Heart surgery in the early days (1970’s) was a very dangerous proposition. The heart lung machines were rudimentary, with risks of complications that were dramatically higher than what we see today. For instance, CABG was the most common operation then (as it is now) and was associated with a risk of postoperative stroke that was around 5-10%. Dramatic improvements in preoperative risk stratification, perfusion strategies, and operative techniques have reduced the risk of stroke after CABG in the modern era to just under 2%.

Two strokes out of every 100 cases still too high for most patients and referring providers. Though the risk has come a long way, it remains one of the major concerns limiting enthusiasm about referral for CABG. The reason for this goes beyond concerns that the patient is gong to die from the stroke – the risk of operative mortality is 10-fold higher after a stroke (1-2% standard vs. 10-20% after CVA). More concerning to patients is that they won’t die because the quality of life after postop stroke is notoriously poor. Stroke after CABG is not usually a benign event that resolves with physical therapy. These unfortunate patients have 3 fold longer length of stay in the ICU and hospital and in many cases remain hospitalized for weeks and even months. The vast majority (85%) that are able to get discharged end up in an inpatient rehab facility where they suffer from a risk of clinical depression that exceeds 40%. The mortality risk over the first year for patients in rehab is around 50%, meaning that most never make it home. This event has an agonizing effect on the families. There is a profound economic impact on patients and families that must take time off work to support care. The tremendous physical and psychologic burden of having to provide this care disproportionately affects those without resources, causing worsening health disparities.

Considering this grim but realistic picture of stroke, it is clear why many referring providers have been sending patients for balloon angioplasty/PCI even when their pattern of coronary disease would indicate CABG is more effective according to the guidelines. Multivessel PCI may not have the longevity of CABG but it has 1/10th the risk of stroke (0.2%). Virtually every paper that has compared the outcomes of CABG vs. PCI contains a statement like: “the survival benefit after CABG needs to be balanced against the risk of stroke.” A typical CABG program performs 250 cases per year. A 2% stroke risk means that there are on average 5 families every year from every program that are put through the gauntlet of suffering alongside a family member with stroke. There are some programs that will have fewer strokes but these are balanced out by those programs that have more. Even when a program goes a few years with a lower risk of stroke, the tendency is to have other years where the risk is higher and end up with a 2% risk in the long run – a phenomenon described as “regression to the mean”.

The only way to reduce the stroke risk below 2% is to somehow improve the procedure. Intraoperative transesophageal echocardiography is increasingly used as routine practice during CABG. This imaging test can detect aortic or intra-cardiac pathology at risk for being disrupted by the surgical manipulation required to perform CABG. This is the source of intra-arterial embolus, which is the main etiology of stroke. The use of a separate epiaortic ultrasonic probe placed directly on the aorta improves the ability to detect focal plaque that evades intraoperative palpation by the surgeon’s finger or TEE imaging. Both of these imaging techniques prompt the surgeon to change intraoperative decisions such as the use of modified aortic clamping or “no touch” techniques that help reduce embolism risk. There have been improvements in bypass circuitry/cannulas including inline arterial filters and intraaortic filters that are better able to capture embolic debris. Improvements in anesthesia care have been introduced such as the routine use of cerebral oximetry probes placed on the scalp which can guide intraoperative hemodynamic management in order to maintain adequate cerebral perfusion during surgery. Despite the wide-spread introduction of these innovations into routine clinical practice, the risk of stroke after CABG reported to the STS National Cardiac Surgical Database has not changed for 2 decades. The annual stroke risk for CABG in the US over the past 20 years seems to be stuck on the asymptotic phase of a curve fixed at 2%.

The steady annual incidence of stroke despite best efforts was a red flag pointing out that further incremental innovations weren’t likely to influence outcomes. It suggests that the heart lung machine and/or need to cross-clamp the aorta causes a fixed risk of embolic debris getting into the circulation. The amount of debris may be less than the early days but adding more and more safeguards is not going to eliminate this issue and improve risk. Economists describe this state as the law of diminishing returns. In the world of business, this phase often spurs thought of shifting the whole paradigm for how things are done. A disruptive innovation was introduced in the late 1990’s that enabled CABG to be performed by without cardiopulmonary bypass or aortic cross-clamping. This technique, called off-pump CABG (OPCAB), was paradigm shifting because it did not try to mitigate the impact of embolic debris. Rather, it avoided the initial triggers so that embolism didn’t occur in the first place. Considerable accumulated evidence shows this approach works. Dozens of randomized controlled trials of on- vs. off-pump CABG have been performed over the past 20 years. Meta-analyses of all these trials have shown that OPCAB results in a 50% reduction in the risk of postoperative stroke compared to conventional, on-pump CABG. It appeared at the time that this paradigm shift fixed one of the most vexing problems of CABG.

Surprisingly, the introduction of OPCAB resulted in no change in the annual risk of stroke in real world practice. It remained at 2%. This time, the failure resulted from problems with how this complex procedure was adopted at new centers. At the time, OPCAB was perceived by many as an incremental upgrade. Only minor tweaks, not wholesale changes, to the mature CABG procedure were required to get this new program off the ground. In retrospect, we now see that OPCAB was a far more disruptive innovation than proposed. Performing heart surgery on a beating heart is not just a technical feat for the surgeon, it greatly challenges the teamwork, communication and leadership skills of the team. The anesthesiologist must learn new ways to maintain the patient’s hemodynamics when the surgeon manipulates the heart to create the bypass grafts on the back of the heart. The support staff must be comfortable with new equipment demands of this procedure. A whole new set of adverse events can occur that need to be anticipated and avoided. Most notably, an OPCAB procedure may have to convert to an on-pump approach as its “bail-out” strategy, usually due to poor hemodynamics. These bail-out conversions are risky and challenging because the patients are often unstable and this maneuver must be performed quickly. Given the common lingo used to describe this event – “slammed onto the pump” – it is not surprising that the postoperative stroke risk exceeds 10% when a conversion was required. It only takes a few such conversions and their associated risks to wipe out the 1% advantage in stroke risk seen with OPCAB cases completed without a conversion.

Expert centers that performed the initial clinical trials reported conversion rates of <3%, perhaps due to the technical prowess of the operative surgeon and/or other innate skills present within the members of these early adopter teams. Unfortunately, there was no translatable strategy that enabled OPCAB to be successful as it diffused through the wider majority of the surgical community. The innate skills and tacit knowledge of early adopters did not translate onto teams adopting OPCAB later on in its life cycle. There were predictable challenges with teamwork and communication as well as problems with credentialing and training of surgeons to do new procedures that compromised results (see blog post about the Scandal of Robotics training). As it turns out, this limitation loomed large. At its peak, OPCAB approached 35% of the CABG market but disastrous results at a few unprepared programs dampened enthusiasm and market penetration of the technique quickly fell to less than 15%. True to the motto of “not always right but never unsure”, cardiac surgeons maintained their hubris. Many of those that failed to adopt during this period point their finger on the innovation itself as a failed concept rather than admitting any problem with their own team’s level of training or ability.

One of the biggest challenges of open chest OPCAB surgery has been trying to graft targets on myocardial zones outside of the anterior myocardial surface (LAD and diagonal). This requires the heart to be torqued and manipulated outside of its in situ position. As a result of cardiac compression and distortion, the blood pressure often drops, which can cause myocardial ischemia and poor hemodynamics. Robotic assisted CABG is another innovation that enables a LIMA graft to be placed onto the LAD while the heart is beating and in its natural in situ position. For this procedure, bypasses that are created cause very little hemodynamic instability and the risk of emergent conversion to on-pump is negligible. In addition, blood loss is minimal which further improves the overall hemodynamics. Treatment of multivessel CAD in robotic cases utilizes a hybrid approach, which means that coronary lesions in vessels on the inferior and lateral surfaces of the heart are treated with angioplasty/PCI. There is no need to manipulate the heart with this approach which greatly reduces the risk of conversion to on-pump CABG.

There are additional advantages of robotic/hybrid coronary revascularization that, in theory, help mitigate the risk of stroke. There is no need to stop anticoagulants prior to robotic surgery due to the reduced risk of bleeding caused by this procedure. Some of the strokes in the published CABG trials have coincided with the preoperative period that antithrombotic agents were withdrawn to reduce surgical bleeding after a full sternotomy. There is a theoretical risk of a “rebound hypercoagulability” in the immediate period that these agents are withdrawn. It is a basic tenant of vascular biology (i.e. Virchow’s triad) that hypercoagulability is a necessary ingredient for any thrombotic event like stroke to be able to occur. In addition, it has been established that a no touch aortic technique reduces the risk of stroke. However, the adoption of this approach has been minimal because it depends on multiarterial grafting using bilateral IMA and or composite radial conduits. This is a complex grafting strategy has been used at only a few expert centers and is not suitable for addressing the coronary anatomy of many patients that are typically referred for CABG. In hybrid revascularization, PCI enables coronary lesions outside of the LAD to be addressed without the need for a complicated grafting strategy. Combined with an off-pump LIMA to LAD, this approach has the theoretical advantage of inducing a negligible risk of either an intraoperative conversion to an on-pump procedure or stroke. This strategy also is not possible in all patients, but it almost certainly expands the number of cases that can be done with a no touch concept well beyond what others have achieved.

This time, the myriad of theoretical advantages of rCABG reconciles with real world evidence. Two large databases have collected information on nearly 10,000 patients that have undergone this robotic procedure over the past 5 years, many of which received the hybrid strategy. The Society of Thoracic Surgeons database and the National Inpatient Sample have demonstrated that the rate of stroke after rCABG is less than 0.5%. (P Cavallaro, JTCVS 2015; Whellan D, ATS 2016). My own experience agrees and includes 553 robotic CABG cases in a row without a postoperative stroke. Cardiologists in the centers where I’ve practiced have recognized this stroke advantage from robotics. Since so few other cardiac surgeons do robotic cases, there has been a referral bias in my practice towards those that are high risk for stroke. My series of 553 cases has included a disproportionate number of elderly patients, those with prior strokes in the past, those with severe atherosclerosis in the ascending aorta or tight occlusions of one or both carotid arteries and a variety of other serious risk factors for stroke.

The power of the robotic strategy comes from mitigating both mechanisms of postoperative stroke with less need to bail out. By expanding the feasibility of off-pump and “no touch” coronary revascularization, the fixed risk of embolic debris generated by aortic manipulation and cardiopulmonary bypass is essentially eliminated. Compared to off-pump CABG via open sternum, there is less need for cardiac manipulation to expose coronary targets. This helps reduce the risk of cerebral hypoperfusion and watershed infarction caused by periods of low blood pressure when the heart is being manipulated. My lab has documented that cerebral oxygenation is significantly better during the course of robotic CABG than a risk matched conventional CABG group (Benjamin Mills Master’s Thesis: “Cerebral oximetry during robotic, less invasive heart surgery”, presented and accepted May 2, 2013). We have also noted reduced serum levels of s100, a biomarker of brain injury, after rCABG as compared to open chest CABG.

Ironically, this success of robotic CABG in stroke was probably caused by its failures. The biggest failure has been the resistance of the cardiac surgery community to endorse its wide-spread adoption. This kept robotics within the hands of the 20-30 expert cardiac surgery programs which contributed the bulk of data that made up the database analyses. The evidence that robotic CABG would reduce a rare complication like stroke almost certainly depended on keeping it focused in the early phase on those few centers motivated to work through its other problems. Now there is evidence of a reduction in stroke risk, so there is now a compelling clinical reason for others to adopt this new procedure in a way that didn’t previously exist.

High risk industries – hospitals, nuclear power plants, the aviation industry – achieve extraordinary improvements in safety when they pursue all strategies that improve outcomes, regardless of their financial or political impact. If conventional approach to CABG is all that is offered, then we should be resigned to accept a 2% risk of stroke as the cost of doing business. There is no reason to believe that the next two decades of using this procedure would be any different than the past two decades. In fact, the stroke risk may actually rise as the baby boomers reach retirement because age is a major risk factor for stroke after CABG. In light of these higher risk cases, now is the time for our profession to be asking the tough questions about robotics and the inconvenient truth that it likely reduces stroke risk. What would happen if we label stroke after CABG as a “never event” – similar to leaving a sponge in the chest or amputating the wrong limb? Would such a mandate prompt an increase the small subset of programs that offer robotic CABG/hybrid revascularization when the number of programs has been flat for the past 10 years? Would it lead programs that don’t offer it refer their CABG patients to centers that do? Hospitals that have embraced the tenants of high reliability organizations are crystal clear in their answers and show the political and financial will to do what they say.