Despite increasing complexity of contemporary procedures at tertiary care hospitals, the relationship between interventional cardiology fellows-in-training (ICFITs) and complications of percutaneous coronary intervention (PCI) has not been reported. We compiled logbooks of 6 ICFITs at an academic hospital and evaluated patient and procedural characteristics of PCIs performed with and without presence of an ICFIT. The primary end point was the composite of all in-hospital PCI complications defined by the American College of Cardiology’s National Cardiovascular Data Registry: (1) catheterization laboratory events such as no-reflow and dissection/perforation, (2) general clinical events such as stroke or cardiogenic shock, (3) vascular and bleeding complications, and (4) miscellaneous complications such as peak troponin or creatinine levels. Logistic regression adjusted for differences in measured confounders between patients treated with and without presence of an ICFIT. All analyses were repeated after excluding PCI for ST-elevation myocardial infarction. Of 2,605 PCI procedures at the academic hospital between July 2007 and April 2010, an ICFIT was present for 1,638 procedures (63%). Despite having worse clinical and procedural characteristics, patients in the ICFIT group experienced similar rates of the composite end point (12.9% vs 14.5% without ICFIT, p = 0.27). Longer mean fluoroscopy times and greater number of stents were noted in the ICFIT group; however, hospital length of stay was shorter and no individual adverse events were increased in the ICFIT procedures. Presence of an ICFIT remained unrelated to the composite end point after multivariable adjustment (odds ratio 0.92, 95% confidence interval 0.71 to 1.20; p = 0.53), and findings were similar after excluding PCI for ST-elevation myocardial infarction. In conclusion, in contemporary practice at a large academic medical center, PCI complication rates were not adversely affected by the presence of an ICFIT.
Despite the identification of multiple factors associated with clinical outcomes during percutaneous coronary intervention (PCI). whether interventional cardiology fellows-in-training (ICFITs) are related to PCI complications has not been evaluated. To address this gap in knowledge, we evaluated the association between presence of an ICFIT during PCI and periprocedural adverse events among patients treated at an established academic training hospital.
Methods
All patients who underwent PCI at Saint Luke’s Hospital in Kansas City, Missouri, between July 1, 2007, and April 3, 2010, were included in our study. One university-affiliated cardiology group with 12 attending interventional cardiologists performed all PCI procedures, and 2 ICFITs per academic year participated as the primary operators in most PCI procedures. Demographic, clinical, and angiographic data were collected in a prospective hospital registry as part of the American College of Cardiology’s National Cardiovascular Data Registry (ACC-NCDR). Presence of an ICFIT was determined by combining the logbooks of 6 ICFITs at Saint Luke’s Hospital over a 3-year period.
The primary end point was defined as the composite of all procedural and in-hospital complications captured by the ACC-NCDR, falling into the following 4 predefined categories : (1) those occurring in the catheterization laboratory (no reflow, dissection, acute closure, perforation), (2) general clinical events (periprocedural myocardial infarction, cardiogenic shock, heart failure, stroke, cardiac tamponade, renal failure, emergency PCI), (3) vascular and bleeding complications (entry-site bleeding, retroperitoneal bleeding, gastrointestinal and genitourinary bleeding, access site occlusion, peripheral embolization, dissection, pseudoaneurysm, arteriovenous fistula), and (4) other miscellaneous complications and measured parameters (emergency coronary artery bypass graft [CABG], salvage CABG, peak creatine kinase-MB fraction, peak troponin, post-PCI creatinine, blood transfusion, in-lab death, in-hospital death). Secondary end points were the individual components of the primary composite end point, resources used (e.g., number of stents, total procedure times), and hospital length of stay. Given the prevailing opinion that procedural complications are more likely during the early learning curve of an operator, we also evaluated whether rates of the primary end point, vascular complications, or the need for emergency CABG varied according to quarters of the academic year (July to September, October to December, January to March, and April to June).
Procedural time was defined as the number of minutes between lidocaine injection of the access site and the departure of the patient from the catheterization table. Devices recorded during PCI included intravascular ultrasound; pressure or flow wire; mechanical or aspiration thrombectomy; rotational, orbital, directional, or laser atherectomy; use of a cutting balloon or covered stent; embolic protection device; extraction catheter; or specialized devices used for chronic total occlusions (e.g., reentry devices). All procedure reporting was performed by the attending interventionalists, irrespective of whether an ICFIT was present, and complications were identified by trained chart abstractors according to standard clinical practice at Saint Luke’s Hospital.
Characteristics of patients at the main academic hospital with and without presence of an ICFIT were compared using Student’s t test for continuous variables, chi-square test for categorical variables, and nonparametric alternatives when appropriate. The relationship between presence of an ICFIT and the primary composite end point was then evaluated using logistic regression after adjusting for variables with nominal statistical significance (p ≤0.1) from bivariate analysis. Multivariable findings are reported as adjusted odds ratios (OR [95% confidence intervals]) and associated p values. Candidate variables were those collected by the ACC-NCDR, according to standard definitions from this data source, including age, gender, body mass index, diabetes, hypertension, peripheral vascular disease, cerebrovascular disease, previous PCI, previous CABG, previous heart failure, renal failure, chronic lung disease, tobacco use, family history of coronary disease, ST-elevation myocardial infarction (STEMI) presentation, minimum stent diameter, and total stent length. Of note, presence of an ICFIT during STEMI was known to be variable over time as a result of changes in ICFIT on-call requirements at our training program, so all analyses were planned a priori to be repeated after excluding patients undergoing primary or rescue PCI for STEMI. In addition, as a sensitivity analysis, the multivariable models were repeated after accounting for clustering by academic training period in case some ICFITs may have affected the primary composite outcome more than other trainees.
Statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, North Carolina), and p values ≤0.05 were considered statistically significant unless otherwise specified. All authors reviewed and contributed to the manuscript as written.
Results
A total of 2,605 patients underwent PCI at the primary academic hospital during the study period: 1,638 with an ICFIT present (63%) and 967 without an ICFIT (37%; Figure 1 ). Patients who underwent PCI with an ICFIT present were slightly older, more likely to have hypertension, more likely to have previous coronary revascularization and less likely to present with STEMI ( Table 1 ). Procedural characteristics were similar between groups, although ICFIT patients had more multivessel interventions and a greater proportion of interventions performed in the left circumflex artery.
| Variable | All Patients | p Value | Excluding STEMI | p Value | ||
|---|---|---|---|---|---|---|
| ICFIT (n = 1,638) | No ICFIT (n = 967) | ICFIT (n = 1,464) | No ICFIT (n = 706) | |||
| Age (yrs) | 65 ± 13 | 63 ± 13 | 0.004 | 65 ± 13 | 65 ± 13 | 0.52 |
| Men | 69% | 66% | 0.13 | 69% | 64% | 0.047 |
| Body mass index (kg/m 2 ) | 30 ± 6 | 30 ± 7 | 0.93 | 30 ± 6 | 30 ± 6 | 0.61 |
| Hypertension | 84% | 79% | 0.002 | 86% | 85% | 0.88 |
| Diabetes mellitus | 34% | 34% | 0.88 | 36% | 38% | 0.21 |
| Current smoker | 28% | 32% | 0.07 | 26% | 24% | 0.22 |
| Family history of coronary artery disease | 17% | 16% | 0.75 | 18% | 17% | 0.85 |
| Previous heart failure | 15% | 14% | 0.39 | 16% | 17% | 0.49 |
| Previous PCI | 46% | 40% | 0.004 | 48% | 47% | 0.58 |
| Previous CABG | 22% | 17% | 0.003 | 23% | 21% | 0.25 |
| History of renal failure | 5% | 5% | 0.96 | 6% | 7% | 0.38 |
| Cerebrovascular disease | 17% | 16% | 0.21 | 18% | 18% | 0.95 |
| Peripheral vascular disease | 15% | 15% | 0.65 | 16% | 17% | 0.81 |
| Chronic lung disease | 18% | 18% | 0.74 | 18% | 19% | 0.56 |
| Clinical indication for PCI ∗ | ||||||
| STEMI | 11% | 27% | <0.001 | — | — | — |
| Non-STEMI | 21% | 22% | 0.58 | 24% | 31% | <0.001 |
| Unstable angina | 25% | 19% | 0.003 | 28% | 26% | 0.27 |
| Stable coronary artery disease | 43% | 32% | <0.001 | 48% | 44% | 0.056 |
| Staged PCI | 6% | 3% | <0.001 | 7% | 5% | 0.027 |
| Angiographic findings | ||||||
| Number of significant coronary narrowings | 1.7 ± 0.9 | 1.6 ± 0.8 | 0.08 | 1.7 ± 0.9 | 1.7 ± 0.9 | 0.79 |
| Vessels treated at PCI procedure | <0.001 | 0.24 | ||||
| 1 | 78% | 85% | 76% | 80% | ||
| 2 | 19% | 13% | 21% | 17% | ||
| ≥3 | 3% | 2% | 3% | 3% | ||
| Most severe coronary stenosis undergoing PCI (%) | 91 ± 9 | 93 ± 9 | <0.001 | 90 ± 9 | 92 ± 10 | <0.001 |
| Minimum stent diameter (mm) | 3.0 ± 0.5 | 3.1 ± 0.5 | 0.55 | 3.0 ± 0.5 | 3.0 ± 0.5 | 0.08 |
| Total stent length (mm) | 35 ± 22 | 33 ± 21 | 0.07 | 35 ± 33 | 35 ± 23 | 0.93 |
| Coronary vessel undergoing PCI ∗ | ||||||
| Left circumflex | 28% | 24% | 0.015 | 30% | 28% | 0.30 |
| Left anterior descending | 44% | 42% | 0.22 | 44% | 42% | 0.26 |
| Right | 39% | 40% | 0.57 | 38% | 39% | 0.71 |
| Left main | 4% | 3% | 0.20 | 4% | 4% | 0.71 |
| Saphenous vein graft | 8% | 7% | 0.43 | 9% | 8% | 0.56 |
∗ May sum to >100% because of multiple indications for PCI and multiple vessels undergoing PCI, respectively, in some patients.
When evaluating clinical events after PCI ( Table 2 ), there was no significant difference in the primary composite end point between patients with and without an ICFIT present during PCI (12.9% vs 14.5%, p = 0.27). Individual complications were also similar between groups, including the frequency of bleeding events or need for emergent CABG. Hospital length of stay was shorter and in-hospital mortality was lower in the ICFIT group. Resource utilization was notable for longer overall procedural time and greater number of stents placed per patient in the ICFIT group. After excluding patients with STEMI from the analyses, clinical characteristics and resource utilization between patients with and without the presence of an ICFIT remained remarkably similar, and clinical events were similar as well except for shorter length of stay in the ICFIT group (see Tables 1 and 2 ). When evaluating complications according to the quarter of treatment (July to September and so forth), there were no differences in the composite primary end point (range 11.6% to 16.8%, trend p = 0.10), vascular complications (range 0.52% to 2.1%, trend p = 0.14), or emergency CABG (range 0.26% to 0.83%, trend p = 0.50).
