For Three-Vessel Disease, Bypass Surgery Has Better Long-Term Outcomes Than Stenting. Here Is the Evidence.

The Scene

The following scene is drawn from the composite of patients I have cared for in clinic and on the hospital floor. All identifying details are changed.

Robert is 66 years old, a retired high school football coach from Champaign, Illinois. He sat across from the interventional cardiologist for 18 minutes after his cardiac catheterization. He came out of the appointment with a single piece of paper and the words “three-vessel disease” written in his own handwriting. Under that: “Left main involvement.” Under that: “SYNTAX score 34.” Under that: a name and a phone number labeled “cardiac surgeon.”

Robert thought coronary bypass surgery was something people had in the 1980s. He associated it with stories his father told about a neighbor who had his chest opened and never quite came back to himself. He did not know anyone his age who had had it. He did not understand what “three-vessel” meant or why the SYNTAX score was relevant or why a stent was not being offered.

The cardiologist had explained. Robert had nodded. He retained very little.

This article is for Robert.

CABG is open heart surgery. That reality needs to be faced plainly. It involves a sternotomy (division of the breastbone), a cardiopulmonary bypass circuit in most cases, and a recovery that takes weeks to months. It also, in patients with complex three-vessel or left main coronary disease, reduces 10-year mortality more than any stenting strategy currently available. That tension is what Robert needs to understand.

What It Is

Coronary artery bypass grafting is a cardiac surgical procedure that creates new routes for blood to bypass blocked coronary arteries. Instead of reopening the diseased native vessel (as PCI does), CABG uses a conduit (a piece of artery or vein harvested from elsewhere in the body) to connect the aorta or another arterial branch to a point on the coronary artery below the obstruction.

The result is a new blood supply to the ischemic myocardium that routes around the blockage entirely. The native diseased artery remains in place with its blockage; the bypass graft provides an alternative conduit.

Who Undergoes CABG

In the United States, approximately 200,000 to 220,000 CABGs are performed annually 5 / Solid . The procedure is used for:

• Three-vessel CAD with complex anatomy (SYNTAX score above 22): CABG provides more complete revascularization than PCI in most of these cases.
• Left main coronary artery disease (stenosis 50% or greater): The left main supplies the LAD and circumflex territories, approximately 75% of the left ventricular myocardium in most patients. Disease here is high-stakes and the revascularization decision requires careful anatomy-based assessment.
• Multivessel CAD with diabetes mellitus: FREEDOM trial data (discussed below) showed CABG is superior to PCI in diabetic patients with multivessel disease.
• Reduced left ventricular function with multivessel disease: The STICH trial evidence defines this population.
• Failed PCI or inability to revascularize percutaneously due to chronic total occlusion anatomy, excessive calcification, or other anatomical barriers.

The Conduits

The choice of conduit determines the long-term patency of the bypass graft and, ultimately, the durability of the operation.

Left internal mammary artery (LIMA): The left internal mammary artery (also called the left internal thoracic artery, LITA) is harvested from the inner chest wall and anastomosed directly to the LAD or a major diagonal branch. LIMA-to-LAD bypass is the single most important surgical advance in CABG: the 10-year patency rate is 90 to 95% 5 / Solid 70321-2). No other conduit or interventional strategy for the LAD comes close to this durability. The LIMA-to-LAD anastomosis became the defining operation of modern CABG after Loop and colleagues published their landmark 1986 paper demonstrating superior 10-year survival with LIMA over saphenous vein grafts 5 / Solid .

Right internal mammary artery (RIMA): Used as a second arterial graft, either as a free graft or as a bilateral mammary strategy. Bilateral internal mammary artery (BIMA) grafting improves 10-year survival compared with single LIMA in younger patients, but increases wound healing complications by two to threefold in diabetics 5 / Solid .

Radial artery: The radial artery from the non-dominant forearm can serve as a second or third arterial conduit. In the RAPCO trial, radial artery grafts showed superior 10-year patency compared with saphenous vein grafts (93% vs. 73%; p < 0.001) 5 / Solid .

Saphenous vein graft (SVG): The great saphenous vein from the leg was historically the most common conduit. SVG patency rates are substantially inferior to arterial grafts: 50% of SVGs are occluded or significantly diseased by 10 years 5 / Solid . SVGs are still used when arterial conduits are insufficient or unavailable, but their use as the primary conduit has declined as the superiority of arterial grafting has become established.

The Mechanism

On-Pump vs. Off-Pump CABG

The traditional CABG uses cardiopulmonary bypass (CPB): the patient’s circulation is temporarily redirected through a heart-lung machine while the heart is arrested (cardioplegia), cooled, and operated on in a motionless, bloodless field. This allows precise suturing of anastomoses on a still heart. The arrest period is called cross-clamp time; typical cross-clamp times for CABG are 45 to 90 minutes.

Off-pump CABG (OPCAB) performs the anastomoses on a beating heart, using tissue stabilizers to immobilize small sections of the epicardium. The rationale for avoiding CPB is to eliminate the inflammatory response associated with the bypass circuit and the embolic risk from aortic manipulation.

The ROOBY trial (Shroyer AL, et al. N Engl J Med. 2009; doi:10.1056/NEJMoa0902321) enrolled 2,203 patients with planned CABG randomized to on-pump versus off-pump. At 1 year, off-pump had significantly lower graft patency (82.6% vs. 87.8%; p < 0.01) and numerically higher 1-year mortality and composite MACE 5 / Solid . The 5-year follow-up confirmed that off-pump CABG resulted in worse composite outcomes, primarily driven by lower graft patency 5 / Solid . Off-pump CABG is now used selectively (highly calcified aortas, renal insufficiency, high neurological risk) rather than as a routine approach.

Pump-Head: The Neurocognitive Question

A feared complication of on-pump CABG is cognitive impairment, colloquially called “pump-head.” Emboli from the cardiopulmonary bypass circuit, inflammatory cytokines, and microemboli from aortic manipulation have been proposed as mechanisms. The COGNITION trial and the Stump 1995 study (Stump DA, et al. Ann Thorac Surg. 1996; doi:10.1016/S0003-4975(96)00070-2) documented short-term cognitive decline in 50 to 60% of patients at hospital discharge 5 / Solid , most of which resolved by 3 months. Long-term cognitive decline at 5 years does not appear to be specifically attributable to CABG itself: the landmark Newman 2001 study in NEJM showed persistent cognitive decline in 42% of CABG patients at 5 years, but this appears to reflect underlying cerebrovascular disease rather than the surgery itself 4 / Promising . Off-pump CABG does not consistently reduce cognitive outcomes in comparison.

How the Graft Works Hemodynamically

The anastomosis connects the conduit to the coronary artery at a site distal to the obstruction. Blood flows from the aorta (or from the mammary artery’s native origin) through the graft and into the coronary artery beyond the blockage, restoring antegrade perfusion. In a LIMA-to-LAD anastomosis, the internal mammary artery originates from the left subclavian artery; its pressure gradient drives flow into the LAD without requiring aortic engagement at all. This anatomical arrangement contributes to the superior patency of the LIMA: it is physiologically appropriate flow in the correct direction through an arterial conduit that is accustomed to pulsatile pressure.

How It Is Used

The SYNTAX Score and Anatomy-Based Decision-Making

The SYNTAX score was developed to quantify coronary anatomy complexity for the PCI vs. CABG decision. It accounts for the number of lesions, their location in the coronary tree, the presence of bifurcation lesions, chronic total occlusions, heavy calcification, and other complexity features. High SYNTAX score (above 32) identifies anatomy in which CABG produces substantially better outcomes than PCI. Low SYNTAX score (below 22) identifies patients who are appropriate candidates for PCI. The intermediate range (23 to 32) represents genuine equipoise requiring Heart Team discussion 5 / Solid .

The Heart Team Model

ACC/AHA guidelines give a Class I recommendation to Heart Team review for all patients with unprotected left main disease and complex multivessel disease 5 / Solid . The Heart Team includes interventional cardiology, cardiac surgery, and ideally a clinical cardiologist with no financial stake in either option. The model exists precisely because the decision between CABG and PCI requires integrating anatomy, comorbidities, patient preferences, and trial evidence in a way no single operator can do without bias.

Pre-Operative Workup

Before CABG:

• Cardiac catheterization with precise coronary anatomy mapping
• Echocardiography to assess LV function, valve disease, and wall motion
• Pulmonary function testing in patients with significant lung disease (critical for estimating post-operative ventilator requirements)
• Carotid Doppler ultrasound for patients above age 65 or with carotid bruit (to risk-stratify for perioperative stroke)
• Peripheral arterial assessment if conduit harvest from the leg is planned
• Renal function (eGFR below 30 significantly increases surgical risk but is not an absolute contraindication)
• Diabetes assessment: HbA1c above 8% is associated with worse wound healing and infection risk

Geographic Access for CABG

CABG requires a full cardiac surgery program with intensive care unit support, perfusionists, and high surgical volume. In central Illinois, CABG is performed at OSF Saint Francis Medical Center in Peoria and at cardiac surgery programs associated with University of Illinois Health and Northwestern Medicine Bluhm Cardiovascular Institute in Chicago. Carle Foundation Hospital refers complex revascularization cases requiring CABG to these regional centers. For patients in rural central Illinois, travel of 90 to 150 miles for cardiac surgery consultation is frequently necessary. Volume matters: centers performing fewer than 100 CABGs annually have higher mortality rates 5 / Solid .

The Evidence

SYNTAX Trial: The Landmark Comparison

SYNTAX (Serruys PW, et al. N Engl J Med. 2009; doi:10.1056/NEJMoa0804626) enrolled 1,800 patients with three-vessel CAD or left main disease randomized to PCI (first-generation paclitaxel-eluting stents) versus CABG. At 12 months, the primary MACCE composite (death, MI, stroke, or repeat revascularization) favored CABG (12.4% vs. 17.8%; p = 0.002) 5 / Solid . At 5 years, CABG remained superior for the overall population (26.9% vs. 37.3%; HR 0.66; p < 0.0001) 5 / Solid 60141-5). Stroke was more common with CABG than PCI (3.7% vs. 2.4% at 5 years; p = 0.09). In the subgroup with high SYNTAX score (above 32), CABG dramatically outperformed PCI for survival.

SYNTAX used first-generation stents. Contemporary second-generation stents perform better; nonetheless, subsequent data have not reversed the CABG advantage in complex anatomy.

EXCEL Trial: Left Main Disease

EXCEL (Stone GW, et al. N Engl J Med. 2016; doi:10.1056/NEJMoa1610227) enrolled 1,905 patients with left main disease and SYNTAX score below 32, randomized to PCI with everolimus-eluting stents versus CABG. At 3 years, the primary composite of death, MI, or stroke did not differ significantly (15.4% PCI vs. 14.7% CABG; p = 0.02 for non-inferiority) 5 / Solid . At 5 years, PCI had a slightly higher rate of the composite (22.0% vs. 19.2%; HR 1.19, 95% CI 0.98 to 1.43), driven primarily by higher repeat revascularization in the PCI arm 4 / Promising . Stroke was more common with CABG (2.9% vs. 1.9% at 5 years). The trial remains contested because of disagreements about MI definition methodology.

NOBLE Trial: Left Main Disease

NOBLE (Makikallio T, et al. Lancet. 2016; doi:10.1016/S0140-6736(16)32052-9) enrolled 1,201 patients with left main disease randomized to PCI versus CABG. At 5 years, PCI was associated with a higher rate of the primary composite (29% vs. 19%; HR 1.48, 95% CI 1.11 to 1.96; p = 0.007) 5 / Solid , driven by higher non-procedural MI and more repeat revascularization. NOBLE, unlike EXCEL, suggested CABG superiority even for low-complexity left main disease.

The NOBLE-EXCEL discordance for left main disease remains unresolved and illustrates why Heart Team decision-making is essential rather than algorithmic.

FREEDOM Trial: CABG vs. PCI in Diabetes

FREEDOM (Farkouh ME, et al. N Engl J Med. 2012; doi:10.1056/NEJMoa1211585) enrolled 1,900 patients with diabetes and multivessel CAD randomized to CABG versus PCI. At 5 years, CABG reduced the primary composite of death, MI, or stroke (18.7% vs. 26.6%; HR 0.71, 95% CI 0.57 to 0.87; p = 0.005) 5 / Solid . Death from any cause: 10.9% CABG vs. 16.3% PCI (HR 0.67, 95% CI 0.52 to 0.86; p = 0.002). Stroke was more common with CABG (5.2% vs. 2.4%; p < 0.001). The FREEDOM data established CABG as the preferred revascularization strategy for diabetic patients with multivessel disease; this is now a Class I recommendation 5 / Solid .

STICH Trial: CABG in Reduced EF

STICH (Velazquez EJ, et al. N Engl J Med. 2011; doi:10.1056/NEJMoa1100356) enrolled 1,212 patients with CAD and LV ejection fraction 35% or below, randomized to CABG plus medical therapy versus medical therapy alone. At a median follow-up of 56 months, all-cause mortality did not differ significantly at the primary analysis (36% CABG vs. 41% medical therapy; HR 0.86, 95% CI 0.72 to 1.04; p = 0.12) 5 / Solid . STICH Extended Follow-Up (STICHES, at 10 years) showed a survival advantage for CABG (58.9% vs. 66.1% mortality; HR 0.84, 95% CI 0.73 to 0.97; p = 0.02) 5 / Solid . The 10-year data established CABG benefit in ischemic cardiomyopathy; the benefit required time to emerge. What STICH did not show: viability testing (nuclear or CMR viability) did not predict differential benefit from CABG versus medical therapy in the prespecified substudy, a finding that challenged the longstanding practice of using viability to guide the CABG decision in reduced EF patients.

Graft Patency and Long-Term Outcomes

The superiority of CABG in complex anatomy rests partly on graft patency. At 10 years, LIMA-to-LAD patency remains 90 to 95%; saphenous vein grafts show 50% occlusion or significant stenosis. The long-term benefit of CABG over PCI in multivessel disease is driven substantially by this durable arterial revascularization, which cannot be replicated with current stent technology 5 / Solid 70321-2).

The Patient Experience

The first three days after CABG are not the recovery. They are the emergence from the operation.

Robert spent 11 hours in the operating room (the procedure itself was 4 hours; prep and anesthesia took the rest). He came out intubated, with a chest tube, urinary catheter, and an arterial line. He was extubated 8 hours after arrival in the ICU. His wife sat with him through the night; he does not remember the first 36 hours clearly.

What patients commonly describe about CABG recovery:

• Sternal discomfort, not always pain. The sternotomy heals over 6 to 8 weeks. Patients describe a deep pressure sensation when coughing or turning. A cardiac surgery unit provides a “heart pillow” (a small cushion held to the chest when coughing) to splint the sternum. Deep breathing exercises begin on post-operative day 1.

• Fatigue. A duration and quality that is categorically different from ordinary tiredness. Most patients describe this as the dominant symptom for 2 to 4 weeks: unable to stay awake through a conversation, unable to walk to the mailbox without resting. This is normal and resolves. The cardiopulmonary bypass inflammatory response and the obligate blood transfusion in many cases contribute.

• Cognitive fog. The “pump-head” phenomenon is real in the first weeks. Most patients notice word-finding difficulties, concentration problems, and what they describe as “thinking through cotton.” In the majority of patients, this resolves within 6 to 8 weeks 5 / Solid 00070-2).

• Leg discomfort from vein harvest. If a saphenous vein was harvested, the leg incision produces swelling, bruising, and a tight sensation for 2 to 4 weeks. Ambulatory compression stockings reduce edema.

• Cardiac rehabilitation. Phase II cardiac rehabilitation (supervised exercise program starting 2 to 4 weeks post-discharge) is a Class I ACC/AHA recommendation and reduces 5-year mortality by 25 to 35% 5 / Solid . It is underutilized: fewer than 25% of eligible patients complete the full program.

What Your Cardiologist Will Not Have Time to Explain

• The grafts are not permanent. The LIMA-to-LAD is durable; saphenous vein grafts are not. By 10 years, half of saphenous vein grafts will be significantly diseased or occluded. The bypass procedure buys time; it does not eliminate atherosclerosis. Aggressive lipid management is required to protect graft patency 5 / Solid .

• Your native coronary arteries will continue to progress. The bypassed vessels and the non-bypassed vessels accumulate plaque on the same timeline as before the operation. Post-CABG patients require the same risk factor management as any patient with established CAD, possibly more aggressively.

• Sternal precautions are real. For 6 to 8 weeks, patients should not push, pull, or lift anything heavier than 10 pounds, not use their arms to push up from a seated position, and not drive. Premature sternal loading causes a “sternal click,” separation of the sternotomy that requires re-wiring and carries a 5 to 10% infection risk.

Sex Differences in CABG Outcomes

Women referred for CABG are older at the time of operation, more likely to have diabetes and hypertension, and have smaller coronary arteries (smaller conduit targets). Women have higher in-hospital mortality after CABG than men (2 to 4% vs. 1 to 2%) after adjustment for comorbidities the gap narrows but does not fully close 5 / Solid . Long-term 5- and 10-year survival after CABG is comparable between sexes in large registry analyses when baseline differences are accounted for.

Decisions and Trade-Offs

The Decision Framework

CABG should be chosen over PCI when:

• SYNTAX score above 32 (Class I, CABG preferred)
• Left main disease with SYNTAX score above 32 or complex anatomy (Class I, CABG preferred)
• Multivessel CAD with diabetes mellitus (Class I, CABG preferred based on FREEDOM)
• CAD with reduced EF (EF below 35%) seeking 10-year survival benefit (Class IIa, CABG reasonable)
• Chronic total occlusion of a major vessel not amenable to PCI (Class IIa)

CABG may be equivalent to or inferior to PCI when:

• SYNTAX score below 22 (Class I, PCI reasonable)
• Left main disease with low-complexity anatomy (SYNTAX below 22) and patient preference for less invasive approach (Class IIa, either reasonable)

The Operative Mortality Discussion

Elective CABG in a 65-year-old with three-vessel disease, normal EF, and no major comorbidities carries an operative mortality of approximately 1 to 2% 5 / Solid . This risk increases substantially with:

• Age above 75: 2 to 4%
• EF below 35%: 3 to 5%
• Reoperation (redo CABG): 5 to 8%
• Emergency operation: 5 to 15%

Patients deserve these numbers stated explicitly before consenting to the procedure. A surgeon who cannot provide a patient-specific operative mortality estimate using Society of Thoracic Surgeons (STS) risk score methodology is not giving adequate informed consent.

Watchful Waiting: The Medical Therapy Alternative

For patients with stable ischemic heart disease and complex anatomy who decline revascularization, aggressive medical therapy with high-intensity statin, antiplatelet therapy, and beta-blockade provides meaningful but incomplete protection. The STICH data showed that CABG’s survival benefit over medical therapy in reduced EF patients did not become statistically significant until 10 years. For patients with preserved EF and stable symptoms, ISCHEMIA showed that medical therapy is non-inferior to PCI at 3 to 5 years; CABG data in stable patients with preserved EF suggest similar short-term outcomes but longer-term graft-driven benefit that accrues over 5 to 10 years.

Cost

CABG hospitalization in the United States carries facility and professional fees ranging from $50,000 to $100,000 before insurance adjustments, with the ICU, perfusionist, anesthesia, and postoperative monitoring contributing to the difference from elective PCI. Medicare and commercial insurance cover CABG for appropriate indications. Patient out-of-pocket is typically $1,000 to $5,000 under most commercial plans. The downstream cost of secondary prevention (statins, antiplatelet therapy, cardiac rehabilitation) is substantial; the CABG itself is not the end of the financial exposure.

The Three Questions Every Patient Should Ask

1. “What is my STS predicted risk score for this operation, and how does your program’s observed mortality compare with the national STS average?” Every cardiac surgery program that performs CABG reports outcomes to the Society of Thoracic Surgeons database; this information is publicly available at CTSurgeryNet.

2. “What conduits will you use, and what is your plan to achieve complete arterial revascularization?” The proportion of arterial (vs. saphenous vein) grafts used is a surrogate for long-term graft durability. Centers with high arterial grafting rates have better 10-year outcomes.

3. “Have my coronary images been reviewed by both the surgeon and the interventional cardiologist as a Heart Team?” If the answer is no, request that meeting before consenting.

Clinical Synthesis

Robert’s SYNTAX score of 34 and left main involvement put him in territory where the trial data are unambiguous. No stent strategy available today matches what the LIMA will do for his LAD over the next 10 to 15 years. His choice is not “surgery vs. something better.” His choice is “surgery vs. suboptimal alternatives.”

The core clinical thesis is built around preventing the situation Robert is in. By the time a SYNTAX score of 34 exists, the disease has been accumulating for 25 to 30 years. The plaque in Robert’s LAD, circumflex, and RCA started forming in his thirties. His LDL-C was probably increased when he was 45. His smoking history, which he quit at 52, contributed. His ApoB, which no one ever measured, was probably discordant from his LDL-C by the time he was 50.

This is not an argument for avoiding surgery. Robert needs CABG. This is an argument for the upstream investment that prevents the next Robert from arriving at the cath table at 66 with three-vessel disease and a SYNTAX score that removes PCI from the options list.

If you have been told you need CABG: a structured cardiovascular assessment reviews the anatomy, the operative risk, the second-opinion question, and the post-operative medical therapy plan. Most important: your statin, your target LDL-C, and your secondary prevention plan after surgery will determine how long your grafts last. Patients who achieve LDL-C below 55 mg/dL post-CABG have significantly lower rates of graft failure at 10 years.

If you are in the decision window between CABG and PCI: the Signal Check provides a structured review of your anatomy, comorbidities, and the trial evidence applicable to your specific case. The decision should never be made in a single cardiologist’s office without surgical input.