Highlights
-
•
There is currently no clinical consensus on standardized approach to risk stratification in patients with complex coronary anatomy and concomitant high bleeding risk undergoing PCI.
-
•
Complex PCI is associated with a higher risk of both ischemic and bleeding complications at 1-year, irrespective of baseline bleeding risk status.
-
•
Comprehensive baseline risk assessment may enhance clinical decision making and facilitate early implementation of preventative strategies to better identify most vulnerable patients at higher risk of adverse events.
The impact of complex coronary anatomy on outcomes in patients at high bleeding risk (HBR) undergoing percutaneous coronary intervention (PCI) remains unclear. Patients from a large PCI referral center were labeled as HBR in presence of ≥1 major or ≥2 minor Academic Research Consortium HBR criteria and undergoing complex PCI if they had any of the following: ≥3 vessels treated, ≥3 stents implanted, ≥3 lesions treated, stent length >60 mm, bifurcation lesion with ≥2 stents, PCI of chronic total occlusion, left main, saphenous venous graft or requiring the use of atherectomy. Primary outcomes at 1 year were major adverse cardiac event (MACE), a composite of all-cause death, myocardial infarction or target vessel revascularization, and major bleeding. Of 16,966 patients, 7,295 (43.0%) were HBR with 3,611 (49.5%) undergoing complex PCI and 9,671 (57.0%) were non-HBR with 4,103 (42.4%) undergoing complex PCI. At 1 year, patients with complex PCI had a higher risk of MACE in both HBR (adjusted Hazard Ratio [adj. HR] 1.78, 95% Confidence Interval [95% CI] 1.54 to 2.06) and non-HBR (adj. HR 1.82 95% CI 1.55 to 2.12, respectively, p int 0.64) groups and a higher risk of major bleeding (adj. HR 1.55, 95% CI 1.29 to 1.87 and adj. HR 1.36, 95% CI 1.03 to 1.78, respectively; p int = 0.46), as compared to patients undergoing noncomplex PCI. In conclusion, complex PCI was associated with significantly higher risk of ischemic and bleeding adverse outcomes at 1-year in both HBR and non-HBR patients.
The proportion of patients with high-risk coronary artery disease (CAD) undergoing percutaneous coronary intervention (PCI) is increasing annually. , Advances in stent morphology, such as thinner strut architecture and highly biocompatible antiproliferative drug films, along with new strategies for antiplatelet therapy have markedly reduced the incidence of adverse events in these patients. ,, However, lesions with complex coronary anatomy and technically challenging procedures continue to be associated with an elevated rate of recurrent ischemic events. , Concurrently, clinical comorbidities associated with high bleeding risk (HBR) have emerged as major determinants of adverse outcomes, with prognostic significance comparable to that of ischemic risk factors. , The Academic Research Consortium on High Bleeding Risk (ARC HBR) was the pioneering effort to provide a standardized definition for HBR, which has been validated in several external cohorts. ,, In contrast, a universally accepted definition for complex PCI is lacking and prior studies have utilized varying combinations of anatomical and angiographic characteristics to identify high ischemic risk patients. ,,,,,,,, This definitional variability has limited the ability to draw consistent conclusions regarding the relationship between PCI complexity and bleeding risk, since heterogeneous classifications impede valid cross-study comparisons and interpretation of clinical outcomes. To address this gap, the current study utilized a comprehensive complex PCI definition by Dangas et al. along with the ARC HBR criteria in a bidirectional stratification to assess the joint prognostic implications of elevated baseline ischemic and bleeding risk in patients undergoing PCI on their 1-year clinical outcomes.
Methods
Study patients
All consecutive patients undergoing PCI with a drug-eluting stent at a tertiary-care high-volume center (Mount Sinai Hospital, New York, NY) between January 01, 2012, and December 31, 2019, were included in the current study. Patients with missing data on complex PCI characteristics or ARC HBR criteria and patients without at least 30-day follow-up were excluded from the final database. Details on data collection and study definitions have been previously reported. Briefly, data from the index procedure up to 1 year of follow-up was prospectively entered by dedicated research personnel into a database. Baseline and procedural characteristics were extracted from electronic health records. Patients were contacted by phone calls at 30 days and 1 year after index PCI by trained research personnel to collect information on adverse events. The study was approved by the Institutional Review Board and proceedings complied with the principles of the Declaration of Helsinki.
Specifications of the ARC HBR criteria and their adaptation to the current database are reported in the Supplementary Table S1 . Patients were categorized as HBR if they met at least 1 major or 2 minor ARC HBR criteria. Patients without any of the ARC HBR criteria or with only 1 minor criterion were categorized as non-HBR. Patients were considered to be at high ischemic risk and have undergone a complex PCI procedure if they met at least one of the following: ≥3 vessels treated, ≥3 lesions treated, ≥3 stents implanted, bifurcation lesion treated with ≥2 stents, total stent length >60 mm, PCI of chronic total occlusion (CTO), left main, saphenous vein graft (SVG) or requiring the use of atherectomy. Patients without any of these were considered to have undergone noncomplex PCI.
Study outcomes
The primary ischemic outcome was major adverse cardiac event (MACE), a composite of all-cause death, myocardial infarction (MI), or target vessel revascularization (TVR) at 1 year after index PCI. Additional ischemic endpoints were the individual components of MACE, target lesion revascularization (TLR) and definite or probable stent thrombosis (def/prob ST) at 1 year. The primary bleeding outcome was major bleeding, composite of in-hospital bleeding and postdischarge bleeding events up to 1 year, the definition of which is similar to Bleeding Academic Research Consortium (BARC) 3 or 5 ( Supplementary Table S2 ). In-hospital bleeding was defined as any bleeding causing hemoglobin drop ≥3 g/dl, requiring blood transfusion, or procedural intervention at the bleeding site, and postdischarge bleeding was defined as any bleeding event requiring hospitalization or blood transfusion, according to the National Cardiovascular Data Registry CathPCI Registry (version 4.4). Subgroup analyses according to clinical presentation were conducted, with patients stratified into acute coronary syndrome (ACS) and chronic coronary syndrome (CCS).
Statistical analyses
Continuous variables were expressed as means ± standard deviations and analyzed using the t-test, while categorical variables were expressed as numbers with frequencies and analyzed using the chi-square test. UpSet plots were used to illustrate the most frequent combinations of HBR criteria for patients undergoing complex and noncomplex PCI. The Kaplan–Meier method was used to estimate 1-year incidence of adverse events, and log-rank test was used for comparisons between groups. A Cox regression model was used to calculate the unadjusted and adjusted risks associated with complex PCI status and 1-year events in HBR and non-HBR patients, with risk estimates expressed as hazard ratio (HR) and 95% Confidence Interval (CI). Covariates for the multivariable models were prespecified based on pathophysiological relevance, previously demonstrated associations with post-PCI outcomes, and clinically meaningful baseline differences between patient groups. Highly collinear variables were not entered simultaneously but clinically essential covariates with established prognostic relevance were retained a priori in the model irrespective of statistical screening to ensure appropriate clinical adjustment. The final multivariable Cox-regression model was adjusted for gender, race, current smoker, hypertension (HTN), diabetes mellitus (DM), prior MI, family history of CAD, ACS presentation, and radial versus femoral access. Multivariable Cox regression models for outcomes in subgroups by clinical presentation were adjusted for gender, race, current smoking status, HTN, DM, prior MI, family history of CAD, and radial versus femoral access. Formal interaction testing was conducted for comparison of outcomes according to HBR status and presence of complex PCI characteristics. All p-values were generated from 2-sided probability testing, and values <0.05 were considered statistically significant. STATA software (version 16; StataCorp, College Station, Texas) and R software (version 4.3.3; R Foundation for Statistical Computing, Vienna, Austria) were utilized for performing the analyses. Supplementary Methods and Supplementary Figure 1 provide additional details regarding the multivariable models used in this study.
Results
Patient characteristics
A total of 16,966 patients were included, of which 7,295 (43.0%) patients were at HBR with 3,611 (49.5%) undergoing complex PCI. The remainder of 9,671 (57.0%) patients were non-HBR with 4,103 (42.4%) undergoing complex PCI ( Supplementary Figure 2 ). The combination of age ≥75 years and mild anemia were the most frequent HBR-qualifying criteria for patients undergoing complex PCI, while the combination of mild chronic kidney disease and mild anemia were the most frequent HBR-qualifying criteria in patients undergoing noncomplex PCI ( Figure 1 ). Moderate/severe anemia was the second most frequent HBR-qualifying criterion for both groups. Use of atherectomy was the most frequent complex PCI characteristic in HBR patients (23.9%) while total stent length >60 mm was the most frequent complex PCI characteristic in non-HBR patients (19.4%) ( Figure 2 ). HBR patients were more likely to have ≥3 vessels treated, PCI of left main and SVG, but less likely to have PCI of CTO. There were no differences in the number of patients receiving ≥3 stents, ≥3 lesions treated or bifurcation lesions treated with ≥2 stents between the 2 groups.
Prevalence of the 40 most frequent HBR-qualifying criteria in patients with complex PCI (A) and in patients with noncomplex PCI (B). UpSet Plots were utilized to demonstrate the prevalence of overlapping HBR criteria. Dots denote the presence of ARC HBR major or minor criterion; lines connecting the dots denote a combination of risk factors. Patients were considered to be at HBR if they had at least 1 major or 2 minor ARC HBR criterion. Prevalence of individual major and minor ARC HBR criterion is denoted in horizontal bar graphs on the left side. ARC = Academic Research Consortium; CKD = chronic kidney disease; CVA = cerebrovascular accident; GI = gastrointestinal; HBR = high bleeding risk; PCI = percutaneous coronary intervention.
Prevalence and distribution of the complex PCI characteristics among HBR and non-HBR patients. HBR: high bleeding risk; PCI: percutaneous coronary intervention.
Patients undergoing complex PCI in both HBR and non-HBR groups were less likely to be female and to have DM but had a higher prevalence of current smoking, peripheral arterial disease, prior MI and prior coronary artery bypass graft ( Table 1 ). The majority of patients were on guideline-directed medical therapy for CAD, including aspirin and lipid-lowering agents. Patients with complex PCI in both groups were less likely to present with unstable angina, but only patients with complex PCI in non-HBR group were more likely to present with stable angina. Patients with complex PCI required use of mechanical support devices such as impella and intra-aortic balloon pump more often in both HBR and non-HBR group. Complex PCI patients had a higher risk of access site bleeding in the HBR group but there was no difference in the non-HBR group. At discharge, dual antiplatelet therapy (DAPT) was prescribed in most patients, with slightly lower rates in patients undergoing complex PCI in HBR group (89.7% vs 92.3%) than in the non-HBR group (98.6% vs 98.1%). In the HBR group, patients undergoing complex PCI were more likely to have concomitant oral anticoagulants (OAC) prescription as compared to patients undergoing noncomplex PCI (19.6% vs 16.8%) ( Table 2 ).
Table 1
Baseline demographic and clinical characteristics
| HBR (N = 7,295) | Non-HBR (N = 9,671) | |||||
|---|---|---|---|---|---|---|
| Characteristics | Complex PCI (N = 3,611) | Noncomplex PCI (N = 3,684) | p-value | Complex PCI (N = 4,103) | Noncomplex PCI (N = 5,568) | p-value |
| Age, y | 72.6 ± 10.8 | 70.4 ± 11.3 | <0.001 | 62.9 ± 10.0 | 61.5 ± 10.0 | <0.001 |
| BMI, kg/m 2 | 28.1 ± 5.5 | 29.0 ± 6.0 | <0.001 | 28.8 ± 5.5 | 29.2 ± 5.5 | <0.001 |
| Female | 1,172 (32.5%) | 1,502 (40.8%) | <0.001 | 800 (19.5%) | 1,405 (25.2%) | <0.001 |
| Race | <0.001 | <0.001 | ||||
| Caucasian | 1,939 (57.4%) | 1,496 (41.3%) | 1,949 (51.6%) | 2,146 (39.5%) | ||
| African American | 357 (10.6%) | 523 (14.4%) | 265 (7.0%) | 561 (10.3%) | ||
| Asian | 393 (11.6%) | 496 (13.7%) | 722 (19.1%) | 1,178 (21.7%) | ||
| Hispanic | 550 (16.3%) | 886 (24.4%) | 642 (17.0%) | 1,156 (21.3%) | ||
| Other | 137 (4.1%) | 225 (6.2%) | 198 (5.2%) | 396 (7.3%) | ||
| Medical history | ||||||
| Current smoker | 735 (20.4%) | 439 (11.9%) | <0.001 | 1,054 (25.7%) | 1,161 (20.9%) | <0.001 |
| Family history of CAD | 662 (18.3%) | 749 (20.3%) | 0.031 | 1,140 (27.8%) | 1,479 (26.6%) | 0.183 |
| Anemia | 2,501 (69.3%) | 2,572 (69.8%) | 0.619 | 752 (18.3%) | 997 (17.9%) | 0.594 |
| Diabetes Mellitus | 1,894 (52.5%) | 2,069 (56.2%) | 0.001 | 1,699 (41.4%) | 2,463 (44.2%) | 0.005 |
| Insulin dependent | 813 (42.9%) | 838 (40.5%) | 0.122 | 526 (31.0%) | 670 (27.2%) | 0.008 |
| Hypertension | 3,392 (93.9%) | 3,509 (95.2%) | 0.013 | 3,691 (90.0%) | 4,920 (88.4%) | 0.014 |
| Hyperlipidemia | 3,263 (90.4%) | 3,400 (92.3%) | 0.003 | 3,742 (91.2%) | 4,990 (89.6%) | 0.010 |
| Atrial fibrillation | 648 (17.9%) | 544 (14.8%) | <0.001 | 85 (2.1%) | 105 (1.9%) | 0.516 |
| Cerebrovascular disease | 7 (0.2%) | 23 (0.7%) | 0.003 | 6 (0.2%) | 12 (0.2%) | 0.404 |
| Chronic kidney disease | 1,913 (53.0%) | 2,006 (54.5%) | 0.207 | 308 (7.5%) | 474 (8.5%) | 0.073 |
| Dialysis | 345 (18.0%) | 269 (13.4%) | <0.001 | 3 (1.0%) | 5 (1.1%) | 1.000 |
| Peripheral arterial disease | 532 (14.7%) | 454 (12.3%) | 0.003 | 274 (6.7%) | 285 (5.1%) | 0.001 |
| Prior PCI | 1,482 (41.0%) | 1,547 (42.0%) | 0.410 | 1,606 (39.1%) | 2,021 (36.3%) | 0.004 |
| Prior MI | 971 (26.9%) | 789 (21.4%) | <0.001 | 914 (22.3%) | 1,058 (19.0%) | <0.001 |
| Prior CABG | 987 (27.3%) | 556 (15.1%) | <0.001 | 775 (18.9%) | 452 (8.1%) | <0.001 |
| LVEF, % | 51.8 ± 12.8 | 54.1 ± 11.4 | <0.001 | 54.5 ± 10.7 | 55.7 ± 9.3 | <0.001 |
| Baseline medications | ||||||
| Aspirin | 2,953 (81.8%) | 3,004 (81.5%) | 0.794 | 3,603 (87.8%) | 4,778 (85.8%) | 0.004 |
| Lipid-lowering medications | 2,959 (81.9%) | 2,915 (79.1%) | 0.002 | 3,346 (81.6%) | 4,204 (75.5%) | <0.001 |
| Beta-blockers | 2,701 (74.8%) | 2,723 (73.9%) | 0.387 | 2,901 (70.7%) | 3,747 (67.3%) | <0.001 |
| CCBs | 1,415 (39.2%) | 1,518 (41.2%) | 0.079 | 1,225 (29.9%) | 1,960 (35.2%) | <0.001 |
| ACEI/ARB | 2,085 (57.7%) | 2,166 (58.8%) | 0.361 | 2,313 (56.4%) | 2,968 (53.3%) | 0.003 |
ACEI/ARBs = Angiotensin-Converting Enzyme Inhibitors/Angiotensin Receptor Blockers; BMI: body mass index; CABG = coronary artery bypass graft; CAD = coronary artery disease; CCBs = Calcium Channel Blockers; HBR = high bleeding risk; LVEF = left ventricular ejection fraction; MI = myocardial infarction; PCI = percutaneous coronary intervention.
Table 2
Procedural angiographic characteristics
| HBR (N = 7,295) | Non-HBR (N = 9,671) | |||||
|---|---|---|---|---|---|---|
| Complex PCI (N = 3,611) | Noncomplex PCI (N = 3,684) | p-value | Complex PCI (N = 4,103) | Noncomplex PCI (N = 5,568) | p-value | |
| PCI presentation | ||||||
| Asymptomatic | 203 (5.6%) | 155 (4.2%) | 0.005 | 223 (5.4%) | 205 (3.7%) | <0.001 |
| Stable angina | 1,805 (50.0%) | 1,862 (50.5%) | 0.634 | 2,267 (55.3%) | 2,838 (51.0%) | <0.001 |
| Unstable angina | 984 (27.3%) | 1,130 (30.7%) | 0.001 | 1,103 (26.9%) | 1,693 (30.4%) | <0.001 |
| NSTEMI | 502 (13.9%) | 408 (11.1%) | <0.001 | 397 (9.7%) | 595 (10.7%) | 0.106 |
| STEMI | 65 (1.8%) | 104 (2.8%) | 0.004 | 97 (2.4%) | 228 (4.1%) | <0.001 |
| Access | ||||||
| Radial access | 578 (16.0%) | 589 (16.0%) | 0.983 | 831 (20.3%) | 1,128 (20.3%) | 0.995 |
| Femoral access | 2,944 (81.5%) | 3,040 (82.5%) | 0.271 | 3,183 (77.6%) | 4,378 (78.6%) | 0.216 |
| PCI vessels | ||||||
| Left main | 607 (16.8%) | 0 (0.0%) | NA | 399 (9.7%) | 0 (0.0%) | NA |
| LAD | 2,227 (61.7%) | 1,700 (46.1%) | <0.001 | 2,567 (62.6%) | 2,694 (48.4%) | <0.001 |
| LCX | 1,685 (46.7%) | 1,219 (33.1%) | <0.001 | 1,807 (44.0%) | 1,759 (31.6%) | <0.001 |
| RCA | 1,532 (42.4%) | 1,081 (29.3%) | <0.001 | 1,794 (43.7%) | 1,560 (28.0%) | <0.001 |
| Multivessel disease | 2,878 (79.7%) | 2,232 (60.6%) | <0.001 | 3,008 (73.3%) | 3,055 (54.9%) | <0.001 |
| Lesions and stents | ||||||
| B2/C lesion | 3,229 (89.4%) | 2,235 (60.7%) | <0.001 | 3,683 (89.8%) | 3,364 (60.4%) | <0.001 |
| Calcification (moderate/severe) | 1,716 (47.7%) | 453 (12.4%) | <0.001 | 1,338 (32.7%) | 381 (6.9%) | <0.001 |
| Bifurcation | 987 (27.3%) | 377 (10.2%) | <0.001 | 1,235 (30.1%) | 705 (12.7%) | <0.001 |
| Chronic total occlusion | 430 (11.9%) | 0 (0.0%) | <0.001 | 827 (20.2%) | 0 (0.0%) | <0.001 |
| Total lesion length, mm | 40.8 ± 24.5 | 20.8 ± 11.5 | <0.001 | 43.6 ± 24.1 | 21.2 ± 11.1 | <0.001 |
| Total stent length, mm | 59.3 ± 43.9 | 25.4 ± 11.9 | <0.001 | 61.6 ± 40.7 | 25.9 ± 11.7 | <0.001 |
| No. of stents implanted | 1.8 ± 0.8 | 1.2 ± 0.4 | <0.001 | 1.8 ± 0.8 | 1.2 ± 0.4 | <0.001 |
| No. of lesions treated | 1.8 ± 0.8 | 1.2 ± 0.4 | <0.001 | 1.8 ± 0.8 | 1.2 ± 0.4 | <0.001 |
| Maximum stent diameter, mm | 3.3 ± 0.5 | 3.0 ± 0.5 | <0.001 | 3.3 ± 0.5 | 3.1 ± 0.5 | <0.001 |
| Use of imaging and devices | ||||||
| Intravascular ultrasonography | 209 (5.8%) | 344 (9.3%) | <0.001 | 200 (4.9%) | 514 (9.2%) | <0.001 |
| Optical coherence tomography | 76 (2.1%) | 99 (2.7%) | 0.104 | 104 (2.5%) | 277 (5.0%) | <0.001 |
| Orbital atherectomy | 236 (6.5%) | 0 (0.0%) | <0.001 | 208 (5.1%) | 0 (0.0%) | <0.001 |
| Rotational atherectomy | 1,519 (42.1%) | 0 (0.0%) | <0.001 | 1,155 (28.2%) | 0 (0.0%) | <0.001 |
| Impella | 187 (5.2%) | 17 (0.5%) | <0.001 | 112 (2.7%) | 13 (0.2%) | <0.001 |
| Intra-aortic balloon pump | 119 (3.3%) | 67 (1.8%) | <0.001 | 92 (2.2%) | 76 (1.4%) | 0.001 |
| Discharge medication | ||||||
| DAPT | 3,223 (89.7%) | 3,405 (92.5%) | <0.001 | 4,034 (98.6%) | 5,458 (98.1%) | 0.076 |
| Aspirin | 3,232 (89.9%) | 3,414 (92.7%) | <0.001 | 4,045 (98.9%) | 5,476 (98.4%) | 0.085 |
| P2Y12-inhibitors | 3,564 (99.2%) | 3,652 (99.2%) | 0.974 | 4,075 (99.6%) | 5,535 (99.5%) | 0.531 |
| Clopidogrel | 2,895 (80.6%) | 3,105 (84.3%) | <0.001 | 2,526 (61.7%) | 3,936 (70.8%) | <0.001 |
| Prasugrel | 210 (5.8%) | 172 (4.7%) | 0.025 | 583 (14.2%) | 603 (10.8%) | <0.001 |
| Ticagrelor | 474 (13.2%) | 381 (10.3%) | <0.001 | 983 (24.0%) | 1,010 (18.2%) | <0.001 |
| Oral anticoagulants | 700 (19.6%) | 614 (16.8%) | 0.002 | 0 (0.0%) | 0 (0.0%) | N/A |
| Dabigatran | 39 (1.1%) | 54 (1.5%) | 0.147 | 0 (0.0%) | 0 (0.0%) | N/A |
| Rivaroxaban | 158 (4.4%) | 146 (4.0%) | 0.360 | 0 (0.0%) | 0 (0.0%) | N/A |
| Apixaban | 194 (5.4%) | 112 (3.1%) | <0.001 | 0 (0.0%) | 0 (0.0%) | N/A |
| Warfarin | 312 (8.7%) | 305 (8.3%) | 0.546 | 0 (0.0%) | 0 (0.0%) | N/A |
| Statin | 3,393 (94.4%) | 3,444 (93.5%) | 0.109 | 3,951 (96.6%) | 5,303 (95.3%) | 0.003 |
| Beta blocker | 2,932 (81.6%) | 2,987 (81.1%) | 0.601 | 3,365 (82.2%) | 4,484 (80.6%) | 0.042 |
| Access site bleeding | 19 (0.5%) | 4 (0.1%) | 0.001 | 6 (0.1%) | 9 (0.2%) | 0.849 |
| Technical success | 3,424 (94.8%) | 3,623 (98.3%) | <0.001 | 3,874 (94.4%) | 5,502 (98.8%) | <0.001 |
| Procedural success * | 3,009 (83.3%) | 3,487 (94.7%) | <0.001 | 3,509 (85.5%) | 5,358 (96.2%) | <0.001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
