. Individuals with 2^nd or 3^rd degree AV block and those with extreme heart rates, <40 or >120 beats per minute, were also excluded (Supplemental Figure 1A). The final derivation sub-cohort consisted of 44,979 individuals.

• Download figure
• Open in new tab

Figure 1.

Study design and flowchart. GWAS: genome-wide association study; PRS: polygenic risk score; P+T: pruning and thresholding; TOPMed: Trans-Omics for Precision Medicine; UKBB: United Kingdom Biobank.

The validation sub-cohort included participants from the UKBB with both imputed genomic data and QTc interval calculated from the resting 12-lead ECG. Again, we excluded participants with poor quality genetic data, and with characteristics that would affect their QT interval, which resulted in a validation sub-cohort comprising 39,800 individuals (Supplemental Figure 1A). Use of UKBB data was performed under application number 17488 and was approved by the local Massachusetts General Hospital institutional review board.

UKBB Cohort

The characteristics of the UKBB derivation and validation sub-cohorts were overall similar with few notable differences (Table 1). Individuals in the derivation sub-cohort were younger than those in the validation sub-cohort (mean age 52.7 ± 5.8 vs. 63.7 ±7.8 years, respectively).

Additionally, individuals in the derivation sub-cohort had a faster heart rate and shorter QT and QTc intervals as compared to individuals in the validation sub-cohort. While individuals of European ancestry constituted the majority of individuals in both the derivation and validation sub-cohorts, the derivation sub-cohort had a higher proportion of individuals of non-European ancestry (9.4% vs. 3.2%).

In the QTc interval GWAS in the derivation sub-cohort, we identified 20 independent loci that exceeded genome-wide significance (Figure 2). We replicated 16 loci previously reported to be associated with the QTc interval and identified 4 novel loci (TOGARAM2, FAM13B, CHD2 and ZNF358) (Supplemental Tables 1, 2). There was no evidence of genomic inflation (λ=1.08, Supplemental Figure 2). In TOPMed we replicated the association at the ZNF358 locus with the QTc interval (p-value=3.7×10⁻⁷) (Supplemental Table 3).

• Download figure
• Open in new tab

Figure 2.

QTc interval genome-wide association results across 22 autosomes in the UKBB derivation sub-cohort. Nearest genes are used for annotation. The dashed grey line represents the threshold for genome-wide significance (P<5×10⁻⁸).

Utilizing the GWAS results we generated 30 candidate PRS in the derivation cohort using pre-specified pruning and thresholding cut-offs and comparatively assessed their performance in the validation cohort. The best performing PRS in the validation cohort comprised 565 variants (linkage disequilibrium r²≤0.6 and p<5×10⁻⁴) and had an r² of 0.11 in the multivariable model including the PRS, age, genetically determined sex, beta blocker use, calcium channel blocker use, history of heart failure, history of myocardial infarction and principal components 1-12 of genetic ancestry (Supplemental Table 4).

TOPMed Cohort

Ancestral diversity was more prominent in TOPMed as compared to the UKBB. Individuals of European (N= 16,074, 59.6%) and African (N=4,887, 18.1%) ancestries constituted the largest ancestry groups in the TOPMed cohort. There was a substantial degree of admixture, with genetically-inferred ancestry unable to be determined for 3,653 (13.5%) individuals. The study population mean QT duration was 408.4 ± 31.6 ms and mean QTc duration was 423.8 ± 22.9 ms (Table 1).

The QTc polygenic risk score, derived and validated in the UKBB and comprising 565 common variants, was normally distributed in TOPMed and was significantly associated with the QTc interval (Pearson correlation coefficient 0.08, 95% CI 0.07 – 0.09; p-value=2.5×10⁻³⁸) (Supplemental Figure 3). Following adjustment for age, sex, beta blocker use, calcium channel blocker use, and principal components 1-12 of genetic ancestry, the QTc PRS remained significantly associated with the QTc interval (ΔQT_{c/decile of PRS} =0.9 ms, 95% CI 0.8 −1.0; p-value=1.1×10⁻⁶⁴).

• Download figure
• Open in new tab

Figure 3. Top PRS Decile and Putative Pathogenic Rare Variant Carrier Distribution Across QTc Strata.

We collapsed putative pathogenic rare variant carrier status into two categories: (1) carriers of rare variants in any of the 10 LQTS genes (N=160) and (2) carriers of rare variants in KCNQ1, KCNH2 or SCN5A (N=58). In multivariable analyses, carriers of putative pathogenic rare variants in any of the 10 LQTS genes and those who were carriers of putative pathogenic rare variants in KCNQ1, KCNH2 or SCN5A had a 10.9 ms (95% CI, 7.4 - 14.4, p-value=1.1×10⁻⁹) and 26.5 ms (95% CI, 20.7 - 32.3, p-value=4.7×10⁻¹⁹) increase in their QTc interval, respectively, as compared to non-carriers (Supplemental Table 5). After excluding heterozygous carriers of TRDN rare variants, carriers of putative pathogenic rare variants in any of the 10 LQTS genes had a 22.7 ms (95% CI, 17.5 – 27.9, p-value 1.2 x10⁻¹⁷) increase in their QTc interval as compared to non-carriers.

We then examined the distribution of polygenic and monogenic risk across the spectrum of QTc intervals in the study population to assess the contribution of genetic determinants to the QTc interval. To approximate monogenic and polygenic risk, we categorized the polygenic risk score into two categories (highest decile vs. lower 90^th percentile). The mean difference in QTc interval between the top and lowest PRS deciles (8.1 ms) approximates the delta QTc observed among carriers of putative pathogenic rare variants in LQTS genes compared to non-carriers. Across increasing strata of QTc intervals in the population, there was a consistent increase in the proportion of individuals with a PRS in the top decile (p-trend=3.8×10⁻²⁷) and the proportion of individuals who were carriers of monogenic putative pathogenic rare variants in LQTS genes (p-trend=4.1×10⁻¹⁴) (Figure 3). However, we found that only 3.4% of individuals with pronounced QTc prolongation (>480 ms) carried a monogenic rare variant in a LQTS gene, 13.6% had a polygenic risk score in the top decile and 16.7% had either a monogenic rare variant or a polygenic risk score in the top decile (Figure 3). Thus, over 80% of individuals with pronounced QTc prolongation (>480 ms) do not have an identified rare genetic variant or burden of polygenic risk equivalent to a monogenic variant to explain the QTc prolongation. Following exclusion of heterozygous carriers of TRDN variants from the putative pathogenic rare variant carrier group, 3.1% of individuals with pronounced QTc prolongation (>480 ms) were carriers of a monogenic rare variant.

Next, we assessed for effect modification between polygenic risk and putative pathogenic rare variant carrier status. Across strata of carrier status for putative pathogenic rare variants in LQTS genes, the QTc polygenic risk score was associated with a larger increase in the QTc interval in carriers (ΔQT_{c/decile of PRS} =2.0 ms, 95% CI 0.2 - 3.8; p-value=0.03) as compared to non-carriers (ΔQT_{c/decile of PRS} =0.9 ms, 95% CI 0.8 - 1.0; p-value=3.9×10⁻⁶⁴) (p-interaction=0.005; Figure 4A). The effect was more pronounced among carriers of putative pathogenic rare variants confined to KCNQ1, KCNH2 and SCN5A where the QTc on average increased by 5.8 ms per decile of the PRS (ΔQT_{c/decile of PRS} =5.8 ms, 95% CI 1.6 - 9.9; p-value=0.008) (p-interaction=0.01; Figure 4B). Consistent with the above findings, we found evidence of effect modification by the PRS on the association of putative pathogenic rare variant carrier status with the extent of QTc prolongation. QTc prolongation associated with putative pathogenic rare variants in LQTS genes consistently increased across increasing tertiles of the polygenic risk score (lowest PRS tertile: ΔQTc_{carrier/non-carrier}= 4.8 ms [-1.2 - 10.7]; intermediate PRS tertile: ΔQTc_{carrier/non-carrier}= 9.4 ms [3.5 - 15.4]; highest PRS tertile: ΔQTc_{carrier/non-carrier}= 18.9 ms [12.8 - 25.1]; p-interaction=0.001). A similar pattern was observed when examining putative pathogenic rare variants in the three genes with definitive evidence for typical LQTS¹³ (p-interaction=0.006) (Figure 5) and after excluding heterozygous carriers of TRDN rare variants (Supplemental Figure 4).

• Download figure
• Open in new tab

Figure 4.

Variation in the Predicted QTc Across Percentiles of the Polygenic Risk Score stratified by Putative Pathogenic Rare Variant Carrier Status in all long QT Syndrome genes (A) and in KCNQ1, KCNH2 and SCN5A (B).

• Download figure
• Open in new tab

Figure 5.

Multivariable Association of Putative Pathogenic Rare Variant Carrier Status with QTc Interval Duration Across Tertiles of QTc Polygenic Risk Score. Low, intermediate and high PRS strata reflect first, second and third tertiles of the PRS in the study sample, respectively.

Most carriers (80.6%) and non-carriers (93.4%) of putative pathogenic rare variants in LQTS genes had normal QTc durations (<460 ms). Among carriers of putative pathogenic rare variants with prolonged QTc interval (≥460 ms), we observed a 2-fold enrichment in the odds for high polygenic risk as compared to carriers with normal QTc interval (<460 ms). Conversely, there was only a 1.2-fold enrichment in the odds for high polygenic risk among non-carriers of putative pathogenic rare variants with prolonged vs. normal QTc interval (Figure 6). Notably, 73% of putative pathogenic rare variant carriers with normal QTc< 460 ms had low/intermediate polygenic risk which potentially contributes to their normal QTc interval despite carrying a putative pathogenic rare variant in a LQTS gene.

• Download figure
• Open in new tab

Figure 6.

Distribution of Low/Intermediate and High Polygenic Risk in Individuals with Normal (<460 ms) and Prolonged QTc Intervals (≥ 460 ms) among Carriers and non-Carriers of Putative Pathogenic Rare Variants in Long QT Syndrome genes. Low, intermediate and high PRS strata reflect first, second and third tertiles of the PRS in the study sample, respectively.

To explore potential mediators of the observed effect modification of the PRS on the association of putative pathogenic rare variants with the QTc interval, we performed a number of sensitivity analyses. First, after excluding common variants in previously described genetic modifiers of LQTS (NOS1AP and KCNE1), our findings remained unchanged suggesting that additional distinct variants underlie the observed effect modification (Supplemental Figure 5A). Second, following omission of common variants within ± 1MB from the 10 evaluated LQTS genes from the PRS, which we assume capture most of the regulatory variation likely to influence gene expression, the interaction between the polygenic risk and putative pathogenic rare variant carrier status persisted (Supplemental Figure 5B). Third, exclusion of the top 10% of common variants by effect size to assess whether our findings were mediated by the top common variants most strongly associated with the QTc interval did not alter our findings (Supplemental Figure 5C). Additionally, we note that carriers of rare variants in each of the 10 LQTS genes were randomly distributed across the spectrum of polygenic risk in the study sample (Supplemental Figure 6) and thus the identified interaction is not mediated by non-random clustering of carriers of rare variants in a particular gene across the polygenic risk distribution. Lastly, we performed a stratified analysis by European vs. non-European ancestry in TOPMed and found an overall consistent pattern of effect modification between the PRS and putative pathogenic rare variants in LQTS genes (beta for interaction 1.23, 95 % CI (−0.30 - 2.75) and 1.20, 95% CI (−0.96 - 3.35) for European and non-European ancestry, respectively); power was limited by the small sample size of putative pathogenic rare variant carriers in each ancestral group (Supplemental Figure 7).

Genetic Determinants of the QT Interval

The QT interval is a complex and dynamic heritable trait with both monogenic and polygenic determinants.^2,15 Heritability estimates in unselected community-based populations range from 30-40%.¹⁶ Comprehensive genetic evaluation of individuals with QTc prolongation without other criteria for LQTS may identify a small subset with a possible genetic explanation for the QTc prolongation. Indeed, nearly 1 in 6 individuals with marked QTc prolongation, defined as QTc > 480 msec, carried a putative pathogenic rare variant in a LQTS gene or high polygenic risk equivalent in this study with a larger contribution from polygenic risk.

Current clinical genetic testing for evaluation of prolonged QTc interval is focused on identification of rare putative pathogenic variants in LQTS genes.¹⁴ Our findings highlight the sizeable genetic contribution of common variant polygenic risk and its role in comprehensive assessment of genetic determinants of the QTc. Nevertheless, over 80% of individuals with prolonged QTc interval did not carry a monogenic putative pathogenic rare variant or a polygenic risk equivalent in this study. Our findings support the possibility that additional genetic contributions to QTc interval duration remain undiscovered. Future increasingly large-scale common and rare variant population-based association studies of the QTc interval may shed light on additional novel genetic determinants of the QTc interval.

Polygenic Risk and Susceptibility to QTc Prolongation

In this study, we show a 4-fold increase in the magnitude of QTc prolongation associated with putative pathogenic rare variants among individuals with high polygenic risk as compared to those with low polygenic risk (ΔQTc 18.9 vs. 4.8 ms). Additionally, putative pathogenic rare variant carriers with prolonged QTc ≥ 460 ms had a 2-fold higher odds of high background polygenic risk as compared to those with QTc intervals <460 ms. With enhanced access to sequencing methods in clinical settings and genomic sequence data emerging from large-scale biobanks, increasingly more unaffected carriers of putative pathogenic rare variants in LQTS genes will come to attention. Our findings demonstrate the role of polygenic risk in modulating phenotypic expression of putative pathogenic rare variants in LQTS genes and may inform why some carriers exhibit QTc prolongation and others do not.

The mechanisms by which polygenic risk modulates the phenotypic expression of putative pathogenic rare variants in LQTS genes remains poorly understood. Repolarization reserve reflects redundancies in biological pathways that govern cardiac repolarization and protect against marked QT prolongation from a single pathologic insult.¹ Akin to the “multiple-hit” hypothesis in cancer biology, pathologic prolongation of the QTc interval occurs in the presence of multiple lesions that converge and significantly impinge on repolarization reserve. Polygenic risk scores of the QTc interval offer an opportunity to collectively assess multiple biologic pathways involved in cardiac repolarization and may serve as a genetic surrogate for repolarization reserve hence, capturing an individual’s susceptibility to QTc prolongation in the presence of a putative pathogenic rare variant in an LQTS gene.

Previous studies have demonstrated that polygenic risk may have important implications for determining susceptibility to QT prolonging drugs as well as expressivity of the LQTS phenotype. A study demonstrated the utility of a QTc PRS, comprising 61 variants, in predicting the variability in QTc prolongation associated with different anti-arrhythmic drugs and the associated risk of Torsade des Pointes.¹² Additionally, QTc polygenic risk has been shown to be higher on average among LQTS cases, as compared to controls, further highlighting the contribution of polygenic risk to disease susceptibility.¹⁷

Genetic Modifiers and Long QT Syndrome Genes

Studies of LQTS families have revealed incomplete penetrance of LQTS.¹⁸ A number of studies have investigated individual common genetic modifiers associated with the QT interval that may underlie incomplete penetrance and variable expressivity in cohorts of patients with LQTS.^3–6 Common variants in NOS1AP, a gene identified to be strongly associated with the QT interval in the general population², have been shown to influence phenotype expression and outcomes in LQTS populations.^3,4 KCNE1-D85N a common polymorphism in KCNE1, has been reported to be more prevalent among patients with congenital LQTS and individuals with acquired drug-induced LQTS as compared to healthy controls.⁵ A number of other genetic modifiers of LQTS, including common variants in AKAP9, KCNH2, KCNQ1 and REM2, have been also described in the literature.⁶

In this study we expand upon previous findings of single common variant genetic modifiers and demonstrate a role for common variant polygenic risk scores in modulating the magnitude of QTc prolongation imparted by pathogenic rare variation in genes known to be associated with LQTS. The effect modification of polygenic risk on phenotypic expression of putative pathogenic rare variants persisted following exclusion of common variants in previously described genetic modifiers (NOS1AP and KCNE1) and common variants with strong effect size, highlighting the incremental role of comprehensive common variant risk assessment using polygenic risk scores in capturing the interplay between common and rare variation.

Additionally, exclusion of common variants within ±1 MB of LQTS genes did not alter our findings suggesting that effect modification is not solely driven by modulation of LQTS gene expression or their gene product. Further studies are needed to examine the biologic pathways and mechanisms underlying the observed effect modification.

Polygenic and Monogenic Contribution to Other Genomic Disorders

Recently, the interplay between monogenic and polygenic risk has been described for a number of genomic disorders. A contemporary study examining a spectrum of diseases including familial hypercholesterolemia, hereditary breast and ovarian cancer and Lynch syndrome found that polygenic risk explained more variation in disease risk among carriers vs. non-carriers of monogenic rare variants and aided in better risk estimation among carriers.¹⁹ Our group has previously reported higher penetrance of atrial fibrillation among TTN loss-of-function carriers within higher tertiles of polygenic risk of atrial fibrillation.²⁰ Expanding these findings to carriers of monogenic rare variants in LQTS genes is of particular relevance given low penetrance in LQTS¹⁸ and the implications of living with sudden cardiac death risk among carriers of putative pathogenic rare variants in LQTS genes.

Limitations

Our results should be interpreted in the context of the study design. The study findings apply to individuals who enrolled in the included population-based cohorts, and do not necessarily apply to patients seeking clinical evaluations for evident or possible LQTS. The TOPMed cohort did not ascertain LQTS diagnosis among study participants which prohibited performing an analysis examining the contribution of polygenic and monogenic risk to LQTS in the general population and our findings strictly apply to the QTc interval. Further population-based studies with LQTS phenotyping are needed to address this question. Additionally, despite studying a large multi-ancestry cohort from the UKBB to derive and validate our PRS, the population was predominantly of European ancestry which may not accurately characterize polygenic risk in the TOPMed cohort. However, in a sensitivity analysis, we found overall similar findings after stratifying the TOPMed cohort by European vs. non-European ancestry. Our analyses of monogenic variation focused on coding variants within established LQTS genes. Additional studies are needed to evaluate the contribution of non-coding rare variation, copy-number variation and structural variation to the QTc interval in the general population. This study focused on exploring genetic determinants of the QTc interval, however, multiple non-genetic factors contribute to the QTc interval, which we were unable to fully account for in this analysis. Lastly, further studies are needed to examine the impact of polygenic risk on clinical outcomes of carriers of monogenic rare variants in LQTS genes in the general population.

Perspectives

Competency in Medical Knowledge: Population based QTc duration is influenced by both monogenic rare variation in established LQTS genes and polygenic risk, as well as non-genetic factors. One in 6 individuals with QTc > 480 ms have an identified genetic risk factor. Polygenic risk modulates the phenotypic expression of monogenic rare variants in established LQTS genes.

Translational Outlook 1: QTc polygenic risk informs the assessment of monogenic rare variant carriers in LQTS genes in the general population and contributes to the understanding of why certain individuals who carry putative pathogenic rare variants in LQTS genes have prolonged QTc intervals and others do not.

Translational Outlook 2: The role of QTc polygenic risk in modulating penetrance in LQTS and cardiovascular outcomes among carriers of monogenic rare variants in LQTS genes in the general population should be investigated.