Percutaneous Patent Ductus Arteriosus (PDA) Closure During Infancy: A Meta-analysis
CONTEXT: Patent ductus arteriosus (PDA) is a precursor to morbidity and mortality. Percutaneous (catheter-based) closure is the procedure of choice for adults and older children with a PDA, but use during infancy (<1 year) is not well characterized.
OBJECTIVE: Investigate the technical success and safety of percutaneous PDA closure during infancy.
DATA SOURCES: Scopus, Web of Science, Embase, PubMed, and Ovid (Medline) were searched through December 2015 with no language restrictions.
STUDY SELECTION: Publications needed to clearly define the intervention as percutaneous PDA closure during infancy (<1 year of age at intervention) and must have reported adverse events (AEs).
DATA EXTRACTION: The study was performed according to the Systematic Reviews and Meta-Analysis checklist and registered prospectively. The quality of the selected studies was critically examined. Data extraction and assignment of AE attributability and severity were independently performed by multiple observers. Outcomes were agreed on a priori. Data were pooled by using a random-effects model.
RESULTS: Thirty-eight studies were included; no randomized controlled trials were found. Technical success of percutaneous PDA closure was 92.2% (95% confidence interval [CI] 88.8–95.0). Overall AE and clinically significant AE incidence was 23.3% (95% CI 16.5–30.8) and 10.1% (95% CI 7.8–12.5), respectively. Significant heterogeneity and publication bias were observed.
LIMITATIONS: Limitations include lack of comparative studies, lack of standardized AE reporting strategy, and significant heterogeneity in reporting.
CONCLUSIONS: Percutaneous PDA closure during infancy is feasible and associated with few catastrophic AEs; however, the limitations constrain the interpretability and generalizability of the current findings.
- ADO —
- Amplatzer ductal occluder
- AE —
- adverse event
- CI —
- confidence interval
- CNS-AE —
- clinically nonsignificant adverse event
- CS-AE —
- clinically significant adverse event
- PDA —
- patent ductus arteriosus
- RCT —
- randomized controlled trial
Patent ductus arteriosus (PDA) is considered a significant precursor to short- and longer-term morbidity.1–3 Percutaneous PDA closure has become the procedure of choice for PDA closure in adults and children4; however, generalizable scientific evidence to support its use during infancy (<1 year) is limited.5,6 Somewhat conflicting results on the safety of percutaneous PDA closure during infancy has led to uncertainty regarding patient selection and optimal timing and indications for percutaneous PDA closure, leaving health care providers with little evidence-based data to guide their clinical management.7
Previous reviews on percutaneous PDA closure have broadly investigated outcomes across all age groups, with most interventions performed outside of infancy.8–10 To our knowledge, no systematic reviews on the feasibility and complication rates among infants undergoing percutaneous PDA closure have been published. Although percutaneous PDA closure is considered a low-risk intervention,11 procedures performed during infancy are more complex than are those performed during childhood or adulthood12; thus, a separate consideration of the potential risks and benefits in this at-risk subgroup is needed. In view of the increasing number of catheter-based closures among infants,5–7,13 we conducted a systematic review and meta-analysis of the use and outcomes of percutaneous PDA closure during infancy, while attempting to characterize potential sources of data heterogeneity.
This study was performed according to the Preferred Items for Systematic Reviews and Meta-Analysis14 and registered with the PROSPERO database, the international prospective registry of systematic reviews (http://www.crd.york.ac.uk/prospero, identifier CRD42016033924). With assistance from a research librarian (A.G.), the authors performed a comprehensive search of Scopus, Web of Science, Embase, PubMed, and Medline for studies investigating percutaneous PDA closure. Search terms are available by request to the corresponding author. All searches were conducted in January of 2016. No date or language restrictions were applied.
Studies that enrolled patients <1 year of age at the time of percutaneous PDA closure were included in this review. To keep the number of studies manageable, studies were excluded if they evaluated <3 infants undergoing attempted percutaneous PDA closure. We excluded studies that did not provide data on patient age at the time of procedure. Published studies that enrolled mixed populations (infants and children or adults) were included if individual outcomes of at least 3 infants could be ascertained. Studies were not excluded for lack of adverse events (AEs), but for lack of mention of safety or AE assessment.
Two reviewers (C.B., B.R.) undertook the application of inclusion/exclusion criteria. The eligibility of the studies was formulated according to Participants, Interventions, Comparator, Outcomes, and Study Design criteria15:
Participants: Infants (postnatal age <12 months) who underwent percutaneous PDA closure.
Intervention: Percutaneous PDA closure, defined as closure with either a device (eg, Amplatzer ductal occluder [ADO] [St Jude Medical, Saint Paul, MN]) or coil (eg, Gianturco [Cook Medical, Bloomington, IN], Flipper [Cook Medical, Bloomington, IN], Nit-Occlud[pfm medical ag, Köln, Germany]).
Comparator: Any; this also included no treatment (conservative management) and any of the currently available treatments (medical therapy, surgical closure).
Outcomes: No restriction was made according to measured outcomes. However, technical success (defined later in this article), overall AEs, and clinically significant AEs (CS-AEs) were the primary outcomes designated a priori.
Study Design: In the absence of randomized controlled trials (RCTs), the inclusion criteria were extended to include trials that were observational (cohort, case series).
Decisions on study inclusion were made independently of the data extraction and before the scrutiny of results. Each identified citation was designated as definitively, possibly, or clearly not meeting inclusion criteria by using a standardized screening tool. Both abstracts and full-text reviews were piloted on sample abstracts or articles, respectively, to ensure reviewer consistency in judging inclusion criteria. For each definitively or possibly eligible citation, full-text articles were obtained. Disagreements were settled through discussion, with involvement of a third reviewer (D.B.) as necessary. When data were unclear or missing, the corresponding author was contacted via e-mail at least twice to obtain additional data to make a final determination of inclusion eligibility. When the same center reported multiple eligible case series, each of the series was included in the review. Multiple publications describing the same or overlapping series of patients were combined when feasible.
For non-English language studies included in the full-text review, independent reviewers with fluency in the article’s language translated and abstracted data from the article. To ensure accurate translations, all foreign-language articles (n = 38) were translated to English by using computer software previously shown to be effective for systematic reviews.16 All citations were imported into an electronic database (EndNote ×4; Thomson Reuters, New York, NY), which was also used for recording screening decisions and data extraction.
Study Quality Assessment
Although the checklist of Jadad et al17 has been widely used to determine study quality in systematic reviews, it was not relevant here, as no RCTs were identified. Two reviewers (C.B., B.R.) independently assessed the methodological quality of studies by using the Newcastle-Ottawa Scale for nonrandomized studies, which uses a star system to assess studies on the basis of (1) selection of study groups, (2) comparability of groups, and (3) ascertainment of exposure/outcome.18 The content validity and interrater reliability of the Newcastle-Ottawa Scale were previously established, and the scale continues to be recommended to assess nonrandomized trials.19 No studies were excluded on the basis of quality.
Authors independently extracted data via an electronic abstraction form, which was pilot tested for consistency among reviewers. Data were collected in a standardized format, as recommended by the Cochrane Non-Randomized Studies Methods Group.20
Consistent with previous reports, technical success was defined as the patient leaving the catheterization laboratory (or alternative setting) with a coil or device in the PDA. Cases in which an implant embolized during the procedure but was retrieved percutaneously and the PDA closed with a larger or different device (during the same procedure) were considered technical successes, but also listed as an AE (described later in this article).11,21 Procedural abandonments were defined as cases in which the infant left the catheterization laboratory (or alternative setting) without a device or coil in the ductus. Technical failures were defined as cases in which the device or coil was placed and the infant left the catheterization suite, but subsequently required surgical or percutaneous removal at a later time. Residual shunting was defined as angiographic or echocardiographic evidence of shunting after device placement at longest reported follow-up. Procedural details of the catheterization, including case duration, access sites (arterial, venous), and type of device/coil were abstracted, when available. When multiple device placements were attempted, only the final implant was recorded. To compare potential changes over time in risk of an AE, including embolization rates, the cohort was divided into the following epochs based on year of study publication: “first epoch” (1994–2009) and “second epoch” (2010–2016).
AEs were recorded and assessed independently by 2 pediatric cardiac catheterization (cardiac interventionalist) physicians (B.B., A.A.) based on previous work by Bergersen and colleagues.22 Consistent with previous work, AEs were stratified according to severity level (1–5).22 AE levels 1 or 2 were considered clinically nonsignificant (CNS-AE), and levels 3 to 5 considered CS-AE, with levels 4 and 5 considered major and catastrophic, respectively (Supplemental Table 4). AEs were further categorized into 4 subheadings: (1) access-related, (2) sedation or airway, (3) general catheterization, and (4) device/coil-related (Supplemental Table 5).23 The degree of association between the intervention and the AE was assessed independently (A.A., B.B.) by using the causality algorithm used by the World Health Organization Collaborating Centre for International Drug Monitoring; terminology was modified for use for a device rather than for a pharmacological product.24 Only AEs adjudicated as probable, probable/likely, or certain were included. Disagreements between reviewers on the assignment of AE, or the degree of causality, were resolved by discussion, and, if necessary, a third party was consulted (D.B.).
Synthesis of Results and Statistical Analysis
A random-effects meta-analysis model was selected a priori based on the assumption that treatment effects were heterogeneous based on expected differences in study designs and patient characteristics among studies. However, by using a fixed effects model, results did not consistently change (data available on request). Denominators were adjusted, where appropriate, to include the number of reported cases or outcome of interest. For each primary outcome, the incidence and 95% confidence interval (CI) were calculated. A forest plot was used to illustrate the individual study findings and the random-effects meta-analysis results for primary outcomes. Although traditional meta-analysis methods for calculating prevalence is based on the inverse variance method, this puts undue weight on the studies with small or large prevalence; therefore, we used MetaXL data analysis software (EpiGear International Pty Ltd, Queensland, Australia) with the double arcsine transformation.25 For presentation, the pooled transformation and its CI were back transformed to a proportion.25
The I2 statistic was used to estimate heterogeneity of effects across studies. Consistent with previous studies, values of ≤25%, 25% to 75%, and ≥75% represented low, moderate, and high heterogeneity, respectively.26 Publication bias was visually assessed with funnel and Doi plots (not shown, data available on request) and quantitatively assessed by using the LFK Index (no bias, index within ±1; minor bias, index exceeds ±1 but within ±2; major bias, index exceeds ±2).27–30
Subgroup analyses (χ2 or Fisher’s exact test) were undertaken to explore potential differences and sources of heterogeneity in outcomes. Consistent with previous studies,31–33 we compared outcomes among infants <6 kg versus those ≥6 kg. Although our objective was to evaluate percutaneous PDA closure as a generic technique, embolization rates from device and coil arms were compared. Differences in the overall AE rates between first and second epochs were also compared. A P < .05 was considered significant for overall effect.
The flow diagram (Fig 1) summarizes the identified, screened, eligible, and included studies. The most common reason for exclusion in the full-text review was <3 infants included in the study. Interrater agreement on the inclusion/exclusion of articles was good (κ = 0.82).
Study characteristics, representing 635 infants, are summarized in Table 1. The sample sizes from the studies meeting inclusion criteria ranged from 3 to 94 patients. No RCTs comparing percutaneous PDA closure with alternative management strategies (surgical ligation, conservative management, drug therapy) were found. Included studies were highly diverse with regard to the participants, interventions, and outcome measures. Interrater agreement on the methodological quality of included articles was good (κ = 0.74). Studies ranged from 4 to 9 stars on the Newcastle-Ottawa Scale (range, 0–9; a lower score indicates methodological weakness).
Aggregate data synthesis of the included studies is shown in Table 2. Included studies reported outcomes from 18 countries, with 9 studies performed in the United States. Although 1 study reported on factors associated with length of stay and hospital charges, an economic evaluation of direct health care utilization costs, or nonmedical costs assumed by affected parties (parents, families) was not performed by any of the included studies.
Technical Success (Feasibility)
Technical success with percutaneous PDA closure was 92.2% (95% CI 88.8%–95.0%) with modest heterogeneity (I2 = 32%, P = .03; Fig 2); minor publication bias was evident (LFK Index = 1.80). Among 40 cases designated as procedural abandonments, the reasons for this included an AE (n = 15), device malposition within the aorta (n = 10), device malposition within the left pulmonary artery (n = 3), technical failure (n = 10), or unknown/undisclosed (n = 2).
Four cases (0.6%) were considered technical failures. In 1 case, the patient had successful catheter placement of coils in the PDA, which were surgically removed 3 days later due to persistent ductal shunting and hemolysis. In the remaining cases, evidence of late embolizations (>24 hours after successful placement) necessitated device (n = 1) or coil (n = 2) retrieval.
Among 28 studies reporting the incidence of residual shunting after coil or device placement, the incidence of immediate ductal occlusion after device or coil placement was 76.7% (95% CI 65.2%–83.3%) with significant heterogeneity among studies (I2 = 70%, P < .01); major publication bias was evident (LFK Index = 2.87). Among cases (n = 83) with residual ductal shunting, most (n = 68/83, 82%) subsequently closed within 24 hours. Eight (10%) cases remained patent at longest reported follow-up (range 3–36 months). Severity of residual shunting was trivial (n = 5) or was unknown/not reported (n = 3).
Overall AE rate was 23.3% (95% CI 16.5%–30.8%; Fig 3). Significant heterogeneity in AE rates was identified among studies (I2 = 82%, P < .01); mild publication bias was evident (LFK Index = 1.16). Among 140 AEs, causality was assessed as probably (n = 3), probably/likely (n = 57), or certain (n = 80). Interrater agreement on causality of AEs was good (κ = 0.86). Among AEs, most were CNS-AE (80/140, 57.1%).
The rate of CS-AE was 10.1% (95% CI 7.8%–12.5%; Fig 4), with low heterogeneity (I2 = 0%, P = .51), and evidence of mild publication bias (LFK Index = 1.44). Most CS-AEs (78.3%, 47/60) were level 3 AEs. Among all procedures, level 4 AE (major) or 5 AE (catastrophic) occurred in 1.6% (10/635) and <0.5% (3/635) of cases, respectively. Most major or catastrophic events (92.3%, 12/13) occurred among infants <6 kg. Additional details on level 4 or 5 AEs are provided in Supplemental Table 6. The prevalence of embolization (both coils and devices) was 5.0% (95% CI 3.5%–8.5%), with moderate heterogeneity (I2 = 34%; P = .02) and evidence of major publication bias (LFK Index = 2.6).
Nature of AEs
Device or Coil-Related AEs (Embolization, Malposition)
Device or coil-related complications were the most frequent AEs, occurring in 12.3% (95% CI 7.9%–17.6%). Moderate heterogeneity in device or coil-related AE rates was identified among studies (I2 = 57%, P < .01) with evidence of major publication bias (LFK Index = 2.18). We observed a higher proportion of coil than device embolizations (21/216, 9.7% vs 11/419, 2.6%; P < .01). Most coil (16/21, 76.2%) and device embolizations (8/11, 72.7%) were retrieved percutaneously. Embolizations (n = 32) were observed to the following: pulmonary arteries (n = 19), aorta (n = 5), internal/external iliac arteries (n = 1), or uncertain/not provided (n = 7). In 10 of these cases (31.3%), the implant was retrieved in the catheterization laboratory and the PDA closed by using a larger device during the same procedure. In 2 cases (Supplemental Table 6 for details), a device embolized and, despite successful percutaneous retrieval, the patients did not recover from the hemodynamic compromise and died.31,65
The rate of access-related complications was 8.2% (95% CI 5.6%–11.2%) with moderate heterogeneity (I2 = 37%, P = .37) among studies. We observed no evidence of publication bias (LFK Index = 0.57). Access-related complications were the second most frequent AEs (50/140, 35.7%), and included hematoma or transient pulse loss not requiring therapy (n = 19), pulse loss or thrombosis requiring therapy (n = 23), or blood transfusion for vascular compromise (n = 8).
Two AEs were sedation/airway-related, occurring in 0.3% (2/635) of attempted PDA closures and comprising <1% of reported AEs. One sedation/airway-related AE was the need to reposition an endotracheal tube during the catheterization,21 whereas a second was need for transient bag-and-mask ventilation for apnea during the procedure.42
General Catheterization AEs
Among 16 general catheterization AEs, 1 death was attributed to the catheterization. In this case, a 1.5-kg premature infant with multiple comorbidities had a suspected cardiac perforation and underwent emergent pericardiocentesis; however, the infant did not respond to resuscitation and died.33
Subgroup Analysis of Outcomes
Subgroup analyses were performed to explore potential sources of heterogeneity in primary outcomes (Table 3). No variables influenced the rate of technical success (feasibility). The incidence of CS-AEs was more than twofold to threefold higher among studies with infants weighing <6 kg (14.0% vs 4.8%). Many comparisons were limited by nonreporting of the variables of interest.
This study reports the largest known meta-analysis among infants treated percutaneously for PDA. In 38 studies encompassing 635 procedures, percutaneous closure was associated with 92.2% technical success, 23.3% overall AE rate, and 10.1% CS-AE rate. Although a better understanding of risks associated with percutaneous closure is an important first step, lack of comparative trials (percutaneous closure versus surgical ligation) precludes determination of the optimal treatment of PDA closure during infancy.69 Pragmatic clinical trials using strict inclusion criteria, well-defined treatment thresholds, standardized protocols for AE surveillance, and long-term follow-up are needed to generate relevant and generalizable data to develop evidence-based standards for PDA treatment during infancy.69–71 This goal is achievable, but will require a high level of interdisciplinary (neonatology, cardiology, interventional medicine) and multi-institutional collaboration.
Traditionally, PDA treatments (eg, prophylactic drug therapy) have been applied broadly, irrespective of markers of disease burden. Rather than an “all-or-none approach,” efforts to develop more individualized approaches to PDA management that take into account markers (clinical, echocardiographic) of adverse ductal sequelae and the natural history of the disease (rates of spontaneous closure) may improve outcomes.72 For example, using conservative management (fluid restriction, diuretics, positive pressure ventilation) to reduce symptoms from the PDA, recent data show that approximately two-thirds of infants spontaneously close their ductus before hospital discharge,73 thereby avoiding the risks of an unnecessary intervention, without evidence of increasing risk associated with conservative management. Targeted use of percutaneous PDA closure in the subset of infants whose ductus fails to close after conservative treatment, and who continue to show evidence of adverse ductal consequences (clinical, echocardiographic, serum biomarkers), would enable clinicians to minimize risk and yield the greatest benefits.72 In the present review, neither the primary indications for PDA closure, nor the nature and extent of management before referral for closure, were reported consistently; thus, optimal timing and thresholds for percutaneous PDA closure remain unknown and can vary greatly according to age, weight, and clinical condition.74
In adults and children with a persistent ductus warranting closure, percutaneous techniques provide clear advantages over surgical ligation and comprise the treatment of choice for PDA closures beyond the first year of life.4,11 Given growing concerns on the merits and safety of surgical ligation during infancy,75 percutaneous PDA closure represents a potentially attractive alternative. However, consistent with previous studies, higher rates of overall AEs and CS-AE were observed among a subgroup of low weight (<6 kg) infants.11 At these lower weights, providers must be careful not to trade the risks of surgical ligation for those associated with percutaneous closure without producing and examining the necessary evidence base. Early (<7 days of life) surgical PDA ligation seems to have fallen out of favor in recent years,76,77 but surgery remains an important option in the treatment of symptomatic low weight infants, particularly in centers without a dedicated pediatric team of cardiac interventionists.78,79
Consistent with previous reports, we observed that access-related injuries are frequently observed in percutaneous PDA closure during infancy.21,80 However, inconsistent reporting precluded a better understanding of the possible link between sheath size and access-related injuries. Approaches that limit or avoid arterial access, such as the use of fluoroscopy and transthoracic echocardiography to guide transvenous PDA closure, will likely reduce such complications.67
Our findings suggest that outcomes for percutaneous PDA closure have changed over time, which is likely attributable to new techniques, approaches, and available technologies. Recent device modifications to the ADO-II AS (St Jude Medical, Minneapolis, MN; not available in the United States)81 and reports on the safety and feasibility of a new, flexible, self-shaping device (Occlutech PDA occluder; Occlutech International AB, Helsingborg, Sweden; not available in the United States)82 suggest that risk/benefit profiles are likely to continue to change. Thus, we encourage investigators to document and publish their results to further the collective knowledge.
The inclusion of data from nonrandomized, noncontrolled, and retrospective studies may have introduced bias in the results. Observational studies may report outcomes in “best-case scenarios,” in which the health care providers feel personally committed to the success of an intervention; thus, reported AE rates may not accurately reflect those events encountered in clinical practice (publication bias). In the absence of therapy randomization, defining any link between percutaneous PDA closure and AEs was not feasible. Although front-line providers (pediatric cardiac interventionalists) determined the attributability and severity level of AEs based on previous criteria with strong interrater agreement, no formal certifying training was provided. One of the critical steps in remedying the gaps identified in this review is the standardization of definitions and research methodologies for AEs after cardiac catheterization.12,23
Within the meta-analysis, heterogeneity and publication bias were observed frequently, which confounded data interpretation. Marked variation in the completeness of data reported among studies limited data synthesis. Limited descriptions of patient-selection procedures, including how infants were drawn from the eligible population, likely increased the risk of selection bias among included studies.
Studies in the present review provided limited or no description of sedation- or anesthesia-related procedures; however, data showing infants to be at the greatest risk for such complications among all pediatric populations83 suggest that thoughtful consideration of optimal anesthesia and sedation practices are necessary. Although we evaluate percutaneous PDA closure among infants <1 year at time of intervention, risk/benefit ratios are likely to be continuous in nature and dependent on a number of patient- and procedural-related factors beyond age at intervention. Given the interrelatedness of health and resource utilization, lack of available data on resource use and cost associated with percutaneous PDA closure is noteworthy.
It is possible that relevant published peer-reviewed evidence was not identified, and disagreements about whether specific articles should have been included may be reasonable. To minimize this risk, we performed a sensitive literature search with assistance from a research librarian, by using a diverse set of databases without language restrictions. Although percutaneous PDA closure may be feasible in some centers, broad generalizability has yet to be demonstrated.
Percutaneous PDA closure during infancy is feasible and is associated with few major or catastrophic AEs; however, the absence of high-quality studies and significant heterogeneity for main outcomes limits interpretability and generalizability of current findings. Large, pragmatic, multicenter studies that systematically evaluate existing PDA treatments (percutaneous closure, surgical ligation) are needed to address the fundamental gaps in knowledge documented by this review.
The authors recognize and thank Dr Michael Borenstein, PhD, a member of the Development Team for Comprehensive Meta-Analysis software and author of Introduction to Meta-Analysis, for his willingness to provide statistical consultation to the current study. The authors also thank Alison Gehred, MS, clinical librarian at Nationwide Children’s Hospital for her assistance in conducting the database searches to locate articles for possible inclusion.
- Accepted October 31, 2016.
- Address correspondence to Carl Backes, MD, Center for Perinatal Research, Nationwide Children’s Hospital, 700 Children’s Dr, Columbus, OH 43205. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: Dr Hijazi is a consultant for NuMED and Occlutech, and has ownership/partnership of the Colibri Heart Valve. He is also chairman of the PICS Foundation. Dr Armstrong reports the following potential conflicts of interest: Medtronic Inc, research grants; Edwards Lifesciences, consultant, proctor, research grant; Siemens Healthcare AX, consultant; St Jude Medical, consultant, proctor, research grant; B. Braun Interventional Systems Inc, proctor; and Pfm Medical, Inc, research grant. Dr Justino is a consultant for St Jude Medical, B-Braun Interventional Systems, and Janssen Pharmaceutical. Dr Bergersen is a consultant for 480 Biomedical. The other authors have indicated they have no potential conflicts of interest to disclose.
- Sellmer A,
- Bjerre JV,
- Schmidt MR, et al
- Evans N,
- Kluckow M
- Oh W,
- Poindexter BB,
- Perritt R, et al; Neonatal Research Network
- Schneider DJ,
- Moore JW
- Moore JW,
- Vincent RN,
- Beekman RH III, et al; NCDR IMPACT Steering Committee
- Pass RH,
- Hijazi Z,
- Hsu DT,
- Lewis V,
- Hellenbrand WE
- El-Said HG,
- Bratincsak A,
- Foerster SR, et al
- O’Connor D,
- Green S,
- Higgins J
- Balk E,
- Chung M,
- Hadar N
- Wells G
- Higgins J,
- Green S
- Backes CH,
- Cheatham SL,
- Deyo GM, et al
- Bergersen L,
- Gauvreau K,
- Marshall A, et al
- World Health Organization–Uppsala Monitoring Center
- Barendregt JJ,
- Doi SA,
- Lee YY,
- Norman RE,
- Vos T
- Higgins JP,
- Thompson SG,
- Deeks JJ,
- Altman DG
- Barendregt JJ,
- Doi SA
- Dimas VV,
- Takao C,
- Ing FF, et al
- Butera G,
- De Rosa G,
- Chessa M, et al
- Fischer G,
- Stieh J,
- Uebing A,
- Grabitz R,
- Kramer HH
- Francis E,
- Singhi AK,
- Lakshmivenkateshaiah S,
- Kumar RK
- Hijazi ZM,
- Geggel RL
- Masura J,
- Walsh KP,
- Thanopoulous B, et al
- Owada CY,
- Teitel DF,
- Moore P
- Roberts P,
- Adwani S,
- Archer N,
- Wilson N
- Thanopoulos BD,
- Hakim FA,
- Hiari A, et al
- Benitz WE; Committee on Fetus and Newborn, American Academy of Pediatrics
- Bose CL,
- Laughon MM
- McNamara PJ,
- Sehgal A
- Rolland A,
- Shankar-Aguilera S,
- Diomandé D,
- Zupan-Simunek V,
- Boileau P
- Chorne N,
- Leonard C,
- Piecuch R,
- Clyman RI
- Brooks JM,
- Travadi JN,
- Patole SK,
- Doherty DA,
- Simmer K
- Brotschi B,
- Hug MI,
- Kretschmar O,
- Rizzi M,
- Albisetti M
- Copyright © 2017 by the American Academy of Pediatrics