Ischemic cardiomyopathy with depressed left ventricular ejection fraction (LVEF) is predictive of death after percutaneous coronary intervention (PCI), but its association with stent thrombosis (ST) and the need for repeat revascularization is less clearly defined. In total 5,377 patients undergoing PCI were retrospectively evaluated. Multivariable Cox proportional hazards regression and competitive outcome analysis were employed. The primary end point was 1-year major adverse cardiac events (all-cause death, Q-wave myocardial infarction, ST, and target lesion revascularization [TLR]). Individual end points of ST and of TLR were also evaluated. Patients with normal LVEF (>50%) were compared to those with mild (41% to 50%), moderate (25% to 40%), and severe (<25%) decreases in LVEF. Patients with abnormal LVEF were older and more commonly diabetic and had renal insufficiency and heart failure syndrome (p <0.001 for all variables). These patients demonstrated more angiographically complex lesions and less frequently received a drug-eluting stent. The primary end point was significantly increased in patients with lower LVEF (9.7% for normal LVEF vs 20.6% for severely decreased LVEF, p <0.001). ST occurred more frequently in these patients (1.4% for normal LVEF vs 6% for severely decreased LVEF, p <0.001), but clinically driven TLR did not significantly change across LVEF categories. After adjustment, only moderate and severe LVEF decreases (i.e., LVEF ≤40%) demonstrated an association with major adverse cardiac events and with the individual outcome of ST. Subgroup analysis of patients receiving only a drug-eluting stent or a bare-metal stent demonstrated no statistically significant differences for the probability of ST. In conclusion, decreased LVEF is not associated with clinically driven TLR but does increase the risk of ST. Patients with LVEF ≤40% appear to be at significantly higher risk for ST and therefore might benefit from interventional and pharmacologic strategies aimed at minimizing this risk.
Percutaneous coronary intervention (PCI) in patients with decreased left ventricular ejection fraction (LVEF) is followed by increased short-term and long-term mortality, by more nonfatal myocardial infarction, and by a greater need for repeat revascularization. More complex clinical and angiographic features treated in this patient subset may account for the greater incidence of adverse outcomes and have imposed a challenge when attempting to demonstrate a benefit associated with PCI. Coronary stenting in such patients decreases the need for repeat revascularization as presumably does the use of drug-eluting stents. An increased restenosis rate assumes a greater importance in patients with decreased LVEF. In these patients a propensity for in-stent restenosis to present as an acute process with attendant morbidity and mortality has been described. Unfortunately, there may also exist a greater risk of stent thrombosis (ST) in patients with systolic dysfunction. Concern over this possibility and the ongoing worry of increased ST associated with drug-eluting stents raise the possibility that any decrease in restenosis may be offset by an increased risk of ST. Thus, data to guide the interventional cardiologist in a patient with low LVEF are incomplete. This study was undertaken to re-evaluate the relation of decreased LVEF on outcomes after PCI in the drug-eluting stent era. Our particular interest was the relation of LV dysfunction to ST and target lesion revascularization (TLR).
Methods
Clinical, procedural, and follow-up data for consecutive patients undergoing PCI with stenting from January 2000 through December 2009 were retrospectively analyzed from an ongoing registry of catheter-based coronary procedures maintained at our institution. Given the focus of this study on LV function, patients presenting with cardiogenic shock or with acute myocardial infarction were excluded because systolic dysfunction in such settings may be transient.
Preprocedure LVEF was visually estimated from 2-dimensional echocardiography or left ventriculography. The most recent determination, including that made at the time of PCI, was recorded. Patients were divided into 4 categories based on LVEF: normal LVEF (>50%), mild decrease in LVEF (41% to 50%), moderate decrease in LVEF (25% to 40%), and severe decrease in LVEF (<25%).
The institutional review board at Washington Hospital Center and MedStar Health Research Institute (Washington, DC) approved this study. Experienced staff at a dedicated data-coordinating center performed all data collection, entry, and analysis. Data on baseline clinical and procedural data and postprocedure inpatient events were obtained from hospital chart review. Clinical follow-up at 30 days, 6 months, and 1 year was conducted by telephone contact or office visits. Primary source documents for all events were obtained and were adjudicated by physicians not involved in the procedures and who were unaware of the objectives of the study.
PCI was performed according to guidelines current at the time of the procedure. In all cases interventional strategies and choice of periprocedural and discharge medications were at the discretion of the responsible physicians. Anticoagulation regimens included bivalirudin 0.75 mg/kg followed by an infusion of 1.75 mg/kg/hour for the duration of the procedure or unfractionated heparin to achieve an activated clotting time of 200 to 300 seconds in all patients. Glycoprotein IIb/IIIa inhibitors (almost exclusively eptifibatide) were used at the operator’s discretion. All patients received an aspirin loading dose of 325 mg and were prescribed 81 to 325 mg 1 time/day indefinitely. After a clopidogrel loading dose of 300 to 600 mg, patients were prescribed 75 mg 1 time/day for ≥6 months in patients receiving drug-eluting stents and ≥1 month in patients receiving bare-metal stents.
The primary end point was defined as major adverse cardiac events (MACEs), the composite of all-cause mortality, Q-wave myocardial infarction, ST, and TLR at 1-year follow-up. Secondary end points included individual outcomes of ST and TLR at 1-year follow-up.
Academic Research Consortium definitions for ST were used and included definite and probable STs. Early STs were those that occurred within the first 30 days of PCI, and late STs were those events from >30 days to 1 year. Q-wave myocardial infarction was defined as an increase of creatine kinase-MB ≥2 times the upper normal value in the presence of new pathologic Q waves (>0.4 second) in ≥2 contiguous leads on electrocardiogram. TLR was defined as clinically driven revascularization of the index lesion. PCI angiographic success was defined as a residual stenosis <30% with Thrombolysis In Myocardial Infarction grade 3 flow. Clinical success was defined as angiographic success plus absence of TLR, Q-wave myocardial infarction, or death before hospital discharge. Major in-hospital cardiac events were defined as the composite of all-cause death, Q-wave myocardial infarction, urgent TLR, and urgent coronary artery bypass surgery before hospital discharge. Major bleeding was defined as a decrease in hemoglobin level >5 g/dl (or ≥15% in hematocrit level) or intracranial hemorrhage using Thrombolysis In Myocardial Infarction criteria. Vascular complications were defined as the presence of a large hematoma (≥5 cm), fistula, or pseudoaneurysm formation, retroperitoneal bleeding, or need for surgical repair.
Statistical analysis was performed using SAS 9.1 (SAS Institute, Cary, North Carolina). Normally distributed continuous variables are presented as mean ± SD. Those that were not normally distributed are shown as median ± interquartile range. Categorical variables are expressed as frequency and percentage. Baseline characteristics were compared using Student’s t test for parametric variables or Mann–Whitney U test when not normally distributed. Categorical variables were compared using chi-square test or Fisher’s exact test as appropriate. One-year outcomes were compared to log-rank test and are presented by Kaplan–Meier percentages.
To evaluate the association of LVEF with 1-year MACE, TLR, and all STs, multivariable Cox regression models were constructed. Covariables for the models were selected based on significant univariable p values and overall clinical relevance. Covariables included LVEF category (mild decrease in LVEF, moderate decrease in LVEF, and severe decrease in LVEF, with normal LVEF as reference), age (per 5-year increase), male gender, African-American race, history of diabetes mellitus, history of chronic renal insufficiency, number of stents implanted, receipt of ≥1 drug-eluting stent, and use of intravascular ultrasound. Only patients with data present for all covariables were included in the final models. A p value <0.05 was considered statistically significant.
To evaluate if death before occurrence of TLR confounded the result for this end point, a competing outcome analysis was performed. Subgroup analysis for patients receiving a drug-eluting stent or a bare-metal stent only was performed for the entire cohort, for patients with at least moderate decrease in LVEF (LVEF ≤40%), and for those with severe systolic dysfunction (LVEF <25%).
Results
In total 6,382 patients underwent PCI with stenting and had LVEF data available; of these 281 were excluded for presentation with cardiogenic shock and 764 patients for presentation with acute myocardial infarction. The study population (n = 5,377) had an average age of 66 ± 11.5 years and was predominantly men (66%). Seventy percent of patients were European-American and 21% were African-American.
In the entire cohort, a normal LVEF (estimated LVEF >50%) was identified in 3,443 patients (64.5%), and 955 patients (17.9%) had mild LVEF decrease (estimated LVEF 41% to 50%), 686 patients (12.9%) had moderate LVEF decrease (estimated LVEF 25% to 40%), and 253 patients (4.7%) had severe LVEF decrease (estimated LVEF <25%).
Table 1 lists baseline characteristics of the study population. As LVEF decreased, patients had more co-morbidity and associated risk factors. Patients with moderate and severe decreases in LVEF were older, more commonly men, more frequently had a history of diabetes mellitus, chronic renal insufficiency, previous myocardial infarction, previous coronary artery bypass surgery, and congestive heart failure (p <0.001 for all variables). Presentation with unstable angina pectoris was less likely in patients with moderate and severe LVEF decreases (p <0.001). Use of β blockers and angiotensin-converting enzyme inhibitors was more frequent in patients with moderate and severe LVEF decreases (p <0.001).
| Left Ventricular Ejection Fraction | p Value | ||||
|---|---|---|---|---|---|
| Normal | Mild Decrease | Moderate Decrease | Severe Decrease | ||
| (n = 3,443) | (n = 955) | (n = 686) | (n = 253) | ||
| Age (years) | 64.97 ± 11.35 | 66.59 ± 11.69 | 68.29 ± 11.16 | 68.62 ± 11.72 | <0.001 |
| Men | 2,182 (63.4%) | 657 (68.8%) | 484 (70.6%) | 191 (75.8%) | <0.001 |
| European-American | 2,436 (70.8%) | 683 (71.5%) | 468 (68.2%) | 164 (65.1%) | 0.126 |
| African-American | 693 (20.1%) | 207 (21.7%) | 151 (22.0%) | 72 (28.6%) | 0.012 |
| Current smoker | 575 (16.7%) | 177 (18.5%) | 120 (17.5%) | 56 (22.2%) | 0.106 |
| Unstable angina pectoris at presentation | 1,967 (57.3%) | 484 (50.9%) | 332 (48.7%) | 104 (41.4%) | <0.001 |
| History of diabetes mellitus | 1,156 (33.9%) | 368 (38.9%) | 303 (44.6%) | 115 (46.0%) | <0.001 |
| History of systemic hypertension ⁎ | 2,990 (84.7%) | 812 (85.4%) | 592 (86.4%) | 221 (87.7%) | 0.424 |
| Previous chronic renal insufficiency † | 348 (10.2%) | 142 (15.1%) | 174 (25.7%) | 70 (28.1%) | <0.001 |
| Creatinine clearance (ml/min) | 72.84 ± 51.81 | 69.26 ± 27.63 | 63.28 ± 28.70 | 59.83 ± 27.17 | <0.001 |
| Previous myocardial infarction | 650 (19.9%) | 296 (33.7%) | 249 (39.8%) | 115 (50.0%) | <0.001 |
| Previous coronary artery bypass graft | 723 (21.2%) | 309 (32.6%) | 274 (40.5%) | 103 (41.0%) | <0.001 |
| Previous congestive heart failure | 386 (11.7%) | 191 (21.0%) | 281 (43.0%) | 144 (61.0%) | <0.001 |
| Left ventricular ejection fraction | 0.57 ± 0.09 | 0.43 ± 0.04 | 0.31 ± 0.04 | 0.18 ± 0.04 | <0.001 |
| Functional class III or IV ‡ | 131 (4.0%) | 80 (8.8%) | 159 (24.3%) | 96 (40.7%) | <0.001 |
| Angiotensin-converting enzyme inhibitor | 1,438 (42.5%) | 517 (55.4%) | 408 (62.0%) | 155 (65.7%) | <0.001 |
| β Blocker | 2,543 (78.9%) | 730 (81.9%) | 522 (84.9%) | 193 (85.4%) | <0.001 |
| Statin | 2,663 (83.2%) | 752 (85.3%) | 495 (83.6%) | 183 (82.1%) | 0.484 |
⁎ History of systemic hypertension diagnosed and/or treated with medication or currently being treated with diet and/or medication by a physician.
† Previously diagnosed or treated with medication, diet, or dialysis by a physician. Diagnosis at admission if a baseline creatinine level >2.0 mg/dl is present or dialysis-dependent renal insufficiency is present.
‡ Based on New York Heart Association scale and refers to the 6 weeks before admission for index percutaneous coronary intervention.
Angiographic and procedural characteristics are presented in Table 2 . Patients with worse LVEF had more diseased coronary vessels (p <0.001). However, the number of implanted stents did not differ among groups. In patients with severely decreased LVEF, bivalirudin was used less frequently (p <0.001), whereas glycoprotein IIb/IIIa inhibitors were used more frequently (p <0.001). In addition, as LVEF decreased, the index target vessel was less commonly the left anterior descending coronary artery (p <0.001), a drug-eluting stent was implanted less frequently (p <0.001), and intravascular ultrasound was less commonly used (p <0.001). Patients with moderately decreased LVEF demonstrated more angiographic complexity because index lesions were more frequently type C according to the American College of Cardiology/American Heart Association classification (p <0.001) and angiography was more commonly performed on a restenotic lesion (p = 0.04) or on a saphenous vein graft (p <0.001). Angiographic success was similar across all evaluated categories of LVEF (p = 0.90).
| Left Ventricular Ejection Fraction | p Value | ||||
|---|---|---|---|---|---|
| Normal | Mild Decrease | Moderate Decrease | Severe Decrease | ||
| (n = 3,443) | (n = 955) | (n = 686) | (n = 253) | ||
| Procedural characteristics | |||||
| Number of diseased vessels | 1.82 ± 0.85 | 2.05 ± 0.87 | 2.22 ± 0.86 | 2.26 ± 0.87 | <0.001 |
| Number of implanted stents | 1.46 ± 0.86 | 1.46 ± 0.87 | 1.44 ± 0.91 | 1.50 ± 1.00 | 0.874 |
| Bivalirudin | 2,491 (72.3%) | 651 (68.2%) | 402 (58.6%) | 79 (64.7%) | <0.001 |
| Glycoprotein IIb/IIIa inhibitor | 378 (11.1%) | 216 (14.3%) | 114 (16.8%) | 38 (15.4%) | <0.001 |
| Angiographic characteristics (lesion based) | |||||
| Angiographic success | 5,681/5,811 (97.8%) | 1,645/1,685 (97.6%) | 1,155/1,182 (97.7%) | 442/450 (98.2%) | 0.9 |
| Left anterior descending coronary artery percutaneous coronary intervention | 2,172/5,877 (37.0%) | 587/1,721 (34.1%) | 361/1,195 (30.2%) | 140/457 (30.6%) | <0.001 |
| Saphenous vein graft percutaneous coronary intervention | 358/5,877 (6.1%) | 191/1,721 (11.1%) | 223/1,195 (18.7%) | 73/457 (16.0%) | <0.001 |
| In-stent restenosis | 459/5,882 (7.8%) | 140/1,721 (8.1%) | 101/1,198 (8.4%) | 20/456 (4.4%) | 0.04 |
| Bare-metal stent implantation | 1,081/5,261 (20.5%) | 453/1,531 (29.6%) | 371/1,066 (34.8%) | 145/422 (34.4%) | <0.001 |
| Drug-eluting stent implantation | 4,009/5,030 (79.7%) | 1,009/1,364 (74.0%) | 640/869 (73.6%) | 264/372 (71.0%) | <0.001 |
| Intravascular ultrasound performed | 3,866/5,744 (67.3%) | 1,013/1,685 (60.1%) | 730/1,178 (62.0%) | 249/439 (56.7%) | <0.001 |
| American College of Cardiology/American Heart Association type C lesion | 1,459/5,582 (26.1%) | 498/1,663 (29.9%) | 349/1,141 (30.6%) | 133/446 (29.8%) | <0.001 |
In-hospital outcomes are presented in Table 3 . Patients with lower LVEF had a higher incidence of major in-hospital cardiac events (p <0.001) and all-cause mortality, ST, myocardial infarction, and postprocedure acute renal failure (p <0.001 for all). They also more frequently required emergency insertion of an intra-aortic balloon pump (p <0.001). Consequently, patients with worse LVEF had less clinical success on discharge (p <0.001). Patients with a severe decrease in LVEF more frequently required urgent in-hospital TLR (p = 0.008) and demonstrated slightly more frequent major bleeding events (p = 0.018) and vascular complications (p = 0.005).
| Left Ventricular Ejection Fraction | p Value | ||||
|---|---|---|---|---|---|
| Normal | Mild Decrease | Moderate Decrease | Severe Decrease | ||
| (n = 3,443) | (n = 955) | (n = 686) | (n = 253) | ||
| Major in-hospital cardiac events ⁎ | 43 (1.2%) | 19 (2.0%) | 24 (3.5%) | 19 (7.5%) | <0.001 |
| Death | 24 (0.7%) | 9 (0.9%) | 18 (2.6%) | 12 (4.7%) | <0.001 |
| Q-wave myocardial infarction | 14 (0.4%) | 5 (0.5%) | 6 (0.9%) | 6 (2.4%) | <0.001 |
| Stent thrombosis | 2 (0.1%) | 2 (0.2%) | 1 (0.1%) | 4 (1.6%) | <0.001 |
| Urgent target lesion revascularization | 10 (0.3%) | 8 (0.8%) | 3 (0.4%) | 4 (1.6%) | 0.008 |
| Clinical success | 3,392 (98.5%) | 930 (97.4%) | 662 (96.5%) | 234 (92.5) | <0.001 |
| Postprocedure acute renal failure | 2 (0.1%) | 2 (0.2%) | 1 (0.1%) | 4 (1.6%) | <0.001 |
| Emergency intra-aortic balloon pump | 33 (1.0%) | 13 (1.4%) | 35 (5.1%) | 38 (15.1%) | <0.001 |
| In-hospital major bleeding | 48 (1.4%) | 25 (2.6%) | 15 (2.2%) | 8 (3.2%) | 0.018 |
| Vascular complications | 77 (3.9%) | 37 (3.9%) | 26 (3.8%) | 11 (4.4%) | 0.005 |
⁎ In-hospital all-cause death, Q-wave myocardial infarction, urgent target lesion revascularization, and urgent coronary artery bypass surgery.
Table 4 presents 1-year unadjusted outcomes. The primary end point of this study, MACEs at 1 year, was 2 times that in patients in the lowest LVEF stratum compared to those classified as having normal LVEF (p <0.001). It can be seen from the Kaplan–Meier curves that rates of patients with moderate versus those with severe LV dysfunction were inseparable ( Figure 1 ).
| Left Ventricular Ejection Fraction | p Value | ||||
|---|---|---|---|---|---|
| Normal | Mild Decrease | Moderate Decrease | Severe Decrease | ||
| (n = 3,443) | (n = 955) | (n = 686) | (n = 253) | ||
| Major adverse cardiac events ⁎ | 333 (9.7%) | 115 (12.1%) | 137 (20.1%) | 52 (20.6%) | <0.001 |
| Death | 129 (3.8%) | 55 (5.8%) | 94 (13.8%) | 43 (17.1%) | <0.001 |
| Q-wave myocardial infarction | 5 (0.1%) | 2 (0.2%) | 3 (0.5%) | 3 (1.3%) | <0.001 |
| Target lesion revascularization | 210 (6.2%) | 60 (6.5%) | 44 (6.9%) | 13 (5.6%) | 0.886 |
| Target vessel revascularization | 305 (9.0%) | 86 (9.2%) | 69 (10.7%) | 15 (6.5%) | 0.258 |
| Stent thrombosis (definite/probable) | 47 (1.4%) | 21 (2.2%) | 25 (3.6%) | 15 (6.0%) | <0.001 |
| Cumulative definite stent thrombosis | 13 (0.4%) | 8 (0.8%) | 4 (0.6%) | 6 (2.4%) | <0.001 |
| Early definite stent thrombosis | 10 (0.3%) | 5 (0.5%) | 4 (0.6%) | 5 (2.0%) | <0.001 |
| Late definite stent thrombosis | 3 (0.1%) | 3 (0.3%) | 0 | 1 (0.4%) | 0.158 |
| Cumulative probable stent thrombosis | 34 (1.0%) | 13 (1.4%) | 21 (3.1%) | 9 (3.6%) | <0.001 |
⁎ Death, Q-wave myocardial infarction, stent thrombosis, and target lesion revascularization.
Similarly, compared to patients with normal LVEF, those in the lowest stratum had an all-cause mortality rate nearly 5 times higher (p <0.001; Table 4 ). Importantly, although unadjusted mortality increased as LVEF decreased, rates were statistically nondifferent between those with moderate and those with severe LV dysfunction ( Figure 2 ).
