Journal of Pediatric Surgery (2013) 48, 2408–2415
www.elsevier.com/locate/jpedsurg
Standardized reporting for congenital diaphragmatic hernia – An international consensus☆ Kevin P. Lally a,⁎,1 , Robert E. Lasky a,1 , Pamela A Lally a,1 , Pietro Bagolan b,1 , Carl F. Davis c,1 , Bjorn P. Frenckner d,1 , Ronald M. Hirschl e,1 , Max R. Langham f,1 , Terry L. Buchmiller g,1 , Noriaki Usui h,1 , Dick Tibboel i,1 , Jay M. Wilson g,1 The Congenital Diaphragmatic Hernia Study Group 2 a
UT Health Medical School and Children's Memorial Hermann Hospital, Houston, TX, USA Bambino Gesu Children's Hospital, Rome, Italy c Royal Hospital for Sick Children, Glasgow, Scotland d Astrid Lindgren Children's Hospital, Stockholm, Sweden e University of Michigan, Ann Arbor, MI, USA f LeBonheur Children's Hospital, Memphis, TN, USA g Children's Hospital Boston, Boston, MA, USA h Osaka University Graduate School of Medicine, Osaka, Japan i Sophia Children's Hospital, Rotterdam b
Received 16 August 2013; accepted 26 August 2013
Key words: Congenital diaphragmatic hernia (CDH); Apgar score; Risk stratification; Staging system
Abstract Background/purpose: Congenital diaphragmatic hernia (CDH) remains a significant cause of neonatal death. A wide spectrum of disease severity and treatment strategies makes comparisons challenging. The objective of this study was to create a standardized reporting system for CDH. Methods: Data were prospectively collected on all live born infants with CDH from 51 centers in 9 countries. Patients who underwent surgical correction had the diaphragmatic defect size graded (A–D) using a standardized system. Other data known to affect outcome were combined to create a usable staging system. The primary outcome was death or hospital discharge. Results: A total of 1,975 infants were evaluated. A total of 326 infants were not repaired, and all died. Of the remaining 1,649, the defect was scored in 1,638 patients. A small defect (A) had a high survival, while a large defect was much worse. Cardiac defects significantly worsened outcome. We grouped patients into 6 categories based on defect size with an isolated A defect as stage I. A major cardiac anomaly (+) placed the patient in the next higher stage. Applying this, patient survival is 99% for stage I, 96% stage II, 78% stage III, 58% stage IV, 39% stage V, and 0% for non-repair.
☆ All assume responsibility for the content and integrity of this article. ⁎ Corresponding author. Department of Pediatric Surgery, UT Health Medical School, 6431 Fannin, Suite 5.258, Houston TX, 77030, USA. Tel.: + 1 713 500 7300; fax: +1 713 500 7296. E-mail address: [email protected] (K.P. Lally). 1 The Members of the Writing Group. 2 Members of the Congenital Diaphragmatic Hernia Study Group are listed in the Appendix.
0022-3468/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpedsurg.2013.08.014
Standardized reporting for CDH
2409
Conclusions: The size of the diaphragmatic defect and a severe cardiac anomaly are strongly associated with outcome. Standardizing reporting is imperative in determining optimal outcomes and effective therapies for CDH and could serve as a benchmark for prospective trials. © 2013 Elsevier Inc. All rights reserved.
Congenital diaphragmatic hernia (CDH) is estimated to occur in approximately 1 in every 3500 live births [1]. While overall survival has improved in the past 2 decades, the mortality for CDH is still over 30% [2,3]. The improvement in survival is thought, in part, to be owing to avoiding iatrogenic lung injury. However, advanced therapies including extracorporeal membrane oxygenation (ECMO), inhaled nitric oxide (iNO) and high frequency oscillation (HFOV) have all been widely used in these patients [4–6]. The combination of the disease severity and widespread use of these costly therapies makes CDH one of the most expensive and deadly neonatal conditions [7]. Evaluation of optimal therapies in patients with CDH is complicated by the wide range of therapies utilized and the small numbers of patients, even in higher volume institutions. Effective evaluation of treatment strategies requires the ability to compare risk adjusted results from different institutions that may have different patient populations. The complexity of patients and range of treatments offered between institutions caring for CDH infants is quite large. Treatments include ECMO, iNO, HFOV and fetal interventions including placement of occlusive tracheal balloons. Currently each of these expensive and risky therapies is widely used in the treatment of CDH. Risk adjustment has proven useful in other rare childhood conditions such as neuroblastoma, retinoblastoma and most other malignancies, and is now being used to evaluate results after correction of complex cardiac anomalies in children [8]. The importance of risk adjustment in CDH has been recognized for some time [9–13]. Creating risk adjustment to allow for comparison of different treatments however, has proven difficult. Many risk stratification systems were designed in an attempt to predict patient outcome early in the hospital course and not to evaluate therapies [9–11]. Others have identified factors associated with worse outcomes including the size of the diaphragm defect, associated anomalies, and prematurity [12–16]. The goal of this study is to use contemporary results from a large international database to develop a risk stratified reporting system. Accomplishing this goal should allow comparison of results for patients undergoing a wide range of treatments offered at international institutions caring for these infants.
1. Methods The congenital diaphragmatic hernia study group (CDHSG) was founded in 1995 to prospectively collect
data on infants with CDH [17]. This is a voluntary registry collecting data on all live born infants with CDH from a participating institution. Data are collected until death or discharge/transfer from the hospital. The data collected include demographic information as well as treatment information and outcomes. Periodic evaluations of data collection forms have resulted in successive revisions of the data collection form. As defect size is a known important risk factor, in 2006, members of the study group agreed upon a standardized classification scheme for defect size. Defects were coded from A to D as shown in Fig. 1. “A” defects were entirely surrounded by muscle, “B” defects had a small (b 50%) and “C” defects a large (N 50%) portion of the chest wall devoid of diaphragm tissue, and “D” patients had complete or near complete absence of the diaphragm. Patients who did not undergo repair were included in the registry and their results recorded. Cardiac anomalies were characterized as major or minor. Major cardiac anomalies included coarctation of the aorta, combined atrial and ventricular septal defects, hypoplastic left heart syndrome, tetralogy of Fallot, double outlet right ventricle, atrioventricular canal defect as well as other hemodynamically significant defects. Minor anomalies were either isolated atrial or ventricular septal defects or hemodynamically non-significant lesions (such as aberrant vessels). Chromosomal anomalies were recorded. Patients with Fryns syndrome were included in the chromosomal anomaly group. In order to create a simple system, birth weight, gestational age and Apgar scores were evaluated as categorical variables. Cutoffs were b 2.5 kg (birth weight), b 37 weeks (gestational age) and b 5 (Apgar score). Major other anomalies included pulmonary (such as cystic adenomatoid malformation), gastro-intestinal (tracheo-esophageal fistula, esophageal and intestinal atresia, imperforate anus, omphalocele), and neurological (agenesis of the corpus callosum, ventriculomegaly, Dandy-Walker syndrome). Data analyzed for this study included patients born from January 1, 2007 through December 31, 2011. Data were collected until October 2012 to maximize inclusion of patients born during the study period. Unfortunately, reliable autopsy data were not available in many patients.
1.1. Statistics The primary outcome variable used to validate the risk score was death prior to hospital discharge. Descriptive statistics of the demographic and clinical variables recorded were calculated for the surgically repaired infants whose defect was scored for the 4 subgroups of these infants defined
2410
Fig. 1
K.P. Lally et al.
Defect diagram. The diagram is drawn with the diaphragm (defect) on the patient's left looking up from the abdomen towards the chest.
by defect size, and for the 256 infants who were not surgically repaired. Logistic regressions were calculated to evaluate whether the 4 defect sizes recorded were associated with any of the demographic and clinical variables collected. We hypothesized that the defect size variable was the risk factor most strongly associated with death. Each of the other risk factors was then included with defect size in a multiple binary logistic regression to determine whether the prediction of death improved significantly with both variables considered. Evidence for an improved model fit was determined by the magnitude of the decrease in the Bayesian Information Criterion (BIC) statistic of the combined model versus a model that included defect size as the sole covariate in the model. A smaller BIC statistic indicates a better model. We used Raftery's criteria: a combined model that decreased the BIC by 2–6 was positive evidence for improvement, 6– 10 strong evidence, and N 10 very strong evidence [18]. All risk factors associated with positive evidence for improvement or better were to be included in a final binary multiple logistic regression. The risk factors included in the final model were used to calculate a 5 category risk factor that can be easily scored. The BIC statistics of the final multiple binary logistic regression and a binary logistic regression with only the simple clinical risk score were then compared following Raftery's criteria to ensure the risk score adequately captured the information in the component risk factors.
Receiver operator curves (ROC) and area under the curve (AUC) were calculated to summarize the classification of death by the regression models. All analyses were conducted using Stata (version 12.0, College Station, TX). The CDHSG registry was approved by the University of Texas–Houston Institutional Review Board (#HSC-MS-03223). Participating centers filed a waiver of consent for data submission or signed a data use agreement for a limited data set. Center identification was kept anonymous.
2. Results A total of 2,296 patients were entered into the registry during the time period of this study. These patients were cared for in 53 participating centers. Patients treated at 2 centers were excluded as greater than 40% of their patients did not have a coded defect, leaving 51 participating centers from 9 countries. Additionally excluded were 41 who had either bilateral or Morgagni type defects, and 2 patients who were transferred from the participating center before repair. Overall survival for the remaining 1,975 patients was 72% (Fig. 2). All 326 patients (16%) who never had surgical correction died. The rate of non-repair ranged from 0 to 44% in the various centers. Survival to discharge for the 1,649
Standardized reporting for CDH
2411
Fig. 2
Patients comprising the study.
patients who underwent surgical repair was 85%. Demographic data are shown in Table 1 along with differences between the operated and non-operated patients. Several variables were associated with poor outcome. The size of the diaphragm defect was clearly the most significant risk factor affecting outcome (BIC statistic = 26). All patients with an “A” defect had a primary repair and all patients with a “D” defect had a patch repair of the defect. Other factors that influence survival were evaluated to determine if they had an added effect when included with the size of the diaphragm defect in a multiple regression model. By Raftery's criteria, only 2 risk factors improved model fit, a major cardiac anomaly (considering minor cardiac anomalies separately did not improve prediction) and the 1 minute Apgar ≤ 4 (5 minute Apgar also improved model fit but was not included in the table because of the correlation with the 1 minute Apgar). While the Apgar scores at 1 and 5 minutes were significant, 7% (104/1585) of infants did not have a record of any Apgar score. This unscored group represented the most severely ill infants with a mortality of 39% as compared with a mortality of 28% among infants receiving one or more Apgar scores. In addition, 85% (200/236) of infants intubated by 1 minute of life received 1 minute Apgar scores and 91% (719/793) of infants intubated by 5 minutes of life received 5 minute scores such that active therapy confounded the ability to accurately assess respiratory effort and color. With prenatal diagnosis and earlier intervention, the Apgar score is becoming less reliable as a measure to
direct care. Because of these factors, the Apgar score was excluded from the final model. Patients were grouped into a reporting system as shown in Table 2. Stage I are infants who have an isolated “A” size defect. For each increase in defect size the stage increases by 1. The presence of a major cardiac anomaly (defect +), increases the stage by 1. The probability for death using this characterization is shown in Table 3 when compared to patients with an isolated “A” defect (stage I). Using these data, a 6 component reporting system is proposed. The addition of a major cardiac anomaly puts the patient in the next stage. The proposed staging and reporting system and actual survival of patients characterized by defect or defect + is shown in Table 2.
3. Discussion Congenital diaphragmatic hernia remains a challenging problem. Mortality remains high and while survival has improved in the past decade, morbidity in the survivors is substantial [19–22]. Because the anomaly occurs infrequently, very few centers treat large numbers of patients. In the past 20 years, there has been a rapid expansion in the number of therapies available for these infants. These include fetal intervention, ECMO, iNO, surfactant, HFOV and numerous pulmonary vasodilators. Given the small numbers of patients per center and the large number of hitherto evidence based
2412 Table 1
K.P. Lally et al. Demographic and clinical characteristics of the study infants successfully scored.
Characteristic
Defect size
Male sex Prenatal diagnosis Defect on left side Major cardiac anomaly Chromosomal anomaly Other anomaly Inborn Gestational age (weeks) (mean ± SD) Birth weight (kg) (mean ± SD
Not repaired
A n = 218
B n = 716
C n = 495
D n = 209
Total n = 1 638
n = 326
124 # [57] (172) 82 [38] (217) 197 [90] (218) 13 [6] (218) 7 [3] (218) 4 [2] (218) 64 [29] (218) 38.1 ± 2.0 (217) 3.1 ± 0.6 (217)
442 [62] (716) 390 [55] (713)*** 624 [87] (716) 27 [4] (716) 20 [3] (716) 10 [1] (716) 262 [37] (715) 37.9 ± 2.0 (712) 3.1 ± 0.6 (716)
316 [64] (495) 362 [73] (495)*** 384 [78] (495)*** 31 [6] (495) 15 [3] (495) 14 [3] (495) 251 [51] (495)*** 37.5 ± 2.2 (491)** 3.0 ± 0.6 (494)**
136 [65] (209) 177 [85] (209)*** 170 [81] (209)** 20 [10] (209) 9 [4] (209) 9 [4] (209) 103 [49] (209)*** 37.5 ± 2.2 (205)** 2.9 ± 0.6 (207)***
1 018 [62] (1 638) 1 011 [62] (1 634) 1 375 [84] (1 638) 91 [6] (1 638) 51 [3] (1 638) 37 [2] (1 638) 680 [42] (1 637) 37.8 ± 2.1 (1 625) 3.0 ± 0.6 (1 634)
191 [59] (326) 260 [80] (326)††† 261 [80] (326) 60 [18] (326)††† 54 [17] (326)††† 44 [14] (326)††† 188 [58] (326)†† 36.2 ± 3.5 (324)††† 2.6 ± 0.8 (321)†††
Note: Columns A through D do not include 11 patients who were repaired but for whom no defect size was documented. * (p b 0.05), ** (p b 0.01), *** (p b 0.001) indicate whether infants in defect size categories (B, C, or D) significantly differed from infants in defect size category A with respect to this characteristic as evaluated by an appropriate regression model (linear or binary logistic). † (p b 0.05), †† (p b 0.01), ††† (p b 0.001) indicate a significant difference between infants repaired and not repaired with respect to this characteristic as evaluated by an appropriate regression model (linear or binary logistic). # Represents number of infants with characteristic, % with characteristic in brackets and total number of infants with data for the characteristic in parentheses.
unproven therapies available, interpretation of which therapies do in fact have real benefit becomes very difficult. Additionally, most of these therapies were adopted without well designed clinical trials. While not unique to CDH, the use of data from small single institution studies may result in widespread adoption of therapies with little benefit, or even real harm. The rapid adoption of hyperventilation for infants with CDH, which has since been abandoned, is a clear example of this problem [5,23,24]. Some recent improvements have been demonstrated with a consensus based protocol in Europe [25]. One important strength of the current study is that it examines a large number of neonates treated during a 5 year time span when there were few significant new therapies available that would have impacted survival. Table 2
These data demonstrate a wide range of severity amongst infants with CDH. In order to accurately evaluate results of treatment for symptomatic newborns with CDH we assert that reporting of defect size, presence or absence of a major cardiac anomaly and inclusion of patients who die without repair is necessary. While we and other authors have suggested other schemes to determine disease severity, these are meant for use early in life. These various predicting schema have not been reported consistently and no current standard for outcome reporting has been adopted. The previous study from the CDHSG attempted to predict outcomes early in the course using a logistic regression equation. It is important to emphasize that the staging system proposed here is intended to serve as a standard for reporting and not as a predictor for individual patients.
Proposed staging system and survival.
Stage
Defect
I II II III III IV IV V
A A B B C C D D
Major cardiac anomaly + + + +
N
Died
Survived
Group survival
Stage survival
164 8 572 18 372 27 144 18
1 1 21 6 81 12 60 11
163 7 551 12 291 15 84 7
99% 88% 96% 67% 78% 56% 58% 39%
99% 96% 78% 58% 39%
Standardized reporting for CDH Ackerman and colleagues have recently proposed a defect classification of CDH for research purposes, but this system is too complex for a simplified staging system [26]. Defect size alone is considered a surrogate by proxy for pulmonary hypoplasia, but there are no validated clinically available methods to directly determine pulmonary hypoplasia. Chest radiographs are of no value in predicting severity of illness in infants with CDH [27]. Ultrasound and MRI have been used to develop prenatal estimates of pulmonary hypoplasia but these measurements have not been studied in a rigorous, multi-institutional fashion or validated against postnatal anatomic information [28,29]. While many of our centers use potentially useful perinatal markers such as MRI calculated lung volumes, standardized methods to determine volumes have not yet been validated across multiple centers and may not be generalizable [30]. This is of particular importance as it makes interpretation of the benefits of fetal intervention in CDH extremely difficult. There are several limitations to this study. The data are derived from a voluntary registry. While there are agreed upon definitions, individual data entry by a large number of individuals can lead to error. The registry is designed as an observational dataset and postnatal management is not standardized between centers. It will be important to validate the proposed staging system prospectively on patients not included in the data set used to derive the results. Finally, the proposed staging system is not applicable to prenatal risk stratification and provides no guidance to those physicians performing prenatal counseling and future parents as to the optimal treatment before delivery of the affected fetus. This report does not attempt to identify surrogate clinical markers for disease severity that can direct therapy in an individual patient in a clinical setting. Rather, it is designed to allow comparison of results between centers, or to evaluate particular therapies or bundles of care in a risk stratified manner. This dataset demonstrates a very variable rate of non-repair by center. The database alone does not provide an explanation for this. Efforts to investigate this phenomenon are presently being planned using alternative research strategies.
4. Conclusion In conclusion, patients with CDH can be effectively stratified into stages based on readily available clinical information. Given the wide variation in therapies for CDH and the broad range of patients seen in different institutions, we believe it is important that all clinical reports of patients with CDH be reported by stage and include the patients who do not receive surgical therapy. It is clear that using current therapy, mortality is low for infants with stage I and II lesions. Our group plans to focus on high stage patients (III– V and non-repair) who currently represent almost all of the deaths from CDH. It is hoped that this staging system will
2413 thereby allow our group and others to better determine best therapy for high risk infants with CDH.
Appendix Participating CDHSG centers. Hospital
City
State or Country province
Arkansas Children's Little Rock Hospital Astrid Lindgren Stockholm Children's Hospital, Q3:03 BC Children's & Vancouver Women's Health Centre Children's Hospital Boston Boston Children's Hospital of Akron Akron Children's Hospital of Peoria Illinois Children's Hospital of Los Los Angeles Angeles Children's Hospital of Milwaukee Wisconsin Children's Hospital Omaha Omaha Children's Hospitals and Minneapolis Clinics (Minneapolis) Children's Memorial Houston Hermann Hospital Duke University Medical Durham Center Emory University Atlanta Georgia Health Sciences Augusta University Golisano Children’s Rochester Hospital at Strong IRCCS Fondazione Ca' Milano Granda Ospedale Maggiore Policlinico Kosair Children's Hospital Louisville Le Bonheur Children’s Memphis Medical Center Legacy Emanuel Portland Children's Hospital Loma Linda University Loma Linda Children's Hospital Lucile Salter Packard Palo Alto Children's Hospital Mattel Children's Hospital Los Angeles at UCLA Miami Valley Hospital Dayton National Center for Child Tokyo Health and Development
AR Sweden
BC
Canada
MA OH IL CA WI NE MN TX NC GA GA NY Italy
KY TN OR CA CA CA OH Japan
(continued on next page)
2414
K.P. Lally et al.
(continued) Hospital
References City
State or Country province
Oespedale Pediatrico Rome Bambino Gesu Oespedale Riuniti Bergamo Bergamo Osaka University Suita Osaka Graduate School of Medicine Palmetto Health Richland Columbia SC Primary Children's Salt Lake City UT Hospital Radboud University Nijmegen Nijmegen Medical Centre Columbus OH Research Institute at Nationwide Children’s Hospital Royal Children's Hospital Parkville Victoria Royal Hospital for Sick Glasgow Yorkhill Children San Diego Children's San Diego CA Hospital Sophia Children's Rotterdam Hospital St. Francis Children's Tulsa OK Hospital St. Joseph's Hospital and Phoenix AZ Medical Center St. Louis Children's St. Louis MO Hospital Stollery Children's Edmonton Alberta Hospital Sydney Children's Randwick NSW Hospital The Children's Hospital of Birmingham AL Alabama The Hospital for Sick Toronto Ontario Children University of Michigan, Ann Arbor MI C.S. Mott Children's Hospital University of Nebraska Omaha NE Medical Center University of Padua Padua University of Texas Galveston TX Medical Branch at Galveston University of Virginia Charlottesville VA Medical School Nashville TN Vanderbilt Children's Hospital Vladivostok State Medical Vladivostok University Winnie Palmer Hospital Orlando FL for Women & Babies Yale New Haven New Haven CT Children's Hospital
Italy Italy Japan
The Netherlands
Australia Scotland
The Netherlands
Canada Australia
Canada
Italy
Russia
[1] Wenstrom KD, Weiner CP, Janson JW. A five-year statewide experience with congenital diaphragmatic hernia. Am J Obstet Gynecol 1991;165:838-42. [2] Langham MR, Kays DW, Ledbetter DJ, et al. Congenital diaphragmatic hernia: epidemiology and outcome. Clin Perinatol 1996;23: 671-88. [3] Downard CD, Jaksic T, Garza JJ, et al. Analysis of an improved survival rate for congenital diaphragmatic hernia. J Pediatr Surg 2003;38:729-32. [4] The Congenital Diaphragmatic Hernia Study Group. Treatment Evolution in High Risk Congenital Diaphragmatic Hernia: Ten Years Experience. Ann Surg 2006;244:505-13. [5] Boloker J, Bateman DA, Wung J-T, et al. Congenital diaphragmatic hernia in 120 infants treated consecutively with permissive hypercapnia/spontaneous respiration/elective repair. J Pediatr Surg 2002;37: 357-66. [6] Dassinger MS, Copeland DR, Gossett J, et al. Early repair of congenital diaphragmatic hernia on extracorporeal membrane oxygenation. J Pediatr Surg 2010;45:693-7. [7] Raval MH, Wang X, Reynolds M, et al. Costs of congenital diaphragmatic hernia repair in the United States – extracorporeal membrane oxygenation foots the bill. J Pediatr Surg 2011;46:617-24. [8] Jacobs JP, O’Brien SM, Pasquali SK, et al. Variation in outcomes for benchmark operations: an analysis of the Society of Thoracic Surgeons Congenital Heart Surgery Database. Ann Thorac Surg 2011;92: 2184-91. [9] Baird R, MacNab YC, Skarsgard ED, et al. Mortality prediction in congenital diaphragmatic hernia. J Pediatr Surg 2008;43:783-7. [10] Hoffman SB, Massaro AN, Gingalewski C, et al. Survival in congenital diaphragmatic hernia: use of predictive equations in the ECMO population. Neonatology 2011;99:258-65. [11] The Congenital Diaphragmatic Hernia Study Group. Estimating disease severity of congenital diaphragmatic hernia in the first 5 minutes of life. J Pediatr Surg 2001;36:141-6. [12] The Congenital Diaphragmatic Hernia Study Group. Defect Size Determines Survival in Congenital Diaphragmatic Hernia. Pediatrics 2007;121:e651-7. [13] Valfrè L, Braguglia A, Conforti A, et al. Long term follow-up in highrisk congenital diaphragmatic hernia survivors: patching the diaphragm affects the outcome. J Pediatr Surg 2011;46:52-6. [14] Graziano JN, for the Congenital Diaphragmatic Hernia Study Group. Cardiac anomalies in patients with congenital diaphragmatic hernia and their prognosis: a report from the Congenital Diaphragmatic Hernia Study Group. J Pediatr Surg 2005;40:1045-50. [15] Tsao K, Allison ND, Harting MT, et al. Congenital diaphragmatic hernia in the preterm infant. Surgery 2010;148:404-10. [16] Colvin J, Bower C, Dickinson JE, et al. Outcomes of congenital diaphragmatic hernia: a population-based study in Western Australia. Pediatrics 2005;116:e356-63. [17] Tsao K, Lally KP. The Congenital Diaphragmatic Hernia Study Group: a voluntary international registry. Semin Pediatr Surg 2008;17:90-7. [18] Raftery A. Bayesian model selection in social research. In: Marsden P, editor. Sociological Methodology, Vol. 25. Oxford: Blackwell; 1995. p. 111-63. [19] Danzer E, Gerdes M, Bernbaum J, et al. Neurodevelopmental outcome of infants with congenital diaphragmatic hernia prospectively enrolled in an interdisciplinary follow-up program. J Pediatr Surg 2010;45: 1759-66. [20] Friedman S, Chen C, Chapman JS, et al. Neurodevelopmental outcomes of congenital diaphragmatic hernia survivors followed in a multidisciplinary clinic at ages 1 and 3. J Pediatr Surg 2008;43:1035-43. [21] Cortes RA, Keller RL, Townsend T, et al. Survival of severe congenital diaphragmatic hernia has morbid consequences. J Pediatr Surg 2005;40:36-45.
Standardized reporting for CDH [22] van den Hout L, Reiss I, Felix JF, et al. Risk factors for chronic lung disease and mortality in newborns with congenital diaphragmatic hernia. Neonatology 2010;98:370-80. [23] Drummond WH, Gregory GA, Heymann MA, et al. The independent effects of hyperventilation, tolazoline, and dopamine on infants with persistent pulmonary hypertension. J Pediatr 1981;98:603-11. [24] Kays DW, Langham Jr MR, Ledbetter DJ, et al. Detrimental effects of standard medical therapy in congenital diaphragmatic hernia. Ann Surg 1999;230:340-8. [25] van den Hout L, Schaible T, Cohen-Overbeek TE, et al. Actual outcome in infants with congenital diaphragmatic hernia: the role of a standardized postnatal treatment protocol. Fetal Diagn Ther 2011;29:55-63. [26] Ackerman KG, Vargas SO, Wilson JM, et al. Congenital Diaphragmatic Defects: Proposal for a New Classification Based on Observations in 234 Patients. Pediatr Dev Pathol 2012;15:265-74.
2415 [27] Holt PD, Arkovitz MS, Berdon WE, et al. Newborns with diaphragmatic hernia: initial chest radiography does not have a role in predicting clinical outcome. Pediatr Radiol 2004;34:462-4. [28] Mullassery D, Ba'ath ME, Jesudason EC, et al. Value of liver herniation in prediction of outcome in fetal congenital diaphragmatic hernia: a systematic review and meta-analysis. Ultrasound Obstet Gynecol 2010;35:609-14. [29] Arkovitz MS, Russo M, Devine P, et al. Fetal lung-head ratio is not related to outcome for antenatal diagnosed congenital diaphragmatic hernia. J Pediatr Surg 2007;42:107-10. [30] Lee TC, Lim FY, Keswani SG, et al. Late gestation fetal magnetic resonance imaging-derived total lung volume predicts postnatal survival and need for extracorporeal membrane oxygenation support in isolated congenital diaphragmatic hernia. J Pediatr Surg 2011;46: 1165-71.
www.elsevier.com/locate/jpedsurg
Standardized reporting for congenital diaphragmatic hernia – An international consensus☆ Kevin P. Lally a,⁎,1 , Robert E. Lasky a,1 , Pamela A Lally a,1 , Pietro Bagolan b,1 , Carl F. Davis c,1 , Bjorn P. Frenckner d,1 , Ronald M. Hirschl e,1 , Max R. Langham f,1 , Terry L. Buchmiller g,1 , Noriaki Usui h,1 , Dick Tibboel i,1 , Jay M. Wilson g,1 The Congenital Diaphragmatic Hernia Study Group 2 a
UT Health Medical School and Children's Memorial Hermann Hospital, Houston, TX, USA Bambino Gesu Children's Hospital, Rome, Italy c Royal Hospital for Sick Children, Glasgow, Scotland d Astrid Lindgren Children's Hospital, Stockholm, Sweden e University of Michigan, Ann Arbor, MI, USA f LeBonheur Children's Hospital, Memphis, TN, USA g Children's Hospital Boston, Boston, MA, USA h Osaka University Graduate School of Medicine, Osaka, Japan i Sophia Children's Hospital, Rotterdam b
Received 16 August 2013; accepted 26 August 2013
Key words: Congenital diaphragmatic hernia (CDH); Apgar score; Risk stratification; Staging system
Abstract Background/purpose: Congenital diaphragmatic hernia (CDH) remains a significant cause of neonatal death. A wide spectrum of disease severity and treatment strategies makes comparisons challenging. The objective of this study was to create a standardized reporting system for CDH. Methods: Data were prospectively collected on all live born infants with CDH from 51 centers in 9 countries. Patients who underwent surgical correction had the diaphragmatic defect size graded (A–D) using a standardized system. Other data known to affect outcome were combined to create a usable staging system. The primary outcome was death or hospital discharge. Results: A total of 1,975 infants were evaluated. A total of 326 infants were not repaired, and all died. Of the remaining 1,649, the defect was scored in 1,638 patients. A small defect (A) had a high survival, while a large defect was much worse. Cardiac defects significantly worsened outcome. We grouped patients into 6 categories based on defect size with an isolated A defect as stage I. A major cardiac anomaly (+) placed the patient in the next higher stage. Applying this, patient survival is 99% for stage I, 96% stage II, 78% stage III, 58% stage IV, 39% stage V, and 0% for non-repair.
☆ All assume responsibility for the content and integrity of this article. ⁎ Corresponding author. Department of Pediatric Surgery, UT Health Medical School, 6431 Fannin, Suite 5.258, Houston TX, 77030, USA. Tel.: + 1 713 500 7300; fax: +1 713 500 7296. E-mail address: [email protected] (K.P. Lally). 1 The Members of the Writing Group. 2 Members of the Congenital Diaphragmatic Hernia Study Group are listed in the Appendix.
0022-3468/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpedsurg.2013.08.014
Standardized reporting for CDH
2409
Conclusions: The size of the diaphragmatic defect and a severe cardiac anomaly are strongly associated with outcome. Standardizing reporting is imperative in determining optimal outcomes and effective therapies for CDH and could serve as a benchmark for prospective trials. © 2013 Elsevier Inc. All rights reserved.
Congenital diaphragmatic hernia (CDH) is estimated to occur in approximately 1 in every 3500 live births [1]. While overall survival has improved in the past 2 decades, the mortality for CDH is still over 30% [2,3]. The improvement in survival is thought, in part, to be owing to avoiding iatrogenic lung injury. However, advanced therapies including extracorporeal membrane oxygenation (ECMO), inhaled nitric oxide (iNO) and high frequency oscillation (HFOV) have all been widely used in these patients [4–6]. The combination of the disease severity and widespread use of these costly therapies makes CDH one of the most expensive and deadly neonatal conditions [7]. Evaluation of optimal therapies in patients with CDH is complicated by the wide range of therapies utilized and the small numbers of patients, even in higher volume institutions. Effective evaluation of treatment strategies requires the ability to compare risk adjusted results from different institutions that may have different patient populations. The complexity of patients and range of treatments offered between institutions caring for CDH infants is quite large. Treatments include ECMO, iNO, HFOV and fetal interventions including placement of occlusive tracheal balloons. Currently each of these expensive and risky therapies is widely used in the treatment of CDH. Risk adjustment has proven useful in other rare childhood conditions such as neuroblastoma, retinoblastoma and most other malignancies, and is now being used to evaluate results after correction of complex cardiac anomalies in children [8]. The importance of risk adjustment in CDH has been recognized for some time [9–13]. Creating risk adjustment to allow for comparison of different treatments however, has proven difficult. Many risk stratification systems were designed in an attempt to predict patient outcome early in the hospital course and not to evaluate therapies [9–11]. Others have identified factors associated with worse outcomes including the size of the diaphragm defect, associated anomalies, and prematurity [12–16]. The goal of this study is to use contemporary results from a large international database to develop a risk stratified reporting system. Accomplishing this goal should allow comparison of results for patients undergoing a wide range of treatments offered at international institutions caring for these infants.
1. Methods The congenital diaphragmatic hernia study group (CDHSG) was founded in 1995 to prospectively collect
data on infants with CDH [17]. This is a voluntary registry collecting data on all live born infants with CDH from a participating institution. Data are collected until death or discharge/transfer from the hospital. The data collected include demographic information as well as treatment information and outcomes. Periodic evaluations of data collection forms have resulted in successive revisions of the data collection form. As defect size is a known important risk factor, in 2006, members of the study group agreed upon a standardized classification scheme for defect size. Defects were coded from A to D as shown in Fig. 1. “A” defects were entirely surrounded by muscle, “B” defects had a small (b 50%) and “C” defects a large (N 50%) portion of the chest wall devoid of diaphragm tissue, and “D” patients had complete or near complete absence of the diaphragm. Patients who did not undergo repair were included in the registry and their results recorded. Cardiac anomalies were characterized as major or minor. Major cardiac anomalies included coarctation of the aorta, combined atrial and ventricular septal defects, hypoplastic left heart syndrome, tetralogy of Fallot, double outlet right ventricle, atrioventricular canal defect as well as other hemodynamically significant defects. Minor anomalies were either isolated atrial or ventricular septal defects or hemodynamically non-significant lesions (such as aberrant vessels). Chromosomal anomalies were recorded. Patients with Fryns syndrome were included in the chromosomal anomaly group. In order to create a simple system, birth weight, gestational age and Apgar scores were evaluated as categorical variables. Cutoffs were b 2.5 kg (birth weight), b 37 weeks (gestational age) and b 5 (Apgar score). Major other anomalies included pulmonary (such as cystic adenomatoid malformation), gastro-intestinal (tracheo-esophageal fistula, esophageal and intestinal atresia, imperforate anus, omphalocele), and neurological (agenesis of the corpus callosum, ventriculomegaly, Dandy-Walker syndrome). Data analyzed for this study included patients born from January 1, 2007 through December 31, 2011. Data were collected until October 2012 to maximize inclusion of patients born during the study period. Unfortunately, reliable autopsy data were not available in many patients.
1.1. Statistics The primary outcome variable used to validate the risk score was death prior to hospital discharge. Descriptive statistics of the demographic and clinical variables recorded were calculated for the surgically repaired infants whose defect was scored for the 4 subgroups of these infants defined
2410
Fig. 1
K.P. Lally et al.
Defect diagram. The diagram is drawn with the diaphragm (defect) on the patient's left looking up from the abdomen towards the chest.
by defect size, and for the 256 infants who were not surgically repaired. Logistic regressions were calculated to evaluate whether the 4 defect sizes recorded were associated with any of the demographic and clinical variables collected. We hypothesized that the defect size variable was the risk factor most strongly associated with death. Each of the other risk factors was then included with defect size in a multiple binary logistic regression to determine whether the prediction of death improved significantly with both variables considered. Evidence for an improved model fit was determined by the magnitude of the decrease in the Bayesian Information Criterion (BIC) statistic of the combined model versus a model that included defect size as the sole covariate in the model. A smaller BIC statistic indicates a better model. We used Raftery's criteria: a combined model that decreased the BIC by 2–6 was positive evidence for improvement, 6– 10 strong evidence, and N 10 very strong evidence [18]. All risk factors associated with positive evidence for improvement or better were to be included in a final binary multiple logistic regression. The risk factors included in the final model were used to calculate a 5 category risk factor that can be easily scored. The BIC statistics of the final multiple binary logistic regression and a binary logistic regression with only the simple clinical risk score were then compared following Raftery's criteria to ensure the risk score adequately captured the information in the component risk factors.
Receiver operator curves (ROC) and area under the curve (AUC) were calculated to summarize the classification of death by the regression models. All analyses were conducted using Stata (version 12.0, College Station, TX). The CDHSG registry was approved by the University of Texas–Houston Institutional Review Board (#HSC-MS-03223). Participating centers filed a waiver of consent for data submission or signed a data use agreement for a limited data set. Center identification was kept anonymous.
2. Results A total of 2,296 patients were entered into the registry during the time period of this study. These patients were cared for in 53 participating centers. Patients treated at 2 centers were excluded as greater than 40% of their patients did not have a coded defect, leaving 51 participating centers from 9 countries. Additionally excluded were 41 who had either bilateral or Morgagni type defects, and 2 patients who were transferred from the participating center before repair. Overall survival for the remaining 1,975 patients was 72% (Fig. 2). All 326 patients (16%) who never had surgical correction died. The rate of non-repair ranged from 0 to 44% in the various centers. Survival to discharge for the 1,649
Standardized reporting for CDH
2411
Fig. 2
Patients comprising the study.
patients who underwent surgical repair was 85%. Demographic data are shown in Table 1 along with differences between the operated and non-operated patients. Several variables were associated with poor outcome. The size of the diaphragm defect was clearly the most significant risk factor affecting outcome (BIC statistic = 26). All patients with an “A” defect had a primary repair and all patients with a “D” defect had a patch repair of the defect. Other factors that influence survival were evaluated to determine if they had an added effect when included with the size of the diaphragm defect in a multiple regression model. By Raftery's criteria, only 2 risk factors improved model fit, a major cardiac anomaly (considering minor cardiac anomalies separately did not improve prediction) and the 1 minute Apgar ≤ 4 (5 minute Apgar also improved model fit but was not included in the table because of the correlation with the 1 minute Apgar). While the Apgar scores at 1 and 5 minutes were significant, 7% (104/1585) of infants did not have a record of any Apgar score. This unscored group represented the most severely ill infants with a mortality of 39% as compared with a mortality of 28% among infants receiving one or more Apgar scores. In addition, 85% (200/236) of infants intubated by 1 minute of life received 1 minute Apgar scores and 91% (719/793) of infants intubated by 5 minutes of life received 5 minute scores such that active therapy confounded the ability to accurately assess respiratory effort and color. With prenatal diagnosis and earlier intervention, the Apgar score is becoming less reliable as a measure to
direct care. Because of these factors, the Apgar score was excluded from the final model. Patients were grouped into a reporting system as shown in Table 2. Stage I are infants who have an isolated “A” size defect. For each increase in defect size the stage increases by 1. The presence of a major cardiac anomaly (defect +), increases the stage by 1. The probability for death using this characterization is shown in Table 3 when compared to patients with an isolated “A” defect (stage I). Using these data, a 6 component reporting system is proposed. The addition of a major cardiac anomaly puts the patient in the next stage. The proposed staging and reporting system and actual survival of patients characterized by defect or defect + is shown in Table 2.
3. Discussion Congenital diaphragmatic hernia remains a challenging problem. Mortality remains high and while survival has improved in the past decade, morbidity in the survivors is substantial [19–22]. Because the anomaly occurs infrequently, very few centers treat large numbers of patients. In the past 20 years, there has been a rapid expansion in the number of therapies available for these infants. These include fetal intervention, ECMO, iNO, surfactant, HFOV and numerous pulmonary vasodilators. Given the small numbers of patients per center and the large number of hitherto evidence based
2412 Table 1
K.P. Lally et al. Demographic and clinical characteristics of the study infants successfully scored.
Characteristic
Defect size
Male sex Prenatal diagnosis Defect on left side Major cardiac anomaly Chromosomal anomaly Other anomaly Inborn Gestational age (weeks) (mean ± SD) Birth weight (kg) (mean ± SD
Not repaired
A n = 218
B n = 716
C n = 495
D n = 209
Total n = 1 638
n = 326
124 # [57] (172) 82 [38] (217) 197 [90] (218) 13 [6] (218) 7 [3] (218) 4 [2] (218) 64 [29] (218) 38.1 ± 2.0 (217) 3.1 ± 0.6 (217)
442 [62] (716) 390 [55] (713)*** 624 [87] (716) 27 [4] (716) 20 [3] (716) 10 [1] (716) 262 [37] (715) 37.9 ± 2.0 (712) 3.1 ± 0.6 (716)
316 [64] (495) 362 [73] (495)*** 384 [78] (495)*** 31 [6] (495) 15 [3] (495) 14 [3] (495) 251 [51] (495)*** 37.5 ± 2.2 (491)** 3.0 ± 0.6 (494)**
136 [65] (209) 177 [85] (209)*** 170 [81] (209)** 20 [10] (209) 9 [4] (209) 9 [4] (209) 103 [49] (209)*** 37.5 ± 2.2 (205)** 2.9 ± 0.6 (207)***
1 018 [62] (1 638) 1 011 [62] (1 634) 1 375 [84] (1 638) 91 [6] (1 638) 51 [3] (1 638) 37 [2] (1 638) 680 [42] (1 637) 37.8 ± 2.1 (1 625) 3.0 ± 0.6 (1 634)
191 [59] (326) 260 [80] (326)††† 261 [80] (326) 60 [18] (326)††† 54 [17] (326)††† 44 [14] (326)††† 188 [58] (326)†† 36.2 ± 3.5 (324)††† 2.6 ± 0.8 (321)†††
Note: Columns A through D do not include 11 patients who were repaired but for whom no defect size was documented. * (p b 0.05), ** (p b 0.01), *** (p b 0.001) indicate whether infants in defect size categories (B, C, or D) significantly differed from infants in defect size category A with respect to this characteristic as evaluated by an appropriate regression model (linear or binary logistic). † (p b 0.05), †† (p b 0.01), ††† (p b 0.001) indicate a significant difference between infants repaired and not repaired with respect to this characteristic as evaluated by an appropriate regression model (linear or binary logistic). # Represents number of infants with characteristic, % with characteristic in brackets and total number of infants with data for the characteristic in parentheses.
unproven therapies available, interpretation of which therapies do in fact have real benefit becomes very difficult. Additionally, most of these therapies were adopted without well designed clinical trials. While not unique to CDH, the use of data from small single institution studies may result in widespread adoption of therapies with little benefit, or even real harm. The rapid adoption of hyperventilation for infants with CDH, which has since been abandoned, is a clear example of this problem [5,23,24]. Some recent improvements have been demonstrated with a consensus based protocol in Europe [25]. One important strength of the current study is that it examines a large number of neonates treated during a 5 year time span when there were few significant new therapies available that would have impacted survival. Table 2
These data demonstrate a wide range of severity amongst infants with CDH. In order to accurately evaluate results of treatment for symptomatic newborns with CDH we assert that reporting of defect size, presence or absence of a major cardiac anomaly and inclusion of patients who die without repair is necessary. While we and other authors have suggested other schemes to determine disease severity, these are meant for use early in life. These various predicting schema have not been reported consistently and no current standard for outcome reporting has been adopted. The previous study from the CDHSG attempted to predict outcomes early in the course using a logistic regression equation. It is important to emphasize that the staging system proposed here is intended to serve as a standard for reporting and not as a predictor for individual patients.
Proposed staging system and survival.
Stage
Defect
I II II III III IV IV V
A A B B C C D D
Major cardiac anomaly + + + +
N
Died
Survived
Group survival
Stage survival
164 8 572 18 372 27 144 18
1 1 21 6 81 12 60 11
163 7 551 12 291 15 84 7
99% 88% 96% 67% 78% 56% 58% 39%
99% 96% 78% 58% 39%
Standardized reporting for CDH Ackerman and colleagues have recently proposed a defect classification of CDH for research purposes, but this system is too complex for a simplified staging system [26]. Defect size alone is considered a surrogate by proxy for pulmonary hypoplasia, but there are no validated clinically available methods to directly determine pulmonary hypoplasia. Chest radiographs are of no value in predicting severity of illness in infants with CDH [27]. Ultrasound and MRI have been used to develop prenatal estimates of pulmonary hypoplasia but these measurements have not been studied in a rigorous, multi-institutional fashion or validated against postnatal anatomic information [28,29]. While many of our centers use potentially useful perinatal markers such as MRI calculated lung volumes, standardized methods to determine volumes have not yet been validated across multiple centers and may not be generalizable [30]. This is of particular importance as it makes interpretation of the benefits of fetal intervention in CDH extremely difficult. There are several limitations to this study. The data are derived from a voluntary registry. While there are agreed upon definitions, individual data entry by a large number of individuals can lead to error. The registry is designed as an observational dataset and postnatal management is not standardized between centers. It will be important to validate the proposed staging system prospectively on patients not included in the data set used to derive the results. Finally, the proposed staging system is not applicable to prenatal risk stratification and provides no guidance to those physicians performing prenatal counseling and future parents as to the optimal treatment before delivery of the affected fetus. This report does not attempt to identify surrogate clinical markers for disease severity that can direct therapy in an individual patient in a clinical setting. Rather, it is designed to allow comparison of results between centers, or to evaluate particular therapies or bundles of care in a risk stratified manner. This dataset demonstrates a very variable rate of non-repair by center. The database alone does not provide an explanation for this. Efforts to investigate this phenomenon are presently being planned using alternative research strategies.
4. Conclusion In conclusion, patients with CDH can be effectively stratified into stages based on readily available clinical information. Given the wide variation in therapies for CDH and the broad range of patients seen in different institutions, we believe it is important that all clinical reports of patients with CDH be reported by stage and include the patients who do not receive surgical therapy. It is clear that using current therapy, mortality is low for infants with stage I and II lesions. Our group plans to focus on high stage patients (III– V and non-repair) who currently represent almost all of the deaths from CDH. It is hoped that this staging system will
2413 thereby allow our group and others to better determine best therapy for high risk infants with CDH.
Appendix Participating CDHSG centers. Hospital
City
State or Country province
Arkansas Children's Little Rock Hospital Astrid Lindgren Stockholm Children's Hospital, Q3:03 BC Children's & Vancouver Women's Health Centre Children's Hospital Boston Boston Children's Hospital of Akron Akron Children's Hospital of Peoria Illinois Children's Hospital of Los Los Angeles Angeles Children's Hospital of Milwaukee Wisconsin Children's Hospital Omaha Omaha Children's Hospitals and Minneapolis Clinics (Minneapolis) Children's Memorial Houston Hermann Hospital Duke University Medical Durham Center Emory University Atlanta Georgia Health Sciences Augusta University Golisano Children’s Rochester Hospital at Strong IRCCS Fondazione Ca' Milano Granda Ospedale Maggiore Policlinico Kosair Children's Hospital Louisville Le Bonheur Children’s Memphis Medical Center Legacy Emanuel Portland Children's Hospital Loma Linda University Loma Linda Children's Hospital Lucile Salter Packard Palo Alto Children's Hospital Mattel Children's Hospital Los Angeles at UCLA Miami Valley Hospital Dayton National Center for Child Tokyo Health and Development
AR Sweden
BC
Canada
MA OH IL CA WI NE MN TX NC GA GA NY Italy
KY TN OR CA CA CA OH Japan
(continued on next page)
2414
K.P. Lally et al.
(continued) Hospital
References City
State or Country province
Oespedale Pediatrico Rome Bambino Gesu Oespedale Riuniti Bergamo Bergamo Osaka University Suita Osaka Graduate School of Medicine Palmetto Health Richland Columbia SC Primary Children's Salt Lake City UT Hospital Radboud University Nijmegen Nijmegen Medical Centre Columbus OH Research Institute at Nationwide Children’s Hospital Royal Children's Hospital Parkville Victoria Royal Hospital for Sick Glasgow Yorkhill Children San Diego Children's San Diego CA Hospital Sophia Children's Rotterdam Hospital St. Francis Children's Tulsa OK Hospital St. Joseph's Hospital and Phoenix AZ Medical Center St. Louis Children's St. Louis MO Hospital Stollery Children's Edmonton Alberta Hospital Sydney Children's Randwick NSW Hospital The Children's Hospital of Birmingham AL Alabama The Hospital for Sick Toronto Ontario Children University of Michigan, Ann Arbor MI C.S. Mott Children's Hospital University of Nebraska Omaha NE Medical Center University of Padua Padua University of Texas Galveston TX Medical Branch at Galveston University of Virginia Charlottesville VA Medical School Nashville TN Vanderbilt Children's Hospital Vladivostok State Medical Vladivostok University Winnie Palmer Hospital Orlando FL for Women & Babies Yale New Haven New Haven CT Children's Hospital
Italy Italy Japan
The Netherlands
Australia Scotland
The Netherlands
Canada Australia
Canada
Italy
Russia
[1] Wenstrom KD, Weiner CP, Janson JW. A five-year statewide experience with congenital diaphragmatic hernia. Am J Obstet Gynecol 1991;165:838-42. [2] Langham MR, Kays DW, Ledbetter DJ, et al. Congenital diaphragmatic hernia: epidemiology and outcome. Clin Perinatol 1996;23: 671-88. [3] Downard CD, Jaksic T, Garza JJ, et al. Analysis of an improved survival rate for congenital diaphragmatic hernia. J Pediatr Surg 2003;38:729-32. [4] The Congenital Diaphragmatic Hernia Study Group. Treatment Evolution in High Risk Congenital Diaphragmatic Hernia: Ten Years Experience. Ann Surg 2006;244:505-13. [5] Boloker J, Bateman DA, Wung J-T, et al. Congenital diaphragmatic hernia in 120 infants treated consecutively with permissive hypercapnia/spontaneous respiration/elective repair. J Pediatr Surg 2002;37: 357-66. [6] Dassinger MS, Copeland DR, Gossett J, et al. Early repair of congenital diaphragmatic hernia on extracorporeal membrane oxygenation. J Pediatr Surg 2010;45:693-7. [7] Raval MH, Wang X, Reynolds M, et al. Costs of congenital diaphragmatic hernia repair in the United States – extracorporeal membrane oxygenation foots the bill. J Pediatr Surg 2011;46:617-24. [8] Jacobs JP, O’Brien SM, Pasquali SK, et al. Variation in outcomes for benchmark operations: an analysis of the Society of Thoracic Surgeons Congenital Heart Surgery Database. Ann Thorac Surg 2011;92: 2184-91. [9] Baird R, MacNab YC, Skarsgard ED, et al. Mortality prediction in congenital diaphragmatic hernia. J Pediatr Surg 2008;43:783-7. [10] Hoffman SB, Massaro AN, Gingalewski C, et al. Survival in congenital diaphragmatic hernia: use of predictive equations in the ECMO population. Neonatology 2011;99:258-65. [11] The Congenital Diaphragmatic Hernia Study Group. Estimating disease severity of congenital diaphragmatic hernia in the first 5 minutes of life. J Pediatr Surg 2001;36:141-6. [12] The Congenital Diaphragmatic Hernia Study Group. Defect Size Determines Survival in Congenital Diaphragmatic Hernia. Pediatrics 2007;121:e651-7. [13] Valfrè L, Braguglia A, Conforti A, et al. Long term follow-up in highrisk congenital diaphragmatic hernia survivors: patching the diaphragm affects the outcome. J Pediatr Surg 2011;46:52-6. [14] Graziano JN, for the Congenital Diaphragmatic Hernia Study Group. Cardiac anomalies in patients with congenital diaphragmatic hernia and their prognosis: a report from the Congenital Diaphragmatic Hernia Study Group. J Pediatr Surg 2005;40:1045-50. [15] Tsao K, Allison ND, Harting MT, et al. Congenital diaphragmatic hernia in the preterm infant. Surgery 2010;148:404-10. [16] Colvin J, Bower C, Dickinson JE, et al. Outcomes of congenital diaphragmatic hernia: a population-based study in Western Australia. Pediatrics 2005;116:e356-63. [17] Tsao K, Lally KP. The Congenital Diaphragmatic Hernia Study Group: a voluntary international registry. Semin Pediatr Surg 2008;17:90-7. [18] Raftery A. Bayesian model selection in social research. In: Marsden P, editor. Sociological Methodology, Vol. 25. Oxford: Blackwell; 1995. p. 111-63. [19] Danzer E, Gerdes M, Bernbaum J, et al. Neurodevelopmental outcome of infants with congenital diaphragmatic hernia prospectively enrolled in an interdisciplinary follow-up program. J Pediatr Surg 2010;45: 1759-66. [20] Friedman S, Chen C, Chapman JS, et al. Neurodevelopmental outcomes of congenital diaphragmatic hernia survivors followed in a multidisciplinary clinic at ages 1 and 3. J Pediatr Surg 2008;43:1035-43. [21] Cortes RA, Keller RL, Townsend T, et al. Survival of severe congenital diaphragmatic hernia has morbid consequences. J Pediatr Surg 2005;40:36-45.
Standardized reporting for CDH [22] van den Hout L, Reiss I, Felix JF, et al. Risk factors for chronic lung disease and mortality in newborns with congenital diaphragmatic hernia. Neonatology 2010;98:370-80. [23] Drummond WH, Gregory GA, Heymann MA, et al. The independent effects of hyperventilation, tolazoline, and dopamine on infants with persistent pulmonary hypertension. J Pediatr 1981;98:603-11. [24] Kays DW, Langham Jr MR, Ledbetter DJ, et al. Detrimental effects of standard medical therapy in congenital diaphragmatic hernia. Ann Surg 1999;230:340-8. [25] van den Hout L, Schaible T, Cohen-Overbeek TE, et al. Actual outcome in infants with congenital diaphragmatic hernia: the role of a standardized postnatal treatment protocol. Fetal Diagn Ther 2011;29:55-63. [26] Ackerman KG, Vargas SO, Wilson JM, et al. Congenital Diaphragmatic Defects: Proposal for a New Classification Based on Observations in 234 Patients. Pediatr Dev Pathol 2012;15:265-74.
2415 [27] Holt PD, Arkovitz MS, Berdon WE, et al. Newborns with diaphragmatic hernia: initial chest radiography does not have a role in predicting clinical outcome. Pediatr Radiol 2004;34:462-4. [28] Mullassery D, Ba'ath ME, Jesudason EC, et al. Value of liver herniation in prediction of outcome in fetal congenital diaphragmatic hernia: a systematic review and meta-analysis. Ultrasound Obstet Gynecol 2010;35:609-14. [29] Arkovitz MS, Russo M, Devine P, et al. Fetal lung-head ratio is not related to outcome for antenatal diagnosed congenital diaphragmatic hernia. J Pediatr Surg 2007;42:107-10. [30] Lee TC, Lim FY, Keswani SG, et al. Late gestation fetal magnetic resonance imaging-derived total lung volume predicts postnatal survival and need for extracorporeal membrane oxygenation support in isolated congenital diaphragmatic hernia. J Pediatr Surg 2011;46: 1165-71.
