| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Surgery, Oregon Health & Science University, Portland, Oregon
b Division of Pediatric Cardiothoracic Surgery, Oregon Health & Science University, Portland, Oregon
c Division of Pediatric Cardiology, Hospital for Sick Children, University of Toronto, Toronto, Ontario, Canada
Accepted for publication June 2, 2008.
* Address correspondence to Dr Welke, Division of Cardiothoracic Surgery L353, Oregon Health and Science University, 3181 SW Sam Jackson Park Rd, Portland, OR 97239-3098 (Email: welkek{at}ohsu.edu).
Presented at the Forty-fourth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 28–30, 2008. Winner of the J. Maxwell Chamberlain Memorial Award for Congenital Heart Surgery. Winner of the Thoracic Surgery Directors Association Resident Research Award.
| In addition to being the J. Maxwell Chamberlain Memorial Paper for Congenital Heart Surgery, this paper has received the Thoracic Surgery Directors Association (TSDA) Resident Research Award. This award was established in 1990 to encourage resident research in cardiothoracic surgery. Abstracts submitted to The Society of Thoracic Surgeons (STS) Program Committee representing research performed by residents were forwarded to the TSDA to be considered for this award. The abstracts were reviewed and the winner selected by the TSDA Executive Committee. In 2008, the recipient of the TSDA Resident Award was Tara Karamlou, MD, a resident of Ross M. Ungerleider, MD, at Oregon Health & Science University. The TSDA makes this award annually. The resident author of the selected study is recognized at the STS meeting.
|
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
Methods: Using the Nationwide Inpatient Sample and International Classification of Diseases, Ninth Revision, procedure and diagnosis codes, we identified all ASD/PFO closures performed from 1988 to 2005. National estimates and their standard deviations were calculated. Logistic regression analysis identified determinants of closure type.
Results: We identified 15,482 secundum ASD/PFO closures between 1988 and 2005, yielding a national estimate of 79,841 cases. Of these, 5,495 (national estimate, 27,554 ± 2,526) were percutaneous, 10,278 (national estimate, 53,710 ± 1,451) were surgical. The ASD/PFO closures per capita increased dramatically from 1.08 per 100,000 population in 1988 to 2.59 per 100,000 population in 2005, an increase of 139%. When analyzed by closure type, surgical closure increased by only 24% (from 0.86 per 100,000 population in 1988 to 1.07 per 100,000 in 2005) whereas percutaneous closure increased by 3,475% (from 0.04 per 100,000 population in 1988 to 1.43 per 100,000 in 2005). A marked shift in repair type occurred in 2001. Estimated mortality rates remained near 1% for both closure types over time.
Conclusions: Nationwide utilization of ASD/PFO repair has increased over time, largely attributable to the dramatic rise in percutaneous closure. Despite the substantial increase in utilization, mortality has remained constant. In the absence of meaningful benchmarks, prospective studies comparing outcomes and criteria for surgical versus percutaneous closure are needed to determine whether the increased rates of ASD closure are justified.
Atrial septal defect (ASD) is a common form of congenital heart disease accounting for approximately 10% of congenital cardiac defects [1]. Surgical closure of secundum ASD has been the standard for the last 45 years, but since the first attempt by King and Mills [2] in 1976, transcatheter closure has become the preferred option in many centers [3, 4].
Despite the known efficacy and safety of surgical repair, the attraction of percutaneous closure (PC) in terms of perceived reduction in morbidity and improved cosmesis has driven utilization of transcatheter repair, and perhaps lowered the threshold for closure beyond what is necessary and appropriate. There has been a movement toward minimally invasive treatments for surgical disease in other surgical specialties, including the rapid adoption of laparoscopic general surgical procedures and endovascular repair of abdominal aortic aneurysm [5–8]. In most cases, this evolution toward minimally invasive surgery has provided a therapeutic option for patients who are at high risk for open surgery, or it has created a less invasive alternative for patients in whom the need for some form of surgical treatment is clearly indicated. Minimally invasive techniques have not generally increased the incidence of surgical indications for treatment. The surgical closure of secundum ASDs and PFOs has been considered indicated for patients with right-side heart dilation or arrhythmias from large shunts or paradoxical emboli related to neurologic symptoms. The use of PC to close ASDs/PFOs would seem to be a useful option for selected patients in this group. It is not known whether the advent of PC has substantially altered the incidence of ASD/PFO closure, and if it has, what population of patients has primarily been affected.
The purpose of this investigation was to describe the evolution of national practice patterns for ASD/PFO closure over an 18-year period and to better delineate how the widespread availability of PC might have influenced the threshold for closure of ASD/PFOs.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
This study's data were derived from the combination of the NIS databases from 1988 to 2005, and our analysis uses sampling weights provided within the NIS database to derive national estimates. The NIS database was searched for the individual years 1988 to 2005, selecting out hospital discharges for which the primary International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes were for all closures of secundum ASD/PFOs (35.51, 35.52, 35.61, or 25.71) without diagnoses of endocardial cushion defects (subcodes of diagnosis code 745.6), and for which a diagnosis of a secundum ASD/PFO was also present (745.5). Discharges with additional cardiac procedures codes were excluded. Surgical closures were identified as those discharges with a procedure of an open repair (35.51), tissue graft (35.61), or cardiopulmonary bypass (39.61). Similarly, device closures were identified by selecting those discharges with a procedure for a closed repair (35.52) or other catheterization procedure. The ASD/PFO repairs that could not be reliably categorized into PC or surgical closure were categorized as undetermined type. Although we report the initial number of undetermined repair type for the sake of completeness, we did not further consider this group in subsequent analyses. Population estimates for calculation of population-normalized procedure rates were obtained from US Census data.
Statistical Analysis
Overall descriptive statistics were computed. Surgical and PC groups were compared using Rao-Scott 
2 for categorical variables or t test for continuous variables. Years were grouped into sextiles, and hospitals were placed into five groups based on their annual ASD/PFO closure volume. Hospitals were grouped into volume categories (using powers of 2) based on their annual volume of all ASD/PFO closures as follows: fewer than 2 cases, 2 to 4 cases, 5 to 8 cases, 9 to 16 cases, and more than 16 cases. We chose the upper threshold as 17 cases or more because there were too few hospitals consistently performing larger numbers of catheterization repairs before 1998. The probability for PC versus surgical closure was investigated using multivariable logistic regression models, adjusted for patient and hospital level characteristics; and c-indices were generated for all logistic regression models to provide insight into model discrimination. The SAS software, version 9.1 (SAS Institute, Cary, North Carolina), was used to fit the models to account for the survey design of the NIS, the potential clustering of outcomes within a hospital, and the decrease in clustering occurring over increasing time intervals. Institutional Review Board approval was obtained, but given the deidentified nature of the NIS, patient consent was waived for this study.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
|
|
|
|
|
|
|
Mortality and Morbidity
We identified 23 (national estimate, 119 ± 31) in-hospital deaths among PC patients and 84 (national estimate, 472 ± 58) deaths among patients undergoing surgical closure. Crude mortality rates were not different among surgical closure type (0.88%) compared with PC (0.60%; p = 0.15), and did not change significantly over time. Median hospital length of stay was 1 day (interquartile range, 1 to 1) among PC patients and 5 days (interquartile range, 3 to 7) among surgical closure patients (p < 0.001). Decreased length of stay in the PC group; however, must be considered in light of the widely disparate patient populations undergoing each closure type. Specifically, PC patients were older and had fewer noncardiac anomalies than those undergoing surgical closure (Table 1).
We also found that mean total hospital charges were lower with PC ($36,052 ± $1,299) as compared with surgical closure ($40,805 ± $785, p = 0.001).
Determinants of Closure Type
Determinants of closure type by multivariable logistic regression analysis included age, year sextile, annual hospital volume group, and the teaching status and location of the hospital. Determinants favoring PC included older patient age (p < 0.001), higher annual hospital volume group (p < 0.001), and more recent year sextile (p < 0.001). Interestingly, there was a significant interaction between age and annual hospital volume group whereupon the influence of age in determining PC repair diminished with increasing annual volume (Table 3). The c-statistic for our full multivariable model was 0.884, indicating excellent discrimination.
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
The increase in utilization of PC may be attributable to lower thresholds for ASD closure. As the adverse consequences of a given intervention diminish (eg, sternotomy scar, cardiopulmonary bypass, hospital length of stay, and total cost), the threshold for that intervention likewise declines. In other words, with the introduction of a perceived minimally invasive endovascular technique, physicians have a greater inclination to recommend ASD/PFO closure, and patients are more willing to accept such a recommendation. It is intriguing to puzzle whether this increased utilization is warranted. Enthusiasm for PC has likely also been sparked by favorable reports regarding the efficacy, safety, and lower morbidity of PC compared with surgical closure [10–12]. Similar paradigm shifts have been described with the advent of laparoscopic surgery and endovascular repair of abdominal aortic aneurysm and peripheral vascular disease [5–8, 13]. However, laparoscopic and endovascular techniques provide alternative treatments for diseases that would otherwise require open surgery. It is not clear that the enormous increase in ASD/PFO closure parallels this same indication, and it is concerning that it might, instead, be related to reimbursement for procedure rather than to best practice for treatment of disease.
Many studies have directly compared the PC and surgical ASD closure [10–12, 14, 15], yet the patient populations studied have been noncontemporaneous and dissimilar, making it difficult to advocate unequivocally for the superiority of either approach. Such comparative studies are further confounded by the heightened era effect for PC wherein the continuous introduction of updated devices renders prior benchmark data irrelevant or immediately outdated.
Most data have suggested equivalent efficacy, with perhaps slightly higher costs for PC [12]. Earlier studies by Jones and colleagues [15] and Hughes and coworkers [11] showed a reduction in complication rates after closure and hospital length of stay favoring transcatheter repair. However, a recent report by Mayer and colleagues [16] showed higher morbidity and mortality among device closures compared with surgical closure.
We have shown, in agreement with other studies [10, 11, 14, 15], a decrease in total hospital length of stay and total hospital charges with PC. We recognize that these comparisons are unfair, considering that surgical closure is undertaken in much younger patients with important additional congenital anomalies; however, the widely publicized decreased expense and reduced morbidity of PC has overshadowed these comparison inequities and resulted in a strong national bias toward PC.
Importantly, our current study determined that the paradigm shift favoring PC has occurred mainly due to increased prevalence of closure in adults over age 40 years rather than an increase in closure in infants or children. Interestingly, surgical closure rates have also increased in older patients since the rise in PC, although certainly not to the same degree. Because of the increased proclivity for PC in adults, we attempted to ascertain whether PC in these adults older than 40 years met established criteria for surgical closure (stroke from paradoxical embolism, atrial arrhythmias, or heart failure), or whether their ASD/PFO closures represented a lowered threshold for PC. We determined that these patients undergoing PC possessed one of these codiagnoses less than 20% of the time, a number that would support the existence of a more liberal trigger for PC and explain the disproportionate increase in PC among middle-aged adults.
We also found that an important number of patients were admitted with a codiagnosis of coronary artery disease. It is likely that asymptomatic secundum ASD/PFOs were incidentally noted in these patients during coronary angiography and repaired with concomitant device closures. Given that morbidity and mortality with PC are not zero and the complication rate of asymptomatic ASD/PFOs in adults is exceedingly low, the practice of pursuing "prophylactic" device closure seems unreasonable. This is especially so in light of recent guidelines published by the American Heart Association/American Stroke Association stating that insufficient data exists to recommend ASD/PFO closure except for recurrent cryptogenic stroke despite maximum medical therapy [17].
We also found that surgical closure rates increased modestly among adults. This variation, though, may be a downstream consequence of enhanced detection of asymptomatic ASDs that are not amenable to device closure.
Limitations
Use of an administrative database has well-known limitations that certainly apply in our current study [18–20]. Additionally, although we attempted to minimize coding errors by requiring both a diagnosis of secundum ASD/PFO and a procedure code for ASD/PFO closure, we cannot determine the influence of these errors. We could not determine precisely the appropriateness of ASD/PFO closure among patients. Our investigation of the indications for ASD/PFO closure was limited to the presence of coexisting diagnoses, which were present in only a minority of the dataset as the overwhelming majority of patients were admitted with a principal diagnosis of ASD/PFO. This could indicate that, perhaps, patients may have met criteria before admission to hospital for ASD/PFO closure, or that they met criteria not captured by the NIS (Qp:Qs > 1.5).
In conclusion, nationwide utilization of ASD/PFO repair has increased over time, largely attributable to the dramatic rise in PC, especially among adults older than 40 years. We speculate that perceived diminished risk of PC coupled with decreased cost and aggressive advocacy by device companies has lowered thresholds for ASD/PFO closure. It is likely that this phenomenon is an inevitable consequence of our current healthcare system: procedure-based remuneration encourages performance of more procedures. Despite the substantial increase in utilization, mortality has remained constant. It is not within the scope of this study to establish whether this increased utilization has resulted in an important benefit to healthcare, or simply more cost. In the absence of meaningful benchmarks, prospective studies comparing both indications, outcomes and criteria for surgical versus PC are needed to determine whether the increased rates of ASD/PFO closure are justified.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
I congratulate Drs Karamlou and colleagues for conducting this most interesting study and for the elegant presentation. The authors obtained data from the National Inpatient Sample, or NIS, database, discharge information based on ICD-9 codes concerning closure of atrial septal defects over an 18-year period. The NIS was developed for research and decision making under the Healthcare Cost and Utilization Project. It contains data from 8 million hospital stays in 1,000 hospitals. This study demonstrated enormous numbers of ASDs being closed nationally. There was a dramatic increase in percutaneous closure with surprisingly high rates in adult patients. Similar changes in ASD closure rates and method have occurred at Texas Children's Hospital. We noted a moderate increase in ASD closures, with significantly increased numbers of percutaneous and total ASD closures coincident with Amplatzer device approval.
Is the increase in ASD closures medically appropriate? In observing that only 19% of adult percutaneous closures met their definition of accepted criteria, including atrial arrhythmia, heart failure, or paradoxical embolus, the authors make highly charged conclusions. They theorize that device availability, medical industry pressure, ease of implantation, and remuneration are driving the increase in ASD closure rather than medical need.
There is conflicting opinion. Just last Friday, the American College of Cardiology "Cardiosource" highlighted the causal role of patent foramen ovale in cryptogenic stroke. The Cardiosource on-line video notes a November 29, 2007, New England Journal of Medicine report from Handke and colleagues documenting increased incidence of PFO in association with stroke in adult patients. They conclude an association exists between the presence of patent foramen ovale and cryptogenic stroke. The Handke study may make it difficult for the practicing cardiologist to disregard an incidentally discovered ASD or PFO. There are several ongoing randomized studies, with catchy acronyms like RESPECT and CLOSURE, evaluating PFO closure.
I would like to ask the following questions. Number one, can the NIS distinguish between secundum ASD and PFO?
Second, are heart failure, dysrhythmia, and history of paradoxical embolus the only acceptable indications for ASD closure, or is presence of any atrial level communication an adequate indication? Is this not the underlying question resulting from your study?
Third, at our own institution there was a "bubble" effect of increased numbers of ASD closures related to FDA approval of the Amplatz device. Patients and families deferred ASD closure in anticipation of device availability. Is it possible that some of the observed increase in ASD closure rates was patient driven? Did improved diagnostic methods have an effect?
Fourth, is the NIS a reasonable platform for similar future studies, especially the expected introduction of costly and potentially higher risk therapies such as percutaneous valve replacement?
And then finally, a bonus question. An asymptomatic young cardiac surgeon less than 49 years old is incidentally discovered to have a 12-mm isolated secundum ASD with mild right heart enlargement. What is your advice for this patient?
Again, I would like to congratulate the authors and Dr Karamlou for a fine study and for the privilege of discussing it. Thank you.
DR KARAMLOU: Thank you, Dr Fraser, for your thoughtful questions and analysis of our manuscript.
Regarding your first question, the NIS does not distinguish between patent foramen ovales and secundum ASDs. However, I think that this fact actually increases the potency of our study in that given the very low risk of an isolated PFO in an asymptomatic adult, are we sure that we are providing value by putting devices in these asymptomatic adults?
The second question regards whether or not all detected ASDs or PFOs should be closed based on this study, and I think this is really the central question posed. Certainly there are other indications that we could not track using the NIS, such as a Qp/Qs ratio. However, I think given that there is a very low risk of complications associated with not closing these, and that the data and the literature are unclear, it would probably be unreasonable to pursue a policy of closing any detected ASD or PFO in an asymptomatic adult.
Regarding the literature, there was a historical study from Shah and colleagues in London where they followed 100 patients actually over 25 years, and they discovered that compared with surgical versus medical management, there was no benefit in terms of mortality, the rate of paradoxical embolism or stroke, or right heart dysfunction among the medically versus surgically treated patients. Subsequent to this, there have been two randomized studies comparing surgical versus medical management, and the data are conflicting. There was a modest survival benefit in the New England Journal of Medicine paper but no benefit in a subsequent study reported in the Journal of the American College of Cardiology. I am unaware of any randomized trials comparing device closure to medical therapy.
So I think we cannot answer that question since there are insufficient data to formulate a consensus-based recommendtaion. But at the present time I think that repairing these in a prophylactic fashion given the risk, which is not zero for percutaneous closure and is about 2% in most series for device embolization and 2% access complications, it would be unreasonable to pursue prophylactic closure in asymptomatic adults.
The third question regarding the increased detection rates, certainly this plays a role. Many of the advancements in echocardiography, however, with the exception of three-dimensional echocardiography, actually predated this prolific rise in the rates of percutaneous closure. Certainly, I think there is a modest influence of increased detection rates on the increase in closure. We could not track echocardiography through the National Inpatient Sample. However, we did look at patients who were admitted with a diagnosis of an acute coronary syndrome or angina and were undergoing cardiac catheterization, and we found more than 10,000 patients in the percutaneous closure arm who were undergoing concomitant cardiac catheterization for an unrelated event. So I think what is likely happening is that these ASDs are incidental findings seen on cardiac catheterization, and many of these patients underwent closure simply because their ASD or PFO was detected.
The fourth question regarding the use of the NIS, I think that it is an ideal database to study national trends over time in an unbiased fashion. In addition, the NIS can be linked to other databases, including the American Hospital Association database. So you can really look at hospital-based characteristics, including bed size, nurse-to-patient ratios, intensive care unit ratios, that may modulate some of the things that you are trying to study, especially when we are looking at new technologies coming forward.
The fourth bonus question, I think this patient actually is not truly asymptomatic, and I think that the data are clear that if you are having symptoms from your ASD and you are an otherwise reasonable risk candidate, it probably would be beneficial to close the ASD or PFO in that patient. Thanks.
DR CHRISTO I. TCHERVENKOV (Montreal, Quebec, Canada): I congratulate you on a very insightful study, and I really appreciated the discussion and Dr Charles Fraser's comments.
Because many of these patients have a prophylactic closure, we have to, or at least the scientific community has to, demonstrate that this prophylaxis really works. I can understand that the closure with the device will remove the hemodynamic effects from the right heart, but what information do we have or assurance that these devices that have a significant component on the left atrial site are themselves not going to be thrombogenic with time, and what evidence do we have that they might actually not be arrhythmogenic? Do you know of any plan for a long-term follow-up of these device-closed patients to assure the patients and the families that this prophylactic measure doesn't itself cause problems long term?
DR KARAMLOU: Well, I am not sure I can answer your question. We cannot use the NIS to track individual patient outcomes. These are deidentified patients, and there is no way to follow the individual patients over time. You can look at trends over time. Certainly there are data that device closures of ASDs do create some arrhythmogenic long-term complications, but looking at this data set, we would be unable to answer that particular question. I think it is central to determine, however, whether or not prophylactic closure, given that there is a finite risk associated with device closure, is something that is actually providing value to our patients, and I think that that is the provocative question that this study is trying to propose. Further studies going forward will be needed, unfortunately, to provide an answer.
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. R. Opotowsky and G. D. Webb Percutaneous patent foramen ovale/atrial septal defect closure: just because we can? Ann. Thorac. Surg., October 1, 2009; 88(4): 1386 - 1386. [Full Text] [PDF]
|
|
|
|
|
|
T. Karamlou, B. S. Diggs, B. W. McCrindle, R. M. Ungerleider, and K. F. Welke Reply. Ann. Thorac. Surg., October 1, 2009; 88(4): 1386 - 1387. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
