Reports from the Registries: NCDR® Research at ACC.14
At the ACC.14 Scientific Sessions, National Cardiovascular Data Registry (NCDR®) researchers presented a total of 20 papers and posters, including one Late-Breaking Clinical Trial. Here, CSWN: Interventions dives into some of the research originating from the preeminent cardiovascular data repository.
One-Year Outcomes from the STS/ACC TVT Registry
Transcatheter aortic valve replacement (TAVR) is being used with increasing frequency for patients with severe aortic stenosis who are inoperable or high risk but operable. At an ACC Late-Breaking Clinical Trial session, David R. Holmes, Jr., MD, Mayo Clinic, Rochester, Minnesota, and colleagues presented the 1-year outcomes from the STS/ACC Transcatheter Valve Therapy (TVT) Registry™, which involves all US patients undergoing TAVR for native aortic stenosis.
Nearly 6,000 patients >65 years of age were enrolled in the registry. Median age was 85 years, with 48% of patients between 85 and 94 years. Half were female, and the mean STS PROM (Society of Thoracic Surgeons Predicted Risk of Mortality) score was 7.1 with 12.2% showing an STS score >15—indicating that the registry included both high-risk and inoperable patients. The vast majority of patients (83.6%) were in NYHA class III or IV heart failure.
Dr. Holmes' and colleagues found:
- At 1 year, the mortality rate stood at 26.2%, stroke rate at 3.6%, and combination death or stroke rate at 28.4%. Just over half (55.6%) of patients were alive at 1 year and had not required repeat hospitalization.
- A statistically significant gradient was noted with regard to mortality and age, with a 31.9% mortality rate in those ≥95 years, dropping to 21.5% for those <75 years (p = 0.016). This pattern was not seen with stroke, but the stroke rates stayed low, ranging from 3.0-4.1% (p = 0.247).
- Men experienced higher mortality at 1 year than women (29.2% vs. 23.0%; p = 0.007), but women exhibited higher stroke rates (4.3% vs. 2.9%; p = 0.012).
- Patients with more severe chronic obstructive pulmonary disease (COPD) demonstrated significantly higher mortality rates than those with no or mild COPD (36.4% vs. 23.2%; p < 0.001).
- Worsening renal function also was associated with higher mortality: patients requiring dialysis had a mortality rate of 46.0%, compared to 24.6% for those with creatinine <2 and not on dialysis (p < 0.001).
- Transfemoral access, which was employed in 63.7% of procedures, was associated with lower 1-year mortality (24.4%) compared to other access sites (28.8%; p < 0.001); stroke rates were numerically but not significantly higher in the transfemoral group as compared to the alternative access site group (3.9% and 3.0%; p = 0.346).
"I think that unmeasured confounding variables are responsible for most of this difference," said Dr. Holmes about the mortality by access site findings.
Not surprisingly, the patients with STS risk scores of >15 had a mortality rate of 40.0%, falling to 20.9% for those with risk scores <8 (p <0.001). This association was not seen with stroke (p = 0.619).
"In this dataset using the same risk factors that [were significantly associated with mortality at 1 year], the only thing that is statistically significant in terms of being associated with stroke was female gender," said Dr. Holmes. "The crucial thing about this dataset is that identification of these associations is going to be essential in developing risk prediction models and will aid us in selecting patients for optimal therapy."
Asked about the lower stroke rate (3.6%) seen in this dataset as compared to PARTNER B (11.2%) or PARTNER A (8.7%; TABLE 1), Dr. Holmes noted that the occurrence of stroke was determined by administrative data and was not neurologically confirmed.
An analysis is underway looking at the association between center and operator experience/volume and 1-year outcomes.
Length of Hospital Stay and Associated Clinical Outcomes in Elderly Patients Following Primary PCI for STEMI
Using data from the CathPCI Registry® and Centers for Medicare & Medicaid Services claims, Rajesh V. Swaminathan, MD, Weill Cornell Medical College, New York-Presbyterian Hospital, and colleagues determined that hospital length of stay (LOS) after primary PCI for STEMI has been declining, and that earlier discharge may be safe for elderly patients.
Among 33,920 STEMI patients aged ≥65 years treated with primary PCI from 2004 to 2009, 26.9% had a short LOS (≤3 days), 46.3% had a medium LOS (4-5 days), and 26.8% had a long LOS (>5 days).
- In general, long LOS patients were older and female, and had more comorbidities, cardiogenic shock, TIMI 0 flow, and multi-vessel disease, with correspondingly worse 30-day clinical outcomes. Short LOS patients had higher ejection fractions, were less likely to have a pre-procedure TIMI grade flow of 0, and exhibited more single-vessel disease.
- Adjusted analyses revealed no difference in 30-day mortality (hazard ratio [HR] = 1.00; p = 0.98) or major adverse cardiovascular events (MACE; HR = 1.04; p = 0.72) for medium versus short LOS.
- There was a significant increase in mortality (HR = 2.36; p < 0.001) and MACE (HR = 1.78; p < 0.001) for long versus short LOS (TABLE 2).
Dr. Swaminathan told CSWN: Interventions, "We found that early—but not very early (<48 hours)—discharge may be safe after primary PCI for STEMI among selected elderly patients without in-hospital complications." So, while about half of elderly STEMI discharges after primary PCI occur within 4 or 5 days, patients discharged as early as 48 hours after PCI have similar outcomes. "One of the important goals of this research is to help reduce the current variability in STEMI discharge practice patterns across the country, and to ensure that the current trend of reductions in length of stay does not translate to unsafe practice for patient care and outcomes," he added.
Interventional Cardiology Board Certification and In-Hospital Outcomes of PCI Patients
What does board certification mean in terms of outcomes for patients seen by interventional cardiologists? The answer in a study presented by Paul Fiorilli, MD, Yale University School of Medicine, raised an interesting question.
The American Board of Internal Medicine (ABIM) has offered a certification exam in interventional cardiology (IC) since 1999. While previous studies have shown better outcomes with certified ICs and many hospitals require board certification for credentialing purposes, the proportion of PCIs completed by non-certified ICs (and whether this influences outcomes) is not known.
In their examination of IC board certification and outcomes from the CathPCI Registry, Dr. Fiorilli and colleagues found:
- More than 5,000 physicians conducted 510,708 procedures. The vast majority of doctors were male (96.8%), the mean age was 49.9 years, and 70.7% were IC-certified. Interestingly, 309 physicians (6.1%) had lapsed certification.
- No differences were seen between IC-certified and non-IC–certified physicians for in-hospital mortality, bleeding, vascular complications, emergency CABG, or composite adverse event rates.
- After adjustment for patient characteristics and physician and hospital volume, rates of death and emergency CABG were "modestly" higher in the non-certified group (odds ratios [OR] = 1.09 [p = 0.034] and 1.23 [p = 0.015], respectively), but still no differences were seen in bleeding, vascular complications, or the composite adverse event endpoint.
- Discharge medications did not differ by certification status, nor did the appropriateness of the procedure.
The authors found that a substantial proportion of PCI procedures (about 30%) are being performed by non-certified interventionalists, but the overall outcomes are "excellent" regardless of certification status. While significantly lower risk of mortality and emergency CABG was seen among certified physicians, "the effect size was modest," they noted.
"These findings suggest the need to include a more direct assessment of procedural skills in the certification program," they concluded.
Inappropriate and Non-Recommended Prasugrel Prescribing
Based on the findings of the TRITON-TIMI 38 trial, prasugrel is contraindicated in patients with prior stroke or transient ischemic attack (TIA). Furthermore, its use is not recommended in patients aged ≥75 years without diabetes or prior MI, where there is concern that its bleeding risk likely exceeds benefit. Ravi S. Hira, MD, Baylor College of Medicine, and colleagues used PINNACLE Registry® data to examine the frequency and practice patterns of inappropriate (contraindicated) and non-recommended prasugrel prescribing.
PINNACLE is cardiology's largest outpatient registry and collects point-of-care data for all patients with CAD, heart failure, atrial fibrillation (AF), and hypertension from participating practices.
Within the PINNACLE Registry's 8,007,927 patient encounters, the researchers identified 27,533 patients who were prescribed prasugrel. Using hierarchical regression models, they assessed the extent of practice-level variation using the median rate ratio (MRR), interpreted as the likelihood of differing treatment between identical patients and two random practices.
The inappropriate (contraindicated) prasugrel prescribing rate reached 13.9%. The rate of non-recommended prasugrel use was 4.4%.
Researchers found substantial practice-level variation in inappropriate prescribing, ranging from zero to 90.1% (median rate = 4.4%; interquartile range [IQR] = 2.1–10.3%) with an MRR of 2.89 (95% CI 2.75–3.03). There was also signiﬁcant practice-level variation in non-recommended prasugrel prescribing, again ranging from zero to 19.8% (median rate = 2.6%; IQR = 1.5–4.1%) with an MRR of 2.29 (95% CI 2.05–2.51). A total of 4,248 patients (15.4%) were receiving triple therapy (i.e., aspirin, warfarin, prasugrel). Of these, 996 (23.4%) were receiving prasugrel for an inappropriate (319 patients) or non-recommended (677 patients) indication.
The authors concluded that almost one-fifth of patients receiving prasugrel were getting it for either an inappropriate (actually contraindicated) or non-recommended indication, with large practice-level variation. These data identify important opportunities to improve appropriate and evidence-driven prescribing of prasugrel.
Dr. Hira was awarded Honorable Mention in the Cardiovascular Health Outcomes and Population Genetics section of the ACC Young Investigator Awards Competition for this study.
Uptake of NOACS in Patients with Valvular AF
Several novel oral anticoagulants (NOACs) have been FDA-approved for stroke prevention in patients with nonvalvular AF, but the clinical trials used to approve these agents excluded patients with valvular AF. Sandeep M. Jani, MD, Medstar Washington Hospital Center, DC, and colleagues used PINNACLE Registry data to examine the off-label adoption of NOACs for valvular AF.
Outpatient prescription rates for dabigatran and rivaroxaban—compared to warfarin—were examined in patients with nonvalvular AF with CHADS2 score ≥2 (n = 2,141) and for all patients with valvular AF (n = 190,227) from January 2011 through December 2012 from 95 clinical practices.
Valvular AF patients were younger (72.7 vs. 75.0 years; p < 0.001) and had a higher rate of prior stroke or TIA (42.9% vs. 26.7%; p < 0.001) compared to nonvalvular AF patients.
Overall, the uptake of NOACs for stroke prevention has been rapid, comprising 14.9% of all prescription for nonvalvular AF anticoagulation and 13.8% of prescription for valvular AF (TABLE 3).
In 2011, off-label uptake of NOACs in valvular AF was slower than that seen for nonvalvular AF, but by the end of 2012, NOAC prescription rates were not significantly different.
The investigators concluded that off-label use of NOACs for valvular AF has been rapidly adopted in clinical practice, despite a lack of evidence of efficacy. Studies are needed confirming the effectiveness of NOACs for stroke prophylaxis in valvular AF.
TABLE 3 Anticoagulant Use in Patients with Valvular versus Nonvalvular AF
< Back to Listings