Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Cardiac contractility modulation wikipedia , lookup
Cardiovascular disease wikipedia , lookup
Remote ischemic conditioning wikipedia , lookup
Cardiac surgery wikipedia , lookup
History of invasive and interventional cardiology wikipedia , lookup
Quantium Medical Cardiac Output wikipedia , lookup
Original Articles Impact of Incomplete Revascularization on Long-Term Mortality After Coronary Stenting Chuntao Wu, MD, PhD; Anne-Marie Dyer, MS; Spencer B. King III, MD; Gary Walford, MD; David R. Holmes, Jr, MD; Nicholas J. Stamato, MD; Ferdinand J. Venditti, MD; Samin K. Sharma, MD; Icilma Fergus, MD; Alice K. Jacobs, MD; Edward L. Hannan, PhD Downloaded from http://circinterventions.ahajournals.org/ by guest on November 20, 2016 Background—The impact of incomplete revascularization (IR) on adverse outcomes after percutaneous coronary intervention remains inconclusive, and few studies have examined mortality during follow-ups longer than 5 years. The objective of this study is to test the hypothesis that IR is associated with higher risk of long-term (8-year) mortality after stenting for multivessel coronary disease. Methods and Results—A total of 13 016 patients with multivessel disease who had undergone stenting procedures with bare metal stents in 1999 to 2000 were identified in the New York State’s Percutaneous Coronary Intervention Reporting System. A logistic regression model was fit to predict the probability of achieving complete revascularization (CR) in these patients using baseline risk factors; then, the CR patients were matched to the IR patients with similar likelihoods of achieving CR. Each patient’s vital status was followed through 2007 using the National Death Index, and the difference in long-term mortality between IR and CR was compared. It was found that CR was achieved in 29.2% (3803) of the patients. For the 3803 pair-matched patients, the respective 8-year survival rates were 80.8% and 78.5% for CR and IR (P⫽0.04), respectively. The risk of death was marginally significantly higher for IR (hazard ratio⫽1.12; 95% confidence interval, 1.01–1.26, P⫽0.04). The 95% bootstrap confidence interval for the hazard ratio was 0.98 to 1.32. Conclusions—IR may be associated with higher risk of long-term mortality after stenting with BMS in patients with multivessel disease. More prospective studies are needed to further test this association. (Circ Cardiovasc Interv. 2011;4:413-421.) Key Words: coronary angioplasty 䡲 coronary artery disease 䡲 long-term follow-up 䡲 mortality 䡲 stenting A though percutaneous coronary intervention (PCI) is a commonly used revascularization procedure for patients with multivessel coronary artery disease, the impact of incomplete revascularization (IR) on adverse outcomes after PCI is still inconclusive.1–21 Some of the studies before the stent era have examined the impact of IR for coronary angioplasty,3–11 which has now been largely replaced by coronary stenting. The more recent studies using data from the stent era examined short- and intermediate-term outcomes with lengths of follow-up usually less than 3 years.12–21 However, there is a lack of data on adverse outcomes such as mortality and major adverse cardiac events (MACE) in the long term, especially 5 or more years. Because the observed short- and intermediate-term effects of IR may be different than the long-term impact, there is a need for studies that examine long-term adverse outcomes data after PCI with IR or complete revascularization (CR). Our group has conducted 2 studies that have found that IR was associated with higher risk of 3-year mortality in the bare metal stent (BMS) era17 and 18-month mortality in the drug-eluting stent era (DES).20 Building on our past research, the present study examines the impact of IR on long-term mortality for coronary stenting with BMS. We tested the hypothesis that incomplete revascularization is associated with higher risk of long-term (8-year) mortality after coronary stenting with BMS. Methods Databases The databases used in this study include the databases of the New York State’s Percutaneous Coronary Intervention Reporting System (PCIRS) and Cardiac Surgery Reporting System (CSRS) and the National Death Index. The PCIRS contains all PCI procedures performed in nonfederal hospitals in New York State. The data collected in the PCIRS include patient demographics, preprocedural Received April 20, 2011; accepted July 27, 2011. From the Penn State Hershey College of Medicine, Hershey, PA (C.W., A.D.); St Joseph’s Health System, Atlanta, GA (S.B.K.); Johns Hopkins Medical Center, Baltimore, MD (G.W.); Mayo Clinic, Rochester, MN (D.R.H.); United Health Services, Binghamton, NY (N.J.S.); Albany Medical College, Albany, NY (F.J.V.); Mt Sinai Medical Center, New York, NY (S.S.); Harlem Hospital, New York, NY (I.F.); Boston Medical Center, Boston, MA (A.K.J.); and University at Albany, State University of New York, Albany, NY (E.L.H.). The views expressed are those of the authors and do not necessarily reflect those of the New York State Department of Health. An abstract based on part of this study was presented at the American Heart Association, 2010 Scientific Sessions, Chicago, IL; November 13–17, 2010. Correspondence to Chuntao Wu, MD, PhD, Penn State Hershey College of Medicine, Academic Support Bldg, Suite 2200, A210, 600 Centerview Dr, Hershey, PA 17033. E-mail [email protected] © 2011 American Heart Association, Inc. Circ Cardiovasc Interv is available at http://circinterventions.ahajournals.org 413 DOI: 10.1161/CIRCINTERVENTIONS.111.963058 414 Circ Cardiovasc Interv October 2011 risk factors, lesions, and vessel information, including preprocedural and postprocedural stenosis in diseased vessels, procedural information including intracoronary devices, postprocedural complications, and discharge status. The completeness of the PCIRS database is maintained by matching the data to the New York State’s hospital discharge data; the accuracy of the data are ensured by data validation conducted by the New York State Department of Health’s review agent. WHAT IS KNOWN ● Several earlier studies have identified significantly higher short-term mortality rates for patients undergoing percutaneous coronary intervention with incomplete revascularization in comparison to patients who received complete revascularization. Downloaded from http://circinterventions.ahajournals.org/ by guest on November 20, 2016 WHAT THE STUDY ADDS ● ● This study examined the impact of incomplete revascularization on 8-year mortality after coronary stenting for multivessel disease. The results show that the 8-year survival rate was marginally significantly lower for incomplete as compared to complete revascularization after coronary stenting for multivessel disease. The CSRS collects detailed information of all cardiac surgery procedures performed in New York. Similar to the PCIRS, patient demographics, preprocedural risk factors, procedural information, postprocedural complications, and discharge status are recorded in the CSRS. The CSRS data were used to identify coronary artery bypass graft (CABG) surgery after index PCI procedures by linking the two databases using patient identifiers including social security numbers, dates of birth, admission, procedure, and discharge. The National Death Index is an index of death records of the United States maintained by the National Center for Health Statistics.22 It was used to ascertain the vital status of patients after being discharged from hospitals. Study Population We used the PCIRS data to identify patients who had undergone stenting procedures with BMS in 1999 and 2000 in New York for this study. The inclusion criteria were (1) patients had multivessel disease (defined as at least 2 major epicardial arteries with stenosis ⱖ70%) but no left main coronary artery disease (stenosis ⱖ50%),(2) had no history of CABG surgery or PCI before the index stenting procedures,(3) had no myocardial infarction (ST-elevation–myocardial infarction [MI] or non–ST-elevation MI) within 24 hours before stenting, and (4) had not undergone CABG surgery in the index admission or within 30 days of discharge. A total of 13 016 patients were included in this study. Definition of CR The completeness of revascularization of stenting was evaluated by comparing the degrees of stenosis before and after stenting procedures. CR was defined as reduction of stenosis to ⬍50% in all diseased (ⱖ70% stenosis) lesions in major epicardial coronary vessels (left anterior descending artery and major diagonals; left circumflex artery and large marginal branches; and right coronary artery and right posterior descending artery) in the index hospitalization or within 30 days after discharge from the index hospitalization before having a new MI. When a CR was not achieved during a stenting procedure, it was defined as a procedure with IR. The 30-day period after discharge was included to enable patients who were staged to receive CR in another admission shortly after the index admission to be considered as CR also. However, if they had an MI before the CR was completed, this was not regarded as CR because of the occurrence of an adverse event before CR was attained. Outcome The outcome variable is mortality after the index stenting procedure. Each patient’s vital status was followed until December 31, 2007, by matching to the National Death Index using social security number, birth date, and sex. The median length of follow-up (from date of procedure to the end of 2007) was 8.0 years. Statistical Analysis The first step was to examine the differences in the distributions of baseline patient characteristics such as demographics, preprocedural risk factors, and diseased vessels between the IR and CR groups. Pearson 2 test was used for categorical variables and the Student t test for continuous variables. Then, a logistic (propensity) model was fit by including all available baseline risk factors as independent variables to predict the probability of receiving CR (instead of IR). Each CR patient was then matched to an IR patient on the number of diseased vessels, the presence of total occlusion, and the value of propensity score (log-odds of the probability of receiving CR). The matching caliper for the propensity score was 0.6 standard deviation of its distribution.23–25 The distributions of baseline risk factors between the matched IR and CR patients were then compared by examining the standardized differences and performing tests of agreement.26,27 Kaplan-Meier survival curves for the matched IR and CR patients were compared using a method described by Klein and Moeschberger.28 A Cox proportional hazards model stratifying on the matched pairs of patients was fit to obtain a hazard ratio to evaluate the relative risk of death between IR and CR.26,29 A bootstrap 95% confidence interval for the hazard ratio was also obtained by creating 1000 resamples of the paired-matched patients through random sampling with replacement.30 In addition, IR was classified into 2 types, namely 1-vessel IR and at least 2-vessel IR, based on the number of incompletely revascularized vessels. The association between IR and long-term mortality was examined for each type of IR. For a group of patient characteristics including age, ejection fraction, history of MI, congestive heart failure, diabetes, left anterior descending artery disease, proximal vessel disease, and the presence of total occlusion were tested, the interactions between IR and such variables were also tested. For each risk factor, a Cox proportional hazards model was fit including IR, the risk factor itself, their interaction term, and other significant (P⬍0.05) risk factors for mortality identified using a backwards selection. To evaluate whether the statistical results were sensitive to the chosen analytical methods, we also compared the risk of death between the IR and CR groups using the entire unmatched patient population by adjusting for significant (P⬍0.05) baseline risk factors identified by a backwards selection approach in Cox proportional hazards models. All statistical analyses were conducted in SAS version 9.1 (SAS Institute, Cary, NC). Results Among the 13 016 patients who underwent stenting procedures in 1999 to 2000, CR was achieved in 3803 (29.2%) patients and IR was obtained in 9213 (70.8%) patients. The median prevalence of IR across hospitals was 72.5%, and the interquartile range was 64.9% to 77.5%. The median prevalence of IR across operators was 73.0%, and the interquartile range was 63.6% to 82.7%. Table 1 shows that the IR patients were older and more likely to be Hispanic or black, had lower ejection fraction Wu et al Table 1. Baseline Patient Characteristics of Stenting Patients Incomplete Revascularization (n⫽9213) Complete Revascularization (n⫽3803) 964 (10.5) 502 (13.2) 50–59 1961 (21.3) 955 (25.1) 60–69 2518 (27.3) 1062 (27.9) 70–79 2558 (27.8) 961 (25.3) 80⫹ 1212 (13.2) 323 (8.5) Risk Factor ⬍0.0001 Age, y, n (%) ⬍50 Sex, n (%) 2895 (31.4) 1180 (31.0) Male 6318 (68.6) 2623 (69.0) Downloaded from http://circinterventions.ahajournals.org/ by guest on November 20, 2016 Non-Hispanic white 7473 (81.1) 3202 (84.2) Non-Hispanic black 599 (6.5) 193 (5.1) Hispanic 636 (6.9) 200 (5.3) Other 505 (5.5) 208 (5.5) 1.99 (0.25) 2.01 (0.25) 18.5–24.99 0.15 780 (8.5) 259 (6.8) 0.001 ⬍0.0001 711 (7.7) 187 (4.9) 114 (3.0) 8013 (87.0) 3502 (92.1) Malignant ventricular arrhythmia, n (%) 129 (1.4) 48 (1.3) 0.54 Chronic obstructive pulmonary disease, n (%) 598 (6.5) 179 (4.7) ⬍0.0001 2504 (27.2) 882 (23.2) ⬍0.0001 Requiring dialysis 107 (1.2) 37 (1.0) 125 (1.4) 26 (0.7) 8981 (97.5) 3740 (98.3) Diabetes, n (%) 864 (22.7) No renal failure 1649 (43.4) 1257 (33.1) ⬍20% 304 (3.3) 147 (3.9) 20–29% 316 (3.4) 67 (1.8) 30–39% 804 (8.7) 198 (5.2) ⱖ40% 7505 (81.5) 3242 (85.2) Missing 284 (3.1) 149 (3.9) *Pearson’s exact test. ⬍0.0001 No. of diseased vessels, n (%) 2 Diseased vessels 6793 (73.7) 3577 (94.1) 3 Diseased vessels 2420 (26.3) 226 (5.9) 3885 (42.2) 503 (13.2) ⬍0.0001 ⬍0.0001 Previous myocardial infarction, n (%) 1–7 d 2087 (22.7) 883 (23.2) 8–20 d 383 (4.2) 113 (3.0) ⱖ21 d 1719 (18.7) 412 (10.8) No myocardial infarction before procedure 5024 (54.5) 2395 (63.0) Cerebrovascular disease, n (%) 794 (8.6) 204 (5.4) ⬍0.0001 Peripheral arterial disease, n (%) 612 (6.6) 155 (4.1) ⬍0.0001 9163 (99.5) 3787 (99.6) 41 (0.4) 13 (0.3) 9 (0.1) 3 (0.1) Hemodynamic state, n (%) 0.003 2 ⬍0.0001 Ejection fraction, n (%) Shock 0.30* 489 (5.3) 2225 (24.2) 3087 (33.5) Unstable 1 (0.0) Before this admission Creatinine ⬎2.5 mg/dL 3811 (41.4) P Value 8 (0.1) This admission 33 (0.9) 30⫹ Stable Cardiopulmonary resuscitation, n (%) 90 (1.0) 25–29.99 Presence of total occlusion, n (%) Risk Factor Complete Revascularization (n⫽3803) Renal failure, n (%) Body mass index, kg/m2, n (%) ⬍18.5 Incomplete Revascularization (n⫽9213) None 0.004 415 Continued Congestive heart failure, n (%) ⬍0.0001 Race, n (%) Table 1. Left ventricular hypertrophy, n (%) 0.66 Female Body surface area, m2, mean (SD) P Value Incomplete Revascularization and Stenting 0.67 (Continued) values, were more likely to have 3-vessel disease and presence of total occlusion, and were more likely to have histories of comorbidities such as myocardial infarction, cerebrovascular disease, peripheral arterial disease, left ventricular hypertrophy, congestive heart failure, chronic obstructive pulmonary disease, diabetes, and renal failure. Table 2 shows that the standardized differences in the prevalence of risk factors between the 3803 pairs of matched IR and CR patients ranged for 0% to 3.1% in absolute values. Tests of agreement were all statistically nonsignificant, indicating that the pairs matched well with regards to the risk factors. Among the 3803 pairs of matched patients, 817 IR patients and 731 CR patients died during the postprocedural follow-up through December 31, 2007, with a median follow-up of 8 years (interquartile range, 7.5– 8.5 years). The risk of death was 12% higher for IR than CR and was marginally significant (hazard ratio [HR]⫽1.12; 95% confidence interval [CI], 1.01–1.26, P⫽0.04). The 95% bootstrap CI for the HR was 0.98 to 1.32. Figure 1 shows that the respective 8-year survival rates were 78.5% and 80.8% for IR and CR, and the difference was statistically significant (P⫽0.04). When IR was divided into 2 types by the number of incompletely revascularized vessels, the HRs were 1.11 (P⫽0.08) and 1.20 (P⫽0.22) for 1-vessel IR (n⫽3268) and multiple-vessel IR (n⫽535), respectively (Table 3). Figure 2 shows that patients with 1-vessel IR had a lower, though statistically nonsignificant, 8-year survival rate compared with the matched CR patients (78.8% versus 81.0%, P⫽0.08). Figure 3 shows that the 8-year survival rates were 416 Circ Cardiovasc Interv Table 2. October 2011 Baseline Patient Characteristics of Propensity-Matched Patients Risk Factor Incomplete Revascularization (n⫽3803) Complete Revascularization (n⫽3803) 506 (13.3) 502 (13.2) 0.3 ⫺1.3 Standardized Difference, % Age, y, n (%) ⬍50 0.90 50–59 934 (24.6) 955 (25.1) 60–69 1082 (28.5) 1062 (27.9) 1.2 70–79 964 (25.3) 961 (25.3) 0.2 80⫹ 317 (8.3) 323 (8.5) ⫺0.6 Sex, n (%) 0.59 Female 1201 (31.6) 1180 (31.0) 1.2 Male 2602 (68.4) 2623 (69.0) ⫺1.2 Non-Hispanic white 3178 (83.6) 3202 (84.2) ⫺1.7 Non-Hispanic black 204 (5.4) 193 (5.1) 1.3 Hispanic 205 (5.4) 200 (5.3) 0.6 Other 216 (5.7) 208 (5.5) 0.9 2.01 (0.26) 2.01 (0.25) 0.8 Race, n (%) Downloaded from http://circinterventions.ahajournals.org/ by guest on November 20, 2016 Body surface area, m2, mean (SD) 0.75 Body mass index, kg/m2, n (%) ⬍18.5 18.5–24.99 36 (0.9) 33 (0.9) 0.8 0.3 868 (22.8) 864 (22.7) 1656 (43.5) 1649 (43.4) 0.4 30⫹ 1243 (32.7) 1257 (33.1) ⫺0.8 141 (3.7) 147 (3.9) ⫺0.8 Ejection fraction, n (%) 0.78 20–29% 68 (1.8) 67 (1.8) 0.2 30–39% 209 (5.5) 198 (5.2) 1.3 ⱖ40% 3245 (85.3) 3242 (85.2) Missing 140 (3.7) 149 (3.9) 0.2 ⫺1.2 No. of diseased vessels, n (%) 1.0 2 Diseased vessels 3577 (94.1) 3577 (94.1) 3 Diseased vessels 226 (5.9) 226 (5.9) 0 503 (13.2) 503 (13.2) 0 Presence of total occlusion, n (%) 0.74 0.97 25–29.99 ⬍20% P Value 0 Previous myocardial infarction, n (%) 1.0 0.31 1–7 d 921 (24.2) 883 (23.2) 2.3 8–20 d 128 (3.4) 113 (3.0) ⱖ21 d 376 (9.9) 412 (10.8) ⫺3.1 2.3 2378 (62.5) 2395 (63.0) ⫺0.9 Cerebrovascular disease, n (%) 197 (5.2) 204 (5.4) ⫺0.8 0.70 Peripheral arterial disease, n (%) 162 (4.3) 155 (4.1) 0.9 0.67 3787 (99.6) 3787 (99.6) 0 0.87 13 (0.3) 13 (0.3) 0 3 (0.1) 3 (0.1) 0 4 (0.1) 1 (0.0) 3.1 269 (7.1) 259 (6.8) 1.0 No myocardial infarction before procedure Hemodynamic state, n (%) Stable Unstable Shock Cardiopulmonary resuscitation, n (%) Left ventricular hypertrophy, n (%) Congestive heart failure, n (%) 197 (5.2) 187 (4.9) Before this admission 134 (3.5) 114 (3.0) 3472 (91.3) 3502 (92.1) 49 (1.3) 48 (1.3) Malignant ventricular arrhythmia, n (%) 0.65 0.46 This admission None 0.18 1.2 3.0 ⫺2.9 0.2 0.92 (Continued) Wu et al Table 2. Incomplete Revascularization and Stenting 417 Continued Incomplete Revascularization (n⫽3803) Complete Revascularization (n⫽3803) Chronic obstructive pulmonary disease, n (%) 186 (4.9) 179 (4.7) Diabetes, n (%) 870 (22.9) Requiring dialysis Creatinine ⬎2.5 mg/dL Risk Factor Standardized Difference, % 0.9 0.69 882 (23.2) ⫺0.7 0.73 36 (0.9) 37 (1.0) ⫺0.3 21 (0.6) 26 (0.7) ⫺1.7 3746 (98.5) 3740 (98.3) Renal failure, n (%) No renal failure P Value 0.36 Downloaded from http://circinterventions.ahajournals.org/ by guest on November 20, 2016 76.6% and 79.5% for the patients with multiple-vessel IR and the matched CR patients, respectively (P⫽0.21). Table 4 shows that across all subgroups of patients, the adjusted HRs were greater than 1. For some of the risk factors, one category was significant (ejection fraction ⱖ40%, no history of MI, no congestive heart failure, left anterior descending disease, proximal vessel disease, no total occlusion), and the other category (eg, ejection fraction ⬍40%) was not. In at least some cases, this was a result of not enough statistical power. Also none of the interaction terms between IR and selected risk factors was statistically significant (all probability values ⱖ0.35). Therefore, the results suggest that the impact of IR on long-term mortality is not highly dependent on the baseline risk factors in the study. The sensitivity analysis comparing the survival between IR and CR in the entire unmatched patient population of 9213 IR and 3803 CR patients showed that 2564 IR patients and 731 CR patients died during the follow-up period. Adjusting for baseline risk factors, IR was associated with higher risk of mortality than CR (adjusted HR⫽1.16; 95% CI, 1.06 –1.26, P⫽0.001). In addition, a higher risk of mortality for IR was observed among the patients with 1 incompletely revascularized vessel (adjusted HR⫽1.15; 95% CI, 1.05–1.26, P⫽0.002) and the patients with multiple incompletely revascularized vessels (adjusted HR⫽1.28; 95% CI, 1.15–1.42, P⬍0.0001). 1.3 Discussion In this large population-based study, we examined the impact of IR on long-term mortality after PCI for patients with multi-vessel disease in real-world clinical practice. We found that IR was associated with a 12% increased risk of death compared with CR for coronary stenting with BMS for patients with multi-vessel disease during follow-up averaged 8 years (HR⫽1.12; 95% CI, 1.01–1.26, P⫽0.04; 95% bootstrap CI, 0.98 –1.32). We also found that the survival disadvantage associated with IR was not limited to any subgroups of patients as determined by age, ejection fraction, history of MI, congestive heart failure, diabetes, left anterior descending artery disease, proximal vessel disease, or the presence of a total occlusion. While the increased risk of death was not statistically significant when the subgroups of 1-vessel IR and multiple-vessel IR were assessed (HR⫽1.11; P⫽0.08, and HR⫽1.20, P⫽0.22, respectively), the nonsignificant probability values for each type of IR were likely a result of the loss of statistical power because the sample size for each type was smaller than the overall sample size. The survival disadvantage of IR found in this study with longer length of follow-up (7–9 years) is relatively consistent with the results of earlier studies, but somewhat weaker.17,20 For example, an earlier New York study reported that the risk of death for IR was 15% higher than CR during 3-years of Figure 1. Kaplan-Meier survival curves for propensity matched-stenting patients. 418 Circ Cardiovasc Interv October 2011 Table 3. Hazard Ratios (IR and IR Subgroups Versus CR) for Mortality Patient Group No. of Cases No. of Deaths Hazard Ratio (95% CI) P Value IR 3803 817 1.12 (1.01, 1.26) 0.04 CR 3803 731 Reference 1 IR vessel 3268 690 1.11 (0.99, 1.25) 0.08 Matched CR patients 3268 625 Multiple IR vessels 535 127 1.20 (0.90, 1.59) 0.22 Matched CR patients 535 106 Subgroups of IR IR indicates incomplete revascularization; CR, complete revascularization; and CI, confidence interval. Downloaded from http://circinterventions.ahajournals.org/ by guest on November 20, 2016 follow-up in the BMS era.17 In this study, the difference in 3-year survival rates between IR and CR was not statistically significant (91.9% versus 92.3%, P⫽0.38). However, our study found that for the 8-year period examined, the survival rate for IR was consistently lower than CR and the differences were statistically significant at years 6 (83.7% versus 85.8%, P⫽0.04), 7 (80.8% versus 83.4%, P⫽0.02), and 8 (78.5% versus 80.8%, P⫽0.04). Whether differences of this magnitude (2.3% survival at 8 years) should be cause for changes in the way PCI is performed is a topic for future debate. Similarly, other studies with shorter follow-up periods have found that IR was associated with increased risks of mortality or MACE.2,13,15,16,18,19,21 It is worth noting that some previous studies have found that there were no differences between IR and CR in the incidence of MACE including mortality after PCI in the stent era.12,14 One possible explanation for these results is that the sample sizes in these two reports were so small that the studies did not have the statistical power to show a difference. For example, Ijsselmuiden et al reported that in a trial that enrolled 111 IR (culprit vessel revascularization) and 108 CR patients, MACE (death, myocardial infarction, or repeat revascularization) rates at 1 year were 32.4% and 26.9% for IR and CR, respectively, but the difference was not statistically significant (P⫽0.37), probably due to the small sample size.14 In another study by Mariani et al, the respective 1-year MACE rates for IR and CR were 11.5% and 11.3% (probability value not significant) in a group that included 159 IR and 49 CR patients.12 Although the sample size of this study is quite small, the magnitude of the difference is also small, so the findings may not be compromised by limited statistical power. A few strengths of our study ensure its validity and generalizability. First, this is a population-based study that included all eligible patients in New York State in the study period therefore the results are unlikely to be influenced by the practice pattern of a few hospitals. Second, the data used in this study are robust. The data in the PCIRS and CSRS are rigorously audited to ensure their completeness and accuracy. Third, the National Death Index was used to ascertain the vital status of patients so the patients lost to follow up is minimal. Last, the study evaluated the impact of IR in real-world practice. Although clinical trials are best at controlling for selection bias and confounding, patients are usually enrolled in trials based on strict selection criteria; therefore they are sometimes not representative of the patient population in real-world practice. In addition, patients in clinical trials are usually periodically monitored for symptoms, and such monitoring can trigger necessary treatment that may reduce the incidence of adverse outcomes. Patients treated in typical clinical practice usually are not monitored as closely as their counterparts in trials.31 Our study also has limitations. First, because this is an observational study, treatment selection bias could undermine the validity of the study. To control for this bias, we conducted a propensity-matched analysis. Through propensity matching, we were able to balance the distributions of baseline risk factor between the matched IR and CR groups. In addition, we conducted a sensitivity analysis using the unmatched patient population and adjusting for baseline risk factors, and the results were consistent with that of the Figure 2. Kaplan-Meier survival curves for patients with 1-vessel incomplete revascularization (IR) and propensity matched patients with complete revascularization. Wu et al Incomplete Revascularization and Stenting 419 Figure 3. Kaplan-Meier survival curves for patients with at least multiple-vessel incomplete revascularization (IR) and propensity matched patients with complete revascularization. Downloaded from http://circinterventions.ahajournals.org/ by guest on November 20, 2016 propensity-matched analysis. Nevertheless, risk factors like lesion length and vessel size were not available in our databases, and therefore could not be used in the propensity models. The inability to use these and any other important Table 4. missing predictors of adverse outcomes could have introduced a bias in our study. Second, the definition of IR in this study was determined by anatomic criteria instead of functional criteria such as Hazard Ratios (IR Versus CR) for Mortality for Selected Subgroups of Patients No. of Deaths/ Total Cases Patient Group Age ⬍80 y ⱖ80 y Ejection fraction ⬍40% ⱖ40% History of myocardial infarction Yes No Congestive heart failure Yes No Diabetes Yes No Left anterior descending artery disease Yes No Proximal vessel disease Yes No Total occlusion Yes No 8-Year KM Survival, % Adjusted* Hazard Ratio for Death CR IR CR IR P Value Hazard Ratio (95% CI) P Value 565/3480 166/323 634/3486 183/317 83.8 47.5 81.8 41.8 0.03 0.20 1.12 (1.00, 1.25) 1.23 (0.99, 1.52) 0.06 0.06 136/412 554/3242 142/418 639/3245 66.7 82.8 65.7 80.3 0.66 0.01 1.14 (0.90, 1.45) 1.18 (1.06, 1.33) 0.28 0.004 309/1408 422/2395 341/1425 476/2378 77.9 82.5 76.2 79.8 0.21 0.03 1.12 (0.96, 1.31) 1.19 (1.05, 1.36) 0.15 0.008 164/301 567/3502 165/331 652/3472 45.1 83.8 50.3 81.2 0.34 0.005 1.09 (0.87, 1.36) 1.19 (1.06, 1.33) 0.45 0.003 234/882 497/2921 259/870 558/2933 73.4 83.0 70.5 80.9 0.12 0.05 1.22 (1.02, 1.46) 1.14 (1.01, 1.28) 0.03 0.04 520/2682 211/1121 603/2734 214/1069 80.4 81.6 78.1 79.5 0.01 0.53 1.19 (1.05, 1.33) 1.11 (0.92, 1.34) 0.005 0.29 466/2374 265/1429 466/2064 351/1739 80.3 81.5 77.1 80.1 0.02 0.25 1.21 (1.06, 1.38) 1.10 (0.93, 1.29) 0.004 0.26 93/503 638/3300 108/503 709/3300 81.6 80.6 78.9 78.4 0.22 0.03 1.18 (0.89, 1.55) 1.16 (1.04, 1.29) 0.25 0.006 P Value for Interaction 0.45 0.80 0.54 0.50 0.51 0.57 0.35 0.94 IR indicates incomplete revascularization; CR, complete revascularization; and CI, confidence interval. *In addition to incomplete revascularization, patient group and interaction between incomplete revascularization and patient group, models adjusted for age, race, body mass index, body surface area, ejection fraction group, number of diseased vessels, cerebrovascular disease, peripheral arterial disease, hemodynamic state, ECG evidence of left ventricular hypertrophy, congestive heart failure, chronic obstructive pulmonary disease, diabetes, and renal failure. 420 Circ Cardiovasc Interv October 2011 Downloaded from http://circinterventions.ahajournals.org/ by guest on November 20, 2016 fractional flow reserve (FFR) measurements. It has been reported that FFR guided PCI was associated with significantly lower risk of mortality and myocardial infarction during a 2-year follow-up when compared with PCI guided by angiography alone.32,33 Although we had no data on FFR in our study, it is unlikely that it was used much given the time frame of our study. The findings of the effectiveness of FFR31,32 in conjunction with our study findings suggest that the increased use of FFR could result in more CR and improved outcomes, although cost-effectiveness would also have to be assessed. Third, the study is limited to BMS in the stent group. This is necessary in order to obtain 8-year follow-up, and there are many areas in which BMS are used extensively. One report estimates that in 2007 the proportion of stented lesions treated with BMS was 25% in the Mayo Clinic in the United States, but it was as high as 58% in Scotland, United Kingdom, 63% in Alberta, Canada, and 68% in Belgium.34 Nonetheless, DES are used more than BMS in many regions, and the results reported here may not be true for DES. Fourth, to obtain longer-term results, we had to incorporate data from more than a decade ago along with more recent data, and stenting practice and technology have changed considerably since then. Also, antiplatelet and other pharmacological therapies have improved significantly. Thus, the relative results for IR and CR patients may have changed in the interim. Also, in the subgroup analyses, CR and IR were not matched for baseline variables. However, the results were risk-adjusted to account for differences in the baseline variables between CR and IR patients. Conclusions IR may be associated with a higher risk of mortality during long-term follow-up that after coronary stenting with BMS in patients with multivessel disease. These findings should be considered in determining how PCI is performed and in selecting the most appropriate means of revascularization, for example, PCI or CABG surgery. Also, more prospective studies are needed to further test this association. Acknowledgments We thank the New York State Cardiac Advisory Committee for their encouragement and support of this study and Kimberly Cozzens, Cynthia Johnson, and the cardiac catheterization laboratories of the participating hospitals for their tireless efforts to ensure the timeliness, completeness, and accuracy of the registry data. Sources of Funding This work was supported by the National Institutes of Health (RC1HL099122). Disclosures Dr Jacobs reports serving as a site Principal Investigator for research projects funded by Abbot Vascular and Accumetrics. Other authors report that there are no potential conflicts of interest. References 1. Teirstein PS. The dueling hazards of incomplete revascularization and incomplete data. Circulation. 2006;113:2380 –2382. 2. Ong ATL, Serruys PW. Complete revascularization: coronary artery bypass graft surgery versus percutaneous coronary intervention. Circulation. 2006; 114:249 –255. 3. Vandormael MG, Chaitman BR, Ischinger T, Aker UT, Harper M, Hernandez J, Deligonul U, Kennedy HL. Immediate and short-term benefit of multilesion coronary angioplasty - influence of degree of revascularization. J Am Coll Cardiol. 1985;6:983–991. 4. Reeder GS, Holmes DR Jr, Detre K, Costigan T, Kelsey SF. Degree of revascularization in patients with multivessel coronary disease: a report from the National Heart, Lung, and Blood Institute Percutaneous Transluminal Coronary Angioplasty Registry. Circulation. 1988;77: 638 – 644. 5. Bell MR, Bailey KR, Reeder GS, Lapeyre AC, Holmes DR. Percutaneous transluminal angioplasty in patients with multivessel coronary disease: how important is complete revascularization for cardiac event-free survival. J Am Coll Cardiol. 1990;16:553–562. 6. Samson M, Meester HJ, Defeyter PJ, Strauss B, Serruys PW. Successful multiple segment coronary angioplasty: effect of completeness of revascularization in single-vessel multilesions and multivessels. Am Heart J. 1990;120:1–12. 7. Faxon DP, Ghalilli K, Jacobs AK, Ruocco NA, Christellis EM, Kellett MA, Varrichione TR, Ryan TJ. The degree of revascularization and outcome after multivessel coronary angioplasty. Am Heart J. 1992;123: 854 – 859. 8. Cowley MJ, Vandermael M, Topol EJ, Whitlow PL, Dean LS, Bulle TM, Ellis SG. Is traditionally defined complete revascularization needed for patients with multivessel disease treated by elective coronary angioplasty. J Am Coll Cardiol. 1993;22:1289 –1297. 9. Bourassa MG, Yeh W, Holubkov R, Sopko G, Detre KM, Registry NP. Long-term outcome of patients with incomplete vs complete revascularization after multivessel PTCA: a report from the NHLBI PTCA registry. Eur Heart J. 1998;19:103–111. 10. Bourassa MG, Kip KE, Jacobs AK, Jones RH, Sopko G, Rosen AD, Sharaf BL, Schwartz L, Chaitman BR, Alderman EL, Holmes DR, Roubin GS, Detre KM, Frye RL, Investigators B. Is a strategy of intended incomplete percutaneous transluminal coronary angioplasty revascularization acceptable in nondiabetic patients who are candidates for coronary artery bypass graft surgery? The Bypass Angioplasty Revascularization Investigation (BARI). J Am Coll Cardiol. 1999;33:1627–1636. 11. Kip KE, Bourassa MG, Jacobs AK, Schwartz L, Feit F, Alderman EL, Weiner BH, Weiss MB, Kellett MA, Sharaf BL, Dimas AP, Jones RH, Sopko G, Detre KM, Investigators B. Influence of Pre-PTCA strategy and initial PTCA result in patients with multivessel disease: the Bypass Angioplasty Revascularization Investigation (BARI). Circulation. 1999; 100:910 –917. 12. Mariani G, De Servi S, Dellavalle A, Repetto S, Chierchia S, D’Urbano M, Repetto A, Klersy C, Grp RS. Complete or incomplete percutaneous coronary revascularization in patients with unstable angina in stent era: are early and one-year results different? Catheter Cardiovasc Interv. 2001;54:448 – 453. 13. van den Brand MJ, Rensing BJ, Morel MA, Foley DP, de Valk V, Breeman A, Suryapranata H, Haalebos MM, Wijns W, Wellens F, Balcon R, Magee P, Ribeiro E, Buffolo E, Unger F, Serruys PW. The effect of completeness of revascularization on event-free survival at one year in the ARTS trial. J Am Coll Cardiol. 2002;39:559 –564. 14. Ijsselmuiden AJJ, Ezechiels JP, Westendorp ICD, Tijssen JGP, Kiemeneij F, Slagboom T, van der Wieken R, Tangelder GJ, Serruys PW, Laarman GJ. Complete versus culprit vessel percutaneous coronary intervention in multivessel disease: a randomized comparison. Am Heart J. 2004;148: 467– 474. 15. Nikolsky E, Gruberg L, Patil CV, Roguin A, Kapeliovich M, Petcherski S, Boulos M, Grenadier E, Amikam S, Linn S, Markiewicz W, Beyar R. Percutaneous coronary interventions in diabetic patients: is complete revascularization important? J Invasive Cardiol. 2004;16:102–106. 16. McLellan CS, Ghali WA, Labinaz M, Davis RB, Galbraith PD, Southern DA, Shrive FM, Knudtson ML, Investigators A. Association between completeness of percutaneous coronary revascularization and postprocedure outcomes. Am Heart J. 2005;150:800 – 806. 17. Hannan EL, Racz M, Holmes DR, King SB, Walford G, Ambrose JA, Sharma S, Katz S, Clark LT, Jones RH. Impact of completeness of percutaneous coronary intervention revascularization on long-term outcomes in the stent era. Circulation. 2006;113:2406 –2412. 18. Kalarus Z, Lenarczyk R, Kowalczyk J, Kowalski O, Gasior M, Was T, Zebik T, Krupa H, Chodor P, Polonski L, Zembala M. Importance of complete revascularization in patients with acute myocardial infarction treated with percutaneous coronary intervention. Am Heart J. 2007;153: 304 –312. Wu et al Downloaded from http://circinterventions.ahajournals.org/ by guest on November 20, 2016 19. Tamburino C, Angiolillo DJ, Capranzano P, Dimopoulos K, La Manna A, Barbagallo R, Tagliareni F, Mangiafico S, Guzman LA, Galassi AR, Bass TA. Complete versus incomplete revascularization in patients with multivessel disease undergoing percutaneous coronary intervention with drug-eluting stents. Catheter Cardiovasc Interv. 2008;72:448 – 456. 20. Hannan EL, Wu C, Walford G, Holmes DR, Jones RH, Sharma S, King SB. Incomplete revascularization in the era of drug-eluting stents impact on adverse outcomes. JACC Cardiovasc Intervent. 2009;2:17–25. 21. Lehmann R, Fichtlscherer S, Schachinger V, Held L, Hobler C, Baier G, Zeiher AM, Spyridopoulos I. Complete revascularization in patients undergoing multivessel PCI is an independent predictor of improved long-term survival. J Interv Cardiol. 2010;23:256 –263. 22. The National Center for Health Statistics. The National Death Index. http://www.cdc.gov/nchs/ndi.htm. 23. Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika. 1983;70:41–55. 24. Rosenbaum PR, Rubin DB. Reducing bias in observational studies using subclassification on the propensity score. J Am Stat Assoc. 1984;79: 516 –524. 25. Gu XS, Rosenbaum PR. Comparison of multivariate matching methods: structures, distances, and algorithms. J Comput Graph Stat. 1993;2: 405– 420. 26. Austin PC. Propensity-score matching in the cardiovascular surgery literature from 2004 to 2006: a systematic review and suggestions for improvement. J Thorac Cardiovasc Surg. 2007;134:1128 –1135. 27. Normand SL. Some old and some new statistical tools for outcomes research. Circulation. 2008;118:872– 884. Incomplete Revascularization and Stenting 421 28. Klein JP, Moeschberger ML. Survival Analysis: Techniques for Censored and Truncated Data II ed. New York, NY: Springer; 2003. 29. Hosmer DW, Lemeshow S, May S. Applied Survival Analysis: Regression Modeling of Time to Event Data. 2nd ed. Hoboken, NJ: Wiley-Interscience; 2008. 30. Efron B. Better bootstrap confidence intervals. J Am Stat Assoc. 1987; 82:171–185. 31. Hannan EL. Randomized clinical trials and observational studies: guidelines for assessing respective strengths and limitations. J Am Coll Cardiol Cardiovasc Interv. 2008;1:211–217. 32. Pijls NH, Fearon WF, Tonino PA, Siebert U, Ikeno F, Bornschein B, van’t Veer M, Klauss V, Manoharan G, Engstrom T, Oldroyd KG, Ver Lee PN, MacCarthy PA, De Bruyne B, Investigators FS. Fractional flow reserve versus angiography for guiding percutaneous coronary intervention in patients with multivessel coronary artery disease: 2-year follow-up of the FAME (Fractional Flow Reserve Versus Angiography for Multivessel Evaluation) study. J Am Coll Cardiol. 2010;56:177–184. 33. Tonino PA, De Bruyne B, Pijls NH, Siebert U, Ikeno F, van’ t Veer M, Klauss V, Manoharan G, Engstrom T, Oldroyd KG, Ver Lee PN, MacCarthy PA, Fearon WF, Investigators FS. Fractional flow reserve versus angiography for guiding percutaneous coronary intervention. N Engl J Med. 2009;360:213–224. 34. Austin D, Oldroyd KG, Holmes DR, Rihal CS, Galbraith PD, Ghali WA, Legrand V, Taeymans Y, McConnachie A, Pell JP, PCI AIBWGIMC, Registry; Scottish C. Drug-eluting stents: a study of international practice. Am Heart J. 2009;158:576 –584. Downloaded from http://circinterventions.ahajournals.org/ by guest on November 20, 2016 Impact of Incomplete Revascularization on Long-Term Mortality After Coronary Stenting Chuntao Wu, Anne-Marie Dyer, Spencer B. King III, Gary Walford, David R. Holmes, Jr, Nicholas J. Stamato, Ferdinand J. Venditti, Samin K. Sharma, Icilma Fergus, Alice K. Jacobs and Edward L. Hannan Circ Cardiovasc Interv. 2011;4:413-421; originally published online October 4, 2011; doi: 10.1161/CIRCINTERVENTIONS.111.963058 Circulation: Cardiovascular Interventions is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231 Copyright © 2011 American Heart Association, Inc. All rights reserved. Print ISSN: 1941-7640. Online ISSN: 1941-7632 The online version of this article, along with updated information and services, is located on the World Wide Web at: http://circinterventions.ahajournals.org/content/4/5/413 Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in Circulation: Cardiovascular Interventions can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for which permission is being requested is located, click Request Permissions in the middle column of the Web page under Services. Further information about this process is available in the Permissions and Rights Question and Answer document. Reprints: Information about reprints can be found online at: http://www.lww.com/reprints Subscriptions: Information about subscribing to Circulation: Cardiovascular Interventions is online at: http://circinterventions.ahajournals.org//subscriptions/