Abstract
Background
The association of bleeding avoidance strategy (BAS) (consisting of a combination of radial access, bivalirudin [rather than heparin +/− glycoprotein GPIIb/IIIa antagonists], and/or vascular closure devices after femoral access) with bleeding and in-hospital outcomes has not been evaluated among elderly patients undergoing percutaneous coronary interventions (PCI).
Methods
We studied BAS use, bleeding and in-hospital mortality among 121,635 patients categorized by age (< 50, 50–59, 60–69, 70–79, and ≥ 80 years) undergoing PCI from the BMC2 registry (1/2010–12/2013).
Results
The use of BAS decreased marginally with age and despite improved utilization over time, remained lower among the elderly. BAS was used in a much lower risk cohort among all age groups. Nonetheless, compared with no BAS, the use of this strategy was associated with lower bleeding (adjusted OR 0.984, 95% CI 0.980–0.985) and in-hospital mortality (adjusted OR 0.996, 95% CI 0.994–0.997) among all age-groups. Similar relative reduction in the risk of bleeding was observed among all age groups with BAS use with lowest risk (thus greatest absolute risk reduction given their highest risk for bleeding) for the oldest cohort.
Conclusions
BAS use decreased with age among patients undergoing PCI despite its association with lower in-hospital mortality. Although overall utilization improved over time, it still remained lower in the elderly cohort, a group likely to benefit most from it. These data identified an opportunity to design strategies to improve BAS use particularly among high-risk elderly patients undergoing PCI so as to decrease bleeding and reduce related adverse events and costs.
1
Introduction
Peri-procedural bleeding is a common complication among patient undergoing percutaneous coronary intervention (PCI) and is associated with a higher morbidity and mortality post-procedure as well as longer hospitalization and increased costs . Older age has been reported as a significant risk factor for bleeding after PCI . Recently the use of bleeding avoidance strategies (BAS) has been shown to be associated with lower risk of peri-PCI bleeding. These strategies consist of a combination of the use of radial access (rather than femoral access), bivalirudin (rather than heparin +/− glycoprotein GPIIb/IIIa antagonists) and/or vascular closure devices after femoral access for PCI . Prior reports suggested a risk-treatment paradox where BAS were utilized in low risk group compared to groups with high risk of bleedings . However, the incidence of the use of BAS with age and its association with age-related bleeding and outcomes remain unknown. We hypothesized that given the higher risk of bleeding in the elderly, BAS use was likely to be higher with increasing age of the patients undergoing PCI. Furthermore, we theorized that given the higher risk of bleeding with older age, BAS were likely to be associated with greatest reduction in bleeding risk and better outcomes in the elderly. To test these hypotheses, we evaluated BAS use with age and its relationship with age-related bleeding and outcomes in patients undergoing PCI enrolled in the Blue Cross Blue Shield of Michigan Cardiovascular Consortium (BMC2), a collaborative statewide, multi-hospital PCI quality improvement .
2
Materials and methods
2.1
Patient population
We evaluated 124,606 patients undergoing PCI enrolled in the BMC2 registry between January 1, 2010 and December 31, 2013. The details of the design of BMC2 registry and the data collection process have been described previously . Briefly, data on all patients undergoing PCI at 44 participating hospitals were collected using standardized data collection forms. All data elements including adverse events were prospectively defined. A dedicated staff member at participating sites collected the data and forwarded it to the coordinating center. Bleeding events were identified by participating sites and were not adjudicated. Sites were given a uniform standard definition of bleeding as described under the Methods section and were asked to categorized bleeding based on this definition. However, to ensure accuracy all participating sites were audited twice yearly. During the audit, 2% of cases were selected at random for review. Medical records of all patients undergoing coronary artery bypass grafting (CABG), and of those who died in the hospital were reviewed by auditors from the coordinating center to ensure accuracy. The choice of medications as well as equipment was at the discretion of the operating physician and encouraged to be consistent with national Guidelines for PCI . Patients with cardiogenic shock ( n = 2971) were excluded from this analysis.
2.2
Definition of complications
These definitions are available on the registry Website ( https://bmc2.org ) as well as published previously . Access site hematoma (regardless of access) was defined as any hematoma requiring transfusion, or prolonged hospital stay or caused a drop in hemoglobin ≥ 3 g/dl. Vascular complications included pseudoaneurysm, arteriovenous fistula, femoral nerve injury, retroperitoneal hematoma, access site hematoma, or any access site complication requiring surgical repair. Gastrointestinal bleeding was considered when the patient had hematemesis or melena associated with a decrease in hematocrit and hemoglobin. Mortality was defined as all-cause death from either cardiac or non-cardiac etiology. Bleeding avoidance strategies were defined as the use of vascular closure devices during femoral access, radial approach, bivalirudin, or a combination of these . The current analysis was funded by a grant from the Blue Cross Blue Shield of Michigan Foundation. The BMC2 registry is funded by Blue Cross Blue Shield of Michigan. The sponsors had no role in the study design, analysis, drafting and editing of the manuscript or decision to publish these results.
2.3
Statistical analysis
Patients were categorized into five age-based groups: < 50 years, 50 to 59 years, 60 to 69 years, 70 to 79 years, ≥ 80 years. Continuous variables are expressed as mean ± SD, and discrete variables as frequency counts and percentages. Missing data were not defaulted to negative and denominators reflect cases reported. Differences in baseline characteristics between patient groups were evaluated by χ 2 tests or by Kruskal Wallis tests as appropriate. Previously developed multivariable logistic regression models with age as categorical variable (reference < 50 years) were used to derive adjusted odds of in-hospital bleeding and in-hospital death in patients with and without the use of BAS . Interaction for age and BAS was tested in the models using likelihood ratio test. Odds ratios (OR) and 95% confidence intervals (CI) were generated to provide an estimate of these associations. All p -values were two-sided with values < 0.05 considered statistically significant. Statistical analysis was performed by the University of Michigan that had Institutional Review Board approval for the BMC2 registry. All analyses were performed using R version 2.15.1 (R Core Team [2012]. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900,051-07-0 software).
2
Materials and methods
2.1
Patient population
We evaluated 124,606 patients undergoing PCI enrolled in the BMC2 registry between January 1, 2010 and December 31, 2013. The details of the design of BMC2 registry and the data collection process have been described previously . Briefly, data on all patients undergoing PCI at 44 participating hospitals were collected using standardized data collection forms. All data elements including adverse events were prospectively defined. A dedicated staff member at participating sites collected the data and forwarded it to the coordinating center. Bleeding events were identified by participating sites and were not adjudicated. Sites were given a uniform standard definition of bleeding as described under the Methods section and were asked to categorized bleeding based on this definition. However, to ensure accuracy all participating sites were audited twice yearly. During the audit, 2% of cases were selected at random for review. Medical records of all patients undergoing coronary artery bypass grafting (CABG), and of those who died in the hospital were reviewed by auditors from the coordinating center to ensure accuracy. The choice of medications as well as equipment was at the discretion of the operating physician and encouraged to be consistent with national Guidelines for PCI . Patients with cardiogenic shock ( n = 2971) were excluded from this analysis.
2.2
Definition of complications
These definitions are available on the registry Website ( https://bmc2.org ) as well as published previously . Access site hematoma (regardless of access) was defined as any hematoma requiring transfusion, or prolonged hospital stay or caused a drop in hemoglobin ≥ 3 g/dl. Vascular complications included pseudoaneurysm, arteriovenous fistula, femoral nerve injury, retroperitoneal hematoma, access site hematoma, or any access site complication requiring surgical repair. Gastrointestinal bleeding was considered when the patient had hematemesis or melena associated with a decrease in hematocrit and hemoglobin. Mortality was defined as all-cause death from either cardiac or non-cardiac etiology. Bleeding avoidance strategies were defined as the use of vascular closure devices during femoral access, radial approach, bivalirudin, or a combination of these . The current analysis was funded by a grant from the Blue Cross Blue Shield of Michigan Foundation. The BMC2 registry is funded by Blue Cross Blue Shield of Michigan. The sponsors had no role in the study design, analysis, drafting and editing of the manuscript or decision to publish these results.
2.3
Statistical analysis
Patients were categorized into five age-based groups: < 50 years, 50 to 59 years, 60 to 69 years, 70 to 79 years, ≥ 80 years. Continuous variables are expressed as mean ± SD, and discrete variables as frequency counts and percentages. Missing data were not defaulted to negative and denominators reflect cases reported. Differences in baseline characteristics between patient groups were evaluated by χ 2 tests or by Kruskal Wallis tests as appropriate. Previously developed multivariable logistic regression models with age as categorical variable (reference < 50 years) were used to derive adjusted odds of in-hospital bleeding and in-hospital death in patients with and without the use of BAS . Interaction for age and BAS was tested in the models using likelihood ratio test. Odds ratios (OR) and 95% confidence intervals (CI) were generated to provide an estimate of these associations. All p -values were two-sided with values < 0.05 considered statistically significant. Statistical analysis was performed by the University of Michigan that had Institutional Review Board approval for the BMC2 registry. All analyses were performed using R version 2.15.1 (R Core Team [2012]. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900,051-07-0 software).
3
Results
3.1
Demographics and clinical features ( Table 1 )
The use of BAS was high and decreased marginally with increasing age ( Fig. 1 ). Their utilization increased over the study period and improved equally in all age groups ( Fig. 2 ). Older patients were more likely to be white with lower body mass index and higher comorbid conditions with the exception of lower prevalence of being current smokers. The prevalence of coronary artery disease (and previous PCI and/or previous CABG) was higher in older patients. Percutaneous coronary interventions were more likely to be performed for non-ST elevation acute coronary syndromes and less likely for ST elevation myocardial infarction in older compared with younger patients. Baseline hemoglobin and glomerular filtration rate were also lower in the older age groups. Lower use of prasugrel and higher use of clopidogrel were seen in older patients. Predicted bleeding risk increased with older age and BAS were used in patients at lower risk of bleeding in all age cohorts.
| Variable | Age | |||||
|---|---|---|---|---|---|---|
| < 50 years | 50–59 years | 60–69 years | 70–79 years | > 80 years | p -Value | |
| n | 12,557 | 27,822 | 37,028 | 28,583 | 15,645 | |
| (%) | 10.32% | 22.87% | 30.44% | 23.50% | 12.86% | |
| Demographics | ||||||
| Age (years) (mean) | 44.0 | 54.9 | 64.5 | 74.2 | 84.0 | NA |
| Race—white | 81.27% | 83.50% | 85.87% | 88.25% | 90.64% | < 0.001 |
| Body mass index (kg/m 2 ) (mean) | 32.1 | 31.7 | 31.3 | 29.9 | 27.4 | < 0.001 |
| Medical history | ||||||
| Hypertension | 72.59% | 81.10% | 86.99% | 90.79% | 91.94% | < 0.001 |
| Diabetes mellitus | 30.50% | 35.47% | 42.12% | 41.68% | 32.99% | < 0.001 |
| Current smoker | 59.58% | 47.41% | 27.37% | 12.95% | 4.83% | < 0.001 |
| CHF | 8.27% | 10.99% | 14.69% | 19.88% | 26.73% | < 0.001 |
| Atrial fibrillation | 2.67% | 4.91% | 9.60% | 16.86% | 24.26% | < 0.001 |
| Previous MI | 33.08% | 34.43% | 35.44% | 36.55% | 35.97% | < 0.001 |
| Previous PCI | 38.19% | 43.02% | 47.06% | 49.23% | 46.05% | < 0.001 |
| Previous CABG | 7.12% | 11.85% | 19.91% | 26.54% | 24.95% | < 0.001 |
| COPD | 10.19% | 16.66% | 20.02% | 22.20% | 20.15% | < 0.001 |
| PVD | 6.81% | 11.68% | 16.78% | 21.71% | 23.15% | < 0.001 |
| Gastrointestinal bleeding | 0.49% | 0.70% | 0.86% | 1.39% | 1.57% | < 0.001 |
| Indication for PCI | ||||||
| Primary PCI-STEMI | 21.86% | 15.14% | 10.42% | 7.61% | 9.81% | < 0.001 |
| NSTE-ACS | 53.58% | 53.14% | 52.48% | 53.79% | 57.16% | < 0.001 |
| Staged | 5.80% | 6.08% | 6.10% | 6.17% | 6.02% | < 0.001 |
| Other | 15.66% | 23.40% | 29.35% | 31.16% | 25.62% | < 0.001 |
| Baseline laboratory | ||||||
| Hemoglobin (gm/dl) | 14.2 | 14.0 | 13.5 | 13.0 | 12.4 | < 0.001 |
| GFR | 138.9 | 117.0 | 94.7 | 72.7 | 51.7 | < 0.001 |
| Baseline medications | ||||||
| Aspirin | 95.59% | 95.66% | 95.56% | 95.54% | 94.90% | 0.009 |
| Clopidogrel | 64.46% | 66.95% | 70.85% | 79.65% | 85.43% | < 0.001 |
| Prasugrel | 26.45% | 24.12% | 19.95% | 9.88% | 2.40% | < 0.001 |
| Predicted bleeding risk based on the NCDR model (mean) | 1.72% | 1.84% | 2.08% | 2.52% | 43.65% | < 0.001 |
| Predicted bleeding risk based on the NCDR model with BAS use | 1.59% | 1.71% | 1.95% | 2.39% | 3.46% | < 0.001 |
| Predicted bleeding risk based on the NCDR model without BAS use | 2.06% | 2.22% | 2.47% | 2.87% | 4.11% | < 0.001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
