Several factors contribute to the risk of percutaneous coronary intervention–related major entry site (MES) complications. We sought to examine the trends in MES among unselected patients during the stent era. Data from the Dynamic Registry including 5 distinct recruitment waves from 1997 to 2006 (n = 10,932) were used to assess baseline characteristics and MES among consecutive patients undergoing percutaneous coronary intervention. MES was defined as bleeding requiring transfusion, pseudoaneurysm, arterial thrombosis or dissection, vascular complication requiring surgery, or retroperitoneal bleed. Uncomplicated hematomas were not included. Several trends were observed in baseline characteristics including an increase from wave 1 to wave 5 in body mass index >30 kg/m 2 (30.2% to 40.4%), renal disease (3.5% to 9.1%), diabetes (28.0% to 34.1%), and hypertension (59.4% to 78%; p trend <0.001 for all). The use of a thienopyridine increased significantly from wave 1 (49.7%) to wave 5 (84%), whereas glycoprotein IIb/IIIa inhibitor use peaked in wave 3 (53.1%) and then decreased (p <0.001). Access site was predominately femoral, but radial access increased over time (0.3% wave 1, 6.6% wave 5, p ≤0.0001). The rates of MES (2.8% to 2.2%, p trend = 0.01) and MES requiring transfusion (2.0% to 0.74%, p trend <0.001) were low and decreased with time. The trend in less risk for MES in later time periods remained after adjustment. In conclusion, MES has decreased over time; however, opportunity for bleeding avoidance strategies still exists.
Local complications at the site of vascular access are among the most common complications after percutaneous coronary intervention (PCI). Robust data link vascular access site complications to both morbidity and mortality after PCI. The influence of access site bleeding events and transfusion on mortality and resource utilization is particularly significant. Multiple risk factors for vascular complications after PCI have been identified, including both patient and procedural characteristics. In an attempt to minimize vascular complications associated with PCI, bleeding avoidance strategies have been advocated. Such measures include use of smaller arterial sheaths, radial artery access, use of safer and more predictable anticoagulation regimens, and closure devices. Data in support of these strategies are emerging. In this analysis, we sought to investigate the rates and trends in major entry site (MES) complications in a large, prospective, multicenter PCI registry to characterize changes in patient and procedural factors and specific types of MES complications.
Methods
The National Heart, Lung, and Blood Institute–sponsored Dynamic Registry was a multicenter prospective study of consecutive patients undergoing PCI in North America and has previously been described. The Dynamic Registry includes 5 distinct recruitment waves of patients enrolled from 1997 to 2006 (n = 10,965). All patients enrolled in this observational study provided informed consent for the inclusion of data used in this analysis. The present analysis excludes 9 patients with missing data on age and 24 patients missing information on the components that define MES complications, leaving a patient population of 10,932.
MES complication was defined as access site bleeding requiring transfusion, development of pseudoaneurysm, arterial thrombosis or dissection, vascular complication requiring surgery, or retroperitoneal bleeding. Uncomplicated hematomas were not included in the primary MES end point. Data collection for closure device use was captured only in waves 4 and 5. Non–access site bleeding was defined as gastrointestinal, genitourinary, or other bleeding remote from the access site. Two of the investigators (JDA and KY) independently determined whether bleeding events were access or non–access site related. In the event of disagreement, a third reviewer (FS) was used.
Patients undergoing PCI were grouped by the recruitment wave, and descriptive statistics are summarized as means for continuous variables and percentages for categorical variables. Differences between proportions were assessed for temporal trend using the Cochran-Mantel-Haenszel or the Jonckheere-Terpstra test, as appropriate. Similar methods were used for the individual components that defined MES bleeding complications. Logistic regression was used to estimate the independent effect of clinically relevant factors on in-hospital risk of MES. Risk factors in the model included age, gender, procedural indication, cardiogenic shock, previous PCI, history of congestive heart failure, chronic kidney disease, peripheral vascular disease, glycoprotein (GP) IIb/IIIa inhibitor use, access site, and recruitment wave. Goodness of fit was assessed using the Hosmer-Lemeshow method, and the model was considered to be adequate (p >0.05). All statistical analyses were performed using SAS software, version 9.3 (SAS Institute Inc., Cary, North Carolina), and a 2-sided p value of ≤0.05 was considered to indicate statistical significance.
Results
Significant trends in baseline characteristics were observed with increasing co-morbidities over time ( Table 1 ). The prevalence of previous myocardial infarction, however, decreased over the 10-year period. Significant trends in procedural characteristics were also observed ( Table 2 ). The entry site location was predominantly femoral with increasing use of radial access over time. Anticoagulation and antiplatelet regimens varied considerably over time. Notably, the use of a thienopyridine significantly increased, whereas the use of a GP IIb/IIIa inhibitor peaked in wave 3 and decreased in wave 5. Overall use of procedural and postprocedural heparin decreased significantly from wave 1 to wave 5. Within the procedural period, the use of unfractionated heparin decreased, whereas more low–molecular weight heparin increased during the study period (p trend <0.001 for both). Bivalirudin use, reported in waves 4 and 5 only, was 31.9% and 32.0%, respectively. Preprocedure aspirin use was high, and there was no significant change over the 10-year period. Closure devices were used in 30.9% of patients in wave 4 and 40.2% in wave 5 (p <0.001).
| Characteristic ∗ | 1 (n = 2,500), % | 2 (n = 2,099), % | 3 (n = 2,045), % | 4 (n = 2,111), % | 5 (n = 2,177), % | p Value |
|---|---|---|---|---|---|---|
| Age (yrs), mean (median) | 62.6 (63) | 63.0 (63) | 64.4 (65) | 63.7 (64) | 63.9 (64) | <0.001 |
| Women | 35.5 | 36.4 | 36.0 | 32.5 | 32.8 | 0.005 |
| White | 80.1 | 78.2 | 79.3 | 77.0 | 75.1 | 0.99 |
| Black | 7.6 | 12.0 | 12.8 | 14.7 | 16.4 | |
| Asian | 4.8 | 3.9 | 3.8 | 3.0 | 2.2 | |
| Hispanic | 7.2 | 5.7 | 4.0 | 5.3 | 5.8 | |
| Other race/ethnicity | 0.2 | 0.2 | 0.1 | 0.0 | 0.7 | |
| Body mass index (kg/m 2 ), mean (median) | 28.2 (28) | 28.8 (28) | 29.3 (28) | 29.2 (28) | 29.6 (29) | <0.001 |
| Previous PCI | 28.5 | 30.5 | 33.0 | 31.6 | 35.2 | <0.001 |
| Previous coronary bypass | 16.6 | 17.1 | 18.6 | 19.8 | 18.0 | 0.03 |
| Previous myocardial infarction | 39.1 | 32.4 | 27.6 | 25.9 | 22.6 | <0.001 |
| Diabetes mellitus | 28.0 | 28.5 | 30.1 | 33.5 | 34.1 | <0.001 |
| Previous heart failure | 9.9 | 9.6 | 12.5 | 9.3 | 10.1 | 0.94 |
| Hypertension | 59.3 | 64.6 | 74.7 | 77.1 | 78.1 | <0.001 |
| Hypercholesterolemia | 61.0 | 63.4 | 71.4 | 74.8 | 79.1 | <0.001 |
| Smoker | ||||||
| Never | 34.5 | 32.5 | 33.2 | 37.6 | 36.0 | 0.04 |
| Current | 25.3 | 26.7 | 24.0 | 22.5 | 25.7 | |
| Former | 40.2 | 40.7 | 42.8 | 39.9 | 38.3 | |
| Renal disease | 3.5 | 5.0 | 7.3 | 8.8 | 9.2 | <0.001 |
| Peripheral vascular disease | 7.2 | 7.3 | 9.2 | 8.8 | 7.4 | 0.24 |
∗ Baseline clinical variables were based on patient self-reports, physician diagnosis, or pharmacologic therapy.
| Characteristic | 1 (n = 2,500), % | 2 (n = 2,099), % | 3 (n = 2,045), % | 4 (n = 2,111), % | 5 (n = 2,177), % | p Value |
|---|---|---|---|---|---|---|
| Device access site | ||||||
| Femoral | 99.4 | 97.5 | 96.0 | 96.1 | 93.2 | <0.0001 |
| Brachial | 0.3 | 0.5 | 0.2 | 0.5 | 0.2 | |
| Radial | 0.3 | 2.0 | 3.8 | 3.4 | 6.6 | |
| Meds <24 h before or during procedure | ||||||
| Aspirin | 91.9 | 81.0 | 94.1 | 92.2 | 86.7 | 0.88 |
| Clopidogrel or ticlopidine | 49.8 | 41.4 | 67.6 | 85.7 | 84.0 | <0.001 |
| GP IIb/IIIa inhibitor | 24.3 | 31.7 | 53.1 | 33.7 | 37.4 | <0.001 |
| Bivalirudin | — | — | — | 31.9 | 32.0 | 0.77 |
| Heparin | 97.4 | 94.6 | 95.8 | 75.6 | 71.9 | <0.001 |
| Unfractionated heparin | NA | 93.6 | 95.0 | 73.2 | 68.8 | <0.001 |
| Low–molecular weight heparin | NA | 3.9 | 3.7 | 4.7 | 7.3 | <0.001 |
| Meds used after the procedure | ||||||
| Heparin | 38.6 | 21.6 | 10.6 | 9 | 10.6 | <0.0001 |
The trends in MES and bleeding outcomes are presented in Figure 1 . The absolute decrease in MES over time was small, but the trend was significant, from 2.8% in wave 1 to 2.2% in wave 5 (p trend = 0.015). Transfusions related to access site MES decreased over time. Of patients who experienced bleeding (n = 426), the access site was the cause in 74.4%, 72.0%, 60.0%, 49.4%, and 71.4% of cases in waves 1 to 5, respectively. Rates of pseudoaneurysm did not vary over time (from 0.71% in wave 1 to 0.97% in wave 5, p trend = 0.45). Complications including arterial thrombosis and surgery of the access site were rare and similar over time (data not shown).
Several variables were identified as independently associated with MES ( Table 3 ). Older age, female gender, acute myocardial infarction as the indication, cardiogenic shock, and procedural GP IIb/IIIa inhibitor use were associated with a greater risk of MES. Previous PCI and radial access, in contrast, were associated with a significantly less risk of MES. Compared with recruitment wave 1, there was a trend toward less risk of MES in subsequent waves, and this reached statistical significance in wave 4.
