Multiarterial Grafting and Survival After Coronary Artery Bypass Grafting: An Instrumental Variable Analysis

Background

Coronary artery bypass grafting, or CABG, is a common surgical treatment for people with severe coronary artery disease. During CABG, surgeons create new pathways for blood to reach the heart muscle by using blood vessels taken from elsewhere in the body. The most common approach is single arterial grafting, in which one artery is used along with vein grafts. A more intensive strategy, multiarterial grafting (MAG), uses two or more arterial conduits.

In theory, MAG may offer better long-term durability because arterial grafts tend to remain open longer than vein grafts. For that reason, many surgeons believe MAG could improve survival over time. However, randomized clinical trials have not clearly shown a long-term survival advantage, while many observational studies have reported that patients receiving MAG live longer. This discrepancy raises an important question: does MAG truly improve survival, or are observational results influenced by patient selection and other hidden differences?

This study addressed that question using a quasi-experimental method called instrumental variable analysis, designed to reduce the impact of unmeasured confounding. The focus was an older U.S. Medicare population, where CABG is common and long-term follow-up is available.

Why the Question Matters

Choosing between MAG and single arterial grafting is not a trivial technical detail. It reflects a broader decision about balancing surgical complexity, operative time, possible risks, and potential long-term benefit. MAG can be more technically demanding, and not every patient is an ideal candidate. Patients with frailty, severe comorbidities, or limited life expectancy may be less likely to receive MAG.

That creates an important challenge for conventional observational research. If healthier or more suitable patients are more likely to receive MAG, then improved survival could partly reflect differences in baseline risk rather than the grafting strategy itself. Even with careful statistical adjustment, some patient characteristics are difficult to measure fully, such as vessel quality, surgical anatomy, functional status, or subtle clinical judgment.

Study Design and Methods

The investigators retrospectively studied Medicare beneficiaries who underwent CABG between 2001 and 2019. The sample included more than 1.29 million patients, making this one of the largest real-world analyses on this topic.

Patients were classified according to whether they received single arterial grafting or multiarterial grafting. The primary outcome was long-term survival, with median survival estimated over time.

The study used two analytic approaches:

First, a conventional risk-adjusted observational model was built. This model accounted for measured factors such as age, sex, comorbid conditions, hospital characteristics, surgeon characteristics, and procedural details.

Second, the researchers used an instrumental variable, or IV, approach. The IV was the surgeon’s MAG rate during the 12 months before each operation. In simpler terms, surgeons who more often used multiarterial grafting were considered more likely to offer it to a given patient, independent of some patient-level factors. This kind of approach can help mimic certain aspects of randomization, assuming the instrument is strongly related to treatment choice and affects outcomes mainly through that treatment.

To estimate survival, the researchers used flexible parametric survival models with time-dependent effects. They also applied regression standardization to calculate standardized survival probabilities and differences between groups. In the IV analysis, they used a two-stage residual inclusion method to account for the instrument while adjusting for the same measured covariates.

Who Was Included

Among 1,291,314 Medicare beneficiaries who underwent CABG, 1,145,760 patients, or 88.7%, received single arterial grafting. Multiarterial grafting was used in 145,554 patients, or 12.3%.

This distribution itself is clinically meaningful. Despite the theoretical advantages of MAG, it was used in a relatively small proportion of patients, suggesting that surgical practice patterns remain conservative or selective. That may reflect differences in surgeon preference, training, institutional culture, and patient eligibility.

Across 4,164 surgeons, the average MAG rate in the year before the index CABG varied widely. In patients who ultimately received single arterial grafting, the surgeon’s prior MAG rate averaged 7.7% ± 9.5%. In patients who received MAG, the surgeon’s prior MAG rate averaged 32.9% ± 25.8%. This strong association supports the idea that surgeon practice patterns influenced whether MAG was used.

Key Findings from the Conventional Analysis

In the non-IV, conventional risk-adjusted model, patients receiving MAG appeared to do modestly better than those receiving single arterial grafting.

The estimated median survival was:

MAG: 10.74 years, with 95% confidence interval of 10.70 to 10.79 years
SAG: 10.33 years, with 95% confidence interval of 10.31 to 10.35 years

This translated into a survival difference of 0.41 years, or about five months.

At first glance, this suggests a real benefit from using multiple arterial grafts. Such a result is consistent with the idea that arterial conduits may remain patent longer, reducing recurrent ischemia and the need for repeat procedures. For clinicians, however, the key issue is whether this observed advantage is causal or only apparent.

Key Findings from the Instrumental Variable Analysis

When the researchers applied the IV approach to address unmeasured confounding, the apparent benefit nearly disappeared.

The estimated median survival in the IV model was:

MAG: 10.38 years, with 95% confidence interval of 10.29 to 10.48 years
SAG: 10.38 years, with 95% confidence interval of 10.35 to 10.40 years

The survival difference was only 0.01 years, which is essentially no difference.

This result suggests that the modest survival advantage seen in conventional observational analyses may not be due to MAG itself. Instead, it may reflect hidden differences between patients selected for MAG and those who receive single arterial grafting. In other words, when the analysis better accounts for treatment-selection bias, the survival benefit largely vanishes.

Clinical Interpretation

The study does not prove that MAG has no value. Rather, it suggests that in a large older Medicare population, the long-term survival benefit of MAG is much smaller than what conventional studies have implied, and may be negligible after accounting for unmeasured confounding.

This distinction matters. A treatment can be associated with better outcomes in routine practice without being the true cause of those outcomes. For CABG, surgeons may preferentially use MAG in patients who are healthier, less frail, anatomically more favorable, or otherwise more likely to survive longer regardless of the grafting strategy. Even sophisticated statistical adjustment can miss these factors.

The findings also reinforce the importance of interpreting observational surgical studies carefully. They are valuable for understanding real-world practice and generating hypotheses, but they may overestimate benefit when treatment selection is influenced by nuanced clinical judgment.

What This Means for Patients and Surgeons

For patients, this study suggests that the decision to use multiarterial grafting should be individualized rather than assumed to improve survival in every case. MAG may still be attractive in selected patients because of potential durability advantages and possibly fewer future reinterventions, but the survival benefit is not clearly established in this older population.

For surgeons and heart teams, the findings support a balanced approach. Considerations include:

Patient age and life expectancy
Frailty and overall surgical risk
Diabetes, kidney disease, obesity, and other comorbidities
Coronary anatomy and target vessel quality
Surgeon experience with MAG
The patient’s preferences and tolerance for a more complex procedure

In practice, MAG may still be appropriate for many patients, but the decision should not rest on an assumption of large survival gains. Instead, it should reflect careful judgment about expected benefit, operative complexity, and patient goals.

Strengths of the Study

This investigation has several strengths. It included a very large national sample with long follow-up, which improves statistical power and generalizability for older U.S. adults. The use of Medicare data allowed the authors to study real-world outcomes over nearly two decades.

The instrumental variable design is also a major strength because it attempts to address unmeasured confounding, a common weakness in observational research. By using surgeon MAG rate as the IV, the authors leveraged practice variation to approximate a more causal comparison between treatment strategies.

Limitations to Keep in Mind

As with any observational study, important limitations remain. Instrumental variable analyses rely on assumptions that cannot be fully proven. In particular, the IV must influence outcomes only through treatment assignment and not through other pathways. If surgeon MAG preference is correlated with other unmeasured aspects of care, the estimate may still be imperfect.

The study population consisted of Medicare beneficiaries, so the findings may apply best to older adults rather than younger CABG patients. Also, claims data do not capture all clinically relevant details, such as exact coronary anatomy, graft quality, operative nuances, or some aspects of frailty and functional status.

Finally, while the study examined long-term survival, it did not fully address other important outcomes such as quality of life, angina relief, repeat revascularization, stroke, or postoperative recovery. A treatment with little survival advantage may still offer benefits in symptoms or durability for selected patients.

Bottom Line

In a large Medicare cohort, multiarterial grafting appeared to offer a small survival advantage in conventional risk-adjusted analysis. But when the investigators used an instrumental variable approach to account for unmeasured confounding, that advantage was no longer evident.

The main message is that traditional observational studies may overstate the survival benefit of MAG after CABG. The true effect may be smaller than previously believed, at least in older adults. Future work should continue to refine patient selection and evaluate not only survival but also quality of life and other clinically meaningful outcomes.

Reference

Schaffer JM, Shih E, Squiers JJ, Banwait JK, Hale S, Gasparini A, Mack MJ, DiMaio JM. Multiarterial Grafting and Survival After Coronary Artery Bypass Grafting: An Instrumental Variable Analysis. Journal of the American College of Cardiology. 2026-05-13. PMID: 42126371.