Author + information
- Published online March 25, 2019.
- Matthew Oster, MD, MPH, FAHA, Evidence Review Committee Member,
- Ami B. Bhatt, MD, FACC, Evidence Review Committee Member,
- Elisa Zaragoza-Macias, MD, MPH, Evidence Review Committee Member,
- Nandini Dendukuri, PhD, Evidence Review Committee Member and
- Ariane Marelli, MD, MPH, FACC, FAHA, Chair, Evidence Review Committee
Secundum atrial septal defect (ASD) is the most common adult congenital heart defect and can present with wide variation in clinical findings. With the intention of preventing morbidity and mortality associated with late presentation of ASD, consensus guidelines have recommended surgical or percutaneous ASD closure in adults with right heart enlargement, with or without symptoms. The aim of the present analysis was to determine if the protective effect of secundum ASD closure in adults could be qualified by pooling data from published studies.
A systematic review and meta-analysis were performed by using EMBASE, MEDLINE (through PubMed), and the Cochrane Library databases to assess the effect of secundum ASD percutaneous or surgical closure in unoperated adults ≥18 years of age. Data were pooled across studies with the DerSimonian-Laird random-effects model or a Bayesian meta-analysis model. Between-study heterogeneity was assessed with Cochran’s Q test. Bias assessment was performed with the Newcastle-Ottawa Scale and the Cochrane Risk of Bias Tool, and statistical risk of bias was assessed with Begg and Mazumdar’s test and Egger’s test.
A total of 11 nonrandomized studies met the inclusion criteria, contributing 603 patients. Pooled analysis showed a protective effect of ASD closure on New York Heart Association functional class and on right ventricular systolic pressure, volumes, and dimensions. Two additional studies comprising 652 patients were reviewed separately for mortality outcome and primary outcome of interest because they did not meet the inclusion criteria. Those studies showed that ASD closure was associated with a weak protective effect on adjusted mortality rate but no significant impact on atrial arrhythmias in patients >50 years of age. Across all studies, there was significant heterogeneity between studies for nearly all clinical outcomes. The overall body of evidence was limited to observational cohort studies, the limitations of which make for low-strength evidence. Even within the parameters of the included studies, quality of evidence was further diminished by the lack of well-defined clinical outcomes.
In conclusion, pooled data analysis on the impact of secundum ASD closure in adults was notably limited because of the lack of randomized controlled trials in patients with only secundum ASD. The few cohort studies in this population demonstrated improvement in functional status and right ventricular size and function as shown by echocardiogram. However, our findings suggest that at the time of this publication, insufficient data are available to determine the impact of ASD repair on mortality rate in adults.
- ACC/AHA Clinical Practice Guidelines
- atrial septal defects
- congenital heart disease
- Evidence Review Committee
ACC/AHA Task Force Members
Glenn N. Levine, MD, FACC, FAHA, Chair
Patrick T. O’Gara, MD, FACC, FAHA, Chair-Elect
Jonathan L. Halperin, MD, FACC, FAHA, Immediate Past Chair∗
Nancy M. Albert, PhD, RN, FAHA∗
Sana M. Al-Khatib, MD, MHS, FACC, FAHA
Joshua A. Beckman, MD, MS, FAHA
Kim K. Birtcher, PharmD, MS, AACC
Biykem Bozkurt, MD, PhD, FACC, FAHA∗
Ralph G. Brindis, MD, MPH, MACC∗
Joaquin E. Cigarroa, MD, FACC
Lesley H. Curtis, PhD, FAHA∗
Anita Deswal, MD, MPH, FACC, FAHA
Lee A. Fleisher, MD, FACC, FAHA
Federico Gentile, MD, FACC
Samuel S. Gidding, MD, FAHA∗
Zachary D. Goldberger, MD, MS, FACC, FAHA
Mark A. Hlatky, MD, FACC, FAHA
John Ikonomidis, MD, PhD, FAHA
José Joglar, MD, FACC, FAHA
Richard J. Kovacs, MD, FACC, FAHA∗
Laura Mauri, MD, MSc, FAHA
E. Magnus Ohman, MD, FACC∗
Mariann R. Piano, RN, PhD, FAAN, FAHA
Susan J. Pressler, PhD, RN, FAHA∗
Barbara Riegel, PhD, RN, FAHA∗
Frank W. Sellke, MD, FACC, FAHA∗
Win-Kuang Shen, MD, FACC, FAHA∗
Duminda N. Wijeysundera, MD, PhD
Table of Contents
Eligibility Criteria 1582
Information Sources and Search Criteria 1582
Data Collection 1582
Risk-of-Bias Assessment 1582
Statistical Analysis 1582
Treatment Effect 1582
Descriptive Plots 1582
Study Selection 1583
Study Characteristics 1583
Synthesis of Results 1583
Functional Capacity and RV Systolic Pressure 1583
Echocardiographic Measures of RV Size and Function 1583
Echocardiographic Measures of LV Size and Function 1584
Mortality Rate 1584
Other Outcomes 1584
Risk of Publication Bias 1585
Table and Figures 1587
Evidence Review Committee Relationships With Industry and Other Entities (Relevant) 1595
Atrial septal defect (ASD) is the most prevalent congenital heart disease lesion in the adult population with congenital heart disease (1), with secundum ASD accounting for >90% of observed cases (2). A significant left-to-right shunt results in right ventricular (RV) volume overload and an increase in pulmonary blood flow. This increase, in turn, results in right heart failure (HF) with fatigue and exercise intolerance (2). Atrial arrhythmias resulting from atrial enlargement develop with age, conferring a higher risk of thromboembolic complications even after repair (3). Early natural history studies of unoperated large ASDs have indicated a risk of severe pulmonary hypertension with significant morbidity and mortality (4). These studies revealed that nearly 25% of patients with unoperated ASDs died just before their 27th year and 90% by their 60th birthday (4). Even patients with surgically repaired ASD may have reduced survival when compared with an age- and sex-matched control population when surgery is performed after age 25 years (5).
Symptoms vary widely with age at presentation and shunt size and typically lag behind objective evidence of cardiopulmonary dysfunction, which suggests that symptoms alone cannot guide therapy (6). The advent of cardiac ultrasonography has enabled early diagnosis. Coupled with the low mortality rate associated with surgical repair, this has led to an increase in ASD closure rates over the past 2 decades (7). Moreover, the availability of percutaneous closure as a safe and cost-effective method of repair has further facilitated the decision to proceed with intervention (8).
With the intention to improve mortality in adults as has been shown in children (5), and with a goal of improving morbidity in patients >40 years of age (9,10), recommendations have been published supporting surgical or percutaneous ASD closure in adults with right atrial and ventricular enlargement with or without symptoms (Class of Recommendation: I, Level of Evidence: B-NR) (11–14).
On the basis of the “ACCF/AHA Clinical Practice Guideline Methodology Summit Report” (15), the American College of Cardiology (ACC)/American Heart Association (AHA) Task Force on Clinical Practice Guidelines recognized the need for an objective review of available randomized controlled trials (RCTs) and observational studies by an independent evidence review committee (ERC) to inform any recommendations in the “2018 AHA/ACC Guideline for the Management of Adults With Congenital Heart Disease” (16). Despite numerous widely cited individual studies, whether pooled results of ASD closure in adults can be shown to document a robust association with a decrease in mortality and morbidity resulting from adverse cardiopulmonary function remains unknown. Therefore, a systematic review and meta-analysis of the literature were initiated to better characterize the current body of evidence on the impact of ASD surgical or percutaneous closure in unoperated adults with secundum ASD.
Relationships With Industry and Other Entities
The ACC and AHA sponsor the guidelines without commercial support, and members volunteer their time. The ACC/AHA Task Force on Clinical Practice Guidelines avoids actual, potential, or perceived conflicts of interest that might arise through relationships with industry or other entities. All ERC members are required to disclose current industry relationships or personal interests, from 12 months before initiation of the writing effort. The ERC chair and all ERC members may not have any relevant relationships with industry or other entities (Appendix 1). For transparency, ERC members’ comprehensive disclosure information is available online. Comprehensive disclosure information for the Task Force is also available online.
In conjunction with discussion from the guideline writing committee members, our clinical question was broadly defined as follows: Are outcomes in patients with unoperated secundum ASD and RV dilatation improved after percutaneous or surgical ASD closure? A systematic review and meta-analysis were performed to compare outcomes in adults with unoperated secundum ASD with outcomes in patients with shunt closure. Specific elements related to Population, Intervention, Comparison, Outcome, Timing, and Setting (PICOTS) comprised our PICOTS question as follows:
1) Population: Patients ≥18 years of age with unrepaired secundum ASD and RV dilation.
2) Intervention: Surgical or percutaneous closure of ASD.
3) Comparison: No intervention.
4) Outcome: Primary outcome of death; secondary outcomes of functional capacity, arrhythmia, congestive HF, hospitalization including heart transplantation, and improvement in RV dilation, and right dysfunction.
5) Timing: From point of assignment (intervention) until intermediate term (1 month to 1 year) and long term (≥1 year).
6) Setting: Hospital and outpatient care.
Studies were included in the systematic review if they met the following eligibility criteria: 1) an RCT or cohort study with at least 20 adult participants; 2) publication in a peer-reviewed journal; 3) inclusion of only adults (≥18 years of age) or separate reporting of outcomes for adult patients; 4) follow-up for at least 1 month after diagnosis (if no intervention) or after intervention; 5) inclusion of at least 1 of the primary or secondary outcomes of interest previously listed; and 6) full-text article available in English. The full search strategy can be found in Online Data Supplement 1.
Information Sources and Search Criteria
A systematic search strategy was developed, and 3 major databases were searched: MEDLINE (through PubMed), EMBASE, and the Cochrane Library. The full search strategy can be found in Online Data Supplement 1, and Figure 1 illustrates the study selection screening process. To determine the eligibility of articles for inclusion in the systematic review, 2 members of the ERC independently reviewed each abstract and full article. Disagreements were resolved by consensus or through the involvement of a third reviewer (A. Marelli). Abstracted data were entered into the Indico Clinical Guideline Platform (Indico Solutions. Ltd., Melbourne, Victoria, Australia), a Web-based software platform. For each included study, the ERC members abstracted data on study author; year of publication; sample size; inclusion and exclusion criteria; study design; setting (outpatient versus inpatient); participant characteristics (age, sex, and presence of structural heart disease); description of the tests/procedures and their results or acute outcomes; long-term outcomes, including sudden cardiac death or arrhythmic death; atrial fibrillation (AF); regular supraventricular tachycardia; all-cause death; quality of life; hospitalization/readmission for cardiovascular events; ablation-related complications; duration of follow-up; and loss to follow-up. Overall study quality was assessed on the basis of risk of bias, relevance to the study question, and fidelity of implementation (15). To evaluate risk of bias, the Cochrane Collaboration Risk of Bias Tool was used for RCTs (17), and the Newcastle-Ottawa Scale was used for cohort studies (18). An RCT was assigned an overall rating of low-to-intermediate risk of bias if the trial was not deemed to be at high risk of bias for any assessed domain of study quality.
Data were extracted independently by at least 2 reviewers and checked for accuracy. Data extraction was sought for clinical outcome measures of interest. These included death, occurrence of AF, New York Heart Association (NYHA) (class I versus other), oxygen uptake (peak ventilatory equivalent of oxygen [o2]), heart transplantation, and inpatient hospitalization. These also included echocardiographic parameters of RV function: RV end-systolic dimension, RV end-diastolic dimension, RV end-systolic volume, RV end-diastolic volume, RV ejection fraction, and RV systolic pressure. Outcomes also encompassed measures of left ventricular (LV) function: LV end-systolic dimensions, LV end-diastolic dimensions, and LV ejection fraction (LVEF).
Summary statistics for each outcome were extracted (mean and standard deviation) before and after the intervention. This was performed for treatment and control groups separately, where applicable.
Risk of bias was assessed with the Newcastle-Ottawa Scale (18) and the Cochrane Risk of Bias Tool (19) for cohort studies and randomized studies, respectively. The Newcastle-Ottawa Scale was developed to assess and incorporate the quality of nonrandomized studies in the interpretation of meta-analytic results.
The treatment effect was measured by calculating the difference in means before and after the intervention and associated 95% confidence interval for the outcomes of interest. This was performed for treatment and control groups separately.
Forest plots were used to visually examine the heterogeneity between studies. The null hypothesis of no heterogeneity between the studies was tested with Cochran’s Q statistic. A p value <0.05 would suggest evidence of heterogeneity between studies.
Formal meta-analyses were performed for those outcomes of interest with results from at least 2 separate studies meeting eligibility criteria. The pooled treatment effect and associated 95% confidence interval across studies were estimated with the DerSimonian-Laird random-effects model (20). When evidence of between-study heterogeneity existed, a 95% prediction interval was also obtained. The confidence interval reflects the uncertainty in the pooled value across the observed studies, whereas the prediction interval reflects the uncertainty in the estimate of the true value in a future study. The impact of heterogeneity on the overall variance was measured by the I2 statistic (21). When the overall pooled measure of effect was significant (P<0.05) and the measure of heterogeneity between studies appeared to be nonsignificant (P>0.05) (the null hypothesis being that there is no heterogeneity between studies as assessed with the frequentist DerSimonian-Laird random-effects model), Bayesian modeling was applied to better estimate the uncertainty in the heterogeneity, which is known to be underestimated by the DerSimonian-Laird model. In the Bayesian analyses, a noninformative prior distribution on the pooled effect size (uniform [-10, 10]) was used. For the between-study standard deviation, which is known to be sensitive to prior choice, 2 different noninformative prior distributions were considered: 1) uniform (0, 10) and 2) half-normal (mean: 0; standard deviation: 5) (22). In instances in which a particular outcome was reported in only 1 included study, a meta-analysis was not performed for that outcome; the findings in the results were summarized. Publication bias was examined visually by funnel plots and/or by using Begg’s test and the weighted regression test of Egger (23,24). The plots and frequentist meta-analyses were carried out with the metafor package in R (25). The Bayesian analyses were carried out with the WinBUGS software package (26).
Six hundred and twenty-seven abstracts met our search criteria and were screened for potential inclusion. Of those, full-text screening was performed on 135 articles (Online Data Supplement 2); 11 articles met full inclusion criteria for the meta-analysis (27–37) (Figure 1). Common factors limiting study eligibility included but were not limited to: studying ASDs other than secundum ASD (or not specifying results for secundum ASD separately), having <20 patients, and including both children and adults without having results available separately for adults. Of the 11 studies selected for meta-analyses, 7 were prospective, none were randomized, and none included multiple centers.
Two additional studies—a prospective, single-center RCT and a retrospective, nonrandomized, multicenter cohort study—were analyzed despite not meeting the PICOTS search criteria, because they were the only studies to report our primary outcome of interest, death in adults (38,39). However, they were excluded from the meta-analyses of our other outcomes because they did not report outcomes for secundum ASD separately from outcomes for other types of ASD.
Eleven studies met full inclusion criteria (Table 1), contributing 603 patients, with ages ranging from 18 to 92 years. Percutaneous closure was included in 10 of the 11 studies, and surgical closure was included in 1 study. Follow-up time ranged from 1 month to 9 years. In the 2 additional studies reporting mortality rate, ages ranged from 40 to >60 years of age, and all interventions were surgical.
Synthesis of Results
Functional Capacity and RV Systolic Pressure
Pooled analysis showed a protective effect of ASD closure on functional capacity, as demonstrated by significantly higher odds of being in NYHA class I after repair, with either intervention or surgery, with subjects having 14 times the odds of being in NYHA class I after versus before ASD closure (Figure 2).
The outcome RV systolic pressure was reported in 3 studies, 1 of which reported results separately by age strata (36) between patients 18 to 50 years of age and those >50 years of age. Stratification by age was done when included studies considered different age groups because age was expected to be an effect modifier. There was an overall improvement in RV systolic pressure after ASD closure, with a net decrease of 6 mm Hg (Figure 3). However, this effect was inconsistent between studies or age groups. In the study that reported stratification by age, RV systolic pressure decreased significantly among those >50 years of age but not among those 18 to 50 years of age (36).
For peak o2, ASD closure was associated with an increase of approximately 4 mL/kg/min. Whereas the pooled effect was statistically significant, the confidence interval around the measure of effect was wide, with the lower limit of 0.23 not being consistent with a meaningful improvement. Furthermore, the high heterogeneity between studies resulted in the prediction interval including zero, indicating the possibility of no effect on the basis of the limited evidence accrued so far. Thus, a clinically meaningful impact on peak o2 could not be demonstrated (Figure 4).
Echocardiographic Measures of RV Size and Function
For outcomes related to parameters of RV size and function, there were consistent reductions in ventricular size but no significant improvement in systolic function. In the 3 studies that reported RV ejection fraction, pooled analysis did not reveal a statistically significant improvement after ASD closure. The wide confidence interval, the small number of studies, and the significant between-study heterogeneity limit the ability of meta-analyses to result in any meaningful observations related to this outcome (Figure 5). The outcome RV end-diastolic dimension was reported in 8 studies, 1 of which reported results separately by age strata (36) between patients 18 to 50 years of age and those >50 years of age, similar to the stratification previously shown for RV systolic pressure. Thus, a net decrease of 7 mm in RV end-diastolic dimension after ASD closure was documented across all studies, although the effect was more significant in patients <50 years of age (Figure 6A). Although only 2 studies reported RV volumes in end diastole and systole, there were significant net decreases of 64 mL and 28 mL, respectively, after ASD closure versus before ASD closure (Figure 6B and 6C). However, the presence of significant heterogeneity between studies and the relatively small number of studies mitigate our interpretation of these findings.
Echocardiographic Measures of LV Size and Function
An improvement in LV function and size was observed in diastole after ASD closure. There was an overall small mean improvement in LVEF of 5% but with a wide prediction interval that included zero (Figure 7). In the 7 studies from which data analysis on LV end-diastolic dimension could be pooled, there was a small increase in LV size of 5 mm (Figure 8A). For both LVEF and LV end-diastolic dimension, there was significant heterogeneity between studies, which attenuated the interpretation of the impact of ASD closure on these outcomes. Supporting this cautious note, the impact of ASD closure on LV systolic diameter showed an overall effect that was significant when the frequentist DerSimonian-Laird method was used but that became attenuated when the Bayesian model was applied (Figure 8B). Bayesian modeling was performed to overcome the lack of heterogeneity between studies that were detected statistically (P=0.175), potentially giving the misleading impression of a robust measure of effect.
To obtain data on mortality rate, 2 studies were analyzed separately from the other studies because they did not meet inclusion criteria for the original PICOTS question (38,39). These were the study by Attie et al. (38), an RCT of 473 patients, 11.8% of whom had sinus venosus defect; and the study by Konstantinides et al. (39), a retrospective cohort study of 179 patients, 9% of whom had ostium primum or sinus venosus defects. Thus, both studies were excluded from the meta-analyses as per the PICOTS question. In both studies, the data were not reported separately on the basis of anatomy; therefore, Figure 9 represents all ASD patients in both studies.
Figure 9 illustrates the results of pooling analyses from both papers. It should be noted that these analyses are for descriptive purposes only because the 2 studies differed in both design and the variables they considered in a multivariate analysis. Figure 9A shows the unadjusted overall mortality rate, with an overall effect that is not significant and evidence of heterogeneity between studies. Figure 9B illustrates the result of pooled analyses on adjusted mortality rate: in Attie et al. (38), adjusted for pulmonary hypertension and cardiac index; in Konstantinides et al. (39), adjusted for NYHA class, systemic hypertension, AF, and shunt size; and in both, adjusted for age. In Konstantinides et al. (39), only the relative risk of death was reported; however, on the basis of the analysis, this was intended to be a hazard ratio for the purpose of generating Figure 9B.
Thus, for the primary outcome intended for the present meta-analysis, a limited analysis of 2 studies that used different study designs and reported the mortality rate in a group of patients was not restricted to those with secundum ASD. In this analysis, ASD closure was associated with a weak overall protective effect on mortality rate adjusted for different variables.
For the outcome of AF, there were limited data, with a meta-analysis performed only on patients >50 years of age from 2 studies (28,36). Those studies found that there was no statistically significant difference in the odds of AF after ASD closure (odds ratio: 1.78; 95% confidence interval [CI]: 0.90 to 3.52). There was not enough high-quality evidence for our other long-term clinical outcomes of transplantation or inpatient hospitalization to include in any analyses.
Results of the statistical analysis for clinical or methodological heterogeneity of treatment effect on outcomes are depicted at the bottom of each of Figures 2, 3, 4, 5, 6, 7, 8, and 9. Statistical heterogeneity is significant when the effect measure being evaluated differs between studies to a greater extent than it does within studies. There was evidence of significant heterogeneity between studies for nearly all outcomes: NYHA class, peak o2, RV ejection fraction, RV end-diastolic dimension, RV end-diastolic volume, LVEF, and LV end-diastolic dimension. The only outcomes in the main analysis for which there was no significant heterogeneity of treatment effect between studies were RV end-systolic pressure, RV end-systolic volume, and LV end-systolic dimension.
Risk of Publication Bias
The results from the analyses of risk of publication bias—for which Begg and Mazumdar’s test and Egger’s test (Online Data Supplement 3) and Cochrane Risk of Bias Tool (Online Data Supplement 4) were used—are illustrated online. For the LVEF outcome, the funnel plot analysis disclosed severe asymmetry, with the study by Khan et al. (31) lying toward the base of the funnel. Consistent with this, as can be seen in the forest plot shown in Figure 7, the study by Khan et al. has a high standard error. Although the study is quite different from the others, it is of similar weight because of the random-effects model used. Had the study been excluded, the results of the pooled analysis would have been even closer to zero. Despite the difference in outcome being reported as zero, we chose to keep the study in the analysis because it met all PICOTS inclusion criteria.
The risk of methodological bias was assessed with the Cochrane Collaboration Risk of Bias Tool and the Newcastle-Ottawa Scale (Online Data Supplements 4 and 5). With the Cochrane Collaboration Risk of Bias Tool applied to only 1 RCT included in our systematic review but not included in our meta-analysis because it did not meet PICOTS criteria, Attie et al. (38) had a low-to-moderate risk of bias. With the Newcastle-Ottawa Scale for cohort studies, the overall quality score varied from 5 to 9 in only 1 study, with a median score of 5, consistent with a low strength of evidence because of study limitations.
To our knowledge, this is the first systematic review to address ASD closure in adults. For the PICOTS question, “Are outcomes in patients with unoperated secundum ASD and RV dilation improved after percutaneous or surgical ASD closure?,” limited data were found. Death was sought as a primary outcome, and functional capacity, arrhythmia, congestive HF, hospitalization, and improvement in RV function were sought as secondary outcomes. Pooled analyses, wherever possible, showed that ASD closure had a protective effect on functional capacity, as demonstrated by a significantly higher chance of being in NYHA class I after repair, whether with intervention or surgery. Significant improvements in RV systolic pressure, volumes, and dimensions were documented after repair. Finally, improvements in LV function and size in diastole were observed after repair. However, a clinically meaningful impact on peak o2 or RV ejection fraction was not found from the pooled data analysis. Using our search criteria, we did not find a sufficient number of studies for a pooled analysis of the impact of ASD closure on death or arrhythmia. Thus, in a limited analysis in which we pooled 2 studies reporting mortality rate in a group of patients that was not restricted to those with secundum ASD, we found that ASD closure was associated with a weak overall protective effect on the outcome of death, adjusted for differing variables. For arrhythmia, we pooled 2 studies reporting AF outcomes only in patients >50 years of age and found no significant difference in the odds of AF after ASD closure.
Functional capacity in patients with unrepaired ASD is substantially impaired, especially with advancing age (6). This is likely a reflection of increased pulmonary blood flow and pulmonary hypertension worsened by impaired LV compliance as a result of increasing age (40). Studies in patients >40 years of age have suggested decreased morbidity, improved NYHA functional class, and better cardiopulmonary function testing, even after ASD closure in asymptomatic adults (38,41). In keeping with the pathophysiology of unrepaired ASD, pooled results showed the beneficial effect of ASD closure on NYHA class and RV systolic pressure. The inability to document a clinically meaningful impact of ASD closure on peak o2 uptake is likely the result of finding only a small number of heterogeneous studies designed to measure this outcome.
Right atrial and RV remodeling that have been demonstrated in several small studies after ASD closure (42) may be age dependent and may impact RV function (43). The systematic search revealed several studies demonstrating a decrease in RV end-diastolic dimension and RV end-diastolic volume after ASD closure; indeed, the meta-analyses confirmed the significant impact of ASD closure on improvement in RV dimensions. The follow-up time in the published studies varied widely—from 1 month to 9 years. It is thus likely that duration of follow-up affects age-related observations related to RV size and function, with the less compliant RV in older patients possibly taking longer to remodel.
Interestingly, the analysis revealed an improvement in LVEF and LV end-diastolic dimension with closure of ASD. Particularly for individuals diagnosed with ASD at an older age, the effect of closure on the right heart may not be the only potential benefit over time. The interaction between RVs and LVs is well described in various congenital and noncongenital anomalies, with reports on the normalization of abnormal septal motion and septal curvature at end diastole with improvement in RV function (44). The beneficial impact of ASD closure on the LV may also be present in patients with secundum ASD who develop late mitral regurgitation (45). In a study of 74 prospectively enrolled patients with ASD undergoing surgical repair, mitral regurgitation increased in older patients who had elevated end-diastolic LV dimensions and mitral annular enlargement (46). Unrepaired ASDs have also been shown to result in decreased LV compliance in the setting of RV volume overload, which suggests that unrepaired ASDs may carry a risk for LV contribution to clinical HF over time (47).
Atrial arrhythmias are an important consequence of atrial-level shunting and occur as a result of right or left atrial enlargement, or both (48,49). Our assessment of the effect of ASD closure on atrial arrhythmia was limited to 2 eligible studies and revealed no significant difference in AF occurrence after closure. This assessment was based largely on a retrospective study of 353 patients who underwent transcatheter ASD closure and for whom long-term AF risk was not attenuated, especially in the subgroup of adults >50 years of age. That study suggested that transcatheter closure of ASD may not provide much electromechanical improvement for older patients, despite providing improvement in ventricular dimensions (36). The second study further documented that AF occurrence in patients >50 years of age with secundum ASD did not decrease significantly after surgical ASD closure, despite significant clinical symptom improvement (28). Other studies have also revealed that surgical ASD closure has little effect on AF incidence (38) and that atrial flutter and AF appear to be more likely in those undergoing surgery after 40 years of age (3,5).
One important confounder in reports of surgical ASD closure in older adults is the inclusion or exclusion of a Cox MAZE procedure, with the impact of an additional surgical procedure that may be incorporated into surgical planning remaining unexamined. Recently, a prospective study of 135 consecutive ASD patients revealed no stroke, pulmonary hypertension, HF, or death and an impressively low incidence of atrial arrhythmia after surgical ASD closure in childhood (50). That study is not included in our present meta-analysis because it included patients with a sinus venosus defect, and all repairs were performed before 15 years of age.
For the intended primary outcome of death, data were not sufficient to perform a meta-analysis comparing the impact of ASD closure with that of medical therapy in adults. The 2 studies in which mortality rate data were available were substantially different: 1 was a retrospective cohort study, 1 was an RCT, and each was adjusted for different covariates. Between the 2 studies, there were only 24 deaths, yet adjustments were made for up to 8 variables. The instability of the statistical models in each of the studies is reflected in the wide confidence interval shown in Figure 9B. Moreover, for the purpose of generating Figure 9, it was assumed that the relative risk reported in the Konstantinides study (39) was intended to be a hazard ratio as opposed to a relative risk. Nonetheless, both studies were adjusted for age and had a similar follow-up duration of 7 to 9 years. Thus, low mortality rates in the ASD population studies to date not only make clinical trials challenging but also limit our ability to clearly demonstrate the impact of ASD closure on risk of death.
Our analyses were substantially limited in the number and quality of studies available. The overall body of evidence was limited to observational cohort studies that constitute a low strength of evidence because of study limitations. Even within the parameters of the included studies, quality of evidence was further diminished by the lack of hard outcomes such as mortality rate, inadequate accounting for confounding, the absence of blinded review of outcomes, and selection bias. Studies were excluded on the basis of inclusion criteria related to type of ASD, age, or insufficient power to detect meaningful differences in the outcomes of interest. The inclusion of adults of all ages in numerous studies also posed an important challenge. Additionally, the hemodynamic and clinical responses to atrial-level shunt differ significantly in terms of duration of follow-up, and this has not been systematically accounted for in the data published to date. This limited our ability to provide more granular information on the impact of age as an effect modifier and to perform age-stratified analyses.
ASDs constitute the most common congenital heart disease lesion in adults. Although ASD closure is recommended in those with right heart enlargement, data informing such a recommendation in adults specifically are limited. The evidence base for outcomes with or without defect closure is scattered among multiple studies with differing observation periods and across populations of individuals who vary in age and anatomic diagnosis. The data that do exist focus predominantly on imaging endpoints and short-term morbidity but not on mortality or long-term morbidity. The few studies in this population that include adults of all ages have demonstrated mild improvements in functional status and RV and LV echocardiographic parameters after ASD closure. The longitudinal influence of ASD closure on RV function and functional status appears to be more clinically apparent in middle-aged and older adults; therefore, future studies will need to be focused on this population and to pay greater attention to well-defined clinical outcomes.
Thus, the lack of sufficient quality data inhibits our ability to reliably answer the PICOTS question in either the affirmative or the negative, particularly with respect to the outcome of death. Although the literature to date in adults is limited, the clinical experience of individual providers and large, experienced centers will continue to impact clinical decision making. Although such clinical practice may lead to a lack of equipoise, the findings of this systematic review underscore the importance of substantiating consensus recommendations with systematic reviews, which achieve the dual purpose of reevaluating common practice and providing guidance for future research, such as rigorous RCTs.
Presidents and Staff
American College of Cardiology
C. Michael Valentine, MD, FACC, President
Cathleen C. Gates, Interim Chief Executive Officer and Chief Operating Officer
William J. Oetgen, MD, MBA, FACC, Executive Vice President, Science, Education, Quality, and Publishing
MaryAnne Elma, MPH, Senior Director, Science, Education, Quality, and Publishing
Amelia Scholtz, PhD, Publications Manager, Science, Education, Quality, and Publishing
American College of Cardiology/American Heart Association
Katherine Sheehan, PhD, Director, Guideline Strategy and Operations
Abdul R. Abdullah, MD, Senior Manager, Guideline Science
American Heart Association
Ivor Benjamin, MD, FAHA, President
Nancy Brown, Chief Executive Officer
Rose Marie Robertson, MD, FAHA, Chief Science and Medical Officer
Gayle R. Whitman, PhD, RN, FAHA, FAAN, Senior Vice President, Office of Science Operations
Prashant Nedungadi, PhD, Science and Medicine Advisor, Office of Science Operations
Jody Hundley, Production and Operations Manager, Scientific Publications, Office of Science Operations
Appendix 1 Evidence Review Committee Relationships With Industry and Other Entities∗,† (Relevant)—Interventional Therapy Versus Medical Therapy for Secundum Atrial Septal Defect: A Systematic Review (Part 2) for the 2018 AHA/ACC Guideline for the Management of Adults With Congenital Heart Disease
↵∗ Former Task Force member; current member during the writing effort.
This document was approved by the American College of Cardiology Clinical Policy Approval Committee in May 2018, the American Heart Association Science Advisory and Coordinating Committee in June 2018, and the American Heart Association Executive Committee in July 2018.
The American College of Cardiology requests that this document be cited as follows: Oster M, Bhatt AB, Zaragoza-Macias E, Dendukuri N, Marelli A. Interventional therapy versus medical therapy for secundum atrial septal defect: a systematic review (part 2) for the 2018 AHA/ACC guideline for the management of adults with congenital heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. J Am Coll Cardiol 2019;73:1579–95.
This article has been copublished in Circulation.
Copies: This document is available on the World Wide Web sites of the American College of Cardiology (www.acc.org) and the American Heart Association (professional.heart.org). For copies of this document, please contact the Elsevier Inc. Reprint Department via fax (212-633-3820) or e-mail ( ).
Permissions: Multiple copies, modification, alteration, enhancement, and/or distribution of this document are not permitted without the express permission of the American College of Cardiology. Requests may be completed online via the Elsevier site (https://www.elsevier.com/about/policies/copyright/permissions).
- 2019 American Heart Association, Inc., and the American College of Cardiology Foundation
- Marelli A.J.,
- Ionescu-Ittu R.,
- Mackie A.S.,
- et al.
- Campbell M.
- Kotowycz M.A.,
- Therrien J.,
- Ionescu-Ittu R.,
- et al.
- Jemielity M.,
- Dyszkiewicz W.,
- Paluszkiewicz L.,
- et al.
- John Sutton M.G.,
- Tajik A.J.,
- McGoon D.C.
- Warnes C.A.,
- Williams R.G.,
- Bashore T.M.,
- et al.
- Halperin J.L.,
- Levine G.N.,
- Al-Khatib S.M.,
- et al.
- Jacobs A.K.,
- Kushner F.G.,
- Ettinger S.M.,
- et al.
- Stout K.K.,
- Daniels C.J.,
- Aboulhosn J.A.,
- et al.
- Higgins J.P.T.,
- Altman D.G.,
- Gøtzsche P.C.,
- et al.
- Wells G.,
- Shea G.,
- O’Connell D.,
- et al.
- ↵Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]. Available at: http://handbook-5-1.cochrane.org. Accessed May 30, 2018.
- Spiegelhalter D.,
- Abrams K.,
- Myles J.
- Egger M.,
- Davey Smith G.,
- Schneider M.,
- et al.
- Viechtbauer W.
- Brochu M.-C.,
- Baril J.-F.,
- Dore A.,
- et al.
- Eroglu E.,
- Cakal S.D.,
- Cakal B.,
- et al.
- Khan A.A.,
- Tan J.-L.,
- Li W.,
- et al.
- Ströker E.,
- Van De Bruaene A.,
- DM P.,
- et al.
- Veldtman G.R.,
- Razack V.,
- Siu S.,
- et al.
- Attie F.,
- Rosas M.,
- Granados N.,
- et al.
- Giardini A.,
- Donti A.,
- Formigari R.,
- et al.
- Haddad F.,
- Doyle R.,
- Murphy D.J.,
- et al.
- Yoshida S.,
- Numata S.,
- Tsutsumi Y.,
- et al.
- Booth D.C.,
- Wisenbaugh T.,
- Smith M.,
- et al.
- Bouchardy J.,
- Marelli A.J.,
- Martucci G.,
- et al.
- Bouchardy J.,
- Therrien J.,
- Pilote L.,
- et al.
- ACC/AHA Task Force Members
- Table of Contents
- Table and Figures
- Presidents and Staff
- Appendix 1 Evidence Review Committee Relationships With Industry and Other Entities∗,† (Relevant)—Interventional Therapy Versus Medical Therapy for Secundum Atrial Septal Defect: A Systematic Review (Part 2) for the 2018 AHA/ACC Guideline for the Management of Adults With Congenital Heart Disease