- Email: [email protected]
Importance of the Postdischarge Interval in Assessing Major Adverse Clinical Event Rates Following Percutaneous Coronary Intervention
Importance of the Postdischarge Interval in Assessing Major Adverse Clinical Event Rates Following Percutaneous Coronary Intervention Warren K. Laskey, MD, Faith Selzer, PhD, Alice K. Jacobs, MD, Howard A. Cohen, MD, David R. Holmes, Jr., MD, Robert L. Wilensky, MD, Katherine M. Detre, MD, DrPH, and David O. Williams, MD, on behalf of the NHLBI Dynamic Registry Investigators In-hospital major adverse clinical event (MACE) rates after percutaneous coronary intervention serve as benchmarks of performance. However, accelerated clinical pathways, decreased lengths of stay, and potential delayed effects of percutaneous coronary intervention may result in an underestimation of this traditional measurement of outcome. Records from patients in the first 3 waves of the National Heart, Lung, and Blood Institute’s Dynamic Registry (n ⴝ 6,676) were reviewed for rates of composite in-hospital MACEs (death, myocardial infarction, and any repeat target vessel revascularization) and postdischarge MACEs (death, myocardial infarction, repeat hospitalization, and repeat target vessel revascularization) through 30 days. Rates for each composite MACE were compared across waves to assess changes over time. Predictors of each MACE category were identified using multivariate analysis. In-hospital MACE decreased significantly (5.4% of wave 1, 4.9% of wave 2,
3.1% of wave 3, p <0.001), whereas stent implantation increased significantly (67.5% of wave 1, 79.1% of wave 2, 86.2% of wave 3, p <0.001). Postdischarge MACE through 30 days remained unchanged (5.1% of wave 1, 5.1% of wave 2, 4.8% of wave 3, p ⴝ 0.6). Mean length of stay decreased (2.7 days for wave 1, 2.2 days for wave 3, p <0.001). Disparate clinical, procedural, and angiographic factors were associated with each MACE. Postdischarge MACE rates through 30 days comprise a significant and unchanging fraction of overall procedurally related MACE rates despite improving in-hospital outcomes. Most postdischarge events derive from pathology related to the controlled vessel. A 30day MACE rate may serve as a more comprehensive measurement of procedural outcome. 䊚2005 by Excerpta Medica Inc. (Am J Cardiol 2005;95:1135–1139)
A
on postdischarge adverse events,14 –16 the potential for inaccurate/incomplete measurements of procedural outcomes exists. Accordingly, we determined rates of in-hospital major adverse clinical events (MACEs) and MACEs through 30 days in all patients who underwent PCI and were entered in the National Heart Lung and Blood Institute (NHLBI) Dynamic Registry from 1997 to 2002.
ssessment of periprocedural and in-hospital outcomes after percutaneous coronary intervention (PCI) has served as an essential measurement of quality and proficiency.1– 4 However, at low event rates, considerable uncertainly exists in the accuracy and utility of this measurement of performance.5–7 Further, a shorter hospital length of stay will decrease the period of exposure during which the risk of procedurally related adverse events can be identified. With decreasing rates of in-hospital morbidity and mortality after PCI8 –13 and a paucity of population-based data From the Division of Cardiology, University of New Mexico School of Medicine, Albuquerque, New Mexico; the Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania; the Division of Cardiology, Boston University School of Medicine, Boston, Massachusetts; the Division of Cardiology, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania; the Cardiovascular Division, Mayo Clinic, Rochester, Minnesota; the Division of Cardiology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania; and Brown University School of Medicine, Providence, Rhode Island. This study was supported in part by Grant HL 33992-14 from the National Heart, Lung and Blood Institute, Bethesda, Maryland. Manuscript received October 26, 2004; revised manuscript received and accepted January 11, 2005. Address for reprints: Warren K. Laskey, MD, Division of Cardiology, Department of Internal Medicine, MSC10-5550, 1 University of New Mexico, Albuquerque, New Mexico 87131-0001. E-mail: [email protected]. ©2005 by Excerpta Medica Inc. All rights reserved. The American Journal of Cardiology Vol. 95 May 15, 2005
METHODS Details of the purpose, structure, and function of the NHLBI Dynamic Registry have been previously described.12,17 Consecutively enrolled patients in wave 1 (n ⫽ 2,524) were recruited from 1997 to 1998, those in wave 2 (n ⫽ 2,105) were recruited in 1999, and those in wave 3 (n ⫽ 2,047) were recruited from 2001 to 2002. Patient-level data from the 3 waves were pooled, and component and composite rates of in-hospital MACEs (death, myocardial infarction, and any repeat target vessel revascularization procedure) were recorded. Component and composite rates of MACEs (death, myocardial infarction, nonscheduled rehospitalization, and repeat target vessel revascularization) from time of hospital discharge through 30 days in all patients alive at discharge were also recorded. 0002-9149/05/$–see front matter doi:10.1016/j.amjcard.2005.01.038
1135
TABLE 1 Differences in Clinical, Demographic, and Procedural Characteristics by Enrollment Wave in the NHLBI Dynamic Registry
Demographic and clinical factors Mean age (yrs)* Age ⬎65 years White Black Other Body mass index (kg/m2)* Previous PCI* None ⱖ1 Severe noncardiac co-morbitity* Diabetes mellitus Insulin prescription Oral medications† Congestive heart failure Hypertension Hyperlipidemia* Angiographic features Abnormal left ventricular function† No. of coronary arteries narrowed* 1 2 3 Any total occlusion† Mean no. of significant narrowings* Procedural features Indications for revascularization‡ Stable angina pectoris Unstable angina pectoris Acute myocardial infarction ⱕ14 days Circumstances of procedure* Elective Urgent Emergency Use of platelet glycoprotein receptor inhibitor* Thienopyridine use* Overall stent use* Lesion length (mm)*
Wave 1 (n ⫽ 2,524)
Wave 2 (n ⫽ 2,105)
Wave 3 (n ⫽ 2,047)
63 45% 80% 8% 12% 28
63 46% 78% 12% 10% 29
64 49% 79% 13% 8% 29
72% 28% 30% 28% 9% 14% 10% 59% 61%
69% 31% 34% 28% 9% 15% 10% 64% 63%
67% 33% 39% 30% 9% 16% 13% 75% 71%
27%
27%
31%
42% 33% 25% 39% 2.9
42% 32% 25% 36% 2.9
25% 47% 19%
20% 45% 26%
66% 23% 11% 24%
57% 32% 11% 32%
50% 67% 12.4
41% 79% 13.0
*p ⬍0.001; †p ⬍0.05; ‡p ⬍0.01.
Patients participating in the NHLBI Dynamic Registry consented to the collection of in-hospital and postdischarge data after the index procedure. Study coordinators at each site contacted patients at 1 year to obtain vital status and information regarding intercurrent hospitalizations, cardiovascular events, or procedures. Dates corresponding to all reported outcomes were recorded and confirmed by requesting relevant records. Whenever possible, angiographic data pertaining to repeat PCI were collected and reviewed. Accordingly, postdischarge follow-up rates were 97.5%, 94.9%, and 90.2% for waves 1, 2, and 3, respectively. Statistical analysis: Continuous data are summarized as means ⫾ 1 SD. Categorical variables are summarized as counts or percentages. Comparisons of (unadjusted) event rates across waves of the NHLBI Dynamic Registry were accomplished using a contingency table and Mantel-Haenszel statistics. All rates are reported as point estimates with their correspond1136 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 95
ing 95% confidence intervals. Rates were also adjusted for imbalances in important covariates among waves using general linear modeling methods. Clinical, demographic, and procedural factors independently associated with in-hospital and 30-day MACE rates were assessed with multivariable logistic regression. Odds ratios and their corresponding 95% confidence intervals are reported. Cumulative event rates were calculated with Kaplan-Meier methods and compared with log-rank statistics. Statistical significance was defined as a p value ⬍0.05. The study sample possessed 99% power to detect a difference at the 0.05 level in event rates of the magnitude reported.
RESULTS
Table 1 presents all salient, statistically significant (p ⱕ0.05) differences in clinical, demographic, and procedural characteristics across waves in the NHLBI Dynamic Registry. Over successive periods, patients tended to 20% be older, have more severe cardiac and 42% noncardiac co-morbidities, more fre28% quently underwent previous PCI, were more likely to present with an acute coronary syndrome and undergo ur49% 40% gent or emergency PCI, more fre10% quently received platelet glycoprotein 53% IIb/IIIa receptor antagonists during PCI, and were more frequently man68% 86% aged with bare metal stent implanta13.4 tion. There were no drug-eluting stents implanted during this period. Mean length of stay decreased significantly over the 3 waves (2.65 ⫾ 3.5 days for wave 1, 2.36 ⫾ 3.9 days for wave 2, 2.21 ⫾ 3.2 days for wave 3, p ⬍0.001). Neither angiographic success rate (92.2% for wave 1, 92.9% for wave 2, 94% for wave 3) nor procedural success rate (94.7% for wave 1, 95.7% for wave 2, 96.2% for wave 3) differed significantly over the 3 waves. Table 2 presents rates of patient and composite adverse in-hospital and 30-day postdischarge outcomes for each wave. In contrast to the significant decrease in in-hospital MACE rate over the 3 waves (5.4% for wave 1, 4.9% for wave 2, 3.1% for wave 3, p ⬍0.001), there was no significant difference in MACE rate at 30 days after discharge. After adjusting for age, vessel disease, co-morbid conditions, diabetes, and acuity, 30-day postdischarge event rates were 5.4% for wave 1, 5.0% for wave 2, and 4.6% for wave 3 (p ⫽ 0.62). Notably, postdischarge MACE rates at 30 days were of the same order of magnitude as the in-hospital MACE rates. Thus, the significant decrease in cumulative MACE rates at 30 days (10.4% for wave 1, 9.5% for wave 2, 7.8% for wave 3, p ⫽ 0.009) 38% 31% 30% 37% 3.1
MAY 15, 2005
TABLE 2 Major Adverse Clinical Event Rates by Enrollment Wave in the NHLBI Dynamic Registry Wave 1 (%) Event Death Death/myocardial infarction Target vessel revascularization Composite MACE rate (95% CI)
Wave 2 (%)
Wave 3 (%)
In-hospital
30 Day
In-hospital
30 Day
In-hospital
30 Day
1.6 4.2 1.3 5.4 (4.5–6.2)
0.3 0.6 2.3 5.1 (4.2–6.1)
1.3 4.0 1.1 4.9 (3.9–5.8)
0.5 0.9 1.7 5.0 (4.0–6.0)
1.0 2.7 0.5 3.1 (2.4–3.9)
0.6 0.9 1.6 4.9 (3.9–6.0)
CI ⫽ confidence interval.
TABLE 3 Categorization and Patient-specific Frequency of 30-day Postdischarge Events Category of Events 1 Event Death Myocardial infarction PCI Coronary bypass surgery Hospitalization 2 Events Death and myocardial infarction Death and hospitalization Myocardial infarction and PCI Myocardial infarction and coronary bypass surgery PCI and hospitalization Coronary bypass surgery and hospitalization 3 Events Death, PCI, and myocardial infarction PCI, coronary bypass surgery, and myocardial infarction PCI, coronary bypass surgery, and hospitalization PCI, myocardial infarction, and hospitalization
No. of Patients 21 9 30 1 122 3 4 9 2 63 14 2 2 1 1
TABLE 4 Type and Number of 30-Day Postdischarge Events Event Type Death Myocardial infarction Coronary bypass surgery PCI Hospitalization Sum
No. of Events 30 27 20 110 216 403
was driven entirely by the decrease in the in-hospital component. Table 3 lists patient-specific events in hierarchical fashion through 30 days after hospital discharge, and Table 4 lists cumulative events. There were 403 events in 284 patients. A significant minority of patients (18.6%) died or sustained a myocardial infarction; 50% of patients underwent repeat hospitalization and 32% of patients required repeat revascularization of the target vessel. Tables 5 and 6 present significant independent predictors of in-hospital and 30-day MACEs, respectively. There was little overlap in the factors associated with each outcome. Factors associated with in-
hospital MACEs tended to reflect the extent and severity of the underlying coronary (and noncoronary) disease, whereas factors associated with 30-day MACEs tended to reflect procedural complexity. Interestingly, the use of balloon angioplasty alone (vs stent implantation) was associated with a lower risk of postdischarge MACEs.
DISCUSSION In this study, we observed a decreasing in-hospital MACE rate and an unchanging 30-day postdischarge MACE rate in patients who underwent PCI from 1997 to 2002 in the NHLBI Dynamic Registry. Over this period there was also a significant decrease in average length of hospital stay. Under these circumstances, a failure to record clinically significant events shortly after hospital discharge may result in inaccurate assessments of procedural safety and outcome. The present results with respect to decreasing inhospital MACE rates in the current PCI era agree with those of numerous previous reports.8 –13 However, the reporting of MACE rates after PCI at 30 days has generally been limited to the randomized clinical trial literature, with a paucity of similar data available from less select patient populations.14 –16,18 Further, claimsbased datasets15,18 generally lack important ancillary clinical information and confine the end point at 30 days to mortality and/or coronary artery bypass surgery. The value of obtaining more complete follow-up data in consecutive patients under “real-world” conditions stems not only from differences in event rates between controlled clinical trials and less restricted clinical practice but also from the need to more completely characterize unanticipated and/or delayed procedurally related adverse events. The latter is particularly germane because of significantly decreased lengths of stay demonstrated in the present study. The bias represented in such abbreviated ascertainments significantly mitigates against solely using in-hospital outcomes as accurate and reliable quality indicators.4 Our data suggest that, by reporting only in-hospital MACE rates, the “actual” procedurally related MACE rate at 30 days could be substantially underestimated. The use of 30-day outcomes as a performance and quality indicator is well entrenched in the cardiac surgical literature.19 –21 Although there are obvious differences between PCI and coronary bypass surgery in the nature of the index procedure and its associated risks, the utility of the 30-day window resides in the ability to capture those sequelae of the index proce-
CORONARY ARTERY DISEASE/30-DAY OUTCOMES FOLLOWING PCI
1137
database, and the requirement for short- and long-term follow-up. NoFactor Odds Ratio 95% CI p Value tably, benchmarks for procedural proficiency, outcomes, and quality Recruitment time indicators in interventional cardiol1999 (vs 1997–1998) 0.9 0.7–1.2 0.5 2000 (vs 1997–1998) 0.6 0.5–0.9 ⬍0.01 ogy are derived from the controlled Age ⬎65 yrs 1.7 1.3–2.2 ⬍0.001 clinical trial literature and from obMultivessel disease 1.4 1.1–1.9 0.01 servational cohort studies.1– 4 HowCardiogenic shock 4.5 2.8–7.4 ⬍0.001 ever, measurements of outcome, i.e., Lesion characteristics Thrombus 1.7 1.3–2.3 ⬍0.001 procedural efficacy and safety, must Total occlusion 1.6 1.2–2.2 ⬍0.01 be accurate, timely, comprehensive, Calcified lesion 1.3 1.0–1.7 0.02 and clinically relevant. That outClass C lesion 1.3 1.0–1.7 0.04 comes at 30 days are primary end No. of attempted lesions 1.4 1.2–1.6 ⬍0.001 points in much of the clinical trial Emergency procedure 1.6 1.1–2.3 0.01 Renal insufficiency 2.0 1.3–3.1 ⬍0.01 literature reflects study design conPeripheral vascular disease 1.7 1.1–2.4 ⬍0.01 siderations and the recognition that Controlled diabetes mellitus 0.7 0.5–0.9 ⬍0.01 delayed, albeit procedurally related, Thienopyridine use before procedure 0.6 0.5–0.8 ⬍0.001 events may be manifest after hospital Abbreviation as in Table 2. discharge. This is currently of relevance to the assessment of postdischarge outcomes after drug-eluting stent implantation,24 where reliance TABLE 6 Factors Associated With 30-day Postdischarge Major Adverse on in-hospital MACE rates as a meaClinical Events surement of procedural safety may Factor Odds Ratio 95% CI p Value be misleading. Congestive heart failure 1.6 1.2–2.3 ⬍0.01 Although it is reassuring that inMultivessel disease (vs single-vessel disease) 1.4 1.1–1.8 ⬍0.01 hospital outcomes continue to imStable angina 0.7 0.5–1.0 0.04 prove, it is somewhat disappointing Ostial lesion 1.4 1.0–2.0 0.06 to observe the static nature of postBalloon angioplasty (vs stent) use 0.7 0.5–1.0 0.06 discharge MACE rates even after adEmergency procedure 1.8 1.3–2.5 0.001 Pulmonary disease 1.6 1.1–2.4 0.01 justing for the increased complexity Procedural aspirin use 0.7 0.5–1.0 0.05 of disease and incremental use of stents and glycoprotein IIb/IIIa anAbbreviation as in Table 2. tagonists in successive waves. Because of the distinct risk profile of dure that may have delayed manifestations or conse- postdischarge MACEs, it remains to be seen whether quences of clinical relevance. In addition, the 30-day increased efforts to address these risk factors will window is unlikely to be tainted by the high-risk result in the same improvements noted for in-hospiperiod for restenosis (generally 3 to 6 months) or the tal MACEs. The reported results depend on the completeness, likelihood of disease progression (generally ⬎1 year). Previous reports have suggested using a 30-day com- accuracy, and reliability of posthospital discharge inposite MACE rate after PCI as a more robust estimate formation. As stated in Methods, clinical follow-up in of procedural outcomes.22,23 Although these latter sin- the entire dataset was 94% and was ⬎90% for each gle-center reports were intended to assess the issue of wave. Although recall and misclassification bias are statistical precision in the relation between volume recognized concerns with telephone follow-up, the and outcome, the proposal that postdischarge MACE systematic retrieval of ancillary supporting data suffirates may serve as a more robust and clinically mean- cient to provide detailed information with regard to ingful measurement of outcome is inherently sound. intercurrent hospitalization, events, and procedures This contention is supported by the nature and severity mitigates these concerns. Further, there is no plausible of postdischarge events reported here. In the present reason to expect differential rates of these factors in analysis, repeat hospitalization with repeat revascular- each wave. ization procedures on the managed vessel were the most frequent adverse events within 30 days of hospital discharge. 1. Hirshfeld JW Jr, Ellis SG, Faxon DP. Recommendations for the assessment maintenance of proficiency in coronary interventional procedures. Statement The NHBLI Dynamic Registry is designed to pro- and of the American College of Cardiology. J Am Coll Cardiol 1998;31:722–743. vide frequent, cross-sectional “looks” at the changing 2. Smith SC Jr, Dove AT, Jacobs AK, Kennedy JW, Keriekes D, Kern MJ, Kuntz landscape of interventional cardiology. Rapid changes RE, Popma JJ, Schaff HV, Williams DO, et al. ACC/AHA guidelines of percucoronary interventions (revision of the 1993 PTCA guidelines)— execin technology and the penetration of that technology taneous utive summary. A report of the American College of Cardiology/American Heart into routine practice demand such frequent analyses. Association Task for on Practice Guidelines (committee to revise the 1993 The strengths of the NHBLI Dynamic Registry reside guidelines for percutaneous transluminal coronary angioplasty). J Am Coll Car2001;37:2215–2239. in the consecutive, unselected nature of the patient diol 3. Ryan TJ, Bauman WM, Kennedy JW, Keriekes D, King SB III, McCallister population, the completeness of the index procedure BD, Smith SC Jr, Ullyot DJ. Guidelines for percutaneous transluminal coronary TABLE 5 Factors Associated With In-hospital Major Adverse Clinical Events
1138 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 95
MAY 15, 2005
angioplasty. A report of the American Heart Association/American College of Cardiology Task Force on Assessment of Diagnostic and Therapeutic Cardiovascular Procedures (Committee on Percutaneous Transluminal Coronary Angioplasty). Circulation 1993;88:2987–3007. 4. Hannan E, Wu C. Assessing quality and outcomes for percutaneous coronary intervention: choosing statistical models, outcomes, time periods and patient populations. Am Heart J 2003;145:571–574. 5. Ellis SG, Omoigui N, Bittl JA, Lincoff M, Wolfe MW, Howell G, Topol EJ. Analysis and comparison of operator-specific outcomes interventional cardiology. Circulation 1996;93:431– 439. 6. Rill V, Brown DL. Practice of coronary angioplasty in California in 1995: comparison to 1989 and impact of coronary stenting. Circulation 1999;99:e12–16. 7. Ho V. Evolution of the volume-outcome relation for hospitals performing coronary angioplasty. Circulation 2000;101:1806 –1811. 8. Popma JJ, Kuntz RE, Baim DS. A decade of improvement in the clinical outcomes of percutaneous coronary intervention for multivessel coronary artery disease. Circulation 2002;106:1592–1594. 9. Hannan EL, Racz MJ, Arani DT, Ryan TJ, Walford G, McCallister BD. A comparison of short- and long-term outcomes for balloon angioplasty and coronary stent placement. J Am Coll Cardiol 2000;36:395– 403. 10. McGrath PD, Malenka DJ, Wennberg DE, Shubrooks SJ Jr, Bradley WA, Robb JF, Kellett MA Jr, Ryan TJ Jr, Hearne MJ, Hettleman B, et al. Changing outcomes in percutaneous coronary interventions: a study of 34,752 procedures in northern New England, 1990 –97. Northern New England Cardiovascular Disease Study Group. J Am Coll Cardiol 1999;34:674 – 680. 11. Altmann DB, Racz M, Battleman DS, Bergman G, Spokojny A, Hannan EL, Sanborn TA. Reduction in angioplasty complications after the introduction of coronary stents: results from a consecutive series of 2242 patients. Am Heart J 1996;132:503–507. 12. Williams DO, Holubkov R, Yeh W, Bourassa MJ, Al-Bassam M, Block PC, Coady P, Cohen H, Cowley M, Dorros G, et al. Percutaneous coronary intervention in the current era compared with 1985– 86. The National Heart, Lung and Blood Institute Registries. Circulation 2000;102:2945–2951. 13. Laskey WK, Kimmel SE, Krone RJ. Contemporary trends in coronary intervention. A report from the Registry of the Society for Cardiac Angiography and Interventions. Catheter Cardiovasc Interv 2000;49:19 –22.
14. Kimmel SE, Sauer WH, Brensinger C, Hirshfeld J, Haber HL, Localio AR.
Relationship between coronary angioplasty laboratory volume and outcomes after hospital discharge. Am Heart J 2002;143:833– 840. 15. Clark MA, Bakhai A, Lacey MJ, Pelletier EM, Cohen DJ. Clinical and economic outcomes of percutaneous coronary intervention in the elderly: an analysis of Medicare claims data. Circulation 2004;110:259 –264. 16. Kastrati A, Neumann F-J, Schomig A. Operator volume and outcome of patients undergoing coronary stent placement. J Am Coll Cardiol 1998;32: 970 –976. 17. Laskey WK, Williams DO, Vlachos HA, Cohen H, Holmes DR, King SB III, Kelsey SF, Slater J, Faxon D, Al-Bassam M, et al. Changes in the practice of percutaneous coronary intervention: a comparison of enrollment waves of the NHLBI Dynamic Registry. Am J Cardiol 2001;87:964 –969. 18. McGrath PD, Wennberg DE, Dickens JD Jr, Siewers AE, Lukas SL, Malenka DJ, Kellett MA Jr, Ryan TJ Jr. Relation between operator and hospital volume and outcomes following percutaneous coronary interventions in the era of the coronary stent. JAMA 2000;284:3139 –3144. 19. Parsonnet V, Dean D, Bernstein AD. A method of uniform stratification of risk for evaluating the results of surgery in acquired adult heart disease. Circulation 1989;79:3–12. 20. Grover FL, Shroyer AL, Hammermeister K, Edwards FH, Ferguson TB Jr, Dziuban SW, Cleveland JC Jr, Clark RE, McDonald G. A decade’s experience with quality improvement in cardiac surgery using the Veteran’s Affairs and Society of Thoracic Surgeons national databases. Ann Surg 2001;234:464 – 472. 21. Ferguson TB, Hammill BG, Peterson ED, DeLong ER, Grover FL. A decade of change-risk profiles and outcomes for isolated coronary artery bypass grafting procedures, 1990 –99: a report from the STS National Database Committee and the Duke Clinical Research Institute. Ann Thorac Surg 2002;73:480 – 490. 22. Lindsay J, Pinnow EE, Pichard AD. Frequency of major adverse cardiac events within one month of coronary angioplasty: a useful measure of operator performance. J Am Coll Cardiol 1999;34:1916 –1923. 23. Lindsay J Jr, Pinnow EE, Pichard AD. Benchmarking operator performance in percutaneous coronary intervention: a novel approach using 30-day events. Catheter Cardiovasc Interv 2001;52:139 –145. 24. US Food and Drug Administration. FDA Public Health web notification: information for physicians on sub-acute thromboses (SAT) and hypersensitivity reactions with use of the Cordis Cypher™ coronary stent. Available at: http:// www.fda.gov/cdrh/safety/cypher.html. Accessed August 11, 2004.
CORONARY ARTERY DISEASE/30-DAY OUTCOMES FOLLOWING PCI
1139
3.1% of wave 3, p <0.001), whereas stent implantation increased significantly (67.5% of wave 1, 79.1% of wave 2, 86.2% of wave 3, p <0.001). Postdischarge MACE through 30 days remained unchanged (5.1% of wave 1, 5.1% of wave 2, 4.8% of wave 3, p ⴝ 0.6). Mean length of stay decreased (2.7 days for wave 1, 2.2 days for wave 3, p <0.001). Disparate clinical, procedural, and angiographic factors were associated with each MACE. Postdischarge MACE rates through 30 days comprise a significant and unchanging fraction of overall procedurally related MACE rates despite improving in-hospital outcomes. Most postdischarge events derive from pathology related to the controlled vessel. A 30day MACE rate may serve as a more comprehensive measurement of procedural outcome. 䊚2005 by Excerpta Medica Inc. (Am J Cardiol 2005;95:1135–1139)
A
on postdischarge adverse events,14 –16 the potential for inaccurate/incomplete measurements of procedural outcomes exists. Accordingly, we determined rates of in-hospital major adverse clinical events (MACEs) and MACEs through 30 days in all patients who underwent PCI and were entered in the National Heart Lung and Blood Institute (NHLBI) Dynamic Registry from 1997 to 2002.
ssessment of periprocedural and in-hospital outcomes after percutaneous coronary intervention (PCI) has served as an essential measurement of quality and proficiency.1– 4 However, at low event rates, considerable uncertainly exists in the accuracy and utility of this measurement of performance.5–7 Further, a shorter hospital length of stay will decrease the period of exposure during which the risk of procedurally related adverse events can be identified. With decreasing rates of in-hospital morbidity and mortality after PCI8 –13 and a paucity of population-based data From the Division of Cardiology, University of New Mexico School of Medicine, Albuquerque, New Mexico; the Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania; the Division of Cardiology, Boston University School of Medicine, Boston, Massachusetts; the Division of Cardiology, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania; the Cardiovascular Division, Mayo Clinic, Rochester, Minnesota; the Division of Cardiology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania; and Brown University School of Medicine, Providence, Rhode Island. This study was supported in part by Grant HL 33992-14 from the National Heart, Lung and Blood Institute, Bethesda, Maryland. Manuscript received October 26, 2004; revised manuscript received and accepted January 11, 2005. Address for reprints: Warren K. Laskey, MD, Division of Cardiology, Department of Internal Medicine, MSC10-5550, 1 University of New Mexico, Albuquerque, New Mexico 87131-0001. E-mail: [email protected]. ©2005 by Excerpta Medica Inc. All rights reserved. The American Journal of Cardiology Vol. 95 May 15, 2005
METHODS Details of the purpose, structure, and function of the NHLBI Dynamic Registry have been previously described.12,17 Consecutively enrolled patients in wave 1 (n ⫽ 2,524) were recruited from 1997 to 1998, those in wave 2 (n ⫽ 2,105) were recruited in 1999, and those in wave 3 (n ⫽ 2,047) were recruited from 2001 to 2002. Patient-level data from the 3 waves were pooled, and component and composite rates of in-hospital MACEs (death, myocardial infarction, and any repeat target vessel revascularization procedure) were recorded. Component and composite rates of MACEs (death, myocardial infarction, nonscheduled rehospitalization, and repeat target vessel revascularization) from time of hospital discharge through 30 days in all patients alive at discharge were also recorded. 0002-9149/05/$–see front matter doi:10.1016/j.amjcard.2005.01.038
1135
TABLE 1 Differences in Clinical, Demographic, and Procedural Characteristics by Enrollment Wave in the NHLBI Dynamic Registry
Demographic and clinical factors Mean age (yrs)* Age ⬎65 years White Black Other Body mass index (kg/m2)* Previous PCI* None ⱖ1 Severe noncardiac co-morbitity* Diabetes mellitus Insulin prescription Oral medications† Congestive heart failure Hypertension Hyperlipidemia* Angiographic features Abnormal left ventricular function† No. of coronary arteries narrowed* 1 2 3 Any total occlusion† Mean no. of significant narrowings* Procedural features Indications for revascularization‡ Stable angina pectoris Unstable angina pectoris Acute myocardial infarction ⱕ14 days Circumstances of procedure* Elective Urgent Emergency Use of platelet glycoprotein receptor inhibitor* Thienopyridine use* Overall stent use* Lesion length (mm)*
Wave 1 (n ⫽ 2,524)
Wave 2 (n ⫽ 2,105)
Wave 3 (n ⫽ 2,047)
63 45% 80% 8% 12% 28
63 46% 78% 12% 10% 29
64 49% 79% 13% 8% 29
72% 28% 30% 28% 9% 14% 10% 59% 61%
69% 31% 34% 28% 9% 15% 10% 64% 63%
67% 33% 39% 30% 9% 16% 13% 75% 71%
27%
27%
31%
42% 33% 25% 39% 2.9
42% 32% 25% 36% 2.9
25% 47% 19%
20% 45% 26%
66% 23% 11% 24%
57% 32% 11% 32%
50% 67% 12.4
41% 79% 13.0
*p ⬍0.001; †p ⬍0.05; ‡p ⬍0.01.
Patients participating in the NHLBI Dynamic Registry consented to the collection of in-hospital and postdischarge data after the index procedure. Study coordinators at each site contacted patients at 1 year to obtain vital status and information regarding intercurrent hospitalizations, cardiovascular events, or procedures. Dates corresponding to all reported outcomes were recorded and confirmed by requesting relevant records. Whenever possible, angiographic data pertaining to repeat PCI were collected and reviewed. Accordingly, postdischarge follow-up rates were 97.5%, 94.9%, and 90.2% for waves 1, 2, and 3, respectively. Statistical analysis: Continuous data are summarized as means ⫾ 1 SD. Categorical variables are summarized as counts or percentages. Comparisons of (unadjusted) event rates across waves of the NHLBI Dynamic Registry were accomplished using a contingency table and Mantel-Haenszel statistics. All rates are reported as point estimates with their correspond1136 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 95
ing 95% confidence intervals. Rates were also adjusted for imbalances in important covariates among waves using general linear modeling methods. Clinical, demographic, and procedural factors independently associated with in-hospital and 30-day MACE rates were assessed with multivariable logistic regression. Odds ratios and their corresponding 95% confidence intervals are reported. Cumulative event rates were calculated with Kaplan-Meier methods and compared with log-rank statistics. Statistical significance was defined as a p value ⬍0.05. The study sample possessed 99% power to detect a difference at the 0.05 level in event rates of the magnitude reported.
RESULTS
Table 1 presents all salient, statistically significant (p ⱕ0.05) differences in clinical, demographic, and procedural characteristics across waves in the NHLBI Dynamic Registry. Over successive periods, patients tended to 20% be older, have more severe cardiac and 42% noncardiac co-morbidities, more fre28% quently underwent previous PCI, were more likely to present with an acute coronary syndrome and undergo ur49% 40% gent or emergency PCI, more fre10% quently received platelet glycoprotein 53% IIb/IIIa receptor antagonists during PCI, and were more frequently man68% 86% aged with bare metal stent implanta13.4 tion. There were no drug-eluting stents implanted during this period. Mean length of stay decreased significantly over the 3 waves (2.65 ⫾ 3.5 days for wave 1, 2.36 ⫾ 3.9 days for wave 2, 2.21 ⫾ 3.2 days for wave 3, p ⬍0.001). Neither angiographic success rate (92.2% for wave 1, 92.9% for wave 2, 94% for wave 3) nor procedural success rate (94.7% for wave 1, 95.7% for wave 2, 96.2% for wave 3) differed significantly over the 3 waves. Table 2 presents rates of patient and composite adverse in-hospital and 30-day postdischarge outcomes for each wave. In contrast to the significant decrease in in-hospital MACE rate over the 3 waves (5.4% for wave 1, 4.9% for wave 2, 3.1% for wave 3, p ⬍0.001), there was no significant difference in MACE rate at 30 days after discharge. After adjusting for age, vessel disease, co-morbid conditions, diabetes, and acuity, 30-day postdischarge event rates were 5.4% for wave 1, 5.0% for wave 2, and 4.6% for wave 3 (p ⫽ 0.62). Notably, postdischarge MACE rates at 30 days were of the same order of magnitude as the in-hospital MACE rates. Thus, the significant decrease in cumulative MACE rates at 30 days (10.4% for wave 1, 9.5% for wave 2, 7.8% for wave 3, p ⫽ 0.009) 38% 31% 30% 37% 3.1
MAY 15, 2005
TABLE 2 Major Adverse Clinical Event Rates by Enrollment Wave in the NHLBI Dynamic Registry Wave 1 (%) Event Death Death/myocardial infarction Target vessel revascularization Composite MACE rate (95% CI)
Wave 2 (%)
Wave 3 (%)
In-hospital
30 Day
In-hospital
30 Day
In-hospital
30 Day
1.6 4.2 1.3 5.4 (4.5–6.2)
0.3 0.6 2.3 5.1 (4.2–6.1)
1.3 4.0 1.1 4.9 (3.9–5.8)
0.5 0.9 1.7 5.0 (4.0–6.0)
1.0 2.7 0.5 3.1 (2.4–3.9)
0.6 0.9 1.6 4.9 (3.9–6.0)
CI ⫽ confidence interval.
TABLE 3 Categorization and Patient-specific Frequency of 30-day Postdischarge Events Category of Events 1 Event Death Myocardial infarction PCI Coronary bypass surgery Hospitalization 2 Events Death and myocardial infarction Death and hospitalization Myocardial infarction and PCI Myocardial infarction and coronary bypass surgery PCI and hospitalization Coronary bypass surgery and hospitalization 3 Events Death, PCI, and myocardial infarction PCI, coronary bypass surgery, and myocardial infarction PCI, coronary bypass surgery, and hospitalization PCI, myocardial infarction, and hospitalization
No. of Patients 21 9 30 1 122 3 4 9 2 63 14 2 2 1 1
TABLE 4 Type and Number of 30-Day Postdischarge Events Event Type Death Myocardial infarction Coronary bypass surgery PCI Hospitalization Sum
No. of Events 30 27 20 110 216 403
was driven entirely by the decrease in the in-hospital component. Table 3 lists patient-specific events in hierarchical fashion through 30 days after hospital discharge, and Table 4 lists cumulative events. There were 403 events in 284 patients. A significant minority of patients (18.6%) died or sustained a myocardial infarction; 50% of patients underwent repeat hospitalization and 32% of patients required repeat revascularization of the target vessel. Tables 5 and 6 present significant independent predictors of in-hospital and 30-day MACEs, respectively. There was little overlap in the factors associated with each outcome. Factors associated with in-
hospital MACEs tended to reflect the extent and severity of the underlying coronary (and noncoronary) disease, whereas factors associated with 30-day MACEs tended to reflect procedural complexity. Interestingly, the use of balloon angioplasty alone (vs stent implantation) was associated with a lower risk of postdischarge MACEs.
DISCUSSION In this study, we observed a decreasing in-hospital MACE rate and an unchanging 30-day postdischarge MACE rate in patients who underwent PCI from 1997 to 2002 in the NHLBI Dynamic Registry. Over this period there was also a significant decrease in average length of hospital stay. Under these circumstances, a failure to record clinically significant events shortly after hospital discharge may result in inaccurate assessments of procedural safety and outcome. The present results with respect to decreasing inhospital MACE rates in the current PCI era agree with those of numerous previous reports.8 –13 However, the reporting of MACE rates after PCI at 30 days has generally been limited to the randomized clinical trial literature, with a paucity of similar data available from less select patient populations.14 –16,18 Further, claimsbased datasets15,18 generally lack important ancillary clinical information and confine the end point at 30 days to mortality and/or coronary artery bypass surgery. The value of obtaining more complete follow-up data in consecutive patients under “real-world” conditions stems not only from differences in event rates between controlled clinical trials and less restricted clinical practice but also from the need to more completely characterize unanticipated and/or delayed procedurally related adverse events. The latter is particularly germane because of significantly decreased lengths of stay demonstrated in the present study. The bias represented in such abbreviated ascertainments significantly mitigates against solely using in-hospital outcomes as accurate and reliable quality indicators.4 Our data suggest that, by reporting only in-hospital MACE rates, the “actual” procedurally related MACE rate at 30 days could be substantially underestimated. The use of 30-day outcomes as a performance and quality indicator is well entrenched in the cardiac surgical literature.19 –21 Although there are obvious differences between PCI and coronary bypass surgery in the nature of the index procedure and its associated risks, the utility of the 30-day window resides in the ability to capture those sequelae of the index proce-
CORONARY ARTERY DISEASE/30-DAY OUTCOMES FOLLOWING PCI
1137
database, and the requirement for short- and long-term follow-up. NoFactor Odds Ratio 95% CI p Value tably, benchmarks for procedural proficiency, outcomes, and quality Recruitment time indicators in interventional cardiol1999 (vs 1997–1998) 0.9 0.7–1.2 0.5 2000 (vs 1997–1998) 0.6 0.5–0.9 ⬍0.01 ogy are derived from the controlled Age ⬎65 yrs 1.7 1.3–2.2 ⬍0.001 clinical trial literature and from obMultivessel disease 1.4 1.1–1.9 0.01 servational cohort studies.1– 4 HowCardiogenic shock 4.5 2.8–7.4 ⬍0.001 ever, measurements of outcome, i.e., Lesion characteristics Thrombus 1.7 1.3–2.3 ⬍0.001 procedural efficacy and safety, must Total occlusion 1.6 1.2–2.2 ⬍0.01 be accurate, timely, comprehensive, Calcified lesion 1.3 1.0–1.7 0.02 and clinically relevant. That outClass C lesion 1.3 1.0–1.7 0.04 comes at 30 days are primary end No. of attempted lesions 1.4 1.2–1.6 ⬍0.001 points in much of the clinical trial Emergency procedure 1.6 1.1–2.3 0.01 Renal insufficiency 2.0 1.3–3.1 ⬍0.01 literature reflects study design conPeripheral vascular disease 1.7 1.1–2.4 ⬍0.01 siderations and the recognition that Controlled diabetes mellitus 0.7 0.5–0.9 ⬍0.01 delayed, albeit procedurally related, Thienopyridine use before procedure 0.6 0.5–0.8 ⬍0.001 events may be manifest after hospital Abbreviation as in Table 2. discharge. This is currently of relevance to the assessment of postdischarge outcomes after drug-eluting stent implantation,24 where reliance TABLE 6 Factors Associated With 30-day Postdischarge Major Adverse on in-hospital MACE rates as a meaClinical Events surement of procedural safety may Factor Odds Ratio 95% CI p Value be misleading. Congestive heart failure 1.6 1.2–2.3 ⬍0.01 Although it is reassuring that inMultivessel disease (vs single-vessel disease) 1.4 1.1–1.8 ⬍0.01 hospital outcomes continue to imStable angina 0.7 0.5–1.0 0.04 prove, it is somewhat disappointing Ostial lesion 1.4 1.0–2.0 0.06 to observe the static nature of postBalloon angioplasty (vs stent) use 0.7 0.5–1.0 0.06 discharge MACE rates even after adEmergency procedure 1.8 1.3–2.5 0.001 Pulmonary disease 1.6 1.1–2.4 0.01 justing for the increased complexity Procedural aspirin use 0.7 0.5–1.0 0.05 of disease and incremental use of stents and glycoprotein IIb/IIIa anAbbreviation as in Table 2. tagonists in successive waves. Because of the distinct risk profile of dure that may have delayed manifestations or conse- postdischarge MACEs, it remains to be seen whether quences of clinical relevance. In addition, the 30-day increased efforts to address these risk factors will window is unlikely to be tainted by the high-risk result in the same improvements noted for in-hospiperiod for restenosis (generally 3 to 6 months) or the tal MACEs. The reported results depend on the completeness, likelihood of disease progression (generally ⬎1 year). Previous reports have suggested using a 30-day com- accuracy, and reliability of posthospital discharge inposite MACE rate after PCI as a more robust estimate formation. As stated in Methods, clinical follow-up in of procedural outcomes.22,23 Although these latter sin- the entire dataset was 94% and was ⬎90% for each gle-center reports were intended to assess the issue of wave. Although recall and misclassification bias are statistical precision in the relation between volume recognized concerns with telephone follow-up, the and outcome, the proposal that postdischarge MACE systematic retrieval of ancillary supporting data suffirates may serve as a more robust and clinically mean- cient to provide detailed information with regard to ingful measurement of outcome is inherently sound. intercurrent hospitalization, events, and procedures This contention is supported by the nature and severity mitigates these concerns. Further, there is no plausible of postdischarge events reported here. In the present reason to expect differential rates of these factors in analysis, repeat hospitalization with repeat revascular- each wave. ization procedures on the managed vessel were the most frequent adverse events within 30 days of hospital discharge. 1. Hirshfeld JW Jr, Ellis SG, Faxon DP. Recommendations for the assessment maintenance of proficiency in coronary interventional procedures. Statement The NHBLI Dynamic Registry is designed to pro- and of the American College of Cardiology. J Am Coll Cardiol 1998;31:722–743. vide frequent, cross-sectional “looks” at the changing 2. Smith SC Jr, Dove AT, Jacobs AK, Kennedy JW, Keriekes D, Kern MJ, Kuntz landscape of interventional cardiology. Rapid changes RE, Popma JJ, Schaff HV, Williams DO, et al. ACC/AHA guidelines of percucoronary interventions (revision of the 1993 PTCA guidelines)— execin technology and the penetration of that technology taneous utive summary. A report of the American College of Cardiology/American Heart into routine practice demand such frequent analyses. Association Task for on Practice Guidelines (committee to revise the 1993 The strengths of the NHBLI Dynamic Registry reside guidelines for percutaneous transluminal coronary angioplasty). J Am Coll Car2001;37:2215–2239. in the consecutive, unselected nature of the patient diol 3. Ryan TJ, Bauman WM, Kennedy JW, Keriekes D, King SB III, McCallister population, the completeness of the index procedure BD, Smith SC Jr, Ullyot DJ. Guidelines for percutaneous transluminal coronary TABLE 5 Factors Associated With In-hospital Major Adverse Clinical Events
1138 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 95
MAY 15, 2005
angioplasty. A report of the American Heart Association/American College of Cardiology Task Force on Assessment of Diagnostic and Therapeutic Cardiovascular Procedures (Committee on Percutaneous Transluminal Coronary Angioplasty). Circulation 1993;88:2987–3007. 4. Hannan E, Wu C. Assessing quality and outcomes for percutaneous coronary intervention: choosing statistical models, outcomes, time periods and patient populations. Am Heart J 2003;145:571–574. 5. Ellis SG, Omoigui N, Bittl JA, Lincoff M, Wolfe MW, Howell G, Topol EJ. Analysis and comparison of operator-specific outcomes interventional cardiology. Circulation 1996;93:431– 439. 6. Rill V, Brown DL. Practice of coronary angioplasty in California in 1995: comparison to 1989 and impact of coronary stenting. Circulation 1999;99:e12–16. 7. Ho V. Evolution of the volume-outcome relation for hospitals performing coronary angioplasty. Circulation 2000;101:1806 –1811. 8. Popma JJ, Kuntz RE, Baim DS. A decade of improvement in the clinical outcomes of percutaneous coronary intervention for multivessel coronary artery disease. Circulation 2002;106:1592–1594. 9. Hannan EL, Racz MJ, Arani DT, Ryan TJ, Walford G, McCallister BD. A comparison of short- and long-term outcomes for balloon angioplasty and coronary stent placement. J Am Coll Cardiol 2000;36:395– 403. 10. McGrath PD, Malenka DJ, Wennberg DE, Shubrooks SJ Jr, Bradley WA, Robb JF, Kellett MA Jr, Ryan TJ Jr, Hearne MJ, Hettleman B, et al. Changing outcomes in percutaneous coronary interventions: a study of 34,752 procedures in northern New England, 1990 –97. Northern New England Cardiovascular Disease Study Group. J Am Coll Cardiol 1999;34:674 – 680. 11. Altmann DB, Racz M, Battleman DS, Bergman G, Spokojny A, Hannan EL, Sanborn TA. Reduction in angioplasty complications after the introduction of coronary stents: results from a consecutive series of 2242 patients. Am Heart J 1996;132:503–507. 12. Williams DO, Holubkov R, Yeh W, Bourassa MJ, Al-Bassam M, Block PC, Coady P, Cohen H, Cowley M, Dorros G, et al. Percutaneous coronary intervention in the current era compared with 1985– 86. The National Heart, Lung and Blood Institute Registries. Circulation 2000;102:2945–2951. 13. Laskey WK, Kimmel SE, Krone RJ. Contemporary trends in coronary intervention. A report from the Registry of the Society for Cardiac Angiography and Interventions. Catheter Cardiovasc Interv 2000;49:19 –22.
14. Kimmel SE, Sauer WH, Brensinger C, Hirshfeld J, Haber HL, Localio AR.
Relationship between coronary angioplasty laboratory volume and outcomes after hospital discharge. Am Heart J 2002;143:833– 840. 15. Clark MA, Bakhai A, Lacey MJ, Pelletier EM, Cohen DJ. Clinical and economic outcomes of percutaneous coronary intervention in the elderly: an analysis of Medicare claims data. Circulation 2004;110:259 –264. 16. Kastrati A, Neumann F-J, Schomig A. Operator volume and outcome of patients undergoing coronary stent placement. J Am Coll Cardiol 1998;32: 970 –976. 17. Laskey WK, Williams DO, Vlachos HA, Cohen H, Holmes DR, King SB III, Kelsey SF, Slater J, Faxon D, Al-Bassam M, et al. Changes in the practice of percutaneous coronary intervention: a comparison of enrollment waves of the NHLBI Dynamic Registry. Am J Cardiol 2001;87:964 –969. 18. McGrath PD, Wennberg DE, Dickens JD Jr, Siewers AE, Lukas SL, Malenka DJ, Kellett MA Jr, Ryan TJ Jr. Relation between operator and hospital volume and outcomes following percutaneous coronary interventions in the era of the coronary stent. JAMA 2000;284:3139 –3144. 19. Parsonnet V, Dean D, Bernstein AD. A method of uniform stratification of risk for evaluating the results of surgery in acquired adult heart disease. Circulation 1989;79:3–12. 20. Grover FL, Shroyer AL, Hammermeister K, Edwards FH, Ferguson TB Jr, Dziuban SW, Cleveland JC Jr, Clark RE, McDonald G. A decade’s experience with quality improvement in cardiac surgery using the Veteran’s Affairs and Society of Thoracic Surgeons national databases. Ann Surg 2001;234:464 – 472. 21. Ferguson TB, Hammill BG, Peterson ED, DeLong ER, Grover FL. A decade of change-risk profiles and outcomes for isolated coronary artery bypass grafting procedures, 1990 –99: a report from the STS National Database Committee and the Duke Clinical Research Institute. Ann Thorac Surg 2002;73:480 – 490. 22. Lindsay J, Pinnow EE, Pichard AD. Frequency of major adverse cardiac events within one month of coronary angioplasty: a useful measure of operator performance. J Am Coll Cardiol 1999;34:1916 –1923. 23. Lindsay J Jr, Pinnow EE, Pichard AD. Benchmarking operator performance in percutaneous coronary intervention: a novel approach using 30-day events. Catheter Cardiovasc Interv 2001;52:139 –145. 24. US Food and Drug Administration. FDA Public Health web notification: information for physicians on sub-acute thromboses (SAT) and hypersensitivity reactions with use of the Cordis Cypher™ coronary stent. Available at: http:// www.fda.gov/cdrh/safety/cypher.html. Accessed August 11, 2004.
CORONARY ARTERY DISEASE/30-DAY OUTCOMES FOLLOWING PCI
1139