Twenty-Year Evolution of Percutaneous Coronary Intervention and Its Impact on Clinical OutcomesCLINICAL PERSPECTIVE
A Report From the National Heart, Lung, and Blood Institute–Sponsored, Multicenter 1985–1986 PTCA and 1997–2006 Dynamic Registries
Background— Percutaneous coronary intervention (PCI) has undergone rapid progress, both in technology and adjunct therapy. However, documentation of long-term temporal trends in relation to contemporary practice is lacking.
Methods and Results— We analyzed PCI use and outcomes in 8976 consecutive patients in the multicenter, National Heart, Lung, and Blood Institute–sponsored 1985–1986 percutaneous transluminal coronary angioplasty (PTCA) and 1997–2006 Dynamic Registries waves (wave 1: 1997–1998, bare-metal stents; wave 2: 1999, uniform use of stents; wave 3: 2001–2002, brachytherapy; waves 4 and 5: 2004–2006, drug-eluting stents). Patients undergoing PCI in the recent waves were older and more often reported comorbidities than those in the balloon era. PCI was more often performed for acute coronary syndromes and, in spite of the greater disease burden, was more often selective. Procedural success was achieved and maintained more often in the stent era. Significant reductions were observed in in-hospital rates (%) of myocardial infarction (PTCA Registry: 4.9; wave 1, 2.7; wave 2, 2.8; wave 3, 1.9; wave 4, 2.6; wave 5, 2; Ptrend<0.001) and emergency coronary artery bypass surgery (PTCA Registry: 3.7; wave 1, 0.4; wave 2, 0.4; wave 3, 0.3; wave 4, 0.4; wave 5, 0; Ptrend<0.001). Compared with the PTCA Registry, risk for repeat revascularization (31 to 365 days after index PCI) was significantly lower in the dynamic waves (adjusted hazard ratio: wave 1, 0.72; wave 2, 0.51; wave 3, 0.51; wave 4, 0.30; wave 5, 0.36; P<0.05 for all).
Conclusions— Percutaneous interventions, in the last 2 decades, have evolved to include more urgent, comorbid cases, despite achieving high success rates with significantly reduced need for repeat revascularization.
Received October 2, 2008; accepted December 9, 2008.
Percutaneous coronary intervention (PCI) was first performed in 19771 and since then has gained rapid acceptance as a treatment option for coronary artery disease. The field has witnessed rapid technological advancements with a parallel progress in adjunct and secondary pharmacological therapy. Concomitantly, the profile of patients (and lesions) undergoing PCI has become heterogeneous to the extent that it now often includes “off-label” or “untested” circumstances.2 As a result, in spite of the dramatic reduction in the need for repeat procedures, questions regarding the safety and appropriate utilization of the procedure have emerged.3 A time-sensitive appraisal of PCI use and outcomes is thus warranted to place these concerns in a historical perspective. Prior studies of temporal trends in PCI were conducted in early technology eras,4,5 or based on single-center experience.6 Our objectives, therefore, are to document long-term trends in patient and procedural characteristics, and to compare procedural outcomes using data from the multicenter, National Heart, Lung, and Blood Institute (NHLBI)–sponsored percutaneous transluminal coronary angioplasty (PTCA)7 and Dynamic Registries4. Taken together, these registries span 2 decades of clinical practice in North America and as such, range from the balloon angioplasty era to contemporary use of drug-eluting stents.
Editorial see p 1
Clinical Perspective see p 6
The NHLBI-sponsored PTCA and Dynamic Registries were prospective multicenter studies that enrolled consenting patients undergoing PCI from clinical centers in North America (Appendix 1).4,7 The 1985–1986 PTCA Registry aimed at recruiting 2000 consecutive patients undergoing angioplasty for the first time. The NHLBI-sponsored Dynamic Registry was initiated following the advent of bare-metal stents (BMS) and recruited patients in waves (1: 1997–1998; 2: 1999; 3: 2001–2002; 4: 2004; 5: 2006), enriched with an oversampling of women and minorities. Consecutive enrollment at each center ended once 125 white men had been enrolled at that site or 1600 white patients were enrolled across all sites. Then, consecutive women and minorities continued to be enrolled at each center until 200 white patients had been enrolled at that site or 1600 white patients were enrolled across all sites. Consecutive minority patients, men and women, continued to be enrolled until 2000 patients had been enrolled across all sites. Information on patient characteristics and detailed angiographic and procedural data were ascertained at baseline. Written informed consent was obtained from participants to be contacted annually after discharge for health status information. Data on events including death from any cause, myocardial infarction (MI), coronary artery bypass surgery (CABG), and any nonstaged repeat PCI were ascertained and hospital records examined to ensure consistency with protocol definitions. Specifically, the definition of MI was revised to match prevailing expert consensus—in the PTCA Registry, it was defined as evidence of ≥2 of the following: (1) typical chest pain >20 minutes not relieved by nitroglycerin, (2) serial ECG recordings showing changes from baseline or serially in ST-T and/or Q-waves in ≥2 contiguous leads, or (3) serum enzyme elevation of CK-MB >5% of total CK (total CK >2× normal; LDH subtype 1>LDH subtype 2); in the Dynamic Registry, cardiac troponin level was incorporated as a major criteria. The study protocol was approved by the Institutional Review Boards of the coordinating center (University of Pittsburgh) and all the clinical sites involved. The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the article as written.
For sake of comparability between both registries, only consecutive patients with no prior PCI were included in this analysis. Temporal trends in patient and procedural characteristics were assessed across the waves using the Cochran-Armitage test for dichotomous variables8 and the Jonckeheere-Terpsta test for continuous and nominal/ordinal variables9. Kaplan–Meier (KM) estimates of event rates for death, combined death/MI, repeat PCI, CABG, and repeat revascularizations (repeat PCI+CABG) were compared using log rank statistics. Risk of these events was assessed using Cox regression models for the Dynamic Registry waves with the PTCA Registry as reference. Hazard ratios and 95% CI were estimated at 2 time points, Early (≤30 days) and Late (31 to 365 days), following the index procedure. Patients, whose first repeat PCI or CABG occurred after 30 days from initial PCI, were censored for early events; analysis of late events included only those who did not undergo any repeat procedure within the first 30 days. For multivariable analysis, only those data available and showing significant trend across the 5 waves were considered; baseline ejection fraction was not included due to substantial missing data (35%). For the final model, the “wave” variable was forced to stay in the model, and the remaining covariates were selected using standard stepwise procedure (Pentry≤0.15, Pstay≤0.10). The list of variables adjusted for in the final models is provided in Appendix 2. All analyses were performed with SAS version 9.1 (SAS Institute Inc).
Patient and Procedural Characteristics
Consecutive patients undergoing PCI in the recent waves were older (mean age in years, from older to recent cohorts: 58, 62, 63, 64, 64, 64; Ptrend<0.001) and more often women (% women, from older to more recent cohorts: 26, 32, 32, 35, 31, 33; Ptrend<0.001) when compared with the PTCA Registry. Additionally, more patients in the Dynamic waves reported concomitant comorbidities (hypertension, diabetes, and severe noncardiac conditions) and prior bypass surgery (Table 1). Acute coronary syndrome (unstable angina or acute MI, which includes both ST-segment elevation and non-ST segment elevation MI) remained the most common reason for revascularization over time, with a concomitant rise in nonelective procedures. Baseline disease burden, as reflected by mean number of lesions and vessels diseased, was greater in the Dynamic Registry. Procedural attempts, on the other hand, more often involved single vessels and lesions, both in the overall cohort (Table 2) and specifically in the setting of multivessel disease (Table 3). The profile of attempted lesions was more severe in the Dynamic Registry with increased attempts on calcified lesions (% from older to more recent cohorts: 11, 30, 27, 24, 27, 33, Ptrend<0.001) and thrombotic lesions (from older to more recent cohorts: 14, 27, 30, 21, 21, 22; Ptrend0.001); this pattern persisted even within the Dynamic Registry with increased attempts on Type C lesions (wave 1, 20%; wave 2, 16%; wave 3, 19%; wave 4, 22%; wave 5, 30%; Ptrend<0.001).
Use of stents, either alone or in conjunction with balloons, increased from 70% in wave 1 (only bare metal) to 97% in wave 5. Use of drug-eluting stents (DES), introduced in wave 4, increased from 71% to 88% in wave 5; atherectomy and brachytherapy were used in fewer than 2% of patients. Procedural use of glycoprotein IIb/IIIa inhibitors increased from 25% in wave 1 to 55% in wave 3 to then dip to 41% in wave 5, coinciding with the introduction of bivalirudin. Procedural use of ticlopidine/clopidogrel increased from 52% in wave 1 to 83% in wave 5 (Ptrend<0.001).
Total angiographic and procedural success were achieved and maintained more often in the Dynamic Registry (Table 4). Rates of in-hospital MI (from older to more recent cohorts: 4.9%, 2.7%, 2.8%, 1.9%, 2.6%, 2.0%; Ptrend<0.001) and emergency CABG (from older to more recent cohorts: 3.7%, 0.4%, 0.4%, 0.3%, 0.4%, 0.0%; Ptrend<0.001) were significantly lower in the more recent waves. On the other hand, compared with the PTCA Registry, in-hospital mortality rates were marginally higher in most of the recent waves with the exception of wave 5 (from older to more recent cohorts: 1.4%, 1.9%, 1.8%, 1.3%, 2.0%, 0.7%; Ptrend0.34; probability value for wave 4 versus wave 5: 0.01). Mean duration of hospital stay (days) decreased over time (from older to more recent cohorts: 4.0, 2.7, 2.6, 2.4, 2.2, 2.0; Ptrend<0.001). Discharge use of recommended secondary pharmacological therapy (aspirin, β-blockers, lipid-lowering therapy, antiplatelet agents) increased significantly across the waves (Table 4).
Early (≤30 Days) and Late (31 to 365 Days) Follow-Up Events
Cumulative unadjusted mortality rates over 1 year were higher in the Dynamic Registry waves, when compared with the PTCA Registry in the overall cohort (Figure 1). This pattern persisted in patients undergoing PCI for both stable angina (from older to recent waves: 2.2%, 3.2%, 3.7%, 1.0%, 4.8%, 4.9%; Plog rank=0.06) as well as acute coronary syndromes (from older to recent waves: 4.2%, 6.0%, 5.7%, 5.2%, 5.7%, 4.3%; Plog rank=0.23). Risk of mortality was assessed both within and after 30 days following index PCI for the Dynamic Registry waves, using PTCA Registry as reference. The adjusted risk of mortality in the Dynamic waves was nonsignificantly lower when compared with the PTCA Registry in both follow-up periods (Figure 2). The adjusted risk for death/MI, on the other hand, reached statistical significance for some recent waves when compared with the PTCA registry (Figure 2A and 2B).
The overall rates of repeat revascularization (repeat PCI or CABG) over 1 year decreased substantially from 28% in the PTCA Registry to 11% in wave 5 (Plog rank<0.001). However, reduction in early CABG rates was initially steep and followed by a plateau in the stent era, whereas rates of late CABG successively reduced over time (Figure 3). On the other hand, cumulative rates of early repeat PCI, though small in number, increased, whereas the need for late repeat PCI was significantly reduced across the cohorts. Adjusted risk for repeat revascularization (CABG or repeat PCI) was significantly lower for the Dynamic waves, when compared with the PTCA Registry, both in the early (hazard ratios [95% CI]: wave 1, 0.44 [0.32 to 0.60]; wave 2, 0.30 [0.21 to 0.45]; wave 3, 0.27 [0.18 to 0.41]; wave 4, 0.36 [0.26 to 0.52]; wave 5, 0.26 [0.18 to 0.39]; P<0.05 for all) as well as late follow-up periods (hazard ratios [95% CI]: wave 1, 0.72 [0.61 to 0.85]; wave 2, 0.51 [0.42 to 0.63]; wave 3, 0.51 [0.41 to 0.62]; wave 4, 0.30 [0.24 to 0.38]; wave 5, 0.36 [0.28 to 0.45]; P<0.05 for all) (also see Supplemental Figure).
Angioplasty, when introduced in 1977, relied solely on balloons for dilatation of coronary arteries. Although this in itself revolutionized the treatment of coronary artery disease, the advent of BMS caused a drastic expansion in the type of patients undergoing the procedure. However, as problems of in-stent restenosis soon became apparent, alternative techniques including atherectomy and brachytherapy were introduced, only to be followed by the more successful drug-eluting stents. Our report provides a snapshot of this evolution and its impact on procedural outcomes over a 20-year period in 2 large, prospective, multicenter registries of clinical practice in North America. These findings are especially noteworthy given that previous reports of temporal trends have been restricted to specific technology eras4,5 or are single-center experiences.6
PCI use in contemporary practice has expanded to include more patients with severe comorbidities, acute coronary syndromes, and multivessel disease. The lower rates of prior MI in the more recent waves, especially in light of the concomitant trend of higher rates of PCI performed for acute coronary syndromes over time, is noteworthy. Although these trends could potentially be a reflection of the revised guidelines favoring primary PCI instead of fibrinolysis in acute syndromes, they are also in keeping with the reported decline in annual rates of recurrent infarctions in community-based settings.10 Attempted lesions, in the recent waves, were more often determined to be calcified or thrombotic, when compared with the early cohorts. These trends, however, could be a reflection of improved or better imaging techniques over time, rather than a true increase in these characteristics. The concept of complete revascularization, which stemmed from early CABG studies, appears to have given way to a more selective approach over time as evidenced by the predominance of single-vessel attempts in multivessel disease patients. One plausible explanation is the greater use of PCI for acute conditions where functional (and not anatomic) revascularization is of priority. A prior report from this Registry11 has also shown that patients with multivessel disease, in whom complete revascularization was unachievable, were older with more comorbidities and lower ejection fraction, when compared with those in whom complete revascularization was achieved or selective revascularization was the method of choice.
Temporal Trends in Procedural Safety
The present report demonstrates a dramatic improvement in procedural outcomes (high success rates and reduced need for in-hospital bypass surgery) with the advent of stents. Furthermore, even within the stent era, there is a significant reduction in the initially high rates of dissections and abrupt closures. Although this certainly reflects the huge impact of technology, it also underscores the importance of operator training and better techniques. Improvements in in-hospital outcomes are also seen to extend over 1 year (lower rates of death/MI and CABG) and are congruent with previously published reports.4,6 Nonetheless, the higher crude mortality rates in the Dynamic Registry, in the overall cohort and by primary indication, is reflective of sicker profile (older patients, more comorbidities, greater disease burden) of patients enrolled in these waves. Prior comparison of 1-year outcomes between patients undergoing PCI for stable versus unstable angina showed little change in mortality rates in the latter cohort over the past 16 years.12 In another report of 2839 patients with complex lesions (defined as a lesion showing evidence of thrombus, calcification, bifurcation or ostial location, or chronic occlusion), both in-hospital and 1-year mortality rates were higher, compared with attempts on simpler lesions.13 Moreover, reports of stent-thrombosis with DES use has led to concerns of use and timing of dual antiplatelet therapy14 and the need to focus on cause-specific, rather than overall, mortality.15 Thus, in spite of the marked overall improvement in the field, certain high-risk subsets continue to pose major challenges and warrant closer attention.16
Temporal Trends in Effectiveness
Development of devices in PCI was primarily aimed at reducing the need for repeat interventions, be it surgical or percutaneous. Our cohorts are representative of key devices available in the respective time periods: PTCA Registry (balloons), wave 1 (early use of BMS), wave 2 (uniform use of BMS), wave 3 (brachytherapy), wave 4 (early use of DES), and wave 5 (established use of DES). The sustained reduction in the need for repeat revascularization, therefore, truly underscores the progress made in the field. The greater need for “early” repeat PCI paralleling the precipitous drop in CABG rates, small sample size notwithstanding, deserves particular mention. We believe that while in the present era, CABG was a more powerful option in the event of failed index PCIs, the advent of stents caused a shift in favor of using PCI for repeat revascularization in these cases. Alternatively, PCI has been increasingly performed in sicker patients with multivessel disease but using a selective treatment strategy while achieving 100% procedural success (see Supplemental Table). The increased use of early PCI, therefore, may be a related fall out of this greater disease burden.
Secondary Medical Therapy in PCI
Pharmacological therapy in atherosclerosis has undergone major improvements over time and their salutary effects in the setting of coronary revascularization have been well documented.17–20 More recently, a comparison of an initial strategy of PCI and optimal medical therapy versus medical therapy alone, in 2287 patients with stable coronary artery disease, revealed no difference in the rates of the composite end point of death and nonfatal MI.21 Although this highlights the notable progress in the field of medical therapy, it also reiterates the need for systemic management of atherosclerosis. After all, PCI treats only angiographically visible stenoses and not the underlying disease mechanism responsible for new lesions and resultant ischemic events. To this end, the increase in the discharge use of evidence-based therapy (aspirin, β-blockers, cholesterol-lowering agents, and antiplatelet therapy), as observed in our report, is encouraging and reflects an improved awareness of the importance of secondary prevention.
Although limitations inherent to use of a registry database must be acknowledged, enrollment of consecutive patients with no exclusion criteria permits representation of real world practice and allows for the timely evaluation of safety and effectiveness. The majority of centers participating in the registries were medium to high-volume hospitals and more often academic centers, thus limiting the generalizability of our findings. The primary objective of this analysis was to assess temporal trends using the “wave” variable as a marker of change in both technology as well as secondary therapy in that particular time period. Therefore, we did not specifically adjust for or evaluate these factors in multivariable models. Also, ascertainment of data was refined in each wave to incorporate prevailing concerns in the field. Thus, only those variables available in all cohorts were considered for multivariable analysis. Information on periprocedural myocardial enzymes was lacking in both the registries, thus, limiting our ability to assess related impact on events.
Our report from the large, prospective, multicenter, NHLBI-sponsored 1985–1986 PTCA (balloon era) and 1997–2006 Dynamic Registries (BMS and DES era) documents the rapid evolution in PCI as a treatment option for atherosclerosis. PCI use, over time, has expanded to include more patients with severe comorbidities, acute coronary syndromes and complex, multivessel disease. However, in spite of the sicker patient profile and greater disease burden at baseline, procedural success was achieved more often and with significant reductions in immediate (in-hospital rates of MI and emergency CABG) as well as 1-year (reduced need for repeat procedures) outcomes. Further studies, however, are warranted to elucidate the impact, or lack thereof, of PCI on long-term mortality.
Sources of Funding
This study was supported by grants (HL-33292-14 through HL-33292-22 from the NHLBI (Bethesda, Md).
Dr Wilensky received grant support from Boston Scientific and has ownership interests in Johnson & Johnson. Dr Williams received grant support from Cordis Corporation, Boston Scientific, and Abbott Vascular and serves on the consultant/advisory board for Cordis Corporation.
Laskey WK, Yancy CW, Maisel WH. Thrombosis in coronary drug-eluting stents: report from the meeting of the Circulatory System Medical Devices Advisory Panel of the Food and Drug Administration Center for Devices and Radiologic Health, December 7–8, 2006. Circulation. 2007; 115: 2352–2357.
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 III, 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.
Laskey WK, Williams DO, Vlachos HA, Cohen H, Holmes DR, King SB III, Kelsey SF, Slater J, Faxon D, Al Bassam M, Block E, Detre KM. Changes in the practice of percutaneous coronary intervention: a comparison of enrollment waves in the National Heart, Lung, and Blood Institute (NHLBI) Dynamic Registry. Am J Cardiol. 2001; 87: 964–969.
Singh M, Rihal CS, Gersh BJ, Lennon RJ, Prasad A, Sorajja P, Gullerud RE, Holmes DR Jr. Twenty-five-year trends in in-hospital and long-term outcome after percutaneous coronary intervention: a single-institution experience. Circulation. 2007; 115: 2835–2841.
Detre K, Holubkov R, Kelsey S, Bourassa M, Williams D, Holmes D Jr, Dorros G, Faxon D, Myler R, Kent K. One-year follow-up results of the 1985–1986 National Heart, Lung, and Blood Institute’s Percutaneous Transluminal Coronary Angioplasty Registry. Circulation. 1989; 80: 421–428.
Margolin BH. Test for trend in proportions. In: Klotz S, Johnson NL, eds. Encyclopedia of Statistical Sciences. New York: Wiley; 1988: 334–336.
Pirie W. Jonckheere tests for ordered alternatives. In: Klotz S, Johnson NL, eds. Encyclopedia of Statistical Sciences. New York: Wiley; 1983: 315–318.
Srinivas VS, Selzer F, Wilensky RL, Holmes DR, Cohen HA, Monrad ES, Jacobs AK, Kelsey SF, Williams DO, Kip KE. Completeness of revascularization for multivessel coronary artery disease and its effect on one-year outcome: a report from the NHLBI Dynamic Registry. J Interv Cardiol. 2007; 20: 373–380.
Grines CL, Bonow RO, Casey DE Jr, Gardner TJ, Lockhart PB, Moliterno DJ, O'Gara P, Whitlow P. Prevention of premature discontinuation of dual antiplatelet therapy in patients with coronary artery stents: a science advisory from the American Heart Association, American College of Cardiology, Society for Cardiovascular Angiography and Interventions, American College of Surgeons, and American Dental Association, with representation from the American College of Physicians. J Am Dent Assoc. 2007; 138: 652–655.
Nordmann AJ, Briel M, Bucher HC. Mortality in randomized controlled trials comparing drug-eluting vs. bare metal stents in coronary artery disease: a meta-analysis. Eur Heart J. 2006; 27: 2784–2814.
Holmes DR Jr, Williams DO. Catheter-based treatment of coronary artery disease: past, present, and future. Circ Cardiovasc Intervent. 2008; 1: 60–73.
Mehta SR, Yusuf S, Peters RJ, Bertrand ME, Lewis BS, Natarajan MK, Malmberg K, Rupprecht HJ, Zhao F, Chrolavicius S, Copland I, Fox KA. Effects of pretreatment with clopidogrel and aspirin followed by long-term therapy in patients undergoing percutaneous coronary intervention: the PCI-CURE study. Lancet. 2001; 358: 527–533.
Kasai T, Miyauchi K, Kurata T, Satoh H, Ohta H, Tanimoto K, Kawamura M, Okazaki S, Yokoyama K, Kojima T, Akimoto Y, Daida H. Long-term (11-year) statin therapy following percutaneous coronary intervention improves clinical outcome and is not associated with increased malignancy. Int J Cardiol. 2007; 114: 210–217.
Boden WE, O'Rourke RA, Teo KK, Hartigan PM, Maron DJ, Kostuk WJ, Knudtson M, Dada M, Casperson P, Harris CL, Chaitman BR, Shaw L, Gosselin G, Nawaz S, Title LM, Gau G, Blaustein AS, Booth DC, Bates ER, Spertus JA, Berman DS, Mancini GB, Weintraub WS. Optimal medical therapy with or without PCI for stable coronary disease. N Engl J Med. 2007; 356: 1503–1516.
We analyzed temporal trends in the use of de novo percutaneous coronary intervention (PCI) and the associated in-hospital and 1-year outcomes in 8976 consecutive patients from the multicenter, National Heart, Lung, and Blood Institute–sponsored PTCA and Dynamic Registries [PTCA registry: 1985–1986, balloon angioplasty; Dynamic Registry wave 1: 1997–1998, bare-metal stents; wave 2: 1999, uniform use of stents; wave 3: 2001–2002, brachytherapy; waves 4 and 5: 2004–2006, drug-eluting stents]. PCI use, over time, has expanded to include more patients with severe comorbidities, acute coronary syndromes, and complex, multivessel disease. However, in spite of the sicker patient profile and greater disease burden at baseline, procedural success was achieved more often and with significant reductions in the in-hospital rates of myocardial infarction and emergency coronary artery bypass surgery. Considerable improvements were also observed in the use of secondary evidence-based pharmacological therapy after PCI. Multivariable analyses of outcomes over 1 year show significant, sustained reduction in the need for repeat revascularization but little impact on mortality. PCI, in the last 2 decades, has been increasingly applied to “sicker” patients and yet has achieved substantial improvements in procedural safety and effectiveness (reduced need for repeat procedures). Further studies are warranted to elucidate the impact, or lack thereof, of PCI on long-term mortality.
Guest Editor for this article was Eric R. Bates, MD.
The online-only Data Supplement is available at http://circinterventions.ahajournals.org/cgi/content/full/CIRCINTERVENTIONS.108.825323/DC1.