Author + information
- Phillip Moore, MD, MBA∗ ()
- Department of Pediatric Cardiology, University of California-San Francisco, San Francisco, California
- ↵∗Reprint requests and correspondence:
Dr. Phillip Moore, Department of Pediatric Cardiology, University of California–San Francisco, 505 Parnassus Avenue, Room M-1235, UCSF Box 0544, San Francisco, California 94143.
In this issue of the Journal, Lewis et al. (1) give significant insight into outcomes and risks associated with pulmonary artery stenting in children with congenital heart disease while highlighting both the potential power of “big data” prospective registries and their significant limitations. The study is unique in our field as it analyzes the largest series of pulmonary artery stent procedures (>1,100) within a recent 3-year period. To put this in a historical context, the first large multicenter study of pulmonary artery stenting was published in Circulation in 1993 by O’Laughlin et al. (2) and reported only 80 pulmonary artery stent implantations collected over a 2.5-year period from the 2 largest U.S. centers performing the procedure at that time. Even in the more recent past, 3 major U.S. centers collecting retrospective data over 5 years ending in June 2011 (3) reported only 60 pulmonary artery stent implantations. A system that can collect >1,100 stent implantations in 3 years shows that we have come a long way, both in the use of pulmonary artery stents in congenital heart disease and in our ability to rapidly acquire prospective data.
The important findings of the study include a favorable procedural success rate of >75%, but a significant overall complication rate of 14%, major adverse event rate of 9%, and procedural mortality rate of 0.2%. The sheer size of the cohort allowed rigorous multivariate analysis to identify ostial stenosis as the only predictor of success and weight <4 kg, single-ventricle physiology, and emergent need for catheterization as risk factors associated with a major adverse event. Although these findings are in the range of findings reported in previous studies, typically retrospective single-center studies, they suggest that both the success rate is lower and the complication rate is higher for this procedure than previously recognized and for the first time indicated that single-ventricle physiology is associated with a higher risk of pulmonary artery stenting. Clearly, the power of “big data” includes a more well-defined analysis of predictors of both success and risk, critical information to move us toward more optimal patient selection for specific interventional techniques.
Perhaps the most significant finding of this study is the critical importance of clear, relevant, a priori, and widely agreed-on definitions of both procedural success and significant adverse events as we move into an era of collaborative registries whose data will be used eventually for comparative analysis, both within and between centers. In this paper, 3 different a priori definitions of procedural success were used and analyzed, 2 definitions (overlapping features) for 2-ventricle physiology, and 1 definition for single-ventricle physiology. Focusing on the results of the procedural success rates for 2-ventricle subjects, not only did the 2 different success definitions give statistically different rates (76% for definition 1 vs. 84% for definition 2), their confidence intervals did not overlap. The specific definition of procedural success critically affects the rates that we measure, which critically affect comparisons that we make and conclusions that we draw.
Likewise, the definitions of major and minor adverse events associated with a specific procedure are critical and must be clinically relevant and universally accepted to allow for accurate measurements and comparisons. In this paper, the major adverse event rate associated with pulmonary artery stent implantation was substantial at 9% and higher than in some previous reports (4). More in-depth evaluation of this 9% major adverse event rate shows that 57% of them were bleeding events. Bleeding defined in the IMPACT registry may include bleeding at the access site; hematoma development (<3 to >10 cm); and retroperitoneal, gastrointestinal, GU (genitourinary), and other bleeding; all are reported independent of the need for transfusion or additional treatment. As one might imagine, reporting of this complication might be quite variable among participating sites, and not all operators might agree that a 3-cm hematoma constitutes a major adverse event. This is not to suggest that the authors overreported the major adverse events associated with pulmonary artery stenting; they reported accurately what the participating sites recorded. However, it again highlights the importance of the definitions that we use in collective registries and how those definitions affect our results and conclusions.
I applaud the authors for an excellent analysis of the largest dataset available on pulmonary artery stent implantation and for providing new and more accurate insights into both the procedural success and complication rates. I strongly encourage those readers not participating in the IMPACT registry to begin participating. The power and potential benefit of large prospective registries in our field have clearly been demonstrated. I also encourage all of us in the congenital interventional cardiology community to begin robust discussion and debate on the proper and relevant procedure-specific definitions of procedural success, long-term outcome, and adverse events so we can begin to reap the true potential benefits of our “big data” efforts.
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
Dr. Moore is an NCDR IMPACT steering committee member; and has received research grant support from St. Jude Medical and W. L. Gore & Associates, Inc.
- American College of Cardiology Foundation
- Lewis M.J.,
- Kennedy K.F.,
- Ginns J.,
- et al.
- O'Laughlin M.P.,
- Slack M.C.,
- Grifka R.G.,
- et al.
- Meadows J.,
- Teitel D.,
- Moore P.