Supported Percutaneous Intervention

Introduction and Rationale: Identifying the high-risk Patient

Complications of balloon angioplasty that threaten coronary blood flow, termed acute and threatened occlusions, usually require urgent surgical intervention and are the main causes of procedure-related morbidity and mortality. Before the advent of stents as a bail-out treatment for impending vessel closure, this complication occurred in approximately 6% of balloon angioplasty procedures. Patients who required emergent surgery in this setting had a 50% likelihood of suffering myocardial infarction (MI), and mortality rates were as high as 10%. In these early studies, patient characteristics that included compromised ventricular function, left main coronary artery (LMCA) disease, multivessel disease, and older age were identified as risk factors for balloon angioplasty–related mortality. With the development of coronary stents and advanced pharmacotherapy to seal dissections and improve blood flow in thrombotic lesions, respectively, the need for urgent surgery was reduced with a concomitant reduction in percutaneous coronary intervention (PCI)-related morbidity and mortality.

In the current era, studies have demonstrated that the need for urgent surgery after PCI has been reduced to less than 1%, with a marked reduction in procedure-related mortality. In one comparison of patients treated in 1997 and 1998 with those treated in 1985 and 1986, the rate of in-hospital deaths, MI, and coronary artery bypass grafting (CABG) fell from 7.9% to 4.9%, despite the treatment of more complex lesions and stent use in only 71% of patients. Most of the differences between these periods were accounted for by the reduction in the need for emergent CABG from 3.7% to 0.4%. Nonetheless, morbidity and mortality among patients who required emergent CABG remained high. Moreover, the increased confidence afforded by stents and improved operator techniques and experience have prompted interventions on more complex lesions and in patients with more severe cardiac and noncardiac diseases. In particular, results of the Synergy Between Percutaneous Coronary Intervention With Taxus and Cardiac Surgery (SYNTAX) trial confirm reasonable outcomes in select patients with LMCA disease or multivessel coronary disease, and completed and ongoing trials of high-risk PCI with cardiac assist device placement indicate that extremely high-risk patient populations are, indeed, being increasingly considered for PCI. American College of Cardiology Foundation (ACCF)/American Heart Association (AHA) clinical guidelines have kept pace and currently support percutaneous unprotected left main stenting in appropriate high-risk surgical patients as a class IIa and IIb indication. The more recent American College of Cardiology (ACC)/AHA/ Society for Cardiovascular Angiography and Interventions (SCAI) 2017 appropriate use document also considers left main stenting appropriate in a subset of patients, depending on the location of the stenosis, disease burden in other territories, and use of antianginal therapy. Given the large amount of the myocardium in jeopardy in these subsets, as well as the baseline comorbidities frequently present (including reduced ventricular reserve), the potential for severe clinical decompensation in such patients is a real concern. It has therefore become essential to precisely identify the predictors of risk and to consider the use of hemodynamic support for patients at high risk for procedural complications and in-hospital mortality.

A number of variables have been defined that contribute to increased procedural risk in PCI. These can be patient related, be lesion related and depend on clinical presentation. Patient-related variables include age, symptoms of heart failure, impaired left ventricular function, prior cardiovascular disease (CVD) such as prior MI, other known coronary artery disease (CAD) (multivessel or left main disease), peripheral arterial disease (PAD), diabetes mellitus, and chronic kidney disease. Lesion-related variables incorporate left main stenosis, bifurcation disease, saphenous vein grafts, last patent vessel, heavily calcified lesions, ostial lesions, and chronic total occlusions. Clinical acuity of presentation also confers a high risk.

Although a number of investigators have used commonsense definitions to classify patients at high risk, two studies systematically developed risk models. In the Mayo Clinic model, clinical and angiographic variables were used to predict in-hospital complications after PCI. The variables were age, shock, renal insufficiency, urgent procedures, heart failure, thrombus, and LMCA or multivessel disease. A score based on these factors predicted the risk of complications and identified a “highest risk” group with an event rate that exceeded 25%. In a similar study of 46,000 procedures in the mandatory New York State Hospital PCI reporting system, investigators included nine factors in a risk score: (1) ejection fraction (EF), (2) previous MI, (3) gender, (4) age, (5) hemodynamic state, (6) PAD, (7) congestive heart failure (CHF), (8) renal failure, and (9) LMCA disease. Using this New York State Hospital model, a graded risk score for in-hospital mortality was derived and validated. Approximately 2% of all patients had a greater than 5% risk for in-hospital death, and approximately 4% of patients had a greater than 3% risk. Other studies have confirmed these risk factors for PCI-related complications, but neither model may accurately represent the higher-risk patient populations undergoing intervention in the current era.

More recently, CathPCI registry data were used to develop “full,” “precatheterization,” and “bedside risk prediction” mortality models. The precatheterization model consisted of age, body mass index (BMI), CVD, PAD, chronic lung disease, prior PCI, glomerular filtration rate (GFR), EF, cardiogenic shock (CS), acuity of PCI, New York Heart Association (NYHA) class symptoms within past 2 weeks, and cardiac arrest within 24 hours. The CathPCI registry risk prediction score performed well, and the study also showed that clinical acuity remains the strongest predictor of PCI procedural mortality. In addition to these clinical risk factors, angiographic factors of lesion complexity—thrombus, calcification, and bifurcations—have been shown to be associated with more dissections, distal embolization, and side-branch occlusions, resulting in a threefold increase for in-hospital death.

Several patient characteristics deserve separate discussion. Although female sex was initially associated with complications in early studies of balloon angioplasty, more recent studies have failed to demonstrate an important effect on outcome. When compared with CABG, PCI has increased rates of mortality in patients with diabetes. However, when looking at the subset of patients undergoing PCI with and without diabetes, patients with diabetes have more complex lesion characteristics and risk factors but no increase in in-hospital mortality after multivariable adjustment.

More recently, baseline left ventricle (LV) dysfunction and the extent of the myocardium in jeopardy during the procedure have reemerged as perhaps the strongest clinical risk factors for intraprocedural decompensation and in-hospital mortality. Accordingly, recent studies on high-risk PCI have used the combination of severe ventricular dysfunction (represented by EF <30% to 35%) and either LMCA disease, last patent conduit, or multivessel disease as high-risk PCI inclusion criteria. Thus it is clear that despite major advances in the technical and procedural performance of modern PCI, clinical and angiographic predictors of significant morbidity and mortality can be identified ( Table 36.1 ). Moreover, it appears likely that increasing numbers of patients will be undergoing high-risk PCI, including the very old and those with LMCA disease, multivessel disease, and significant ventricular dysfunction. In many cases, bypass surgery is not a viable or palatable option, with prolonged recovery, leaving PCI as the only remaining possible mechanism to improve ventricular function and reduce ischemic symptoms.

Finally, patients undergoing PCI in the setting of acute myocardial infarction (AMI) with CS represent an especially high-risk group in whom hemodynamic support devices have been used, and national guidelines currently support their use after initial attempts at medical stabilization. Lately, there has been emphasis on the timing of hemodynamic support devices in AMI complicated by CS. The concept of door-to-support time has been developed and seems to have significant effect on outcomes in preclinical studies. As the hemodynamic instability in CS progresses, the reduced cardiac output and elevated filling pressures lead to decreased systemic perfusion, lactic acidosis, multiorgan ischemia, hepatic and venous congestion, and multiorgan failure. This trail of events subsequently transitions CS from a potentially reversible hemodynamic problem to a more complex hemometabolic problem. Initiating mechanical circulatory support (MCS) early (i.e., prior to PCI or pharmacotherapy for hemodynamic support) may theoretically lead to improved survival. A retrospective study from the continuous ventricular assist device (cVAD) Registry found that early implantation of Impella (Abiomed, Danvers, MA) before PCI and before requiring inotropes/vasopressors was associated with increased survival. Furthermore, they demonstrated that with increasing time from shock onset to MCS, the survival decreased (66%, 37%, and 26% when MCS was initiated <1.25 hours, 1.25 to 4.25 hours, and after 4.25 hours, respectively; P = .017). A smaller subset of patients from the cVAD Registry who underwent unprotected left main PCI as a culprit lesion in patients with MI complicated by CS was then looked at. The number of patients was small ( n = 36); however, the data showed that patients who received Impella prior to PCI had better early survival as compared with the ones who received the Impella after the PCI. Similar results were also noted in the USPella Registry that showed significantly higher in-hospital survival (63% higher) when Impella was initiated prior to revascularization. This concept will be further evaluated in an ongoing randomized clinical trial: Door to Unloading With Impella CP System in AMI trial.

Together, these data provide the rationale and impetus for the increased use of hemodynamic support during complex or high-risk PCI in various clinical settings in the current era, including CS. The remainder of this chapter will discuss the approach to such patients, historical and current data, the devices currently available to provide support, and the results that may be achieved by using them.

Approach to the Patient

Mechanical circulatory support at the time of PCI has historically been instituted in one of two settings: electively for presumed high-risk intervention and emergently for periprocedural hemodynamic instability. The latter may occur during a planned elective PCI or occur on presentation, such as with AMI and CS. However, specific indications remain unclear because of limitations in performing large-scale randomized trials and evaluating specific devices in individual patient subsets, especially for the latter indication. Nevertheless, a review of existing literature and ongoing clinical trials provides a framework for both patient and device selection when evaluating patients for elective or emergent support.

Electively placed mechanical support is aimed at improving procedural success by minimizing myocardial ischemia and maintaining hemodynamic stability, thereby reducing clinical decompensation and resultant mortality in high-risk preselected patient subsets, as discussed in the previous section. In this setting, prophylactic insertion of the intraaortic balloon pump (IABP) or the Impella ventricular assist device (VAD) appears to have the most robust observational data for improving procedural success with a minimal increase in complications. Despite encouraging registry data, however, a randomized controlled trial (RCT) failed to show clinical benefit to routine prophylactic IABP insertion in patients undergoing high-risk PCI. The more powerful Impella device was subsequently studied in patients deemed to need hemodynamic support for high-risk PCI. Whereas the device was successful in providing superior hemodynamic support, in comparison to routine IABP use (based on magnitude of drop in cardiac power output [CPO] during the case), no difference was reported in major adverse events at 30 days. However, secondary end points in each of these trials were hypothesis generating, in that procedural complications and long-term mortality were improved with routine IABP use in the Balloon Pump-Assisted Coronary Intervention Study (BCIS-1) trial, and repeat revascularization, total length of stay, and 90-day major adverse events were improved with Impella, especially in patients who did not undergo rotational atherectomy.

The use of mechanical circulatory support in the emergent setting for patients with documented hemodynamic instability or CS is more familiar to interventional cardiologists. Instability may be present before PCI (as in acute MI with compromised ventricular function) or may develop as a consequence of procedural complications such as coronary dissection, poor coronary reflow, or thromboembolism. A common underlying finding in most of these patients is ventricular dysfunction. Although there is paucity of RCT data in this arena, a pooled meta-analysis of IABP use in patients with acute ST-elevation myocardial infarction (STEMI) found no benefit to IABP use in conjunction with primary angioplasty, but some benefit was seen in those receiving thrombolytic therapy. The more recent Intraaortic Balloon Pump in Cardiogenic Shock II (IABP-SHOCK II) trial similarly found no benefit to routine IABP use in patients presenting with CS, prompting a downgrading of the European guidelines to class III. Available percutaneous ventricular assist devices (pVADs) appear to improve hemodynamic parameters in patients with acute clinical decompensation, but only registry data are available to suggest improved outcome. As a result, determining whether and which device to use in specific settings remains controversial and is primarily guided by experience and expert consensus, with the field as a whole shifting toward more powerful support earlier in the course of clinical deterioration.

Device selection is based on several factors; these include ease and rapidity of institution, level of invasiveness and complications, physician familiarity, requisite technical expertise, level of anticipated circulatory support, available supportive clinical trial data, and, in the case of the Impella device, Food and Drug Administration (FDA) approval for both high-risk PCI and CS. An IABP is the least invasive and most familiar device and may be instituted rapidly, but it also provides the least support, averaging 0.5 L/min augmentation in cardiac output. It may be left in place for several days and has a low vascular complication rate. Conversely, full cardiopulmonary support (CPS) is significantly more invasive, and it requires timely surgical and perfusionist collaboration for institution and removal, but it can produce greater improvement in cardiac output, approximating normal physiology. However, CPS cannot be maintained indefinitely because hematologic and pulmonary complications increase as bypass time approaches 6 hours and, in the absence of concomitant LV venting or unloading, its deleterious effect on myocardial oxygen consumption remains a concern. pVADs, such as the TandemHeart or the Impella 2.5 or CP, provide an intermediate level of support that approaches 2.5 to 4 L/min; in addition, these devices can be placed emergently in the catheterization laboratory (cath lab) without surgical backup, which has prompted their increased use recently. Unlike the Impella, the TandemHeart requires transseptal puncture to deliver the inflow cannula into the left atrium, and it requires a somewhat larger arterial cannula. Thus only patients with a larger femoral arterial diameter are able to accommodate device placement, which can be performed only by those skilled in the transseptal technique. Consequently, compared with the Impella device, cath lab utilization rate of the TandemHeart appears to have stabilized or decreased in recent years. In both devices, the cannulae are larger than with an IABP and may result in significant vascular morbidity. However, unlike full CPS, the use of a pVAD has been successful for intermediate lengths of time (up to 14 days). Relatively smaller cannulae, as with the Impella, are likely to reduce femoral complications.

Historically, elective high-risk PCI has been performed safely with either provisional or prophylactic IABP support. However, as described earlier, RCT data have suggested that routine prophylactic IABP support may offer little meaningful benefit in these patients over the short term. pVAD has been shown to have superior hemodynamic parameters (e.g., mean arterial pressures) as compared with IABP. In those select patients who appear to require additional circulatory support, as defined by the inclusion criteria of completed and ongoing randomized trials, a pVAD may be considered, with the caveats that the inherent increase in delay and invasiveness (particularly access-site complications) may partially offset the benefit and that definitively supportive RCTs remain absent. However, some observational and retrospective data do suggest significant reduction in mortality with pVADs when compared with IABPs in patients undergoing PCI. For patients who develop severe hemodynamic instability, CS, or frank arrest during PCI, bail-out use of the IABP appears beneficial and is certainly the most familiar and rapid strategy; however, the support provided may be insufficient. Cath labs experienced in rapid pVAD placement may opt for these larger, more powerful devices as either an initial or rapidly escalating strategy. Developing data on hemodynamic parameters that might predict meaningful recovery, such as CPO (measured in Watts), would at least theoretically support their use. Although full CPS may be considered in cath labs equipped and staffed for timely initiation, its clinical use had been relegated to anecdotal experience over the past decade until more compact CPS units were recently developed. Consequently, there has been a rise in their utilization in some cath labs. Nevertheless, in emergent settings, pVADs appear most promising and consequently have become increasingly prominent, but they require specialized technical expertise for optimal patient selection, device placement, and postprocedural management. Although two pVADs currently exist, the TandemHeart and the Impella devices, the vast majority of the data and clinical experience have centered around the Impella device, based on the large amount of observational data and subsequent FDA approvals.

Description of Devices and Available Clinical Data