Combining Clinical and Polygenic Risk Improves Stroke Prediction Among Individuals With Atrial Fibrillation
Circulation: Genomic and Precision Medicine
Abstract
Background:
Atrial fibrillation (AF) is associated with a five-fold increased risk of ischemic stroke. A portion of this risk is heritable; however, current risk stratification tools (CHA2DS2-VASc) do not include family history or genetic risk. We hypothesized that we could improve ischemic stroke prediction in patients with AF by incorporating polygenic risk scores (PRS).
Methods:
Using data from the largest available genome-wide association study in Europeans, we combined over half a million genetic variants to construct a PRS to predict ischemic stroke in patients with AF. We externally validated this PRS in independent data from the UK Biobank, both independently and integrated with clinical risk factors. The integrated PRS and clinical risk factors risk tool had the greatest predictive ability.
Results:
Compared with the currently recommended risk tool (CHA2DS2-VASc), the integrated tool significantly improved Net Reclassification Index (2.3% [95% CI, 1.3%–3.0%]) and fit (χ2 P=0.002). Using this improved tool, >115 000 people with AF would have improved risk classification in the United States. Independently, PRS was a significant predictor of ischemic stroke in patients with AF prospectively (hazard ratio, 1.13 per 1 SD [95% CI, 1.06–1.23]). Lastly, polygenic risk scores were uncorrelated with clinical risk factors (Pearson correlation coefficient, −0.018).
Conclusions:
In patients with AF, there appears to be a significant association between PRS and risk of ischemic stroke. The greatest predictive ability was found with the integration of PRS and clinical risk factors; however, the prediction of stroke remains challenging.
Introduction
Atrial fibrillation (AF) is the most common cardiac arrhythmia, and its prevalence is increasing.1 Atrial fibrillation itself can cause substantial morbidity, including a 5-fold increased risk of ischemic stroke.2
To help prevent the thromboembolic complications of AF, selected patients are offered prophylactic anticoagulation. This prophylaxis is highly effective in the right patient,3–5 but the selection of these patients remains difficult.6,7 The current gold standard risk stratification tool is an amalgamation of clinical risk factors (CHA2DS2-VASc).4 However, there are limitations in the development, validation, and performance of CHA2DS2-VASc. Most notably, the small number of AF patients in the development (n=1084),6 and short follow-up, small numbers, and conflicting performance in validation studies.8 Additionally, CHA2DS2-VASc tool does not include family history or genetic risk of ischemic stroke, despite evidence suggesting the risk of ischemic stroke is heritable (≈40% heritability).9 Previous research has shown that polygenic risk scores are comparable to clinical risk factors in the prediction of ischemic stroke in the general population10; however, this has not been extended into patients with AF, nor did it examine CHA2DS2-VASc.
Given the known heritability of ischemic stroke, and the apparent need to improve the existing gold standard risk tool (CHA2DS2-VASc), we set out to construct a polygenic risk score (PRS), and then an integrated genetic and clinical risk tool (CHA2DS2-VASc+PRS) to help predict which patients with AF will go on to develop ischemic stroke.
Methods
In keeping with journal policy, the full methods are in the Data Supplement. Anonymized PRS data have been made publicly available at The Polygenic Score Catalog and can be accessed at http://www.pgscatalog.org/. In short, we followed a similar study design to previously published PRS articles,10–14 in line with recommended methodological15 and reporting guidance.16 We will briefly describe the five broad steps we completed in this paragraph (Figure 1); we elaborate on each of these steps individually in the Data Supplement. The 5 steps were (1) curation of previously published genome-wide association study (GWAS) summary statistics, (2) accounted for linkage disequilibrium in GWAS summary statistics, using the R package lassosum,17 (3) construction of PRS (see Methods in the Data Supplement) in our UK Biobank Prevalent cohort. Eighty different PRS were constructed across the lassosum hyperparameters (λ and s). (4) Determined the most accurate PRS in the UK Biobank Prevalent Cohort. (5) The PRS with the greatest predictive accuracy (from step 4) was then validated in the UK Biobank incident cohort. The UK Biobank received ethical approval from the National Health Service’s National Research Ethics Service North West (11/NW/0382). This research was conducted using the UK Biobank under Application Numbers 24 983 and 22 282.
Results
The characteristics of the UK Biobank incident cohort are reported in Table 1; there were 15 929 participants with AF (Figure I in the Data Supplement), of which 684 suffered an ischemic stroke, and 15 245 did not, over follow-up. Participants were followed up for a median of 7 years (interquartile range, 5.9; 25th percentile, 3.6; 75th percentile, 9.5). After reweighting, 530 933 single nucleotide variants (SNVs) had a nonzero effect size and were included in our PRS.
| All (n=15 929) | Ischemic stroke (n=684) | No ischemic stroke (n=15 245) | |
|---|---|---|---|
| Age at recruitment, mean (SD) | 62.8 (5.9) | 64.3 (5.3) | 62.7 (5.9) |
| Age at AF diagnosis, mean (SD) | 63.7 (8.8) | 65.5 (7.4) | 63.6 (8.8) |
| Time to ischemic stroke (or censor), mean (SD) | 6.4 (3.3) | 1.9 (2.4) | 6.6 (3.2) |
| Sex, % (male) | 10 620 (66.7%) | 461 (66.2%) | 10 159 (66.6%) |
| CHA2DS2-VASc=0 | 4270 (26.8%) | 127 (18.6%) | 4143 (27.2%) |
| CHA2DS2-VASc=1 | 5535 (34.8%) | 241 (35.2%) | 5294 (34.7%) |
| CHA2DS2-VASc=2 | 3880 (24.4%) | 188 (27.5%) | 3692 (24.2%) |
| CHA2DS2-VASc=3 | 1595 (10.0%) | 94 (13.7%) | 1501 (9.9%) |
| CHA2DS2-VASc=4 | 528 (3.3%) | 27 (4.0%) | 501 (3.3%) |
| CHA2DS2-VASc=5 | 110 (0.7%) | 5 (0.7%) | 105 (0.7%) |
| CHA2DS2-VASc=6 | 11 (0.1%) | 2 (0.3%) | 9 (0.1%) |
| HTN, N | 6690 (42.0%) | 314 (45.9%) | 6376 (41.8%) |
| Vascular disease, N | 1852 (11.6%) | 105 (15.4%) | 1747 (11.5%) |
| Heart failure, N | 1544 (9.7%) | 76 (11.1%) | 1468 (9.6%) |
| Diabetes, N | 1678 (10.5%) | 91 (13.3%) | 1587 (10.4%) |
| Warfarin, N | 2310 (14.5%) | 93 (13.6%) | 2217 (14.5%) |
Descriptive characteristics of UK Biobank Incident Cohort at AF diagnosis. AF indicates atrial fibrillation; and HTN, hypertension.
The logistic regression analyses showed an area under the receiver operating characteristics for CHA2DS2-VASc of: 0.60 (95% CI, 0.58–0.62; Table II in the Data Supplement), with the addition of PRS this rose to 0.61 (95% CI, 0.59–0.63) and corresponded to a PRS odds ratio of 1.14 per SD (95% CI, 1.06–1.23). These results were largely consistent when CHA2DS2-VASc was modeled collectively or individually (Table II in the Data Supplement). The analysis using PRS as the sole predictor revealed a PRS odds ratio, 1.14 per SD (95% CI, 1.06–1.23; Figure 2), and an area under the receiver operating characteristics of 0.60 (95% CI, 0.58–0.62). Warfarin was prescribed to 2326 AF participants at UK Biobank recruitment. The sensitivity analyses both adjusting for warfarin and removing participants prescribed warfarin revealed modestly improved discrimination (Table II and Figure II in the Data Supplement), albeit with a loss of power. Further, we found that PRS and an individual’s CHA2DS2-VASc score was significantly, but poorly correlated (Pearson correlation coefficient, −0.019 [95% CI, −0.035 to −0.0035], P=0.02), Table III in the Data Supplement.
The Cox regression analysis using PRS as the sole predictor revealed a hazard ratio (HR) of 1.13 (95% CI, 1.04–1.21) per 1 SD, and a C statistic of 0.56 (95% CI, 0.54–0.58; Figure 3). The same analysis adjusting for age violated the Cox proportional hazards assumption, but nevertheless showed a similar PRS HR (1.14 [95% CI, 1.01–1.23]), and higher C statistic (0.63 [95% CI, 0.61–0.65]). The analysis that adjusted for warfarin prescription violated the Cox proportional hazards assumption but nevertheless showed a similar HR and C statistic. The analysis that removed participants prescribed warfarin did not violate the Cox proportional hazards assumption and showed an HR of 1.13 (95% CI, 1.04–1.22) and a C statistic of 0.57 (95% CI, 0.54–0.59). The sensitivity analyses using age as time scale for all aforementioned analyses showed largely consistent results (Sensitivity Analysis 2, Table IV in the Data Supplement). All models were well-calibrated (P>0.2 for all models via the Greenwood-Nam-D’Agostino χ2 test).
The C statistic for the integrated PRS and CHA2DS2-VASc model was 0.61 (95% CI, 0.58–0.63; Figure 3). Compared with the currently recommended CHA2DS2-VASc only model, the integrated PRS and CHA2DS2-VASc risk model showed a significantly improved statistical fit (χ2 P=0.002), modestly improved discrimination (Figure 3) and improved overall Net Reclassification Index: 2.3% (95% CI, 1.3%–3.0%; Table V in the Data Supplement). The Net Reclassification Index (method 1 in Methods) was significantly improved for noncases (2.3% (95% CI, 0.6%–5.4%) and no different for cases (ischemic strokes; 0.01% (95% CI, −0.4% to 0.1%; Table V in the Data Supplement). Both models were well calibrated (P>0.3 for both models via the Greenwood-Nam-D’Agostino χ2 test). The sensitivity analysis using a risk threshold of 5% showed similar significance overall, but lower Net Reclassification Index (1.91% [0.12%–6.4%]). This Net Reclassification Index was significant for noncases, but not for cases (Sensitivity Analysis 2, Table VI in the Data Supplement). Tables VII, VIII, and IX in the Data Supplement provide further visualization of results and sensitivity analyses. Lastly, the model where the individual components of the CHA2DS2-VASc score were considered (Cox regression method 6 above) violated the Cox proportional hazards assumption, but nevertheless showed consistent results (Figure III in the Data Supplement).
For the genomically enhanced CHA2DS2-VASc scores (CHA2DS2-VASc-G), the proportion of strokes observed in participants that were upclassified (moved from below the anticoagulation threshold with CHA2DS2-VASc to above the anticoagulation threshold with CHA2DS2-VASc-G) was similar (statistically no different) to the proportion of strokes in participants who were classified as high risk in both CHA2DS2-VASc and CHA2DS2-VASc-G (above anticoagulation threshold with both): 5.8% versus 5.4%, P=0.7, Table 2, Figure 4, and Table X in the Data Supplement (method 2 in Methods). This indicated that the upclassified participants had a similar stroke risk to those at shared high risk. Similarly, the proportion of strokes in participants downclassified (moved from above the anticoagulation threshold with CHA2DS2-VASc to below the anticoagulation threshold with CHA2DS2-VASc-G) was no different from the proportion of strokes in participants who were classified as low risk in both CHA2DS2-VASc and CHA2DS2-VASc-G (below anticoagulation threshold for both): 3.9% versus 3.7%, P=0.9, Table 2, and Figure 4 (method 2 in Methods). This indicated that the stroke risk between those downclassified and those at shared low risk was similar. These results were comparable to that observed when the reclassification table was calculated using the 4% risk threshold (Table VII in the Data Supplement). These results were almost identical when participants that were prescribed warfarin were removed (Sensitivity Analysis 3, Table VIII in the Data Supplement) and similar when we recalculated this table without subtracting a point from participants who were in the bottom 25% PRS risk (Sensitivity Analysis 4, Table IX in the Data Supplement). Furthermore, congruent results were found when comparing the risk of those in the top 25% PRS risk versus the bottom PRS 25% (Table XI in the Data Supplement).
| Category | CHA2DS2-VASc (%) | CHA2DS2-VASc-G (%) | CHA2DS2-VASc-G breakdown (%) | n Ischemic stroke over f/u (%) |
|---|---|---|---|---|
| Anticoagulation recommended | 4863 (30.5%) | 5310 (33.3%) | 4083 (77.0%) → | 221 (5.4%) |
| 1227 (23.0%) ↑ | 71 (5.8%) | |||
| Anticoagulation not recommended | 11 066 (69.5%) | 10 619 (66.7%) | 9839 (92.7%) → | 362 (3.7%) |
| 780 (7.4%) ↓ | 30 (3.9%) |
Reclassification table comparing classification using conventional risk prediction tool (CHA2DS2-VASc) compared with our integrated genetic and clinical risk factors tool (CHA2DS2-VASc-G). Up and down arrows denote up or downclassified participants respectively: Up arrow denotes participants who were moved from below the anticoagulation threshold (with CHA2DS2-VASc) to above the anticoagulation threshold (with CHA2DS2-VASc-G). Down-arrow denotes participants who were moved from above the anticoagulation threshold (with CHA2DS2-VASc) to below the anticoagulation threshold (with CHA2DS2-VASc-G). Horizontal arrows represent participants who stay in the same category for both risk tools. The last column shows the observed number of ischemic strokes in the different reclassification groups over follow-up.
Discussion
We constructed a PRS for predicting ischemic stroke in patients with an established diagnosis of AF. Our analysis was based on over 15 000 participants in a well-conducted, prospective national biobank (UK Biobank). We extracted SNVs from the largest GWAS18 and combined over half a million SNVs to construct our PRS. Additionally, we built an integrated genomic and clinical risk tool, integrating our PRS with the current gold standard risk tool (CHA2DS2-VASc).
Our results show that a PRS is individually predictive of ischemic stroke in patients with an established diagnosis of AF, and this predicted risk appears independent of established clinical risk factors. The combined PRS and clinical risk tool shows significantly improved risk prediction over the current gold standard risk tool (CHA2DS2-VASc). These improvements, when applied to the large number of people with AF, translate to improved risk classification in thousands of people in the United States. Nevertheless, the prediction of ischemic stroke remains challenging.
Polygenic risk scores have been produced for many cardiovascular diseases.10,11,14,19,20 A large number of PRS have been produced for coronary artery disease11,14,19 and also for ischemic stroke.10 All of these have been constructed for the population at large, compared with our PRS, which predicts an outcome (ischemic stroke) within a group of people with an established diagnosis (of AF). A minority of these previous articles have compared PRS to clinical risk factors, and even fewer have integrated PRS with clinical risk factors. Abraham et al10 constructed a PRS for ischemic stroke for all people (not just AF) and found an HR of 1.26 (95% CI, 1.22–1.31) and C statistic of 0.58 (95% CI 0.57–0.59). The inclusion of all participants may explain the slightly higher HR observed by Abraham et al.10 Aside from their inclusion of all participants, our study differs as we compared (and adjusted for) clinical risk factors specific to stroke in AF (eg, CHA2DS2-VASc), we integrated our PRS with CHA2DS2-VASc, and determined reclassification between models.
Implications for Patients and Clinicians
Our results have implications for patients, researchers, and policymakers.21,22 First, the integration of any new innovation in clinical medicine should center around patients. The integrated PRS and CHA2DS2-VASc tool significantly improves risk classification of patients; this means the tool can help identify which AF patients will likely benefit from prophylactic anticoagulation and also potentially prevent people from unnecessary anticoagulation prescription. This latter point spares people from an increased bleeding risk, as well as the costs associated with prescription (for an individual) and the costs associated with bleeding complications (for hospital and society). In the United States, there are an estimated 5.1 million people with AF23,24; using the improved, integrated PRS and CHA2DS2-VASc tool, 117 000 patients with AF would have improved risk stratification.
Implications for Policy Makers
Second, our results may be of value to policymakers. Risk tools are continually updated, mainly when new covariates are identified that improve model fit and prediction. This is evident through the history of the currently recommended CHA2DS2-VASc score. It was initially developed as the CHADS2 tool,25 and with the emergence of evidence suggesting new covariates that improve the model, the tool was updated to CHA2DS2-VASc.6,7 Our article shows that the addition of PRS improves CHA2DS2-VASc, and future iterations of guidelines may benefit from considering its addition. This is particularly relevant due to the decreasing costs of genetic testing,26 genetic data is attained (and paid for) once, and there are multiple clinical uses once genetic data has been attained.
Implications for Researchers
Third, our study design may be of interest to other researchers. We chose to construct a PRS that (1) could lead to actionable clinical changes and (2) exists within current clinical practice. Currently, if a patient with AF has a CHA2DS2-VASc of score of ≥2 (in men) or ≥3 (in women), guidelines suggest they should be recommended anticoagulation.4 This represents an actionable decision threshold, based on an accepted threshold of risk. We modeled the addition of a new covariate (PRS; method 1) into an existing risk tool (CHA2DS2-VASc) at an accepted percentage risk threshold (≥2 in men or ≥3 in women) to initiate an actionable outcome (prescription of prophylactic anticoagulation). Additionally, we recalculated each participant’s CHA2DS2-VASc score with the addition of their PRS (method 2). Lastly, we plan to make our PRS available upon publication at http://www.pgscatalog.org/.
Study Limitations
Our study should be interpreted with an understanding of its limitations. Our study was limited by the demographics of the UK Biobank. Most notably, the UK Biobank is of primarily European ancestry, and we only included those of European ancestry in this study; the UK Biobank recruited participants aged 40 to 69 years old and who are healthier and more affluent than the general UK population.27 Most studies that used the UK Biobank were only able to follow participants for 7 years; we were able to overcome this limitation as our outcome of interest (ischemic stroke) is algorithmically defined centrally by the UK Biobank using an amalgamation of non-UK Biobank data (eg, electronic health records from primary care, etc). However, primary care data has only been released from around half of the UK Biobank participants (≈230 000 participants).
Furthermore, it is plausible that participants were prescribed an anticoagulant after recruitment. Anticoagulant prescription could affect our study results by lowering the risk of ischemic stroke; however, we think our results remain robust for 2 reasons: (1) Our sensitivity analyses, both adjusting for and removing participants prescribed anticoagulation at baseline, showed almost identical results to our main analysis, and this was true for both logistic and Cox regression, and (2) the prescription of an anticoagulant is likely to lead to an underestimation of the predictive power of PRS, as polygenic risk seems to mirror anticoagulation prescription (Figure II in the Data Supplement). This limitation is also observed in PRS for CAD (with statins).11,14 Furthermore, novel anticoagulants (NOACs, also known as direct anticoagulants) were not approved for stroke prophylaxis in patients with AF in the UK at the time of UK Biobank recruitment (dabigatran and rivaroxaban were approved in 2008 and apixaban in 2011),28 nor were they used commonly in the UK until after 2015 (the first time their use was greater than warfarin).28 Thus, NOACs were not listed in the medications code (https://biobank.ctsu.ox.ac.uk/crystal/coding.cgi?id=4&nl=1) and no participants were on NOACs at recruitment. Given NOACs have been shown to be more effective at preventing strokes in patients with AF,5 their absence from our study may actually be advantageous; we were able to capture more closely the nonintervention incidence of stroke, without the mitigating effects of NOACs. A further limitation is the inclusion of nonspecific phenotype outcomes; we used ischemic stroke as our outcome rather than cardioembolic ischemic stroke (specific to AF). Unfortunately, the UK Biobank does not stratify ischemic stroke outcomes into subtypes. This may further explain why we observe slightly lower HR per 1 SD than was observed by Abraham et al.10 Similarly, stroke phenotyping presents unique challenges. Hemorrhagic and ischemic strokes can present similarly and ischemic strokes can transform into hemorrhagic strokes. It thus seems plausible that some patients are misphenotyped (at the UK Biobank level).
Furthermore, we used GWAS summary statistics originally derived on patients with stroke (MEGASTROKE18). There are no GWAS data available that are specific to patients with AF who go on to suffer from an ischemic stroke. However, we did use the cardioembolic GWAS summary statistics within MEGASTROKE. Our results align with other studies that used the UKBiobank to test a PRS for stroke,10,29 although we focused exclusively on patients with AF and used a larger number of SNVs than most previous studies. Additionally, our sample size is smaller than previous PRS studies that have used the UK Biobank. This is because we only included participants with AF. A more broad weakness is the limitations of CHA2DS2-VASc; its discriminative ability is poorer than other cardiovascular risk tools (eg, American Heart Association/American College of Cardiology’s pooled cohort equation) and hence improving on it is somewhat expected. Nevertheless, we feel integrating PRS with the current gold standard risk tool is important, even if the gold standard is sub-optimal. Lastly, our reclassification method 2 used the top and bottom 25% risk to add and subtract a point, respectively, from participant’s CHA2DS2-VASc score. The decision to use 25% is based on previous research, but future research in larger cohorts may reveal a different threshold. A further weakness of reclassification method 2 is this categorization of PRS into top 25%, middle 50%, and bottom 25% risk. We performed reclassification method 2 to align with the currently recommended CHA2DS2-VASc score (which adds up risk factors and then anticoagulation is recommended above a threshold). The categorization of PRS was necessary for this analysis, although diminishes the power of the analysis. We acknowledge this limitation and performed reclassification method 1 to address this (which does not categorize PRS).
Future Research
Our study identifies several research priorities. First, even with the addition of PRS, ischemic stroke remains difficult to predict. There are numerous factors that likely contribute to this. Namely, heterogeneous phenotyping and unidentified risk factors. Future research resources should focus on improving phenotyping, electronic health records, and the identification of new stroke risk factors (eg, biomarkers, transcriptomics). Second, it would be advantageous for formal health economics studies to evaluate the cost-effectiveness of potentially implementing a combined PRS and clinical risk factors tool. Third, a tool that integrates the genetic risk of both ischemic stroke and bleeding is likely to be most useful to clinicians and patients, as they weigh up the risk and benefits of prophylactic anticoagulation. Lastly, it is vital that future GWAS, polygenic risk scores, and large biobanks include non-European populations. We were unable to include non-Europeans in our study as <5% of UK Biobank participants are non-European. Thus the number of participants with AF that are non-European is likely to be less than 1000 and the number of participants that are non-European with AF who had an ischemic stroke is likely to be <15.
Conclusions
Our PRS of over half a million SNVs is individually predictive of ischemic stroke in patients with an established diagnosis of AF, and this predicted risk appears independent of established clinical risk factors. The combined PRS and clinical risk tool (our proposed, CHA2DS2-VASc-G) shows significantly improved risk prediction over the current gold standard risk tool (CHA2DS2-VASc); however, the prediction of ischemic stroke remains challenging.
Acknowledgments
A condition of the free download of the MEGASTROKE data is inclusion of the following statement: The MEGASTROKE project received funding from sources specified at http://www.megastroke.org/acknowledgments.html. We have also listed all MEGASTROKE authors appearing in the main author byline in our Data Supplement, however no authors were involved in the design or conduct of this study. This research was conducted using the UK Biobank Resource under application number 24 983, and 22 282
Footnote
Nonstandard Abbreviations and Acronyms
- AF
- atrial fibrillation
- GWAS
- genome-wide association study
- NOAC
- novel anticoagulants
- PRS
- polygenic risk score
- SNV
- single nucleotide variant
Supplemental Material
References
1.
Virani SS, Alonso A, Benjamin EJ, Bittencourt MS, Callaway CW, Carson AP, Chamberlain AM, Chang AR, Cheng S, Delling FN, et al. Heart disease and stroke statistics—2020 update: a report from the American Heart Association [Internet]. Circulation. 2020;141:e139–e596. doi: 10.1161/CIR.0000000000000757
2.
Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke. 1991;22:983–988. doi: 10.1161/01.str.22.8.983
3.
Kirchhof P, Benussi S, Kotecha D, Ahlsson A, Atar D, Casadei B, Castellá M, Diener HC, Heidbuchel H, Hendriks J, et al. 2016 ESC guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Rev Esp Cardiol (Engl Ed). 2017;70:50. doi: 10.1016/j.rec.2016.11.033
4.
January CT, Wann LS, Calkins H, Chen LY, Cigarroa JE, Cleveland JC, Ellinor PT, Ezekowitz MD, Field ME, Furie KL, et al. 2019 AHA/ACC/HRS focused update of the 2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society in Collaboration With the Society of Thoracic Surgeons. Circulation. 2019;140:e125–e151. doi: 10.1161/CIR.0000000000000665
5.
López-López JA, Sterne JAC, Thom HHZ, Higgins JPT, Hingorani AD, Okoli GN, Davies PA, Bodalia PN, Bryden PA, Welton NJ, et al. Oral anticoagulants for prevention of stroke in atrial fibrillation: systematic review, network meta-analysis, and cost effectiveness analysis. BMJ. 2017;359:j5058. doi: 10.1136/bmj.j5058
6.
Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation. Chest. 2010;137:263–272. doi: 10.1378/chest.09-1584
7.
Ntaios G, Lip GY, Makaritsis K, Papavasileiou V, Vemmou A, Koroboki E, Savvari P, Manios E, Milionis H, Vemmos K. CHADS2, CHA2S2DS2-VASc, and long-term stroke outcome in patients without atrial fibrillation. Neurology. 2013;80:1009–1017. doi: 10.1212/WNL.0b013e318287281b
8.
van Doorn S, Debray TPA, Kaasenbrood F, Hoes AW, Rutten FH, Moons KGM, Geersing GJ. Predictive performance of the CHA2DS2-VASc rule in atrial fibrillation: a systematic review and meta-analysis. J Thromb Haemost. 2017;15:1065–1077. doi: 10.1111/jth.13690
9.
Bevan S, Traylor M, Adib-Samii P, Malik R, Paul NL, Jackson C, Farrall M, Rothwell PM, Sudlow C, Dichgans M, et al. Genetic heritability of ischemic stroke and the contribution of previously reported candidate gene and genomewide associations. Stroke. 2012;43:3161–3167. doi: 10.1161/STROKEAHA.112.665760
10.
Abraham G, Malik R, Yonova-Doing E, Salim A, Wang T, Danesh J, Butterworth AS, Howson JMM, Inouye M, Dichgans M. Genomic risk score offers predictive performance comparable to clinical risk factors for ischaemic stroke. Nat Commun. 2019;10:5819. doi: 10.1038/s41467-019-13848-1
11.
Khera AV, Chaffin M, Aragam KG, Haas ME, Roselli C, Choi SH, Natarajan P, Lander ES, Lubitz SA, Ellinor PT, et al. Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations. Nat Genet. 2018;50:1219–1224. doi: 10.1038/s41588-018-0183-z
12.
Elliott J, Bodinier B, Bond TA, Chadeau-Hyam M, Evangelou E, Moons KGM, Dehghan A, Muller DC, Elliott P, Tzoulaki I. Predictive accuracy of a polygenic risk score-enhanced prediction model vs a clinical risk score for coronary artery disease. JAMA. 2020;323:636–645. doi: 10.1001/jama.2019.22241
13.
Mosley JD, Gupta DK, Tan J, Yao J, Wells QS, Shaffer CM, Kundu S, Robinson-Cohen C, Psaty BM, Rich SS, et al. Predictive accuracy of a polygenic risk score compared with a clinical risk score for incident coronary heart disease. JAMA. 2020;323:627–635. doi: 10.1001/jama.2019.21782
14.
Inouye M, Abraham G, Nelson CP, Wood AM, Sweeting MJ, Dudbridge F, Lai FY, Kaptoge S, Brozynska M, Wang T, et al; UK Biobank CardioMetabolic Consortium CHD Working Group. Genomic risk prediction of coronary artery disease in 480,000 adults: implications for primary prevention. J Am Coll Cardiol. 2018;72:1883–1893. doi: 10.1016/j.jacc.2018.07.079
15.
Choi, SW, Mak, TSH, O’Reilly, PF. Tutorial: a guide to performing polygenic risk score analyses. Nat Protoc. 2020;15:2759–2772. Available at: https://doi.org/10.1038/s41596-020-0353-1
16.
Wand, H, Lambert, SA, Tamburro, C, Iacocca, MA, O’Sullivan, JW, Sillari, C, Kullo, IJ, Rowley, R, Brockman, D, Venner, E, et al. Improving reporting standards for polygenic scores in risk prediction studies. Nature. 2021;591(7849):211–219. doi: 10.1038/s41586-021-03243-6
17.
Mak TSH, Porsch RM, Choi SW, Zhou X, Sham PC. Polygenic scores via penalized regression on summary statistics. Genet Epidemiol. 2017;41:469–480. doi: 10.1002/gepi.22050
18.
Malik R, Chauhan G, Traylor M, Sargurupremraj M, Okada Y, Mishra A, Rutten-Jacobs L, Giese AK, van der Laan SW, Gretarsdottir S, et al; AFGen Consortium; Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE) Consortium; International Genomics of Blood Pressure (iGEN-BP) Consortium; INVENT Consortium; STARNET; BioBank Japan Cooperative Hospital Group; COMPASS Consortium; EPIC-CVD Consortium; EPIC-InterAct Consortium; International Stroke Genetics Consortium (ISGC); METASTROKE Consortium; Neurology Working Group of the CHARGE Consortium; NINDS Stroke Genetics Network (SiGN); UK Young Lacunar DNA Study; MEGASTROKE Consortium. Multiancestry genome-wide association study of 520,000 subjects identifies 32 loci associated with stroke and stroke subtypes. Nat Genet. 2018;50:524–537. doi: 10.1038/s41588-018-0058-3
19.
Riveros-Mckay F, Weale ME, Moore R, Selzam S, Krapohl E, Sivley RM, Tarran WA, Sørensen P, Lachapelle AS, Griffiths JA, et al. Integrated polygenic tool substantially enhances coronary artery disease prediction. Circ Genom Precis Med. 2021;14:e003304. doi: 10.1161/CIRCGEN.120.003304
20.
Dikilitas O, Schaid DJ, Kosel ML, Carroll RJ, Chute CG, Denny JA, Fedotov A, Feng Q, Hakonarson H, Jarvik GP, et al. Predictive utility of polygenic risk scores for coronary heart disease in three major racial and ethnic groups. Am J Hum Genet. 2020;106:707–716. doi: 10.1016/j.ajhg.2020.04.002
21.
Ashley EA, Butte AJ, Wheeler MT, Chen R, Klein TE, Dewey FE, Dudley JT, Ormond KE, Pavlovic A, Morgan AA, et al. Clinical assessment incorporating a personal genome. Lancet. 2010;375:1525–1535. doi: 10.1016/S0140-6736(10)60452-7
22.
Dewey FE, Grove ME, Pan C, Goldstein BA, Bernstein JA, Chaib H, Merker JD, Goldfeder RL, Enns GM, David SP, et al. Clinical interpretation and implications of whole-genome sequencing. JAMA. 2014;311:1035–1045. doi: 10.1001/jama.2014.1717
23.
Virani SS, Alonso A, Benjamin EJ, Bittencourt MS, Callaway CW, Carson AP, Chamberlain AM, Chang AR, Cheng S, Delling FN, et al; American Heart Association Council on Epidemiology and Prevention Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics-2020 update: a report From the American Heart Association. Circulation. 2020;141:e139–e596. doi: 10.1161/CIR.0000000000000757
24.
O’Sullivan JW, Grigg S, Crawford W, Turakhia MP, Perez M, Ingelsson E, Wheeler MT, Ioannidis JPA, Ashley EA. Accuracy of smartphone camera applications for detecting atrial fibrillation: a systematic review and meta-analysis. JAMA Netw Open. 2020;3:e202064. doi: 10.1001/jamanetworkopen.2020.2064
25.
Gage BF, Waterman AD, Shannon W, Boechler M, Rich MW, Radford MJ. Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA. 2001;285:2864–2870. doi: 10.1001/jama.285.22.2864
26.
Genome.gov. The Cost of Sequencing a Human Genome [Internet]. [cited 2020 May 18]; Available from: https://www.genome.gov/about-genomics/fact-sheets/Sequencing-Human-Genome-cost.
27.
Fry A, Littlejohns TJ, Sudlow C, Doherty N, Adamska L, Sprosen T, Collins R, Allen NE. Comparison of sociodemographic and health-related characteristics of UK Biobank participants with those of the general population. Am J Epidemiol. 2017;186:1026–1034. doi: 10.1093/aje/kwx246
28.
Loo SY, Dell’Aniello S, Huiart L, Renoux C. Trends in the prescription of novel oral anticoagulants in UK primary care. Br J Clin Pharmacol. 2017;83:2096–2106. doi: 10.1111/bcp.13299
29.
Rutten-Jacobs LC, Larsson SC, Malik R, Rannikmäe K, Sudlow CL, Dichgans M, Markus HS, Traylor M; MEGASTROKE consortium; International Stroke Genetics Consortium. Genetic risk, incident stroke, and the benefits of adhering to a healthy lifestyle: cohort study of 306 473 UK Biobank participants. BMJ. 2018;363:k4168. doi: 10.1136/bmj.k4168
30.
Goode, EL. Linkage disequilibrium. Schwab, M, editor. In: Encyclopedia of Cancer. Springer Berlin Heidelberg; 2011. pp 2043–2048.
31.
Vilhjalmsson B, Yang J, Finucane HK, Gusev A, Lindstrom S, Ripke S, Genovese G, Loh P-R, Bhatia G, Do R, et al. Modeling linkage disequilibrium increases accuracy of polygenic risk scores. Am J Hum Genet. 2015 Oct 1;97(4):576–92. doi: 10.1016/j.ajhg.2015.09.001
32.
Allegrini AG, Selzam S, Rimfeld K, von Stumm S, Pingault JB, Plomin R. Genomic prediction of cognitive traits in childhood and adolescence. Mol Psychiatry. 2019;24:819–827. doi: 10.1038/s41380-019-0394-4
33.
Berisa T, Pickrell JK. Approximately independent linkage disequilibrium blocks in human populations. Bioinformatics. 2016;32:283–285. doi: 10.1093/bioinformatics/btv546
Information & Authors
Information
Published In
Copyright
© 2021 American Heart Association, Inc.
History
Received: 7 August 2020
Accepted: 5 May 2021
Published ahead of print: 24 May 2021
Published in print: June 2021
Published online: 15 June 2021
Keywords
Subjects
Authors
Disclosures
Disclosures E.A. Ashley reports the following: founder, advisor Personalis; founder, advisor Deepcell; advisor SequenceBio; advisor Foresite Labs; advisor Apple. The other authors report no conflicts.
Sources of Funding
The lead author (Dr O’Sullivan) was supported by an NIH T32 grant, otherwise, there is no specific funding.
Metrics & Citations
Metrics
Citations
Download Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Select your manager software from the list below and click Download.
- The Inclusion of Underrepresented Populations in Cardiovascular Genetics and Epidemiology, Journal of Cardiovascular Development and Disease, 11, 2, (56), (2024).https://doi.org/10.3390/jcdd11020056
- Polygenic Risk Scores Driving Clinical Change in Glaucoma, Annual Review of Genomics and Human Genetics, 25, 1, (287-308), (2024).https://doi.org/10.1146/annurev-genom-121222-105817
- Longer and better lives for patients with atrial fibrillation: the 9th AFNET/EHRA consensus conference, Europace, 26, 4, (2024).https://doi.org/10.1093/europace/euae070
- Integration of risk factor polygenic risk score with disease polygenic risk score for disease prediction, Communications Biology, 7, 1, (2024).https://doi.org/10.1038/s42003-024-05874-7
- Combined polygenic scores for ischemic stroke risk factors aid risk assessment of ischemic stroke, International Journal of Cardiology, 404, (131990), (2024).https://doi.org/10.1016/j.ijcard.2024.131990
- Polygenic risk scores in atrial fibrillation: Associations and clinical utility in disease prediction, Heart Rhythm, 21, 6, (913-918), (2024).https://doi.org/10.1016/j.hrthm.2024.02.006
- Donor genetic burden for cerebrovascular risk and kidney transplant outcome, Journal of Nephrology, (2024).https://doi.org/10.1007/s40620-024-01973-0
- A review of disease risk prediction methods and applications in the omics era, PROTEOMICS, 24, 18, (2024).https://doi.org/10.1002/pmic.202300359
- Genetics in Ischemic Stroke: Current Perspectives and Future Directions, Journal of Cardiovascular Development and Disease, 10, 12, (495), (2023).https://doi.org/10.3390/jcdd10120495
- Culture intelligent workflow, structure, and steps, Frontiers in Artificial Intelligence, 6, (2023).https://doi.org/10.3389/frai.2023.985469
- See more
Loading...
View Options
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Personal login Institutional LoginPurchase Options
Purchase this article to access the full text.
eLetters(0)
eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.
Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.