Outcomes of patients who present with ST-segment. Original Article

Transcription

1 Original Article Lower Hospital Volume Is Associated With Higher In-Hospital Mortality in Patients Undergoing Primary Percutaneous Coronary Intervention for ST-Segment Elevation Myocardial Infarction A Report From the NCDR Michael C. Kontos, MD; Yongfei Wang, MS; Sarwat I. Chaudhry, MD; George W. Vetrovec, MD; Jeptha Curtis, MD; John Messenger, MD; on behalf of the NCDR Background Current guidelines recommend >36 primary percutaneous coronary interventions (PCIs) per hospital per year. Whether these standards remain valid when routine coronary stenting and newer pharmacological agents are used is unclear. Methods and Results We analyzed patients who underwent primary PCI from July 2006 through June 2009 included in the CathPCI Registry. Hospitals were separated into 3 groups: low ( 36 primary PCIs/y, current guideline recommendation), intermediate (>36 60 primary PCIs/y), and high volume (>60 primary PCIs/y). In-hospital mortality and door-to-balloon time were examined for each group. A total of patient visits for patients from 738 hospitals were included. There were 278 low- (38%), 236 (32%) intermediate-, and 224 (30%) high-volume hospitals. The majority of patients with primary PCI (54%) were treated at high-volume hospitals, with 15% at low-volume hospitals. Unadjusted mortality was significantly higher in low-volume hospitals compared with high-volume hospitals (5.6% versus 4.8%; P<0.001), which was maintained after multivariate adjustment (1.20; 95% confidence interval, ; P=0.001). In contrast, mortality was not significantly different between intermediate-volume and high-volume hospitals (4.8% versus 4.8%; adjusted odds ratio, 1.02; 95% confidence interval, ; P=0.61). Door-to-balloon times were significantly shorter in high-volume hospitals compared with low-volume hospitals (median, 72 minutes; interquartile range, [53 91] versus 77 [57 100] minutes; P<0.0001). Conclusions Higher annual hospital volume of primary PCI continues to be associated with lower mortality, with higher mortality in hospitals performing 36 primary PCIs/y. (Circ Cardiovasc Qual Outcomes. 2013;6: ) Key Words: myocardial infarction outcomes assessment percutaneous coronary intervention Outcomes of patients who present with ST-segment elevation myocardial infarction (STEMI) are time dependent, with increasing time to reperfusion associated with worse outcomes. 1 Important additional factors associated with outcomes include patient characteristics, 2 physician volume, 3 5 and hospital procedural volume Hospital procedural volume has been a consistent predictor of outcomes for multiple cardiac procedures, 11,12 including primary percutaneous coronary intervention (PCI) for STEMI. 4 8 The point at which volume affects outcomes has varied among studies. Initial studies, performed before primary PCI was the preferred reperfusion method, demonstrated a significant reduction in mortality when yearly primary PCI hospital volume exceeded 13 6 or 33 7 patients/y. Most, 9,11 but not all, 13 recent studies, in which coronary stenting and use of newer pharmacological treatments are routine, have continued to demonstrate improved outcomes with higher hospital volumes. Current guidelines for primary PCI recommend a minimum 36 primary PCIs per hospital per year. However, recent data supporting this volume recommendation are lacking. 14 In addition, the number of hospitals performing PCI has increased, 15 despite no significant change in the annual rate of PCI, 16 suggesting that the number of PCIs per hospital has decreased. Because procedural volume is increasingly being used as a surrogate for quality of care 10 by several organizations and agencies, additional examination of the volume relationship with outcomes is warranted. We hypothesized that higher volume primary PCI hospitals would be associated with lower mortality and shorter doorto-balloon (D2B) times compared with low-volume hospitals. Received March 19, 2013; accepted September 27, From the Division of Cardiology, Pauley Heart Center, Virginia Commonwealth University, Richmond, VA (M.C.K., G.W.V.); Section of Cardiovascular Medicine, Department of Internal Medicine, Yale University School of Medicine, New Haven, CT (Y.W., J.C.); Section of General Internal Medicine, Yale University, New Haven, CT (S.I.C.); and Division of Cardiology, University of Colorado School of Medicine, Aurora, CO (J.M.). This article was handled by Gregg C. Fonarow, MD, as Guest Editor. The Editors had no role in the evaluation of this manuscript or in the decision about its acceptance. Correspondence to Michael C. Kontos, MD, Room 285 Gateway Bldg, 2nd Floor Gateway, PO Box , 1200 E Marshall St, Richmond, VA mkontos@mcvh-vcu.edu 2013 American Heart Association, Inc. Circ Cardiovasc Qual Outcomes is available at DOI: /CIRCOUTCOMES

2 660 Circ Cardiovasc Qual Outcomes November 2013 WHAT IS KNOWN Hospital procedural volume has been a consistent predictor of outcomes for primary percutaneous coronary intervention (PCI) for ST-segment elevation myocardial infarction, resulting in the current competency statements and guidelines that recommend >36 primary PCISs per hospital per year. Whether these standards remain valid when routine coronary stenting and newer pharmacological agents are used is unclear. WHAT THE STUDY ADDS In a large cohort of consecutive patients who underwent primary PCI included in the CathPCI Registry, the majority (54%) were treated at high-volume hospital (>60 primary PCIs/y), with 15% at low-volume hospitals (<36 primary PCIs/y). Unadjusted and adjusted mortality (1.20; 95% confidence interval, ; P=0.001) was significantly higher in low-volume compared with intermediate-volume (36 60 primary PCIs/y) and high-volume hospitals. Higher annual hospital volume of primary PCI continues to be associated with lower mortality, supporting current volume recommendations. Methods Data Source The CathPCI Registry is a large, ongoing registry of clinical data and procedural outcomes associated with diagnostic cardiac catheterization and PCI including >1400 participating sites sponsored by the American College of Cardiology and the Society for Cardiovascular Angiography and Interventions. 17 Data are captured by the National Cardiovascular Data Registry (NCDR)-certified databases at participating hospitals and include patient and hospital characteristics, procedural findings, interventions, and outcomes. The NCDR also has a data quality program, including data abstraction training, data quality thresholds for inclusion, site data quality feedback reports, independent auditing, and data validation. 18 Data auditing has demonstrated accurate representation with agreement with chart review of >93%. 19 Data elements and definitions are available at WebNCDR/ELEMENTS.ASPX#1. Patient Population We included all visits for patients presenting with STEMI from July 2006 to June 2009 using version 3 of the data-collection form. Visits were excluded if a patient did not undergo PCI during the visit or was not the first PCI during the hospital admission and if the PCI was not classified as emergent or salvage or was not primary PCI for STEMI (Figure 1). Patients who were transferred in for primary PCI were excluded, as this may represent selection bias and a higher risk group. 7 Patients in whom the D2B time was either unknown, <15 minutes, or >6 hours or received thrombolytics before PCI were excluded. In addition, hospitals were excluded if they treated >20% of their patients with thrombolytics because this may identify hospitals that are preferentially treating lower risk patients 20 or transferred out >10% of their patients because of inability to assess mortality. Finally, consistent with prior studies, we excluded visits in hospitals with an annual volume of <5 cases/y. 6,7,13 The final study sample included admissions for patients from 738 hospitals. Main Interested Variable, Outcomes, and Subgroups Hospital annual volume was calculated as: (number of admissions)/ (number of quarters the hospital contributed data to the registry in the 3 years) multiplied by 4. Hospitals were separated into 3 groups based on annualized primary PCI volume: low volume (n=278), hospitals performing 36 primary PCIs a year (current guidelines). 14 The remaining 460 (62%) hospitals were separated into 2 relatively equal groups: intermediate volume (n=236), 37 to 60 primary PCIs/y, and high volume (n=224), >60 primary PCIs/y. The primary end points of this study were D2B time and in-hospital mortality. Secondary outcomes included complications (vascular complications, emergent bypass surgery, bleeding, and transfusion) during the admission. Two specific subgroups were predefined based on the likelihood of worse outcomes: patients presenting during off hours (presentation from 5 pm to 7 am Monday to Friday or on weekends) and patients presenting with cardiogenic shock at the time of admission (a sustained episode of hypotension, cardiac index <2.2 L/min per m 2 secondary to cardiac dysfunction, or the requirement for inotropic or vasopressor agents or mechanical support). Statistical Analysis We evaluated the baseline patient clinical characteristics, hospital characteristics, and outcomes among the hospital groups determined Figure 1. Exclusion criteria and study population. Numbers in parentheses represent number of hospitals. D2B indicates door-to-balloon; PCI, percutaneous coronary intervention; and STEMI, ST-segment elevation myocardial infarction.

3 Kontos et al Hospital Volume, Primary PCI, and Mortality 661 Table 1. Hospital Characteristics Stratified by Hospital Annual Primary PCI Volume All < >60 P Value Hospital No No. of PCIs 756± ± ± ±651 <0.001 Patient beds 383± ± ± ±217 <0.001 Surgical backup 625 (85) 215 (77) 207 (88) 203 (91) <0.001 present Hospital location Rural 111 (15) 53 (19) 36 (15) 22 (10) Suburban 251 (34) 92 (33) 77 (33) 82 (37) Urban 376 (51) 133 (48) 123 (52) 120 (54) Hospital type Government 11 (15) 4 (1.4) ) 2 (0.9) Private/community 658 (89) 237 (85) 214 (91) 207 (92) University 69 (9) 37 (13) 17 (7.2) 15 (7) Data presented as mean±sd or number (percent). PCI indicates percutaneous coronary intervention. by the hospital annual volume using ANOVA for continuous variables and the χ 2 test for categorical variables. Summary of n (%) for categorical variables and mean (SD) as well as median (interquartile range) for continuous variables are reported. To examine the independent and declining effect of hospital annual volume on the primary outcomes, logistic regression models were used adjusting sequentially for patient characteristics: patient demographics, risk factors, medical history (age, renal function [glomerular filtration rate if nondialysis dependent], body mass index, history of heart failure, valvular surgery or valvular disease, cerebrovascular disease, chronic lung disease, previous PCI, diabetes mellitus [noninsulin or insulin dependent]), cardiac status (angina class IV, cardiogenic shock on admission), coronary anatomy, the PCI procedure (intra-aortic balloon pump placement, ejection fraction, coronary lesion 50%, subacute thrombosis [yes or no], highest risk lesion [Society for Cardiovascular Angiography and Interventions (SCAI) lesion class 2 or 3 versus class 4], highest risk preprocedure Table 2. Patients Characteristics Stratified by Hospital Annual Volume Thrombolysis in Myocardial Infarction (TIMI) flow, PCI vessel and segment, and salvage PCI, and D2B time. Variables included were those used for the NCDR mortality equation, 21 with the addition of D2B time, but different modeling was used as different end points were evaluated. D2B time was analyzed using Cox regression, with censoring at 90 minutes, with similar variables included. Odds ratios for low- and intermediate-volume hospitals were reported using high-volume hospitals as the reference group. The Human Investigation Committee of the Yale University School of Medicine approved the use of data from the NCDR Registry for research purposes. CathPCI Registry is an initiative of the American College of Cardiology Foundation and the Society for Cardiovascular Angiography and Interventions. Results Baseline Characteristics A total of admissions for patients presenting to 738 hospitals were included in the analysis. There were 278 hospitals (38%) classified as low-volume, 236 (32%) as intermediate-volume, and 224 (30%) as high-volume hospitals. The majority of patients with primary PCI (54%) were treated at high-volume hospitals, with lesser percentages at intermediate- (30%) and low-volume hospitals (15%; Table 1). Low-volume hospitals performed fewer total PCIs and were less likely to have cardiac surgery backup. High-volume hospitals were more likely to be urban than rural or suburban and were more likely to be a private/community hospital than an academic hospital. Distributions of age and sex were comparable and not significantly different across the 3 groups (Table 2). Absolute differences were relatively small among most of the other variables, although they were statistically different, including diabetes mellitus, renal function, prior myocardial infarction, and prior revascularization. Patients with heart failure on admission were more common at low-volume hospitals (10.6%) compared with intermediate-volume (9.2%) and high-volume (8.5%) hospitals (P<0.0001; Table 2). A total of All Patients 36 >36 60 >60 P Value All patients Age >70 y (24) 3263 (25) 6324 (24) (24) 0.12 Women (28) 3840 (29) 7523 (28) (28) 0.21 Hypertension (61) 8104 (61) (61) (62) 0.02 Diabetes mellitus (19) 2640 (20) 5153 (19) 8963 (19) GFR categories <0.001 Unknown 7663 (8.8) 1079 (8.1) 2279 (8.6) 4305 (9.1) <30 ml/min per m (2.8) 390 (2.9) 726 (2.7) 1289 (2.7) 30 <60 ml/min per m (23) 2935 (22) 5870 (22) (23) 60 ml/min per m (66) 8863 (67) (67) (65) <0.001 History of CHF 3706 (4.2) 627 (4.7) 1142 (4.3) 1937 (4.1) Prior PCI (21) 2675 (20) 5288 (20) 9952 (21) Prior CABG 4955 (5.7) 693 (5.2) 1458 (5.5) 2804 (5.9) Prior MI (>7 d) (18) 2356 (18) 4621 (17) 8753 (18) <0.001 CHF at admission 7857 (9.0) 1409 (10.6) 2436 (9.2) 4012 (8.5) <0.001 Cardiogenic shock 8840 (10.1) 1423 (10.7) 2554 (9.6) 4863 (10.2) <0.001 CABG indicates coronary artery bypass surgery; CHF, congestive heart failure; GFR, glomerular filtration rate; MI, myocardial infarction; and PCI, percutaneous coronary intervention.

4 662 Circ Cardiovasc Qual Outcomes November 2013 Table 3. Coronary Angiographic Results in Patients Undergoing Primary PCI All Patients 36 >36 60 >60 P Value IABP placed 9940 (11) 1541 (12) 3038 (11) 5361 (11) 0.58 EF, % 47±13 47±13 47±13 47±13 <0.001 EF<40% (17.0) 2059 (15.5) 4236 (15.9) 8592 (18.1) <0.001 Infarct-related artery Right coronary (48) 6413 (48) (48) (48) 0.40 Left main 581 (0.7) 87 (0.7) 155 (0.6) 339 (0.7) 0.11 Left anterior descending (40) 5424 (41) (40) (40) Left circumflex (16) 2087 (16) 4139 (16) 7607 (16) 0.22 Ramus 793 (0.9) 115 (0.9) 248 (0.9) 430 (0.9) 0.81 Data presented as mean±sd or number (percent). EF indicates ejection fraction; IABP, intra-aortic balloon pump; and PCI, percutaneous coronary intervention patients (60%) presented off hours, and 8840 patients (10%) had cardiogenic shock. Angiographic Results The infarct-related artery was most commonly the right coronary artery (48%) followed by the left anterior descending artery (40%; Table 3). Ejection fraction was significantly lower and more likely to be <40% (18.1 versus 15.5%; P<0.001) in high-volume hospitals compared with low-volume hospitals. Outcomes Patients in high-volume hospitals had shorter median D2B times (72 [53 91] versus 77 [57 100] minutes; P<0.0001; Figure 2; Table 4), with significantly more patients having a D2B time 90 minutes (72 versus 66%; P<0.0001) when compared with patients in low-volume hospitals. Differences in D2B times between intermediate- and high-volume hospitals (73 [55 94] versus 72 [53 91] minutes) were not significant. Unadjusted in-hospital mortality was significantly lower in high-volume hospitals compared with low-volume hospitals (4.8% versus 5.6%; P<0.0001; Figure 3; Table 5). However, mortality between intermediate- and high-volume hospitals (4.8% versus 4.8%) was similar. The correlation with hospital volume and mortality (r 2 =0.02; Figure 4A) and D2B times (r 2 =0.04; Figure 4B) was very low. Bleeding complications and need for blood products (all P<0.0001) were significantly higher in high-volume hospitals (Table 5). D2B times were slower for patients presenting off hours versus overall by 6 minutes, a finding consistent across all volume groups. Off-hour D2B times remained significantly shorter in high- compared with low-volume hospitals (median, 78 [60 96] versus 83 [63 103] minutes; P<0.0001). Overall mortality in patients presenting off hours was similar to that of all patients and remained significantly lower in high-volume hospitals compared with low-volume hospitals (4.9% versus 5.6%; P=0.01). High-volume hospitals also had shorter D2B times than low-volume hospitals in patients with cardiogenic shock (76 [55 98] versus 80 [54 106]; P=0.01), with no significant difference between high- and intermediate-volume hospitals. Mortality for cardiogenic shock in all volume groups was substantially increased, averaging >10 that of patients without shock. Although mortality was lower in high-volume compared with low-volume hospitals (27.6 versus 29.3%; P=0.31), the results were not significant. Multivariate Analysis After multivariate adjustment for patient and hospital characteristics, mortality was significantly higher (1.22; 95% confidence interval, ; P<0.0001) in low-volume compared with high-volume hospitals. After further adjustment for D2B time, mortality remained significantly higher in low-volume centers (1.20; 95% confidence interval, ; Table 6). In contrast, mortality was not significantly different between intermediate-volume and highvolume hospitals (adjusted odds ratio, 1.02; 95% confidence interval, ; P=0.61). A sensitivity analysis was performed in which transfer-in patients were included; mortality remained significantly different for low-volume versus highvolume centers (odds ratio, 1.14; 95% confidence interval, ; P<0.04). Results were similar in the subgroups of patients who presented off hours, in which mortality was significantly higher in low-volume hospitals, with no significant difference between intermediate-volume and high-volume hospitals (Table 6). However, there was no significant difference in mortality in patients with cardiogenic shock among the 3 groups. In the subgroup of patients without cardiogenic Figure 2. Door-to-balloon (D2B) times among hospital volume groups, for all patients, patients presenting off hours, and patients with cardiogenic shock. Gray bars indicate low volume ( 36 primary percutaneous coronary interventions [PCIs]/y); blue bars, intermediate volume (37 59 primary PCIs/y); and red bars, high-volume hospitals ( 60 primary PCIs/y). *P<0.001 compared with intermediate-volume and high-volume hospitals.

5 Kontos et al Hospital Volume, Primary PCI, and Mortality 663 Table 4. Door-to-Balloon Times Among All Patients, Patients Presenting Off Hours, and Patients With Cardiogenic Shock Description All Patients 36 >36 60 >60 P Value All patients No. of patients D2B time: mean 80±41 84±45 80±41 79±40 < D2B time: median 73 (54 93) 77 (57 100) 73 (55 94) 72 (53 91) < D2B time 90 min (71) 8797 (66) (72) (72) < Off hours No. of patients D2B time: mean 86±40 91±44 87±40 85±39 < D2B time: median 79 (62 97) 83 (63 103) 80 (63 98) 78 (60 96) < D2B time 90 min (65) 4685 (60) (65) (67) < Cardiogenic shock No. of patients D2B time: mean 86±44 89±47 87±46 84± D2B time median 77 (54 100) 80 (54 106) 78 (56 101) 76 (55 98) 0.01 D2B time 90 min 5734 (65) 857 (60) 1642 (64) 3235 (67) < Data presented as mean±sd, number (percent) or number (interquartile range). D2B indicates door-to-balloon time. shock, mortality remained higher in low-volume centers (2.7%) compared with intermediate-volume and high-volume centers (2.2% for both; P<0.004), which was significant after multivariate analysis (1.24; 95% confidence interval, ; P=0.003). After multivariate adjustment, low-volume hospitals were significantly less likely to achieve a D2B time 90 minutes compared with high-volume hospitals for all groups (Table 7). Intermediate-volume hospitals were less likely to achieve a D2B time 90 minutes, when considering all patients and those presenting off hours, but not those with cardiogenic shock (Table 7). Discussion Our results indicate that higher hospital primary PCI volume is associated with lower in-hospital mortality. After adjusting for patient and procedural variables, the risk of mortality was increased by 20% in low-volume as compared with Figure 3. Unadjusted mortality among hospital volume groups, for all patients, patients presenting off hours, and patients with cardiogenic shock. Gray bars indicate low volume ( 36 primary percutaneous coronary interventions [PCIs]/y); blue bars, intermediate volume (37 59 primary PCIs/y); and red bars, high-volume hospitals ( 60 primary PCIs/y). *P<0.001; #P=0.01, compared with intermediate-volume and high-volume hospitals. high-volume hospitals. In addition, D2B time was significantly shorter in high-volume hospitals. PCI Volume and Outcomes Current guidelines recommend that primary PCI for STEMI should be performed in institutions that perform >200 elective PCIs per year and >36 primary PCI procedures for STEMI per year. 14 When the initial recommendations were developed, it was recognized that threshold activity level standards for institutions and operators was controversial because of the limited available data. 22 Since the publication of the initial guidelines, the relationship between hospital volume and outcomes in patients undergoing PCI has been extensively investigated. Although the specific volume cut point and the method of assessing volume (provider, hospital) have varied, most studies have found that higher volumes are associated with lower mortality at the hospital 4,5,9 11 and operator 3 5,23 levels. This relationship has persisted in more recent studies. 5,9,10 Primary PCI Studies Data from clinical registries including patients undergoing primary PCI, such as the New York Coronary Angioplasty Reporting System Registry 4 and the National Registry of Myocardial Infarction (NRMI)-2, 6 8 have mirrored those of results of studies including a broad spectrum of patients undergoing PCI, with significantly lower mortality 7,8 as well as shorter D2B times 7 in higher volume hospitals, defined as >33 8 or 36 7 primary PCIs per year. Despite demonstrating a volume/mortality relationship, several factors potentially limit the ability to extrapolate these studies to current practice. Most were performed before primary PCI became the preferred treatment for STEMI, many patients underwent treatment with fibrinolytics, even at high-volume primary PCI hospitals, and there was less emphasis on achieving rapid reperfusion. 6 8 As a

6 664 Circ Cardiovasc Qual Outcomes November 2013 Table 5. Complications and Outcomes Among All Patients, Patients Presenting Off Hours, and Patients With Cardiogenic Shock All Patients 36 >36 60 >60 P Value All patients No. of patients Bleeding 4412 (5.1) 504 (3.8) 1210 (4.5) 2698 (5.7) < Vascular complications 739 (0.8) 102 (0.8) 222 (0.8) 415 (0.9) 0.49 Emergency CABG 1248 (1.4) 182 (1.4) 364 (1.4) 702 (1.5) 0.39 Blood products 7604 (8.7) 1030 (7.8) 2191 (8.2) 4383 (9.2) < In-hospital mortality 4283 (4.9) 737 (5.6) 1269 (4.8) 2277 (4.8) Off hours No. of patients Bleeding 2641 (5.0) 292 (3.7) 723 (4.5) 1626 (5.6) < Vascular complications 453 (0.9) 68 (0.9) 132 (0.8) 253 (0.9) 0.84 Emergency CABG 707 (1.3) 90 (1.1) 194 (1.2) 423 (1.5) Blood products 4520 (8.6) 593 (7.6) 1304 (8.1) 2623 (9.1) In-hospital mortality 2617 (5.0) 440 (5.6) 763 (4.8) 1414 (4.9) Shock No. of patients Bleeding 4166 (5.1) 460 (3.7) 1145 (4.6) 2561 (5.7) < Vascular complications 702 (0.9) 95 (0.8) 214 (0.9) 393 (0.9) 0.52 Emergency CABG 1178 (1.4) 169 (1.4) 340 (1.4) 669 (1.5) 0.31 Blood products 6999 (8.5) 914 (7.4) 1992 (8.0) 4093 (9.1) < In-hospital mortality 2497 (28) 417 (29) 738 (29) 1342 (28) 0.31 CABG indicates coronary artery bypass surgery. result, D2B times were substantially longer than in current practice, with median times >40 minutes longer 7 than those currently routinely achieved. 24 In addition, contemporary practice and management have changed substantially, with increased use of evidence-based therapies, standardization of care, and reduced practice variation. 25 The routine use of stenting has further decreased the risk of complications and adverse outcomes. 14 Two recent studies have reported on the hospital volume/ outcome relationship in patients undergoing primary PCI. Srinivas et al 3 examined the relationship between volume and mortality using the New York PCI database in 7321 patients undergoing primary PCI and found that lower volume for physicians (<10 primary PCIs/y) and hospitals (<50 primary PCIs/y) was associated with a significantly increased in-hospital mortality. In contrast, Kumbhani et al, 13 using data from the Get With the Guidelines Registry, found that mortality was higher in low-volume centers, and it was not significant. In this study, patients underwent primary PCI at 166 hospitals. Low-volume primary PCI hospitals (defined as <36 primary PCIs/y) had significantly longer D2B times (98 versus 88 minutes; P<0.001) compared with high-volume hospitals (>70 primary PCIs/y). However, unadjusted (3.9% versus 3.0%; Figure 4. Relationship of unadjusted (A) in-hospital mortality and (B) proportion of patients with door-to-balloon time exceeding 90 minutes based on hospital annual primary percutaneous coronary intervention volume.

7 Kontos et al Hospital Volume, Primary PCI, and Mortality 665 Table 6. Multivariate Analysis for Mortality 36<Primary PCI Volume Odds Ratio P Value Odds Ratio P Value All patients Unadjusted 1.17 ( ) < ( ) 0.86 Adjusted for clinical, angiographic variables 1.22 ( ) < ( ) 0.48 Adjusted, clinical, angiographic variables, and D2B time 1.20 ( ) ( ) 0.61 Off hours Unadjusted 1.16 ( ) ( ) 0.50 Adjusted for clinical, angiographic variables 1.22 ( ) ( ) 1.00 Adjusted, clinical, angiographic variables, and D2B time 1.20 ( ) ( ) 0.83 Shock patients Unadjusted 1.09 ( ) ( ) 0.24 Adjusted for clinical, angiographic variables 1.14 ( ) ( ) 0.26 Adjusted, clinical, angiographic variables, and D2B time 1.12 ( ) ( ) 0.36 Odds ratios reflect comparison with hospitals that have an annual primary percutaneous coronary intervention (PCI) volume >60 cases/y. P values for the comparison with hospital primary PCI volume >60/y. P=0.26) and adjusted mortality (odds ratio, 1.22; ) were not significantly different. The relative and absolute differences in the mortality found in the 2 studies were similar; however, because smaller numbers of hospitals (166 versus 738) and patients ( versus ) who were at lower risk (overall in-hospital mortality 3.2 versus 4.9%) were included by Kumbhani et al 13 compared with the current study may have limited the ability to demonstrate a significant difference in outcomes. D2B Times We found that D2B times were significantly longer in lowvolume as compared with intermediate-volume and highvolume hospitals. However, this difference, a median of 3 to 4 minutes, with an additional 5% to 6% of patients with a D2B >90 minutes, is unlikely to explain the mortality difference with high-volume hospitals. 26 Rather, shorter D2B times likely reflect improved processes of care at these hospitals. In addition, the low correlation that D2B times and mortality had with hospital volume indicates that other factors contribute to the reduced mortality that cannot be easily Table 7. Multivariate Analysis for Door-to-Balloon Times measured. These may include hospital processes of care, infrastructure for quality improvement, and interdepartmental and interdisciplinary co-ordination and communication. Study Implications Our data, which include 85% of cardiac catheterization laboratories performing PCI in the United States, 17 and therefore are representative of current clinical practices, indicate that hospital primary PCI volume continues to be an important predictor of in-hospital STEMI mortality. It is possible that activity below a specific threshold would limit the ability of catheterization laboratory and intensive care staff to gain the experience and expertise required to rapidly and effectively treat complicated and high-risk patients with STEMI undergoing primary PCI, particularly in emergent situations. 27 This may explain the lower mortality in higher volume centers, despite increased complications, such as bleeding, that have been associated with higher mortality. The relationship between volume and outcomes seems to be continuous. Our data indicate that current guideline recommended volume numbers seem reasonable, 14 as Door-to-Balloon Time <90 min Relative Risk P Value Relative Risk P Value All patients Unadjusted 0.90 ( ) < ( ) Adjusted for clinical, angiographic variables 0.93 ( ) < ( ) 0.02 Off hours Unadjusted 0.83 ( ) < ( ) < Adjusted for clinical, angiographic variables 0.84 ( ) < ( ) < Shock patients Unadjusted 0.87 ( ) < ( ) 0.06 Adjusted for clinical, angiographic variables 0.89 ( ) ( ) 0.06 Data reflect the relative risk of having a door-to-balloon time 90 min as compared with hospitals that have an annual primary PCI volume >60 cases/y.

8 666 Circ Cardiovasc Qual Outcomes November 2013 hospitals with <36 primary PCIs/y had significantly higher mortality than those with 36/y. The lack of a mortality difference between centers with a primary PCI volume of 30 to 60/y and those >60/y suggests that selection of a higher volume requirement may not lead to significant differences in mortality. Despite a significant increase in hospitals capable of performing PCI (from 1176 in 2001 to 1695 in 2006; relative 44% increase), the proportion of Americans estimated to live within 60 minutes of a PCI center has not changed. 15,28 The PCI utilization rate has remained stable and not significantly changed since 2001, 16 consistent with most of the newer PCI hospitals being near ones that already exist, rather than in areas of need, further reducing hospital procedure volumes. 29,30 Limitations The volume criteria we chose, although different from prior studies, were chosen to reflect current guideline recommendations. We did not analyze data using provider volume so we were unable to assess the role of provider experience on outcomes. We did not include symptom onset-to-balloon time and total hospital PCI volume in the mortality model. We were limited to in-hospital mortality; it is possible that differences we found would have increased over time, particularly in patients with cardiogenic shock. 31,32 We excluded transferin patients, as these patients may represent selection bias and a higher risk group. 7 However, inclusion did not change the overall results. An unequal geographic distribution of participating hospitals could lead to selection bias, which limits the proportion of the STEMI population that was evaluated during the study period. We included only hospitals participating in Cath/PCI; however, this registry includes 85% of cardiac catheterization laboratories performing PCI in the United States 17 and therefore is likely representative of current clinical practices. Because data used for our study are self-reported, there is a potential for data errors. However, before inclusion in the registry, data are filtered through the registry-specific algorithms that require predetermined levels of completeness and consistency for submitted data fields as part of the data quality report. A recent report indicated a high degree of agreement. 17 Conclusions Despite improvements in D2B times and reductions in overall mortality, our data show that hospitals with a primary PCI volume that <36 primary PCIs per year has a significantly higher mortality than hospitals with higher primary PCI volume. Sources of Funding This research was supported by the American College of Cardiology Foundation s National Cardiovascular Data Registry (NCDR). The views expressed in this article represent those of the author(s) and do not necessarily represent the official views of the NCDR or its associated professional societies identified at Disclosures Dr Curtis reports receiving salary support from the NCDR and Medicare and owns Medtronic stock. The other authors report no conflicts. References 1. Gersh BJ, Antman EM. Selection of the optimal reperfusion strategy for STEMI: does time matter? Eur Heart J. 2006;27: Peterson ED, Dai D, DeLong ER, Brennan JM, Singh M, Rao SV, Shaw RE, Roe MT, Ho KK, Klein LW, Krone RJ, Weintraub WS, Brindis RG, Rumsfeld JS, Spertus JA; NCDR Registry Participants. Contemporary mortality risk prediction for percutaneous coronary intervention: results from 588,398 procedures in the National Cardiovascular Data Registry. J Am Coll Cardiol. 2010;55: Srinivas VS, Hailpern SM, Koss E, Monrad ES, Alderman MH. Effect of physician volume on the relationship between hospital volume and mortality during primary angioplasty. J Am Coll Cardiol. 2009;53: Vakili BA, Kaplan R, Brown DL. Volume-outcome relation for physicians and hospitals performing angioplasty for acute myocardial infarction in New York state. Circulation. 2001;104: McGrath PD, Wennberg DE, Dickens JD Jr, Siewers AE, Lucas FL, 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: Magid DJ, Calonge BN, Rumsfeld JS, Canto JG, Frederick PD, Every NR, Barron HV; National Registry of Myocardial Infarction 2 and 3 Investigators. Relation between hospital primary angioplasty volume and mortality for patients with acute MI treated with primary angioplasty vs thrombolytic therapy. JAMA. 2000;284: Canto JG, Every NR, Magid DJ, Rogers WJ, Malmgren JA, Frederick PD, French WJ, Tiefenbrunn AJ, Misra VK, Kiefe CI, Barron HV. The volume of primary angioplasty procedures and survival after acute myocardial infarction. National Registry of Myocardial Infarction 2 Investigators. N Engl J Med. 2000;342: Cannon CP, Gibson CM, Lambrew CT, Shoultz DA, Levy D, French WJ, Gore JM, Weaver WD, Rogers WJ, Tiefenbrunn AJ. Relationship of symptom-onset-to-balloon time and door-to-balloon time with mortality in patients undergoing angioplasty for acute myocardial infarction. JAMA. 2000;283: Hannan EL, Wu C, Walford G, King SB III, Holmes DR Jr, Ambrose JA, Sharma S, Katz S, Clark LT, Jones RH. Volume-outcome relationships for percutaneous coronary interventions in the stent era. Circulation.2005;112: Epstein AJ, Rathore SS, Volpp KG, Krumholz HM. Hospital percutaneous coronary intervention volume and patient mortality, 1998 to 2000: does the evidence support current procedure volume minimums? J Am Coll Cardiol. 2004;43: Post PN, Kuijpers M, Ebels T, Zijlstra F. The relation between volume and outcome of coronary interventions: a systematic review and metaanalysis. Eur Heart J. 2010;31: Freeman JV, Wang Y, Curtis JP, Heidenreich PA, Hlatky MA. Physician procedure volume and complications of cardioverter-defibrillator implantation. Circulation. 2012;125: Kumbhani DJ, Cannon CP, Fonarow GC, Liang L, Askari AT, Peacock WF, Peterson ED, Bhatt DL; Get With the Guidelines Steering Committee and Investigators. Association of hospital primary angioplasty volume in ST-segment elevation myocardial infarction with quality and outcomes. JAMA. 2009;302: Harold JG, Bass TA, Bashore TM, Brindis RG, Brush JE Jr, Burke JA, Dehmer GJ, Deychak YA, Jneid H, Jollis JG, Landzberg JS, Levine GN, McClurken JB, Messenger JC, Moussa ID, Muhlestein JB, Pomerantz RM, Sanborn TA, Sivaram CA, White CJ, Williams ES. ACCF/AHA/SCAI 2013 update of the clinical competence statement on coronary artery interventional procedures: a report of the American College of Cardiology Foundation/American Heart Association/American College of Physicians Task Force on Clinical Competence and Training (Writing Committee to Revise the 2007 Clinical Competence Statement on Cardiac Interventional Procedures). J Am Coll Cardiol. 2013;62: Concannon TW, Kent DM, Normand SL, Newhouse JP, Griffith JL, Cohen J, Beshansky JR, Wong JB, Aversano T, Selker HP. Comparative effectiveness of ST-segment-elevation myocardial infarction regionalization strategies. Circ Cardiovasc Qual Outcomes. 2010;3: Epstein AJ, Polsky D, Yang F, Yang L, Groeneveld PW. Coronary revascularization trends in the United States, JAMA. 2011;305: Dehmer GJ, Weaver D, Roe MT, Milford-Beland S, Fitzgerald S, Hermann A, Messenger J, Moussa I, Garratt K, Rumsfeld J, Brindis RG. A contemporary view of diagnostic cardiac catheterization and percutaneous coronary intervention in the United States: a report from the CathPCI Registry of the National Cardiovascular Data Registry, 2010 through June J Am Coll Cardiol. 2012;60:

9 Kontos et al Hospital Volume, Primary PCI, and Mortality Krone RJ, Shaw RE, Klein LW, Block PC, Anderson HV, Weintraub WS, Brindis RG, McKay CR; ACC-National Cardiovascular Data Registry. Evaluation of the American College of Cardiology/American Heart Association and the Society for Coronary Angiography and Interventions lesion classification system in the current stent era of coronary interventions (from the ACC-National Cardiovascular Data Registry). Am J Cardiol. 2003;92: Messenger JC, Ho KK, Young CH, Slattery LE, Draoui JC, Curtis JP, Dehmer GJ, Grover FL, Mirro MJ, Reynolds MR, Rokos IC, Spertus JA, Wang TY, Winston SA, Rumsfeld JS, Masoudi FA; NCDR Science and Quality Oversight Committee Data Quality Workgroup. The National Cardiovascular Data Registry (NCDR) Data Quality Brief: the NCDR Data Quality Program in J Am Coll Cardiol. 2012;60: Fazel R, Krumholz HM, Bates ER, French WJ, Frederick PD, Nallamothu BK; National Registry of Myocardial Infarction (NRMI) Investigators. Choice of reperfusion strategy at hospitals with primary percutaneous coronary intervention: a National Registry of Myocardial Infarction analysis. Circulation. 2009;120: Peterson ED, Dai D, DeLong ER, Brennan JM, Singh M, Rao SV, Shaw RE, Roe MT, Ho KK, Klein LW, Krone RJ, Weintraub WS, Brindis RG, Rumsfeld JS, Spertus JA; NCDR Registry Participants. Contemporary mortality risk prediction for percutaneous coronary intervention: results from 588,398 procedures in the National Cardiovascular Data Registry. J Am Coll Cardiol. 2010;55: Smith SC Jr, Dove JT, Jacobs AK, Kennedy JW, Kereiakes D, Kern MJ, Kuntz RE, Popma JJ, Schaff HV, Williams DO, Gibbons RJ, Alpert JP, Eagle KA, Faxon DP, Fuster V, Gardner TJ, Gregoratos G, Russell RO, Smith SC Jr; American College of Cardiology; American Heart Association Task Force on Practice Guidelines. Committee to Revise the 1993 Guidelines for Percutaneous Transluminal Coronary Angioplasty. ACC/AHA guidelines of percutaneous coronary interventions (revision of the 1993 PTCA guidelines)--executive summary. J Am Coll Cardiol. 2001;37: Ellis SG, Weintraub W, Holmes D, Shaw R, Block PC, King SB III. Relation of operator volume and experience to procedural outcome of percutaneous coronary revascularization at hospitals with high interventional volumes. Circulation. 1997;95: Krumholz HM, Herrin J, Miller LE, Drye EE, Ling SM, Han LF, Rapp MT, Bradley EH, Nallamothu BK, Nsa W, Bratzler DW, Curtis JP. Improvements in door-to-balloon time in the United States, 2005 to Circulation. 2011;124: Venkitachalam L, Kip KE, Selzer F, Wilensky RL, Slater J, Mulukutla SR, Marroquin OC, Block PC, Williams DO, Kelsey SF; Investigators of NHLBI-Sponsored PTCA and Dynamic Registries. Twenty-year evolution of percutaneous coronary intervention and its impact on clinical outcomes: a report from the National Heart, Lung, and Blood Institute-sponsored, multicenter PTCA and Dynamic Registries. Circ Cardiovasc Intervent. 2009;2: Rathore SS, Curtis JP, Chen J, Wang Y, Nallamothu BK, Epstein AJ, Krumholz HM; National Cardiovascular Data Registry. Association of door-to-balloon time and mortality in patients admitted to hospital with ST elevation myocardial infarction: national cohort study. BMJ. 2009;338:b Jollis JG. The New York State Primary Angioplasty Registry and procedural volume. J Am Coll Cardiol. 2009;53: Nallamothu BK, Bates ER, Wang Y, Bradley EH, Krumholz HM. Driving times and distances to hospitals with percutaneous coronary intervention in the United States: implications for prehospital triage of patients with ST-elevation myocardial infarction. Circulation. 2006;113: Horwitz JR, Nichols A, Nallamothu BK, Sasson C, Iwashyna TJ. Expansion of invasive cardiac services in the United States. Circulation. 2013;128: Concannon TW, Nelson J, Kent DM, Griffith JL. Evidence of systematic duplication by new percutaneous coronary intervention programs. Circ Cardiovasc Qual Outcomes. 2013;6: Hochman JS, Sleeper LA, Webb JG, Dzavik V, Buller CE, Aylward P, Col J, White HD; SHOCK Investigators. Early revascularization and longterm survival in cardiogenic shock complicating acute myocardial infarction. JAMA. 2006;295: Webb JG, Sanborn TA, Sleeper LA, Carere RG, Buller CE, Slater JN, Baran KW, Koller PT, Talley JD, Porway M, Hochman JS; for the SHOCK Investigators. Should We Emergently Revascularize Occluded Coronaries for Cardiogenic Shock. One-year survival following early revascularization for cardiogenic shock. JAMA. 2001;285: