Hybrid Coronary Revascularization vs Complete Coronary Artery Bypass Grafting for Multivessel Coronary Artery Disease: A Systematic Review and Meta-Analysis

Abstract: Background. Hybrid coronary revascularization (HCR) has emerged as a potential alternative to complete coronary artery bypass graft (CABG) surgery. However, the efficacy and safety of HCR vs CABG remain unclear. We therefore conducted a systematic review and meta-analysis to compare these interventions. Methods. We systematically searched PubMed, MEDLINE (via Ovid), EMBASE (via Ovid), Cochrane Library of Clinical Trials, and the Web of Science for studies comparing HCR to CABG in patients with multivessel coronary artery disease. The primary outcome was major adverse cardiovascular and cerebrovascular events (MACCE) and its components (myocardial infarction, stroke, mortality, and target-vessel revascularization [TVR]) at ≥1 year. Secondary outcomes included MACCE at ≤30 days, its components, and postoperative safety outcomes (renal failure, blood transfusion, new-onset atrial fibrillation, and infection). Results. One randomized controlled trial and 9 cohort studies were included in our systematic review. Pooled results indicate that HCR is associated with a lower risk for postoperative blood transfusion (odds ratio [OR], 0.43; 95% confidence interval [CI], 0.27-0.68) and infection (OR, 0.19; 95% CI, 0.04-0.98), and a shorter hospital stay (6.0 days for HCR vs 7.8 days for CABG) and intensive care unit (ICU) stay (25.4 hours for HCR vs 45.7 hours for CABG). Long-term outcome data showed an association between HCR and long-term TVR (OR, 3.10; 95% CI, 1.39-6.90). Conclusions. Our results suggest that compared to CABG, HCR is associated with a lower risk of postoperative blood transfusion and infection, as well as a shorter ICU stay and hospital stay. HCR was also associated with a higher risk of long-term TVR.  

J INVASIVE CARDIOL 2018;30(12):E131-E149.

Key words: coronary artery bypass grafting, hybrid coronary revascularization, multivessel coronary artery disease, percutaneous coronary intervention

Hybrid coronary revascularization (HCR) has emerged as a potential minimally invasive alternative to sternotomy coronary artery bypass grafting (CABG) for multivessel coronary artery disease (MVCAD).^1,2 HCR typically incorporates a minimally invasive approach for the grafting of the left internal mammary artery (LIMA) to the left anterior descending (LAD) artery along with percutaneous coronary intervention (PCI) in the non-LAD vessels.³ This approach combines the respective advantages of CABG and PCI. The LIMA to LAD graft is associated with superior long-term patency compared to PCI,⁴ and long-term patency of drug-eluting stents may approach that of saphenous vein grafts.^4,5 HCR has been further enhanced by the use of robotic technology to harvest the LIMA and by endoscopic techniques for the anastomosis to the LAD.⁶ A minimally invasive approach may reduce morbidity, pain, scarring, and recovery time compared to the conventional CABG technique.^7-9 However, the effect of HCR relative to CABG on clinical outcomes remains unclear, as individual studies comparing the two have not been adequately powered to examine these outcomes. Therefore, we conducted a systematic review and meta-analysis to synthesize the available evidence from studies comparing HCR to CABG to assess the efficacy and safety of HCR. 

Methods 

This systematic review and meta-analysis was conducted following a prespecified protocol and is reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.¹⁰ 

Search strategy. We conducted a systematic search of PubMed, MEDLINE (via Ovid), EMBASE (via Ovid), Cochrane Library of Clinical Trials, and the Web of Science from their inception until July 25, 2016. Details of the database-specific search strategies can be found in Appendices 1-5. Briefly, a combination of keywords, Medical Subject Headings (MeSH) terms (for the PubMed, MEDLINE, and Cochrane Library searches), and EMTREE terms (for the EMBASE search) were used for two main concepts: “hybrid coronary revascularization” and “coronary artery bypass grafting.” This search was supplemented by a hand-search of the references of included studies and previous reviews to identify additional relevant studies not captured by our electronic search. Finally, we searched clinicaltrials.gov for any ongoing or recently completed potentially relevant trials. 

Appendices 1-5

Study selection. Titles and abstracts of the identified articles were screened by two independent reviewers (SN and RA). Any article deemed potentially relevant by either reviewer was carried forward to full-text review. Disagreements during full-text review were resolved by consensus or, when necessary, a third reviewer (KBF). 

In order to be included, studies had to: (1) be randomized controlled trials (RCTs) or observational studies (cohort or case-control studies); (2) compare HCR to CABG; (3) include patients with MVCAD or left main CAD; (4) report data regarding the outcomes of interest (defined below) by treatment group; (5) include a minimum of 10 participants; (6) be conducted in humans; and (7) be published in English or French. In instances where studies reported data from the same database, registry, or institution during overlapping time periods, only the study presenting the longest available follow-up or the most comprehensive report of the outcomes of interest was included. Similarly, studies solely assessing a specific patient subgroup (eg, patients with diabetes) that were part of an already included study were excluded. Abstracts, conference proceedings, letters to the editor, commentaries, editorials, reviews, meta-analyses, clinical practice guidelines, and cross-sectional studies were excluded.  

Data abstraction, primary outcomes, and secondary outcomes. For each study that met our inclusion criteria, data were independently abstracted by the reviewers (SN and RA). Disagreements were resolved by consensus or by a third reviewer (KBF). Data on study and baseline patient characteristics, as well as the primary and secondary outcomes, were extracted. The primary outcome was a long-term (≥1 year) composite of major adverse cardiovascular and cerebrovascular events (MACCE). Secondary outcomes included the individual components of long-term MACCE (myocardial infarction [MI], stroke, mortality, and target-vessel revascularization [TVR]), short-term (≤30 days) MACCE and its individual components, short-term safety outcomes (renal failure, blood transfusion, new-onset atrial fibrillation, infection rates), and resource utilization outcomes (intensive care unit length of stay [LOS], hospital LOS, operating room time). For short-term outcomes, 30-day results were used when in-hospital outcomes were not available. Consequently, our short-term analysis included either in-hospital or 30-day outcomes.

Quality assessment. Quality assessment was performed using the Risk of Bias in Non-Randomized Studies of Interventions (ROBINS-I) tool¹¹ for observational studies or the Cochrane Risk of Bias tool for RCTs.¹² The ROBINS-I tool assesses the risk of bias arising in several domains: confounding; selection of study participants; classification of interventions; deviations from intended interventions; missing data; measurement of outcomes; and selection of the reported results. Observational studies were deemed to be at low, moderate, serious, critical, or unknown risk of bias in each of these domains. Although the ROBINS-I tool requires an outcome-specific assessment for each study,¹¹ with all of our key outcomes considered cardiovascular in nature and thus having the same set of confounders, the quality of each study was only assessed once. The Cochrane Risk of Bias tool for RCTs assigns a low, high, or unclear risk of bias in terms of: sequence generation; allocation concealment; blinding of participants, personnel and outcome assessors; incomplete outcome data; and selective outcome reporting. Studies were not excluded on the basis of the results of their quality assessment.   

Statistical analysis. Data were pooled across studies using the DerSimonian-Laird approach with inverse variance weighting, which considers both within-study (random error) and between-study variability. Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated for each of our primary and secondary outcomes. The amount of heterogeneity that was present was estimated using the I² statistic. Studies with zero events in either group were included using a 0.5 continuity correction.

Three sensitivity analyses were performed (Appendices 6-10). The first examined long-term MACCE and long-term TVR stratified by study design (RCT vs observational study). We chose long-term MACCE as it was our primary outcome. We also conducted these analyses for long-term TVR, as it appeared to have a strong effect on the composite MACCE outcome. The second stratified analysis by adjusted vs unadjusted ORs was meant to examine the amount of residual confounding that was present. The third sensitivity analysis included data from the 2014 study by Harskamp and colleagues¹³ that used a United States national database, rather than their 2015 single-center study.¹⁴ All statistical analyses were performed using R version 3.2.2. (R Core Team [2015], R Foundation for Statistical Computing). 

Appendices 6-10

Results

Literature search. Our electronic search yielded 1213 results (Figure 1), with no additional results identified via hand-searching. After removal of duplicates, 638 publications underwent title and abstract screening. Of these, 186 publications underwent full-text review, of which 11 met our inclusion criteria.^13-22,32 Two were from the same institution, with one providing more updated long-term results¹⁶ following the previously published short-term outcomes.¹⁷ Another used a country-wide database,¹³ which overlapped with patients in another study.¹⁴ Thus, our systematic review included 10 studies (n = 201,064), of which 9 (n = 2442) were included in our meta-analysis. 

Study characteristics. One RCT and 9 cohort studies were included in our systematic review. The numbers of patients in the HCR and CABG groups ranged from 15 to 950, and from 18 to 197,672, respectively (Table 1). Five studies used some form of matching to control for a variety of baseline and prognostic characteristics^14,16,20,21 and 2 studies used propensity scores.^14,16 HCR strategies varied across studies in terms of single-stage vs staged approach, and in terms of the techniques used for the LIMA to LAD harvest and anastomosis. In all but 1 study,²¹ drug-eluting stents were the most commonly used type of stent. Patients in the CABG reference groups underwent on-pump or off-pump procedures. Follow-up duration ranged from 30 days to 3 years. Inclusion and exclusion criteria were similar across most studies, and all included only patients with MVCAD or left main CAD. 

Quality assessment. Included observational studies were classified as being at moderate or serious risk of bias based on the ROBINS-I tool (Appendix 12). In general, the domain with the most significant risk of bias was confounding. Studies that used methods to account for confounding were assigned a moderate risk of bias given their observational nature, while those that did not had more serious risks of bias. Since no studies made reference to a preregistered protocol in which they outlined the outcomes to be reported, they were all deemed to be at a moderate risk of bias in the domain of selection of reported results. Most or all studies had a low risk of bias in the other domains. As per the Cochrane Risk of Bias tool, the sole included RCT was deemed to be at low risk of bias in half of the domains and at an unclear risk of bias in the other half due to lack of information (Appendix 13). 

Appendices 11 and 12

It is generally accepted that high-quality non-randomized studies can provide valuable evidence in the absence of RCTs. Our selection of studies consisted mostly of observational studies, but these were all deemed to be at a moderate risk of bias according to the ROBINS-I tool, which means that the studies were considered sound for non-randomized studies but not comparable to a well-performed randomized trial. 

Baseline patient characteristics. Baseline characteristics were generally similar between treatment groups, particularly in studies that used matching, as well as across included studies (Supplemental Table S1). Most patients were male and in their sixties. There was some variation across studies in terms of cardiovascular risk factors. Patients in Shen 2013¹⁶ had lower prevalences of hypertension, diabetes, and dyslipidemia, but a higher prevalence of smoking (60.7% vs a range of 19.6%-41.7% in all other studies). Compared to most studies, patients in de Canniere 2001²¹ and Delhaye 2010²² had lower prevalences of diabetes and hypertension, respectively, whereas those in Harskamp 2015¹⁴ had a higher prevalence of cerebrovascular disease. 

Short-term safety outcomes. The need for intraoperative and immediately postoperative blood transfusion was consistently higher with CABG in all but 1 study (Table 2).¹⁵ CABG was associated with higher postoperative blood transfusion rates than HCR (OR, 0.43; 95% CI, 0.27-0.68) (Figure 2). Moreover, HCR was associated with a lower risk for postoperative infection (OR, 0.19; 95% CI, 0.04-0.98). Of note, there were no infections with HCR in the 3 studies reporting this outcome. Results were inconclusive in regard to renal failure (OR, 0.72; 95% CI, 0.34-1.51) and atrial fibrillation (OR, 0.72; 95% CI, 0.31-1.67). In terms of resource utilization, intensive care unit hospital stay (25.4 hours for HCR vs 45.7 hours for CABG) and hospital LOS (6.0 days for HCR vs 7.8 days for CABG) were consistently lower with HCR in all studies that reported these outcomes (Table 2). 

Long-term and short-term MACCE. Short-term MACCE (≤30 days) ranged from 0.0%-8.8% with HCR vs 0.0%-23.3% with CABG (Table 3). When data from the 9 studies included in the meta-analysis were pooled using a random-effects model, the difference in short-term MACCE between HCR and CABG was inconclusive due to a wide 95% CI (OR, 1.05; 95% CI, 0.62-1.79) (Figure 3). Results for most components of short-term MACCE were inconclusive, as 95% CIs were wide and results were non-significant. 

Long-term MACCE (≥1 year) ranged from 6.4% to 11.1% with HCR, and from 5.6% to 23.3% with CABG (Supplemental Table S2). Results for long-term MI and mortality were inconclusive, as 95% CIs were wide and the results were non-significant (Figure 4). However, HCR was significantly associated with a higher risk for long-term TVR (OR, 3.10; 95% CI, 1.39-6.90).

Tables 1-3; Supplemental Tables S1 and S2

Figures 1-4

Discussion

This systematic review and meta-analysis was designed to assess the short- and long-term safety and efficacy of HCR vs CABG in MVCAD patients. When data were pooled across studies, we found a significant association between HCR and lower risk for postoperative blood transfusion and infection, as well as a shorter intensive care unit and hospital LOS. In regard to long-term outcomes, HCR was significantly associated with a higher risk of long-term TVR. 

While previous meta-analyses^28-30 have also assessed the efficacy and safety of HCR vs CABG in MVCAD patients, several studies have since been published in the literature. Most notably, we included an updated observational study from Emory University (n = 1224) and the first published RCT of HCR vs CABG (n = 200), which helped bolster the findings from observational studies. Furthermore, some of these previous meta-analyses had some methodological issues. For instance, a meta-analysis published in 2015³⁰ seems to have included multiple studies that had partially overlapping populations within their analysis.     

The only RCT to date was published in Poland by Gasior et al¹⁸ in 2014. This trial randomized a total of 200 patients and was insufficiently powered to examine clinical outcomes. In this trial, all study patients had MVCAD involving the LAD and at least one other major epicardial vessel, were eligible for either PCI or CABG, and were randomized in a 1:1 ratio to undergo either HCR or CABG. Its primary endpoint included components of MACCE as well as other safety and feasibility outcomes at 1-year follow-up. 

There are sound scientific arguments supporting both HCR and CABG as revascularization strategies in MVCAD patients. On one hand, CABG remains the gold standard procedure for the superior patency of the LIMA to LAD graft.¹ It also allows for the avoidance of multiple diseased coronary vessels, thereby reducing the need for future revascularization in the non-LAD vessels while providing long-term relief from anginal episodes.²⁴ On the other hand, HCR offers the LIMA to LAD graft that is responsible in large part for the survival advantage associated with CABG.²⁵ Unlike CABG, it does not rely on saphenous venous grafts for the non-LAD vessels, which are associated with a high failure rate (up to 10%-15% at 1 year and 50% at 10 years post surgery).²⁶ Instead, HCR utilizes PCI for non-LAD vessels, with restenosis occurring in <10% of cases⁴ and in <5% of cases at 1 year with later-generation everolimus and zotarolimus drug-eluting stents.²⁷ In addition, the minimally invasive nature of HCR allows for faster recovery and shorter hospital LOS. Nevertheless, some evidence suggests that the quality of the LIMA to LAD anastomosis in the minimally invasive HCR procedure may be compromised as compared to the same anastomosis performed in a conventional CABG technique.²⁸ We observed that rates of long-term TVR increased in patients with HCR. These arguments warrant further investigation in a large, definitive, well-designed RCT. 

With only 1 RCT published to date, the currently available evidence is largely observational. Such studies are inherently affected by confounders, and it is possible that healthier patients may have been channeled toward HCR. Nonetheless, this represents the totality of the available evidence regarding the efficacy and safety of HCR relative to CABG. The results of this systematic review and meta-analysis suggest that HCR is promising and underscore the need for large, well-conducted RCTs to investigate the efficacy and safety of HCR vs CABG in patients with MVCAD.     

Study limitations. Our study has several potential limitations. First, 9 of the 10 studies included in our systematic review were observational. By virtue of their design, they presented a moderate to serious risk of bias due to confounding. Second, the definitions of the MACCE composite outcome differed slightly between studies. We used the definition that was most commonly used in the studies, which included MI, stroke, mortality, and TVR. However, certain studies did not include TVR or stroke in their composite measure. Since the number of strokes and TVR events was extremely low, this heterogeneity is unlikely to have had a substantial impact on our findings. Third, there was some heterogeneity in the techniques used for HCR across the different studies, with single-staged or multi-staged procedures and a variety of techniques used for the LIMA to LAD harvest and anastomosis. This heterogeneity further supports the rationale for well-conducted RCTs to compare the efficacy and safety of HCR vs CABG.

Conclusion 

Our study was designed to assess the short-term and long-term safety and efficacy of both HCR and CABG in MVCAD patients. Our results suggest that there is an increased risk of long-term TVR with HCR. However, HCR was also associated with lower risks for postoperative blood transfusion and infection, as well as with a shorter hospital LOS and intensive care unit LOS. With the available evidence mostly derived from observational studies, there remains a need for a large, definitive, well-designed RCT comparing these two treatment options for patients with MVCAD. 

References

1.    Fihn SD, Gardin JM, Abrams J, et al. 2012 ACCF/AHA/ACP/AATS/PCNA/SCAI/STS guideline for the diagnosis and management of patients with stable ischemic heart disease: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines, and the American College of Physicians, American Association for Thoracic Surgery, Preventive Cardiovascular Nurses Association, Society for Cardiovascular Angiography and Interventions, and Society of Thoracic Surgeons. J Am Coll Cardiol. 2012;60:e44-e164.

2.    Verhaegh AJ, Accord RE, van Garsse L, Maessen JG. Hybrid coronary revascularization as a safe, feasible, and viable alternative to conventional coronary artery bypass grafting: what is the current evidence? Minim Invasive Surg. 2013;2013:142616.

3.    Byrne JG, Leacche M, Vaughan DE, Zhao DX. Hybrid cardiovascular procedures. JACC Cardiovasc Interv. 2008;1:459-468.

4.    Harskamp RE, Zheng Z, Alexander JH, et al. Status quo of hybrid coronary revascularization for multi-vessel coronary artery disease. Ann Thorac Surg. 2013;96:2268-2277.

5.    Green KD, Lynch DR Jr, Chen TP, Zhao D. Combining PCI and CABG: the role of hybrid revascularization. Curr Cardiol Rep. 2013;15:351.

6.    Currie ME, Romsa J, Fox SA, et al. Long-term angiographic follow-up of robotic-assisted coronary artery revascularization. Ann Thorac Surg. 2012;93:1426-1431.

7.    Ishikawa N, Watanabe G. Robot-assisted cardiac surgery. Ann Thorac Cardiovasc Surg. 2015;21:322-328.

8.    Modi P, Hassan A, Chitwood WR Jr. Minimally invasive mitral valve surgery: a systematic review and meta-analysis. Eur J Cardiothorac Surg. 2008;34:943-952.

9.    Bonatti J, Lehr E, Vesely MR, Friedrich G, Bonaros N, Zimrin D. Hybrid coronary revascularization: which patients? When? How? Curr Opin Cardiol. 2010;25:568-574.

10.    Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ. 2009;339:b2700.

11.    Sterne JA, Hernan MA, Reeves BC, et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ. 2016;355:i4919.

12.    Higgins JP, Altman DG, Gotzsche PC, et al. The Cochrane Collaboration’s tool for assessing risk of bias in randomised trials. BMJ. 2011;343:d5928.

13.    Harskamp RE, Brennan JM, Xian Y, et al. Practice patterns and clinical outcomes after hybrid coronary revascularization in the United States: an analysis from the Society of Thoracic Surgeons Adult Cardiac Database. Circulation. 2014;130:872-879.

14.    Harskamp RE, Vassiliades TA, Mehta RH, et al. Comparative effectiveness of hybrid coronary revascularization vs coronary artery bypass grafting. J Am Coll Surg. 2015;221:326-334.e1.

15.    Leacche M, Byrne JG, Solenkova NS, et al. Comparison of 30-day outcomes of coronary artery bypass grafting surgery verus hybrid coronary revascularization stratified by SYNTAX and EuroSCORE. J Thorac Cardiovasc Surg. 2013;145:1004-1012.

16.    Shen LZ, Hu SS, Wang HR, et al. One-stop hybrid coronary revascularization versus coronary artery bypass grafting and percutaneous coronary intervention for the treatment of multivessel coronary artery disease 3-year follow-up results from a single institution. J Am Coll Cardiol. 2013;61:2525-2533.

17.    Hu SS, Li Q, Gao PX, et al. Simultaneous hybrid revascularization versus off-pump coronary artery bypass for multivessel coronary artery disease. Ann Thorac Surg. 2011;91:432-438.

18.    Gasior M, Zembala MO, Tajstra M, et al. Hybrid revascularization for multivessel coronary artery disease. JACC Cardiovasc Interv. 2014;7:1277-1283.

19.    Bachinsky WB, Abdelsalam M, Boga G, Kiljanek L, Mumtaz M, McCarty C. Comparative study of same sitting hybrid coronary artery revascularization versus off-pump coronary artery bypass in multivessel coronary artery disease. J Interv Cardiol. 2012;25:460-468.

20.    Kon ZN, Brown EN, Tran R, et al. Simultaneous hybrid coronary revascularization reduces postoperative morbidity compared with results from conventional off-pump coronary artery bypass. J Thorac Cardiovasc Surg. 2008;135:367-375.

21.    de Canniere D, Jansens JL, Goldschmidt-Clermont P, Barvais L, Decroly P, Stoupel E. Combination of minimally invasive coronary bypass and percutaneous transluminal coronary angioplasty in the treatment of double-vessel coronary disease: two-year follow-up of a new hybrid procedure compared with “on-pump” double bypass grafting. Am Heart J. 2001;142:563-570.

22.    Delhaye C, Sudre A, Lemesle G, et al. Hybrid revascularization, comprising coronary artery bypass graft with exclusive arterial conduits followed by early drug-eluting stent implantation, in multivessel coronary artery disease. Arch Cardiovasc Dis. 2010;103:502-511.

23.    Anzueto A. Clinical course of chronic obstructive pulmonary disease: review of therapeutic interventions. Am J Med. 2006;119:46-53.

24.    Hawkes AL, Nowak M, Bidstrup B, Speare R. Outcomes of coronary artery bypass graft surgery. Vasc Health Risk Manag. 2006;2:477-484.

25.    Puskas JD, Halkos ME, DeRose JJ, et al. Hybrid coronary revascularization for the treatment of multivessel coronary artery disease: a multicenter observational study. J Am Coll Cardiol. 2016;68:356-365.

26.    Harskamp RE, Lopes RD, Baisden CE, de Winter RJ, Alexander JH. Saphenous vein graft failure after coronary artery bypass surgery: pathophysiology, management, and future directions. Ann Surg. 2013;257:824-833.

27.    Alexander JH, Hafley G, Harrington RA, et al. Efficacy and safety of edifoligide, an E2F transcription factor decoy, for prevention of vein graft failure following coronary artery bypass graft surgery: PREVENT IV: a randomized controlled trial. JAMA. 2005;294:2446-2454.

28.    Harskamp RE, Bagai A, Halkos ME, et al. Clinical outcomes after hybrid coronary revascularization versus coronary artery bypass surgery: a meta-analysis of 1,190 patients. Am Heart J. 2014;167:585-592.

29.    Phan K, Wong S, Wang N, Phan S, Yan TD. Hybrid coronary revascularization versus coronary artery bypass surgery: systematic review and meta-analysis. Int J Cardiol. 2015;179:484-488.

30.    Zhu P, Zhou P, Sun Y, Guo Y, Mai M, Zheng S. Hybrid coronary revascularization versus coronary artery bypass grafting for multivessel coronary artery disease: systematic review and meta-analysis. J Cardiothorac Surg. 2015;10:63.

31.    Zembala M, Tajstra M, Zembala M, et al. Prospective randomised pilOt study evaLuating the safety and efficacy of hybrid revascularisation in MultI-vessel coronary artery DisEaSe (POLMIDES) - study design. Kardiol Pol. 2011;69:460-466.

32.    Xia Y, Katz AN, Forest SJ, Pyo RT, Greenberg MA, DeRose JJ Jr. Hybrid coronary revascularization has improved short-term outcomes but worse mid-term reintervention rates compared to CABG: a propensity matched analysis. Innovations (Phila). 2017;12:174-179. 

From the ¹Centre for Clinical Epidemiology, Lady Davis Institute, Jewish General Hospital/McGill University, Montreal, Quebec, Canada; ²Faculty of Medicine, McGill University, Montreal, Quebec, Canada; ³Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Montreal, Quebec, Canada; ⁴Department of Medicine, McGill University, Montreal, Quebec, Canada; ⁵Division of Cardiac Surgery, Jewish General Hospital/McGill University, Montreal, Quebec, Canada; and the ⁶Division of Cardiology, Jewish General Hospital/McGill University, Montreal, Quebec, Canada.

Funding: Dr Filion holds a Canadian Institutes of Health Research (CIHR) New Investigator Award.

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report no conflicts of interest regarding the content herein.

Manuscript submitted August 17, 2018, final version accepted August 25, 2018.

Address for correspondence: Mark J. Eisenberg, MD, MPH, Professor of Medicine, Jewish General Hospital/McGill University, 3755 Côte Ste-Catherine Road, Suite H-421.1, Montreal, Quebec, Canada H3T 1E2. Email: mark.eisenberg@mcgill.ca