Twenty-Five–Year Trends in In-Hospital and Long-Term Outcome After Percutaneous Coronary Intervention
A Single-Institution Experience
Background— Little is known about the impact of technological and pharmacological advances on long-term outcome after percutaneous coronary intervention in general clinical practice.
Methods and Results— We analyzed in-hospital and long-term outcome of 24 410 percutaneous coronary interventions among 18 575 unique patients who underwent percutaneous coronary intervention at Mayo Clinic over 25 years. The study population was divided into group 1 (n=3708), coronary interventions from 1979 to 1989; group 2 (n=7020), interventions from 1990 to 1996; group 3 (n=10 952), interventions from 1996 to 2003; and group 4 (n=2730), interventions from 2003 to 2004. Despite the fact that patients in groups 3 and 4 were significantly older, sicker, and had greater prevalence of comorbid conditions, heart failure, and previous revascularization than those in groups 1 and 2, procedural success in groups 3 and 4 improved significantly (94%) versus groups 2 (89%) and 1 (78%) (P<0.001). Significant reduction in in-hospital mortality (groups 4 to 1: 1.8%, 1.7%, 2.6%, 3.0%; P<0.001) and need for emergency bypass surgery (groups 4 to 1: 0.4%, 0.5%, 1.6%, 5%; P<0.001) was noted in groups 3 and 4 compared with groups 1 and 2. Better adherence to currently recommended evidence-based medications for secondary prevention was seen in the recent time periods. After adjustment, significant reduction in follow-up mortality (hazard ratio, 0.81 and 0.74 for groups 3 and 4, respectively); death or myocardial infarction (hazard ratio, 0.80 and 0.75 for groups 3 and 4, respectively); death, myocardial infarction, or revascularization (hazard ratio, 0.76 and 0.58 for groups 3 and 4, respectively) was noted in recent time periods.
Conclusions— Despite higher-risk profiles of patients who underwent percutaneous coronary intervention in recent time periods, procedural success as well as in-hospital and long-term outcomes improved significantly over the last 25 years.
Received April 7, 2006; accepted April 10, 2007.
Percutaneous coronary intervention (PCI) has evolved since the introduction of balloon angioplasty by Gruntzig et al in 1979.1 The current procedural success rates are high, with improvements in outcomes attributed to operator experience, technology refinement, and availability of improved stent designs. Moreover, advances in adjunctive antiplatelet therapies, refinements in medical treatment for secondary prevention, and the arrival of drug-eluting stents (DES) likely have had a major salutary effect as well.2 These improvements have affected coronary revascularization practice so that PCI is now the preferred strategy for the majority of patients, which includes high-risk patients, for whom it was previously contraindicated.3–5 Previous studies that analyzed the temporal trends of coronary angioplasty do not reflect current practice patterns, such as outcomes after DES.
Clinical Perspective p 2841
To place the improvement in angiographic outcomes and periprocedural results in a more comprehensive perspective and in the context of recent pharmacotherapeutic advances, the aims of the present analysis were to characterize outcome trends in coronary interventional practice at a large-volume tertiary-care referral center over the last 25 years, to define the risk profile of the patients referred for the procedure, and to determine the relation of risk-adjusted outcome of PCI to the introduction of these important advances.
At the time of PCI, patient-specific data are entered into a prospective database that includes clinical, procedural, and angiographic information. A clinical research nurse contacts all patients at 6 months, 12 months, and annually thereafter. Medical records of all patients who required hospitalization at Mayo Clinic or elsewhere are reviewed to further characterize any clinical events during follow-up. Approval for the present study was obtained from the Institutional Review Board of Mayo Foundation. Patients who denied research authorization were excluded from the study in accordance with Minnesota law.
To reflect PCI practice patterns, 24 410 PCIs performed on 18 575 unique patients were stratified into 4 groups according to the time of the intervention: group 1, October 1979 to December 1989 (n=3708); group 2, January 1990 to May 1996 (n=7020); group 3, June 1996 to February 2003 (n=10 952); and group 4, March 2003 to September 2004 (n=2730). Group 1 consisted of patients who principally underwent percutaneous transluminal coronary angioplasty alone (percutaneous transluminal coronary angioplasty). Group 2 consisted of patients in whom stents were used mainly as a bailout strategy, with aggressive periprocedural anticoagulation. Group 3 included patients who regularly received bare metal stents and frequent adjunctive glycoprotein IIb/IIIa inhibitors, accompanied by dual oral antiplatelet therapy. Group 4 consisted of patients whose PCI reflected contemporary practice and included treatment with DES.
The outcomes of interest were major adverse cardiovascular events (MACEs), defined as 1 or more of the following: (1) in-hospital death, (2) Q-wave myocardial infarction (MI), (3) urgent or emergent coronary artery bypass graft surgery (CABG) during the index hospitalization, and (4) cerebrovascular accident. MI was diagnosed in the presence of 2 of the following 3 criteria: (1) typical chest pain for at least 20 minutes, (2) increased creatine kinase (or the MB fraction) level >2×normal, and (3) a new Q wave on electrocardiography. In-hospital deaths included all deaths during the index hospital admission. Procedural success was defined as a decrease in residual luminal diameter stenosis to <50% without in-hospital death, Q-wave MI, or need for emergency CABG. Long-term outcomes studied included all-cause mortality; death or any MI; and death, MI, or any revascularization. Other procedural complications, such as non–Q-wave MI and vascular access site, were not included in the analysis.
Continuous variables, summarized as mean±SD, were compared between groups with the Student 2-sample t test. Discrete variables are presented as group percentages and compared with Pearson χ2 test. Missing values were excluded from the denominator in the calculation of percentages. The Kaplan-Meier method was used to estimate event-free survival rates, with log-rank tests used for curve comparison. Survival analyses were conducted only on the first successful PCI per patient. All hypotheses tests are 2-sided with a 0.05 type-I error rate. Some variables were not collected at the initiation of the angioplasty registry and were only added at later dates. Thus, some characteristics could not be compared for group 1 and portions of group 2.
Logistic regression models were used to estimate partial odds ratios (ORs) for in-hospital events. We used the previously validated Mayo Clinic Risk Score (MCRS)6 to adjust the ORs for different levels of risk. This score is based on the weighted average of 5 clinical and 3 angiographic variables, and it categorizes patient risk from PCI from very low to high risk categories. Multivariable Cox proportional hazards models were developed for patients who had successful interventions, for the end point of postdischarge death, death or MI, and the composite end point of death, MI, or any revascularization. These were developed with backward selection on a set of univariately significant risk factors. The variables that represented the year groups then were forced into the model. Results are presented as hazard ratios (HRs) with corresponding 95% confidence intervals (CIs).
All authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
Baseline Clinical and Angiographic Characteristics
Baseline characteristics of the 4 groups are shown in Table 1. The more recent cohorts were “sicker” on the basis of the number of baseline variables. The mean age of groups 3 and 4 (66.3 and 67.1 years, respectively) was significantly older than that of groups 1 and 2 (62.2 and 64.8 years, respectively) (P<0.001). The proportion of patients >75 years old was significantly higher in groups 3 and 4 (26% and 30%, respectively). The prevalence of comorbid conditions (diabetes mellitus, hypertension), MI within 24 hours, congestive heart failure, and previous revascularization was significantly higher for the recent time periods than for the earlier periods (most comparisons of groups 3 and 4 versus 1 and 2 were statistically significant). Also, the American College of Cardiology/American Heart Association type C lesion classification, lower ejection fraction, and multivessel disease were more frequent in groups 3 and 4 versus groups 1 and 2 (P<0.001). The MCRS increased over time.
The currently recommended evidence-based pharmacotherapy for secondary prevention after PCI has improved over the last 25 years, with significantly higher (P<0.001) predismissal adherence to aspirin (97%), clopidogrel (96%), β-blockers (84%), lipid-lowering agents (88%), and angiotensin-converting enzyme inhibitors (62%) in the most recent time period (Table 2). The use of balloon angioplasty alone during PCI decreased from 95% to 9% across the 4 time periods, and the use of stents increased from zero before 1990 to 91% of all the procedures in the most recent time periods (groups 3 and 4). In group 4, 94% of all stent procedures (87% of all procedures) used DES. The use of rotational atherectomy was low in all the time periods and was used in 1% of the procedures in group 4 and in 5% in group 3. The use of glycoprotein IIb/IIIa inhibitors, such as eptifibatide and abciximab, which had increased during the late 1990s, has decreased slightly in the last 2 years; they were used in 48% of patients in group 4.
The in-hospital outcomes are displayed in Figure 1. The MACE rates decreased significantly, with significant improvement noted in the procedural success rates in groups 3 and 4 (P<0.001 for both outcomes). The in-hospital mortality was 1.7% for group 3 and 1.8% for group 4 and was significantly lower (P<0.001) in comparison with the rates of groups 1 and 2 (3% and 2.6%, respectively). The rates of emergency CABG decreased significantly over time (P<0.001). The incidence of emergency bypass surgery was 5% for group 1, 1.6% for group 2, 0.5% for group 3, and 0.4% for group 4. This translated into a 90% reduction in the need for emergency CABG in the most recent time period, with a major decrease between groups 1 and 2. However, the incidence of Q-wave MI increased from 1.4% in group 3 to 3% in group 4 among patients who underwent emergent PCI, as did the incidence of stroke (1.1% in group 3 versus 1.5% in group 4).
Risk Profile of Patients Who Underwent PCI
To evaluate the demographic and angiographic risk profiles of the patients, we used the validated MCRS and calculated the mean score of patients in the 3 recent time periods.6,7 For this analysis, group 1 was excluded because of the unavailability of all the variables included in the model. Higher risk scores generally predict higher rates of complications after PCI. The mean MCRS increased from 5.8±2.9 in group 2 to 6.5±3.0 in group 3 and 6.4±2.8 in group 4 (P<0.001), which indicates a more adverse risk profile of the patients in the recent groups (Table 1, Figure 2). The MCRS was used also to make risk adjustments for the composite in-hospital outcome (death, Q-wave MI, stroke, or need for emergent or urgent bypass surgery). Before 1993, the components of this score were not consistently available; therefore, models were constructed with data from 1993 onward. As compared with group 2, there was a significant reduction in the MCRS-adjusted outcomes in the recent groups of patients who underwent PCI (for group 4: OR, 0.74 [95% CI, 0.58 to 0.94; P=0.014]; for group 3: OR, 0.61 [95% CI, 0.51 to 0.73; P<0.001]). This corresponds to adjusted risk estimates of 3.7% (95% CI, 2.8 to 3.9) and 4.0% (95% CI, 3.1 to 5.0) for groups 3 and 4, respectively, versus 5.3% (reference group) for group 2. The adjusted risk of in-hospital mortality was significantly lower in group 3, with a 49% and 29% reduction noted in groups 3 and 4, respectively (for group 4: OR, 0.71 [95% CI, 0.49 to 1.03; P=0.07]; for group 3: OR, 0.51 [95% CI, 0.44 to 0.75; P<0.001]).
Event-Free Survival After Hospital Discharge
Follow-up data were analyzed for patients who had a successful coronary intervention. The median follow-up was 17.1 years for group 1, 11.1 years for group 2, 5.1 years for group 3, and 1.1 years for group 4. Figures 3, 4, and 5⇓⇓ show the Kaplan-Meier curves for overall survival; survival free of death or MI; and survival free of death, MI, or any revascularization for the 4 groups. Significant improvement was only noted in the most recent group for survival free of MI or any revascularization. After adjustment for the other significant factors, the risk of death was lower for the 2003 to 2004 cohort (group 4) than for the 1979 to 1996 cohort (groups 1 and 2) (HR, 0.74; 95% CI, 0.59 to 0.92; P=0.008); the risk for the 1996 to 2003 cohort (group 3) was also lower (HR, 0.81; 95% CI, 0.75 to 0.87; P<0.001). Similar results were seen for the risk of death or MI (group 4: HR, 0.75 [95% CI, 0.63 to 0.89; P=0.001]; group 3: HR, 0.80 [95% CI, 0.75 to 0.85; P<0.001]) and risk of death, MI, or any revascularization (group 4: HR, 0.58 [95% CI, 0.52 to 0.65; P<0.001]; group 3: HR, 0.76 [95% CI 0.73 to 0.80, P<0.001]).
The present study of a large dataset at a single institution provides a unique perspective of the major impact that the changes in technology and pharmacology have had on early and late outcomes of PCI since its introduction in 1979.
The data demonstrate a marked improvement over the last 25 years in the in-hospital outcome of patients who underwent PCI. Despite a “sicker” patient profile (as evidenced by a higher MCRS), significant recent declines were observed in the rates of in-hospital death and emergency CABG. We and others have previously demonstrated significant reductions in the need for emergent CABG after failed PCI.8–10 The overall improvement in the mortality and other MACE end points after PCI found in the present study is in agreement with results of other studies that have reported significant improvements in outcomes for various high-risk clinical and angiographic subsets (ie, older age, unstable angina, angiographic thrombus, and myocardial infarction).3,4,11 Similar improvements in results have been reported by other observational registries,10,12,13 which underscores the likely contribution of important technological and pharmacological advancement in conjunction with operator experience and technical improvements in angioplasty equipment.
Although we observed a significant progressive decrease in the in-hospital MACE rates over time, especially during the stent era, an additional decrease in MACE since the introduction of DES could not be demonstrated. The lack of further decline in the outcomes of PCI since the introduction of stents could be explained by the fact that no major changes occurred in the interventional practice since that time except for the replacement of thienopyridine (clopidogrel instead of ticlopidine) in 1998 and the introduction of DES in 2003. Moreover, no increase in the MCRS was observed in group 4 compared with group 3, which indicates similar risk profiles in the 2 recent eras. In the coming years, a larger number of patients will provide us with a more meaningful comparison.
The unadjusted outcomes, which include death and death or MI, were similar in the 2 recent groups. After adjustment, however, we observed a significant reduction in death, death or MI, and death, MI, or any revascularization on follow-up after successful PCI. The trends in improved long-term outcome, in contrast to the in-hospital events, have continued in the DES era. The reduction in adverse cardiovascular events on follow-up likely reflects the beneficial effect of DES on the rates of revascularization, and improved adherence to evidence-based medications, according to the current era standards, was noted during the index hospitalization. The 1-year rates of coronary revascularization at Mayo Clinic decreased from 15% in group 3 to 11% in group 4. The median follow-up of the most recent group is only 1.1 years. HRs reported between groups are only valid as long as patients are being followed in both groups. Because the most recent group has little follow-up after 2 years, our comparison between that group and the other 3 groups reflects only the first 2 years after PCI. Longer follow-up of patients treated with DES will shed more light on the incidence and impact of late stent thrombosis on mortality and myocardial infarction.
Use of a Validated Instrument for Risk Categorization
It is important to note that we used the MCRS to define the patient risk and to make risk adjustments for some of the in-hospital outcomes from the PCI procedure. The mean MCRS and the MACE rates in the present study have stabilized over the course of the 2 recent time periods. The present study shows that, despite the higher risk profile of the recent cohorts, MACE rates have decreased. The long-term outcome after successful PCI was determined by comorbid conditions, congestive heart failure, severity of angina, history of MI, previous MI, and previous CABG (data not presented), which underscores the need to differentiate them from variables that affect the outcome immediately after PCI (as included in the MCRS). The recognition and management of these variables and risk factors for coronary artery disease and plaque rupture will likely improve the long-term outcome of PCI. Current research efforts are focused not only on the development of newer stent technologies but also on aggressive secondary prevention, and, in this respect, the results of the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation, Bypass Angioplasty Revascularization Investigation 2 Diabetes,14 and Future Revascularization Evaluation in Patients with Diabetes Mellitus: Optimal Management of Multivessel Disease trials will be crucial. Cutlip et al15 recently concluded that progressive disease rather than restenosis is the major cause of late events—DES will not change that, but we are at an early stage in the vigor with which secondary prevention is pursued.15
The present study is a retrospective analysis of a study population enrolled over a period of 25 years, so the limitations of such a study need to be considered. The study setting involved a single-center tertiary-care referral center, and the results of the present study may not be generalizable to low-volume community hospitals. We may not have accounted for some unmeasured confounders; however, that would bias the results toward the null. It is difficult to determine the relative importance of better operator skills, improvement in technology, use of stents, and improved antiplatelet and other adjunctive therapy in the improvement of results in the most recent period. We realize that the increase in levels of cardiac biomarkers is an important prognostic marker; however, because of the inherent complexity of this large dataset, the nonavailability of these markers in the early time periods, and changing definitions, we chose not to elaborate on the temporal trends of MI across various time periods. Our institution has not adopted a strategy of routine use of vascular closure devices or alternate anticoagulation therapies; hence, we are not able to assess the impact of these changes on the outcome.
The present study has demonstrated significant improvement in the in-hospital and long-term outcomes over time among patients who undergo PCI since the advent of coronary angioplasty. The improvement has occurred despite the use of PCI in higher-risk patients.
Editing, proofreading, and reference verification were provided by the Section of Scientific Publications, Mayo Clinic (Rochester, Minn).
Hochman JS, Sleeper LA, Webb JG, Sanborn TA, White HD, Talley JD, Buller CE, Jacobs AK, Slater JN, Col J, McKinlay SM, LeJemtel TH; SHOCK Investigators. Early revascularization in acute myocardial infarction complicated by cardiogenic shock: should we emergently revascularize occluded coronaries for cardiogenic shock? N Engl J Med. 1999; 341: 625–634.
Singh M, Rihal CS, Selzer F, Kip KE, Detre K, Holmes DR. Validation of Mayo Clinic risk adjustment model for in-hospital complications after percutaneous coronary interventions, using the National Heart, Lung, and Blood Institute dynamic registry. J Am Coll Cardiol. 2003; 42: 1722–1728.
Loubeyre C, Morice MC, Berzin B, Virot P, Commeau P, Drobinski G, Ethevenot G, Moquet B, Marco J, Labrunie P, Cattan S, Coste P, Aubry P, Ferrier A. Emergency coronary artery bypass surgery following coronary angioplasty and stenting: results of a French multicenter registry. Catheter Cardiovasc Interv. 1999; 47: 441–448.
Seshadri N, Whitlow PL, Acharya N, Houghtaling P, Blackstone EH, Ellis SG. Emergency coronary artery bypass surgery in the contemporary percutaneous coronary intervention era. Circulation. 2002; 106: 2346–2350.
McGrath PD, Malenka DJ, Wennberg DE, Shubrooks SJ Jr, Bradley WA, Robb JF, Kellett MA Jr, Ryan TJ Jr, Hearne MJ, Hettleman B, O’Meara JR, VerLee P, Watkins MW, Piper WD, O’Connor GT; Northern New England Cardiovascular Disease Study Group. Changing outcomes in percutaneous coronary interventions: a study of 34 752 procedures in northern New England, 1990 to 1997. J Am Coll Cardiol. 1999; 34: 674–680.
Williams DO, Holubkov R, Yeh W, Bourassa MG, Al-Bassam M, Block PC, Coady P, Cohen H, Cowley M, Dorros G, Faxon D, Holmes DR, Jacobs A, Kelsey SF, King SB 3rd, Myler R, Slater J, Stanek V, Vlachos HA, Detre KM. Percutaneous coronary intervention in the current era compared with 1985–1986: the National Heart, Lung, and Blood Institute Registries. Circulation. 2000; 102: 2945–2951.
Sobel BE, Frye R, Detre KM. Bypass Angioplasty Revascularization Investigation 2 Diabetes Trial. Burgeoning dilemmas in the management of diabetes and cardiovascular disease: rationale for the Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI 2D) Trial [published correction appears in Circulation. 2003;108:500]. Circulation. 2003; 107: 636–642.
Cutlip DE, Chhabra AG, Baim DS, Chauhan MS, Marulkar S, Massaro J, Bakhai A, Cohen DJ, Kuntz RE, Ho KK. Beyond restenosis: five-year clinical outcomes from second-generation coronary stent trials. Circulation. 2004; 110: 1226–30.
We analyzed the in-hospital and long-term outcome of patients who underwent primary coronary intervention at the Mayo Clinic over 25 years. The study population was divided into 4 groups, with the most recent group (2003 to 2004) treated with drug-eluting stents. Despite the fact that recent groups were significantly older, sicker, and had a greater prevalence of comorbid conditions, heart failure, and previous revascularization than the earlier groups, we noted higher procedural success (94%) and significant reduction in in-hospital mortality and almost 90% reduction in the need for emergency bypass surgery. Better adherence to currently recommended evidence-based medications for secondary prevention was seen in the recent time periods. After adjustment, a significant reduction in follow-up mortality, myocardial infarction, death, or revascularization was noted in recent time periods. Longer follow-up is needed to truly assess the effect of drug-eluting stents on the incidence of myocardial infarction related to late stent thrombosis. Despite higher-risk profiles of patients who underwent primary coronary intervention in recent time periods, procedural success as well as in-hospital and long-term outcomes improved significantly over the last 25 years.