Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Angiographic Lesion Complexity Score and In-Hospital Outcomes after Percutaneous Coronary Intervention

Angiographic Lesion Complexity Score and In-Hospital Outcomes after Percutaneous Coronary Intervention

  • Ayaka Endo, 
  • Akio Kawamura, 
  • Hiroaki Miyata, 
  • Shigetaka Noma, 
  • Masahiro Suzuki, 
  • Takashi Koyama, 
  • Shiro Ishikawa, 
  • Susumu Nakagawa, 
  • Shunsuke Takagi, 
  • Yohei Numasawa
PLOS
x

Abstract

Objective

We devised a percutaneous coronary intervention (PCI) scoring system based on angiographic lesion complexity and assessed its association with in-hospital complications.

Background

Although PCI is finding increasing application in patients with coronary artery disease, lesion complexity can lead to in-hospital complications.

Methods

Data from 3692 PCI patients were scored based on lesion complexity, defined by bifurcation, chronic total occlusion, type C, and left main lesion, along with acute thrombus in the presence of ST-segment elevation myocardial infarction (1 point assigned for each variable).

Results

The patients’ mean age was 67.5 +/- 10.8 years; 79.8% were male. About half of the patients (50.3%) presented with an acute coronary syndrome, and 2218 (60.1%) underwent PCI for at least one complex lesion. The patients in the higher-risk score groups were older (p < 0.001) and had present or previous heart failure (p = 0.02 and p = 0.01, respectively). Higher-risk score groups had significantly higher in-hospital event rates for death, heart failure, and cardiogenic shock (from 0 to 4 risk score; 1.7%, 4.5%, 6.3%, 7.1%, 40%, p < 0.001); bleeding with a hemoglobin decrease of >3.0 g/dL (3.1%, 11.0%, 13.1%, 10.3%, 28.6%, p < 0.001); and postoperative myocardial infarction (1.5%, 3.1%, 3.8%, 3.8%, 10%, p = 0.004), respectively. The association with adverse outcomes persisted after adjustment for known clinical predictors (odds ratio 1.72, p < 0.001).

Conclusion

The complexity score was cumulatively associated with in-hospital mortality and complication rate and could be used for event prediction in PCI patients.

Introduction

Percutaneous coronary intervention (PCI) is a reliable and effective therapeutic option for patients with coronary artery disease (CAD) and has become one of the most widely applied treatments in present-day cardiology. However, although periprocedural complications have declined over time, the risk of complications for patients with complex lesions in coronary vessels remains high. The definition of complex lesions includes vessel bifurcation, the presence of thrombus, involvement of the left main trunk, and the increasing number of “difficult” lesions that are now treated (e.g., those that are heavily calcified or diffuse, lesions in vessels with excessive tortuosity, or chronic total occlusive lesions) [19]. Therefore, it is important to evaluate the risk of complications in patients undergoing PCI for complex lesions.

No simple and user-friendly risk scoring system based on angiographic information has yet been established. The SYNTAX Score was devised to evaluate the angiographic characteristics that make a lesion suitable for either PCI or coronary artery bypass grafting (CABG), but it is not exactly a comprehensive bedside risk prediction tool [910]. Clinicians need a risk scoring system that will help predict short-term (i.e., in-hospital) outcomes, allowing informed clinical decisions to be made. The identification and quantification of the clinical factors associated with the complication risk would also facilitate observational research into the comparative effectiveness of therapeutic approaches. Further, at the policy-making level, predicted risk estimates can help “level the playing field” of provider outcome metrics, helping to adjust for potential differences in cases treated.

Based on the above considerations, we devised a modern PCI scoring system based on simple criteria of angiographic lesion complexity. Utilizing data from a multicenter Japanese registry, we assessed its association with in-hospital mortality and complications, as a means of facilitating more precise risk prediction.

Methods

Study design

The Japan Cardiovascular Database (JCD) is a large, ongoing, prospective multicenter cohort study designed to collect clinical background and outcome data on PCI patients. Data consisting of approximately 200 variables were collected for each patient. Participating hospitals were instructed to record data from consecutive hospital visits for PCI and to record them in an Internet-based database system. This system performs checks to ensure that the reported data are complete and internally consistent. PCI with any commercially available coronary device could be included. The decision to perform PCI was made according to the investigators’ clinical assessment of their patients. The study did not mandate specific interventional or surgical techniques, such as vascular access, or use of specific stents or closure devices. Although the size of the sheath and guiding catheter were not protocol-mandated in this cohort, the commonly used size was 6-Fr to 8-Fr when a transfemoral approach was used and 6-Fr for transradial interventions. The majority of clinical variables in the JCD were defined according to the National Cardiovascular Data Registry (NCDR), which was sponsored by the American College of Cardiology (ACC) to conduct comparative research in order to determine the factors leading to disparities in PCI management. The NCDR is a large PCI registry system with over 1,000,000 entries for ischemic heart disease and over 500,000 entries for PCI, collected from more than 500 institutions in the US [11].

The study was approved by the institutional review board of Keio University School of Medicine. The patient record was anonymized and de-identified prior to analysis. Major teaching hospitals within the metropolitan Tokyo area were selected for the pilot phase of this study, and the study protocol was approved by the institutional review board at each site. The written consent was obtained by the study participants. Patients were enrolled at the event; all the consecutive PCI procedures during the study period, including failure cases, were registered. Patients aged <18 years were excluded.

The present study was funded by the Kakenhi (Grant-in-Aid for Scientific Research) (No. 21790751). The JCD Steering Committee was responsible for overall study guidance, including the study protocol, data analysis, and interpretation of the results. The Department of Healthcare Quality Assessment of Tokyo University managed the database independently. Keio University School of Medicine Interhospital Cardiology Study Group managed the participating sites and provided a monthly on-site monitoring service to assure data accuracy and completeness throughout the study. During the planning, implementation and reporting of this study, there were no issues such as conflict of interest, conflict of responsibility, or intellectual property rights.

Information disclosure

Before the launch of the JCD, information about the objectives of the present study, its social significance, and an abstract were provided for clinical trial registration with the University Hospital Medical Information Network, which is recognized by the International Committee of Medical Journal Editors as an “acceptable registry” according to a statement issued in September 2004 (UMIN R000004736).

Definition of lesion complexity risk score and clinical outcomes

A complex lesion was defined as a treated lesion possessing at least one of the following high-risk angiographic lesion characteristics: bifurcation, chronic total occlusion (CTO), Type C, unprotected left main trunk (UPLMT), and thrombus formation. The size of the main or side branch vessel had to be at least 1.5 mm in diameter, as assessed by diagnostic angiogram, and significant stenosis was defined as a reduction of at least 50% in luminal diameter, by visual assessment. Bifurcation lesions were defined as a division of a vessel into at least two branches, with the plaque extending from at least one of the limbs to the branching point. A CTO lesion was defined as a 100% occluded lesion with complete interruption of antegrade flow (TIMI flow grade 0) that had been present for at least 3 months. Type C lesions were as defined by the American Heart Association/American College of Cardiology (AHA/ACC). UPLMT lesions were left main trunk lesions without patent coronary artery bypass grafts in the left anterior descending artery or the left circumflex artery. Thrombotic lesions were defined as the presence of new ST-segment elevation myocardial infarction (STEMI) as clinical presentation, with cardiac biomarkers exceeding the upper limit of normal, according to the individual hospital’s laboratory parameters. The operator determined the presence of these characteristics at the time of the coronary angiography. Patients were stratified both by absolute complex lesion status (yes/no), and by the total number of the five complex lesion criteria that were present (score 0: no complex lesion, score 1: one complex lesion, score 2: two complex lesions, score 3: three complex lesions, score 4: four complex lesions, and score 5: all complex lesions).

The primary outcome was all-cause in-hospital mortality and secondary outcomes included all in-hospital complications. In-hospital complications collected in the data system included in-hospital death from any cause after PCI; the combined cardiac events included in-hospital cardiac death (in-hospital death, heart failure or cardiogenic shock) [12], periprocedural myocardial infarction (defined as an increase in serum creatine kinase to above normal levels, associated with positive isoenzymes, routinely measured for all patients on the day after the PCI procedure), bleeding with a hemoglobin decrease of more than 3.0 g/dL or blood transfusion, contrast nephropathy, and persistent coronary flow reduction (TIMI flow less than grade 3). Contrast nephropathy was defined, according to the established definition in the literature, as an increase in serum Cr level ≥0.5 mg/dL or ≥25% above baseline values at 30 days after administration of contrast media, in the absence of any other identifiable major kidney insult. Major bleeding was defined as: 1) bleeding requiring a blood transfusion; 2) a decrease in the Hb level by ≥3.0 g/dL due to bleeding from any site, including the percutaneous entry site, retroperitoneum, gastrointestinal tract, genitourinary tract, and other/unknown sites; and 3) the need for intervention/surgery at the bleeding site to reverse/stop or correct the bleeding (such as surgical closure/exploration of the arteriotomy site, balloon angioplasty to seal an arterial tear, or endoscopy with cautery of a gastrointestinal bleed). The latter definition is equivalent to Bleeding Academic Research Consortium Type 3 bleeding.

Statistical analysis

Patients were stratified by the complexity score (0–5) of the target lesion(s), as described above. Descriptive statistics were calculated on the basis of clinical characteristics and treatment information for the registered patients. Statistical analysis was performed using SPSS version 15 (SPSS, Chicago, IL, USA). When continuous variables were assumed to show a normal distribution, the data were expressed as mean ± SD. When normality was not assumed, the data were expressed as median and interquartile range. Categorical data were summarized in terms of frequency and proportion. The 95% confidence intervals of the mean, median, and proportion values were also calculated. The baseline clinical characteristics of patients were compared using chi-square tests for categorical variables and analysis of variance for continuous variables. A two-tailed p-value <0.05 was considered statistically significant. Multivariate logistic regression analysis was performed to assess the association of increments in each lesion complexity score with in-hospital mortality and complications.

Results

Patients and baseline characteristics

Among the 3692 patients who underwent PCI procedures between September 1, 2008, and August 31, 2011, and were recorded in the JCD Registry, 2218 (60.1%) underwent revascularization of at least one complex lesion (Fig 1). Of the total of 3264 complex lesions, 894 (24.2%) were bifurcation, 266 (7.2%) were CTO, 1000 (27.1%) were type C, 270 (7.3%) were UPLMT, and 834 (22.6%) were STEMI. Thus, the most common types of complex lesion undergoing revascularization were type C, bifurcation, and STEMI lesions. The distribution of the study patients by complexity score is shown in Fig 1. In total, 1474 patients (39.9%) had no complex lesion attempted (score 0 group), 1375 patients (37.2%) had one complex lesion (score 1 group), 650 patients (17.6%) had 2 complex lesions (score 2 group), 183 patients (5.0%) had 3 complex lesions (score 3 group), and 10 patients (0.3%) had 4 complex lesions attempted (score 4 group). No patients had lesions with all types of complexity (score 5 group).

thumbnail
Fig 1. Flowchart showing the patients included in the present analysis.

A total of 3692 patients were evaluated.

https://doi.org/10.1371/journal.pone.0127217.g001

The clinical and angiographic characteristics of all the groups are shown in Table 1. The mean age of the entire cohort was 67.5 ± 10.8 years and 2935 (79.5%) were male. Overall, 1857 (50.3%) of the patients presented with acute coronary syndromes (ACS) and 1557 (42.2%) underwent non-elective procedures. The relationship between lesion complexity score and baseline variables had an inverted U-shape for male sex and a U-shape for age, especially for patients over 70 years old. Compared to patients in the score 0 group, patients in higher score groups had a lower left ventricular ejection fraction and were more likely to have a multi-vessel lesion in the left anterior descending artery. Patients in higher score groups also tended to present with ACS and heart failure, underwent more emergency PCI procedures, and were treated more frequently with a femoral artery approach, with intra-aortic balloon pump support, and with drug-eluting stents (DES).

Clinical outcomes

Details of in-hospital complications are shown in Table 2. Although 3265 (88.4%) procedures were completed successfully, 427 (11.6%) were associated with at least one complication. The rates of in-hospital mortality and complications were higher in the higher score groups compared with the lower score groups. Importantly, patients in higher score groups had significantly higher rates of in-hospital events, including death, heart failure and cardiogenic shock, postoperative myocardial infarction, and bleeding with a hemoglobin decrease of more than 3.0 g/dL or transfusion (Fig 2).

The presence of these variables predicted the in-hospital outcome after adjustment for known clinical predictors (Tables 3 and 4; results of the univariate analyses is available in S1 and S2 Tables). Importantly, the in-hospital mortality increased by 1.38 per unit increment in complexity score (odds ratio, OR 1.38; p < 0.001). Furthermore, the risk of an in-hospital complication increased by 1.73 per unit increment in complexity score (OR 1.73; p < 0.001). Of note, there is a partial overlap in the Type C and CTO lesions, and we performed a secondary analysis excluding CTO from our scoring system (S3 and S4 Tables); however, this did not alter our main results and complexity score remained an independent predictor for in-hospital death (OR 1.48; p<0.001) and complication (OR 1.90; p<0.001).

Discussion

We developed a simple PCI scoring system, based on angiographic lesion complexity, for predicting the risk of in-hospital mortality and complications. We tested the system against data from 3692 patients enrolled in a multicenter Japanese registry between 2008 and 2011. This complexity scoring system correlated well with in-hospital mortality and complication rates: patients with higher scores exhibited higher event rates compared with lower score groups, while the mortality and complication rate increased by 1.38 and 1.74, respectively, per unit rise in complexity score. The results of this study suggest that quantification of these angiographic characteristics could be of assistance in in-hospital risk stratification and that patients with a high complexity score warrant special attention.

During the last decade, there has been remarkable development in novel devices for PCI, such as first or second generation DES, along with their delivery systems. Hence, the contemporary management of coronary artery diseases has become increasingly dependent on PCI, rather than CABG. However, successful PCI of difficult lesions requires advanced techniques, and the learning curve increases steeply along with the need for greater skill and experience on the part of the operator. Therefore, accurate risk assessment is an essential part of the evaluation in patients undergoing complex PCI. In the present study, each variable showed a tendency to predict in-hospital events (S1 and S2 Tables). When each variable was assigned as a factor to construct a single ‘scoring system’, each unit increment of the score was cumulatively associated with risk for in-hospital events. Our basis for the selection of each of the 5 angiographic variables was primarily clinical. Chosen variables had to be readily available and clinically relevant when performing complex PCI. These variables are thought to directly reflect the complexity of angiographic lesions seen in practice, and are frequently discussed at the bedside and/or catheterization lab when performing the procedure. This in contrast to previously established “complexity” scores (e.g., the SYNTAX score) that aimed to predict the long-term results of PCI compared to CABG [8,10]. This concept is close to that of CHADS2 score for atrial fibrillation.

It is worth emphasizing that the in-hospital complication rate increased approximately twofold in the score 1 group, threefold in the score 2 and 3 groups, and more than tenfold in the score 4 group, compared to patients with simple lesions (score 0 group). Clearly, the presence of complex lesions must always be taken into account in any attempt to forecast the probability of final procedural success and safety. It may also be possible to assign operators appropriately, according to their level of skill, based on the difficulty of dealing with these complex lesions, or to consider referring the patient for CABG. Complex lesions with a low complexity score can usually be treated successfully, and would be good candidates for training purposes.

Previously published studies have indicated a similar trend in the clinical predictors for in-hospital mortality and complications in PCI patients, based on multivariate logistic regression analysis [3,4,7,1317]. For example, old age, female sex, current or past heart failure symptoms, renal failure, and peripheral artery disease were included in all of the studies. It is noteworthy that our angiographic complexity score continued to be a significant predictor of in-hospital events, even after adjusting for known clinical predictors. Therefore, lesion complexity should be recognized as an important risk factor, in addition to variables that are related to the patient’s background.

In this study, the overall in-hospital mortality rate was 1.2%. This mortality rate is relatively high compared with previous studies, which reflects the high percentage (about 40%) of patients with ACS, cardiogenic shock, or cardiopulmonary arrest. A comparison of the results among patients with stable CAD and those with ACS, shock, or cardiopulmonary arrest will be the subject of further analysis.

Several important limitations of the present analysis should be discussed. The first is sample size. The high score groups contained only small numbers of patients and the full score group had none. Additional validation in the high score groups with larger samples might be needed. A second limitation is that we did not analyze the potential relationship between hospital or operator procedure volume and in-hospital complications. In particular, there may be a relationship between an institution’s or operator’s procedure experience and volume and the outcomes of PCI for complex lesions. Third, it may not have been possible to distinguish Type C lesions from bifurcation lesions and CTO. However, the number of patients with Type C lesions was small, less than 10%. Fourth, we did have clear discrimination between Type C and heavily calcified lesions. In usual practice, defining universal ‘heavy’ calcification nor its quantification would be a major challenge. Because our goal was to generate a bedside-friendly tool to assess complexity of PCI that would predict the clinical outcome, we chose to incorporate type C lesion as a variable in the scoring system, which is widely appreciated among interventional cardiologists. Lastly, the odds ratio for each complex lesion in our scoring system was not evaluated separately using multivariate logistic regression analysis. We believe that the risk weights for each type of complex lesion are likely to be slightly different. Further analysis, multiplication of the predicted risk for each risk score by the odds ratio for the in-hospital mortality and complications could lead to an improved risk score for evaluation in future studies.

Conclusions

Accurate risk assessment can aid in the identification of patients who are at high risk of an in-hospital event. The proposed complexity score was cumulatively associated with in-hospital mortality and complication rate and could be used for event prediction in PCI patients. PCI operators should take special care in order to perform PCI successfully in these complex lesions.

Supporting Information

S1 Table. Univariable predictors for in-hospital mortality.

https://doi.org/10.1371/journal.pone.0127217.s001

(DOCX)

S2 Table. Univariable predictors of any complications.

https://doi.org/10.1371/journal.pone.0127217.s002

(DOCX)

S3 Table. Multivariable predictors of in-hospital mortality without CTO lesion.

https://doi.org/10.1371/journal.pone.0127217.s003

(DOCX)

S4 Table. Multivariable predictors of any complications without CTO lesion.

https://doi.org/10.1371/journal.pone.0127217.s004

(DOCX)

Acknowledgments

We would like to thank all the investigators, clinical coordinators and institutions involved in the JCD-KICS study.

Investigators: Yohei Numasawa, Toshiki Kuno, Hiroyuki Motoda (Ashikaga Red Cross Hospital), Yutaka Okada (Eiju General Hospital), Soushin Inoue、Iwao Nakamura (Hino Municipal Hospital), Shunsuke Takagi, Takashi Matsubara (Hiratsuka City Hospital), Masashi Takahashi, Keishu Li, Koichiro Sueyoshi (Kawasaki City Municipal Hospital), Taku Inohara, Atsushi Anzai, Kentaro Hayashida, Takashi Kawakami, Hideaki Kanazawa, Takahide Arai, Shunsuke Yuasa, Jun Fujita, Yuichiro Maekawa, Akio Kawamura, Shun Kohsaka (Keio University School of Medicine), Masahiro Suzuki, Keisuke Matsumura (National Hospital Organization Saitama National Hospital) Yukinori Ikegami, Yukihiko Momiyama (National Hospital Organization Tokyo Medical Center), Ayaka Endo, Tasuku Hasegawa, Toshiyuki Takahashi, Susumu Nakagawa (Saiseikai Central Hospital), Takashi Yagi, Kenichiro Shimoji, Sonhan Yun, Shigetaka Noma (Saiseikai Utsunomiya Hospital), Masahito Munakata, Shiro Ishikawa (Saitama City Hospital), Atsushi Mizuno, Yutaro Nishi (St Luke’s International Hospital Heart Center), Masaru Shibata, Takashi Koyama (Tachikawa Kyosai Hospital), Kimi Koide, Masaki Kodaira, Yoshinori Mano, Takahiro Ohki (Tokyo Dental College Ichikawa General Hospital), Daisuke Shinmura, Kouji Negishi (Yokohama Municipal Citizens Hospital)

Clinical Coordinators: Junko Susa, Ayano Ishikawa, Hiroaki Nagayama, Miho Umemura, Itsuka Saito, and Ikuko Ueda

Author Contributions

Conceived and designed the experiments: AE SK HM. Performed the experiments: SK HM. Analyzed the data: AE SK HM. Contributed reagents/materials/analysis tools: AE SK HM AK S. Noma MS TK SI S. Nakagawa ST YN. Wrote the paper: AE SK KF.

References

  1. 1. Wilensky RL, Selzer F, Johnston J, Laskey WK, Klugherz BD, Block P, et al. Relation of percutaneous coronary intervention of complex lesions to clinical outcomes (from the NHLBI Dynamic Registry). Am J Cardiol 2002; 90:216–21. pmid:12127606
  2. 2. Harrell L, Schunkert H, Palacios IF. Risk predictors in patients scheduled for percutaneous coronary revascularization. Catheter Cardiovasc Interv 1999; 48:253–60. pmid:10525222
  3. 3. Moscucci M, Kline-Rogers E, Share D, O'Donnell M, Maxwell-Eward A, Meengs W. L, et al. Simple bedside additive tool for prediction of in-hospital mortality after percutaneous coronary interventions. Circulation 2001; 104:263–8. pmid:11457742
  4. 4. Resnic FS, Ohno-Machado L, Selwyn A, Simon DI, Popma JJ. Simplified risk score models accurately predict the risk of major in-hospital complications following percutaneous coronary intervention. Am J Cardiol 2001; 88:5–9. pmid:11423050
  5. 5. Rihal CS, Grill DE, Bell MR, Berger PB, Garratt KN, Holmes D. R. Jr.. Prediction of death after percutaneous coronary interventional procedures. Am Heart J 2000; 139:1032–8. pmid:10827384
  6. 6. Chowdhary S, Ivanov J, Mackie K, Seidelin PH, Dzavík V. The Toronto score for in-hospital mortality after percutaneous coronary interventions. Am Heart J 2009; 157:156–63. pmid:19081413
  7. 7. Wu C, Hannan EL, Walford G, Ambrose JA, Holmes DR, King S. B. 3rd, et al. A risk score to predict in-hospital mortality for percutaneous coronary interventions. J Am Coll Cardiol 2006; 47:654–60. pmid:16458151
  8. 8. Sianos G, Morel MA, Kappetein AP, Morice MC, Colombo A, Dawkins K, et al. The SYNTAX Score: an angiographic tool grading the complexity of coronary artery disease. EuroIntervention 2005; 1:219–27. pmid:19758907
  9. 9. Ong AT, Serruys PW, Mohr FW, Morice MC, Kappetein AP, Holmes D. R. Jr, et al. The SYNergy between percutaneous coronary intervention with TAXus and cardiac surgery (SYNTAX) study: design, rationale, and run-in phase. Am Heart J 2006; 151:1194–204. pmid:16781219
  10. 10. Serruys PW, Onuma Y, Garg S, Sarno G, van den Brand M, Kappetein A. P, et al. Assessment of the SYNTAX score in the Syntax study. EuroIntervention 2009; 5:50–6. pmid:19577983
  11. 11. Roe MT, Messenger JC, Weintraub WS, Cannon CP, Fonarow GC, Dai D, et al. Treatments, trends, and outcomes of acute myocardial infarction and percutaneous coronary intervention. J Am Coll Cardiol 2010; 56:254–63. pmid:20633817
  12. 12. Smith SC Jr., Feldman TE, Hirshfeld JW Jr., Jacobs AK, Kern MJ, King S. B. 3rd, et al. ACC/AHA/SCAI 2005 Guideline Update for Percutaneous Coronary Intervention-Summary Article: A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (ACC/AHA/SCAI Writing Committee to Update the 2001 Guidelines for Percutaneous Coronary Intervention). J Am Coll Cardiol 2006; 47:216–35. pmid:16386696
  13. 13. Hannan EL, Arani DT, Johnson LW, Kemp HG Jr, Lukacik G. Percutaneous transluminal coronary angioplasty in New York State. Risk factors and outcomes. JAMA 1992; 268:3092–7. pmid:1433740
  14. 14. Kimmel SE, Berlin JA, Strom BL, Laskey WK. Development and validation of simplified predictive index for major complications in contemporary percutaneous transluminal coronary angioplasty practice. The Registry Committee of the Society for Cardiac Angiography and Interventions. J Am Coll Cardiol 1995; 26:931–8. pmid:7560620
  15. 15. Ellis SG, Omoigui N, Bittl JA, Lincoff M, Wolfe MW, Howell G, et al. Analysis and comparison of operator-specific outcomes in interventional cardiology. From a multicenter database of 4860 quality-controlled procedures. Circulation 1996; 93:431–9. pmid:8565159
  16. 16. Jollis JG, Peterson ED, DeLong ER, Mark DB, Collins SR, Muhlbaier L. H, et al. The relation between the volume of coronary angioplasty procedures at hospitals treating Medicare beneficiaries and short-term mortality. N Engl J Med 1994; 331:1625–9. pmid:7969344
  17. 17. Hannan EL, Racz M, Ryan TJ, McCallister BD, Johnson LW, Arani D. T, et al. Coronary angioplasty volume-outcome relationships for hospitals and cardiologists. JAMA 1997; 277:892–8. pmid:9062327