Congenital diaphragmatic hernia: a systematic review and summary of best-evidence practice strategies

STATE OF THE ART Congenital diaphragmatic hernia: a systematic review and summary of best-evidence practice strategies JW Logan 1, HE Rice 2, RN Goldberg 1 and CM Cotten 1 (2007) 27, 535 549 r 2007 Nature Publishing Group All rights reserved. 0743-8346/07 $30 www.nature.com/jp 1 Division of Neonatology, Department of Pediatrics, Duke University Medical Center, Durham, NC, USA and 2 Department of Surgery, Duke University Medical Center, Durham, NC, USA Objectives: Recent reports suggest that specific care strategies improve survival of infants with congenital diaphragmatic hernia (CDH). This review presents details of care from centers reporting high rates of survival among CDH infants. Study Design: We conducted a MEDLINE search (1995 to 2006) and searched all citations in the Cochrane Central Register of Controlled Trials. Studies were included if they contained reports of >20 infants with symptomatic CDH, and >75% survival of isolated CDH. Result: Thirteen reports from 11 centers met inclusion criteria. Overall survival, including infants with multiple anomalies, was 603/763 (79%; range: 69 to 93%). Survival for isolated CDH was 560/661 (85%; range: 78 to 96%). The frequency of extracorporeal membrane oxygenation (ECMO) use for isolated CDH varied widely among reporting centers 251/622 (40%; range: 11 to 61%), as did survival for infants with isolated CDH placed on ECMO: 149/206 (73%; range: 33 to 86%). There was no suggestion of benefit from use of antenatal glucocorticoids given after 34 weeks gestation or use of postnatal surfactant. Low mortality was frequently attributed to minimizing lung injury and adhering to center-specific criteria for ECMO. Conclusion: Use of strategies aimed at minimizing lung injury, tolerance of postductal acidosis and hypoxemia, and adhering to center-specific criteria for ECMO were strategies most consistently reported by successful centers. The literature lacks randomized clinical trials of these or other care strategies in this complex patient population; prospective studies of safety and long-term outcome are needed. (2007) 27, 535 549; doi:10.1038/sj.jp.7211794; published online 19 July 2007 Keywords: congenital diaphragmatic hernia; pulmonary hypertension; pulmonary hypoplasia; inhaled nitric oxide; extracorporeal membrane oxygenation; systematic review Correspondence: Dr JW Logan, Brody School of Medicine, Division of Neonatal-Perinatal Medicine, East Carolina University, 600 Moye Blvd, Greenville, NC 27858, USA. E-mail: wellslogan@usa.net Received 18 March 2007; revised 18 June 2007; accepted 19 June 2007; published online 19 July 2007 Introduction Congenital diaphragmatic hernia (CDH) is a rare malformation observed in approximately 1 in 3000 live births, with estimates of postnatal survival ranging from 50 to 70% despite advances in neonatal care. 1 The combination of pathophysiologic derangements associated with CDH, including lung hypoplasia, lung dysmaturity and pulmonary hypertension (PPHN), requires a multidisciplinary approach to care. 2,3 The complex nature of clinical management and the relatively infrequent presentation of infants with this rare malformation to single centers has lead to difficulty identifying consistent strategies for CDH management. A recent report from the Canadian Neonatal Research Network indicates better than predicted survival in Network centers, with significant improvement in the higher mortality odds strata, and 90% survival in the three centers that cared for at least 12 infants over a 22-month period. Details of management strategies are not provided, but the report suggests that survival is best achieved by caring for CDH infants at centers that have developed a standardized approach to care, with adequate numbers of CDH infants to develop expertise. 4 Literature review reveals a limited number of controlled clinical trials examining interventions that improve survival of infants with CDH; 5 14 however, several centers reporting relatively high rates of survival for CDH have described their management strategies in varying detail. 15 25 The objective of this report is to provide clinicians with a summary of specific details of clinical care strategies from these benchmark centers. In the absence of clinical trials, this summary of practice strategies provides a starting point for development of an approach to care for infants with CDH. Study Design Search strategy We conducted a systematic review of the pediatric literature to identify intensive care strategies associated with improved survival among infants with CDH. We performed a MEDLINE search of all clinical reports and a review of articles published between January

536 Congenital diaphragmatic hernia 1995 and December 2006, using the MeSH search terms diaphragmatic hernia and newborns, restricting the search to human subjects and reports in English. We also reviewed all available citations in the Cochrane Central Register of Controlled Trials. Selection criteria Reports with X20 CDH infants were included if they reported survival for isolated CDH (without other identified defects) or all infants with CDH (including those with anomalies) X75%, and included a detailed description of at least one clinical care strategy concordant with the reported survival >75%. One center included in this report is a non-ecmo center, which may have impacted clinical care strategy and/or outcomes. 16 Quality of included studies Our search identified retrospective, single-center case series or case vs historical control comparisons, raising the possibility of both selection and performance bias. 16 27 The improvement in outcomes noted in reports comparing serial epochs are accompanied by reports of changes in clinical practice. Details of various practices (specifics of antenatal care, ventilation strategy, PPHN management strategy and use of ECMO) were not uniformly recorded in all studies. Data collection and analysis The primary measures considered were survival and use of ECMO. The secondary measures considered included specific descriptions of aspects of clinical management, including antenatal care, early postnatal care, management of PPHN, use of inhaled nitric oxide (ino), use of sedation and analgesia, and timing of surgery. In order to detect trends in each of these aspects of clinical practice, the rationale and the frequency of the cited practice strategy was recorded and analyzed. Results We identified 13 reports from 11 centers. Outcomes and practices at each center are summarized in Tables 1 to 7. There are two reports from each of two centers, Columbia University Hospitals in New York and Children s Hospital/Harvard University in Boston. While there is no overlap in the reports from the Boston group, there is a small overlap in the outcomes reports from the Columbia group (August 1992 through December 1995). 15,17,25,26 This overlap was not considered in the analysis, as data could not be appropriately extracted from the reports and omission of either or both of the reports seemed inappropriate. Survival Our analysis considered survival from the most recent epoch at each center as the basis for identifying successful practice strategies. The 13 reports included 763 infants with symptomatic CDH. Overall survival for the combined reports was 603/763 (79%; range: 69 to 93%). Survival for isolated CDH for the successful centers was 560/661 (85%; range: 78 to 96%). 15 20,22 25,27 Reported survival for isolated CDH using conventional medical management, excluding ECMO but including high-frequency ventilation was 331/367 (90%; range: 83 to 100%). The frequency of ECMO use for isolated CDH was 251/622 (40%; range: 11 to 61%), and overall survival for infants with isolated CDH placed on ECMO was 149/206 (73%; range: 33 to 86%; Table 1). Most of the successful centers adopted a strategy of gentle ventilation or permissive hypercapnia. 15 25 Mortality was generally attributed to the presence of additional anomalies, iatrogenic lung injury, severe pulmonary hypoplasia and/or PPHN, and bleeding complications (among ECMO-treated infants). One report suggested that after excluding CDH infants with additional anomalies, hopeless pulmonary hypoplasia occurred in no more than 15% of isolated CDH infants. 25 Clinical care strategies Antenatal diagnosis and management. Among centers reporting antenatal care, 354/630 CDH infants (56%) were prenatally diagnosed, and 247 of 763 infants (33%) were inborn. 15 18,20 23,25,27 Reports varied on whether or not prenatal diagnosis or site of delivery was reported or associated with improved survival, but the combined reports suggest that delivery at a tertiary center was beneficial. For most centers, survival increased for both inborns and prenatally diagnosed infants in the most recent epoch of care. High-risk infants, infants with additional congenital anomalies and infants with an early prenatal diagnosis were often inborn. 16 18,23 Overall, the data suggest that infants with a prenatal diagnosis had a better chance of survival if they were born in a tertiary center (Table 2). Histology consistent with lung immaturity has suggested a possible therapeutic role for prenatal glucocorticoids after 34 weeks postmenstrual age, and animal studies have supported this hypothesis. 28 31 Only 3 of the 11 identified successful centers routinely used antenatal steroids after 34 weeks gestation, with no clear benefit. 18,20,27 A recent report from the CDH Study Group examining the role of antenatal steroids did not support their use for fetuses with CDH. 32 Antenatal care also includes consideration of possible fetal interventions. Reports identified in this review did not include maternal-fetal surgery as part of antenatal care. Although animal models suggest that tracheal occlusion may improve pulmonary function, clinical trials have not demonstrated efficacy. 6,33 35 Surfactant. Animal studies and human clinical trials suggest that infants with CDH may have surfactant deficiency, and surfactant composition may be altered. 36 40 Some centers initiated surfactant therapy in the first few hours of life, while others used

Congenital diaphragmatic hernia 537 Table 1 Survival and ECMO use by report* (year) Overall survival (recent epoch) Isolated CDH Survival (recent epoch) Survival predictors (recent epoch) ECMO for isolated CDH ECMO survival Isolated CDH CMM (non-ecmo) survival Isolated CDH Bagolan 26/29 (90%) (1996 2001) a Gentle ventilation+permissive hypercapnia led to mm in survival in 3rd epoch Boloker (1992 1999) 91/120 (75.8%), (See Wung et al.) a Including infants with additional anomalies Downard (00-02) (See Wilson et al.) a Finer (1989 1995) Frenckner (1990 1995) 36/39 (93%) 27/28 (96%) Extracted from data 26/28 (93%) (1) Prenatal diagnosis (2) No preoperative pneumothorax (3) All lethal anomalies died B82/96 a (85.4%) OI<20 at 6 h: 35/35 (100%) survival. OI>20 at 6 h: 20/24 died (83%). Chest tube: 11/11 died Birth weight and 5 min APGAR (CDHSG) b 51/65 (78%) 51/65 (78%) No currently measurable factors that predict outcome 39/43 (91%) Symptomatic in first 6 postnatal hours No ECMO 26/29 (89%) 3rd epoch P ¼ 0.02 16/97 (16%) Total ECMO survival (Isolated+anomalies): 10/16 (62%) a 72/82 (88%) Total: 14/36 (39%) a Total ECMO survival: 12/14 (86%) a Total non-ecmo survival: 24/25 (96%) a 38/65 (58%) 26/38 (68%) 25/27 (93%) 39/42 (93%) 7/42 (17%) 6/7 (85%) (1 late death) 32/35 (91%), 3 prematures Kays 47/60 (78%) 47/53 (89%) 23/53 (43%) 19/23 (83%) 28/30 (93%) (1992 1998) Osiovich 52/59 (88%) 52/59 (88%) 31/59 (53%) 25/31 (81%) 27/28 (96%) (1993 1999) Reyes 18/22 (82%) 16/18 (89%) 3/18 (17%) Total ECMO 15/18 (83%) (1993 1995) survival: 3/4 (75%) Somaschini 25/28 (89%) 25/28 (89%) 3/28 (11%) 1/3 (33%) 24/25 (96%) (1994 1998) Ssemakula 71/98 (72%) 64/82 (78%) 60/98 (61%) (1985 1996) a Weber 44/56 (79%) 44/56 (79%) 32/56 (57%) 21/32 (65%) 23/24 (96%) (1988 1997) Wilson (1991 1994) 24/43 (56%) 17/24 (71%) 19/19 (100%) 51/74 (69%) Mortality attributed to associated anomalies (39%) and iatrogenic barotrauma 36/43 (84%) Prenatal dx correlated with m in survival (last epoch) Wung 52/70 (74%) 52/63 (82%) 14/63 (22%) 9/14 (65%) 43/49 (88%) (1983 1995) Total 603/763 (79%) 560/661 (85%) 251/622 (40%) 149/206 (73%) 331/367 (90%) Abbreviations: CDH, congenital diaphragmatic hernia; CDHSG, CDH Study Group; CMM, conventional medical management (non-ecmo survival); ECMO, extracorporeal membrane oxygenation;, not applicable; OI, oxygenation index. a Data for isolated CDH survival was not directly included in the report; available data allowed calculation of a number that appears to reflect isolated CDH survival at these sites. b CDHSG, CDH Study Group: probability of survival ¼ 1 1/(1+e x ), where x ¼ 5.0240+0.9165 (BW)+0.4512 (APGAR at 5 min). surfactant only as a rescue therapy for the most severely affected infants with evidence of surfactant deficiency (Table 3). 18,21,22,27 Somaschini et al. 22 used surfactant on 12 of 28 infants, with only four responders, while Osiovich 27 used surfactant in 47 of 59 infants (80%), with no comment on response to therapy. Finer et al. 18 noted occasional sudden, irreversible deterioration of CDH infants given a bolus of surfactant, necessitating emergent use of ECMO; the slow administration of surfactant is recommended to avoid such complications. When combined, these reports seem to agree with a recent report by Van Meurs 41 suggesting that

538 Congenital diaphragmatic hernia Table 2 Birthplace (inborn vs outborn), antenatal diagnosis and effects on survival Inborn Prenatal diagnosis Survival (most recent epoch) Comments Bagolan (1996 2001) Boloker (1992 2000) Downard (2000 2002) Finer (1989 1995) Frenckner (1990 1995) Kays (1992 1998) Osiovich (1993 1999) Reyes (1993 1996) Somaschini (1994 1998) Ssemakula (1985 1996) Weber (1988 1997) Wilson (1991 1994) 18/29 (62%) 18/29 (62%) Inborn: 16/18 (89%) Inborn ¼ prenatal diagnosis Outborn: 10/11 (91%) 67/120 (56%) 92/120 (77%) Inborn 67/67 (100%) Outborn 25/53 (47%) Inborn 41/47 (87%) Outborn 47/53 (89%) 28/39 (72%) 39/39 (100%) See Wilson et al. (same center). 23/65 (35%) 21/65 (32%) Inborn 21/23 (91%) With prenatal dx: 17/20 (85%) Outborn 32/42 (76%) With prenatal dx: 13/21 (62%) 0/43 (All outborn) Survival increased for inborns (high-risk group) with each epoch: 1st epoch 8/15 (53%), 3rd epoch 16/18 (89%). All outborns intubated within 2 h of birth All outborns intubated before transfer. More anomalies in inborns (higher risk patients). 66% of outborn deaths used high (>25 cm H 2 O) PIPs. The 2 deaths in the inborn group required ECMO, one early and one late death. Antenatal diagnosis before 25 weeks gestation was associated with a 60% survival. 10/43 (23%) All infants outborn. The authors report that delivery was performed with extra surveillance for those with a prenatal diagnosis (babies transferred immediately). 25/60 (42%) 22/53 (42%) (Treated patients) 27/59 (46%) 41/59 (69%) Inborn 27/27 Outborn 14/32 8/22 (36%) 9/22 (41%) Inborn 6/8 Outborn 3/14 3rd epoch: (Isolated CDH): Prenatal dx: 20/22 (91%) Inborn: 23/25 (92%) Outborn: 22/25 (88%) Inborn: 26/27 (96%) Outborn: 26/32 (81%) Overall survival: 18/22 (82%) Isolated CDH: 16/18 (89%) Prenatal dx: 7/18 (39%) Inborn: 6/8 (75%) Outborn: 1/3 (33%) 16/28 (57%) 16/28 (57%) Total: 25/28 (89%) Inborn: 15/16 (94%) Outborn: 10/12 (83%) Prenatal dx: 15/16 (94%) 34/98 (35%) 44/98 (45%) Non-ECMO survival 35/38 (92%) Isolated CDH: 64/82 (78%) Inborn: 17/34 (51%) Outborn: 40/64 (62%) 0/56 (0%) (All outborn) 0/74 (0%) (All outborn) Survival improved with increasing epochs of care in both prenatally diagnosed babies and inborn babies. 53 of the total 60 patients were treated (7 lethal) Data suggest that survival odds are better if born in tertiary care center Infants with a prenatal diagnosis had better survival if they were born at a tertiary center. 2 infants in the survivor group had congenital heart anomalies (VSD and double outlet right ventricle+vsd). 3 of the 12 outborns went to ECMO and 2 of these died. Suggestion: high-risk patients may do better if born at a tertiary center All 15 patients with polyhydramnios survived. 29/98 (30%) had a prenatal diagnosis before 25 weeks gestation with 6 survivors (21%) All infants are outborn. Referral center only 42/74 (56%) All infants were outborn. No significant difference in survival between those with and without a prenatal diagnosis. Prenatal diagnosis alone does not predict death See Boloker et al. Wung (1983 1995) 0/70 (0%) (All outborn) Total 247/763 (33%) 354/630 (56%) Inborn: 164/198 (83%) Outborn: 188/242 (78%) Prenatal dx: 88/115 (77%) These numbers do not include data from the entire cohort. Reflects inborn and prenatal diagnosis data available in reports Abbreviations: CDH, congenital diaphragmatic hernia; ECMO, extracorporeal membrane oxygenation;, not applicable; PIPs, peak inspiratory pressures.

Congenital diaphragmatic hernia 539 Table 3 Early management strategies a Author Initial stabilization Surfactant Initial settings or limits HFV limits/settings Blood gas goals: Ideal Blood gas goals: acceptable Bagolan (1996 2001) Boloker (1992 1999) (See Wung et al, same center) Finer (1989 1995) Frenckner (1990 1995) Kays (1992 1998) Osiovich (1993 1999) Reyes (1993 1995) Somaschini (1994 1998) Immediate endotracheal intubation. PIP maintained at 24 cm H 2 O in delivery room Nasotracheal intubation, spontaneous respiration Immediate intubation, limit PIP to 24 cm H 2 O in delivery room, sedation, paralysis, surfactant (test dose first) I-time 0.35, e-time 0.65 PEEP 3 4 cm H 2 O Max PIP 24. Max rate 65. If PIP>24 - HFOV Frequency 10 Hz, I:E ratio 0.33 MAP 2 cm>cmv No PIP up to 25, PEEP 5, i-time 0.3 s, IMV 20 40 Yes PIP up to 24 (usually 20 22) HFCV For severe hypoxia/ hypercarbia IMV 100, PEEP ¼ 0 with conventional ventilation HFOV, rescue PaCO 2 >60 torr ph>7.25, PCO 2 <60 torr Preductal SaO 2 90 95% PaCO 2 <60, if can not attain with conventional, switch to HFOV. ph>7.25, PCO 2 60 65 torr First 2 h SaO 2 >70% was acceptable. Next 2 h SaO 2 75 85% was acceptable Preductal SaO 2 : 80 90% use only minimum FiO 2 to maintain SpO 2 in this range First 2 h, preductal SaO 2 >70%, 2 4 h: preductal SaO 2 : 75 85% No PIP up to 35 HFOV, rescue Postductal SaO 2 >90% PaCO 2 60 65 torr Intubation, ventilation at lowest pressure that provides adequate chest rise Intubation, sedation, paralysis, NG tube, and surfactant (47/59) Those requiring intubation: HFOV No Adequate chest movement (usually 20 24 cm H 2 O PIP; PEEP 4 5 cm H 2 O; IMV for patient comfort; FiO 2 started at 1.0 occasional brief use of PIP ¼ 28 if poor chest rise HFCV For severe Hypoxia Or Hypercarbia IMV 100, PEEP ¼ 0 with conventional ventilation Yes (not reported) HFOV as rescue for PaCO 2 >65 torr Yes, if premature and/or RDS are evident Yes, if FiO 2 >0.80 needed to keep PaO 2 >60 mm Hg HFOV range of MAP 5 25 (only two required>20, mean ¼ 12) FiO 2 ¼ 1, MAP 13 15 cm H 2 O, Amp 30 35 cm H 2 O freq ¼ 10 Hz, insp:expiratory ratio ¼ 0.33 PaCO 2 of 40 65 torr; postductal SaO 2 >97%, with PaO 2 80 100 PaCO 2 <65 torr Postductal SaO 2 60%, and PaO 2 >30 tolerated as long as preductal SaO 2 >85% and satisfactory perfusion. PaCO 2 >65 tolerated if patient otherwise stable. PaCO 2, 65 torr HFOV, first 3 h Normal range PaCO 2 Surfactant to 8, with 6 survivors HFOV, initial ph 7.4 7.5 PaO 2 80 120 torr SaO 2 >95%, PaCO 2 25 35 torr pre- or postductal, and adequate chest expansion Ssemakula No (1985 1996) Weber No Not used (1988 1997) Wilson (1991 1994) b No PIP up to 30, PEEP 5, MAP 12 HFCV for severe hypoxia/hypercarbia Preductal SaO 2 >90%. If unable to attain, or

540 Congenital diaphragmatic hernia Table 3 Continued Author Initial stabilization Surfactant Initial settings or limits HFV limits/settings Blood gas goals: Ideal Blood gas goals: acceptable Wung (1983 1995) Nasotracheal intubation, spontaneous respiration No PIP up to 25, PEEP 5, IMV 20 40 IMV 100, PEEP ¼ 0 with conventional ventilation HFCV for severe hypoxia/hypercarbia IMV 100, PEEP ¼ 0 conventional ventilation PaCO 2 >60, switch to HFV Preductal SaO 2 >90%. Preductal SaO 2 >80% Abbreviations: CMV, conventional mechanical ventilation; FiO 2, fraction of inspired O 2 ; HFCV, high-frequency conventional ventilation; HFOV, high-frequency oscillatory ventilation;, not applicable; MAP, mean airway pressure; OI, oxygenation index; PaCO 2, partial pressure CO 2 ; PaO 2, partial pressure of O 2 ; PEEP, positive end expiratory pressure; PIP, peak inspiratory pressure. a Wung et al. (see Boloker for early treatment strategies). b Downard et al. (see Wilson for early treatment strategies). Table 4 Ventilator weaning a Begin weaning Bagolan (2000 2001) ph>7.25 and PCO 2 <60 only when FiO 2 <70% Bolokor (1992 1999) If preductal SaO 2 >90%, maintain preductal SaO 2 >90%. Assure that infant is comfortable and other organ function is adequate Finer (1989 1995) Preductal PaO 2 was >150, PaCO 2 <60 and ph>7.25. Start reducing FiO 2 by no more than 0.03/h. Reduce pressures when FiO 2 <0.70 with acceptable blood gases. Frenckner (1990 1995) Adjust settings to maintain postductal SaO 2 >90% Kays (1992 1998) Start at 6 h of life. FiO 2 weaned with goal of maintaining postductal SaO 2 >97%. For non-ideal patients, tolerated marginal gas exchange Osiovich (1993 1999) CMV used as initial therapy. A strategy of gentle ventilation with low peak inspiratory pressure and permissive hypercapnia was used. Details not specified Reyes (1993 1995) Adjust settings to maintain PaCO 2 in normal range, adequate lung volumes (8 ribs by X-ray) and acceptable blood gases Somaschini (1994 1998) HFOV (parameters Table 3). Switch to conventional ventilation when FiO 2 was <0.30 Ssemakula (1985 1996) Weber (1988 1997) Wilson (1991 1994) If preductal SaO 2 >90%, maintain preductal SaO 2 >90%. Ignore postductal SpO 2 as long as there is no evidence of evolving metabolic acidosis Wung (1983 1995) If preductal SaO 2 >90%, maintain preductal SaO 2 >90% Abbreviations: CMV, conventional mechanical ventilation; FiO 2, fraction of inspired O 2 ; HFOV, high-frequency oscillatory ventilation; MAP, mean airway pressure; PaCO 2, partial pressure CO 2 ; PaO 2, partial pressure of O 2 ; SaO 2, oxygen saturations. a Downard et al., see Wilson et al. for ventilator weaning strategies. surfactant therapy for CDH is associated with increased risk of mortality, even after correcting for use of ECMO, ino and postnatal steroids. This raises concerns over the routine use of surfactant, as its administration is not without risk in term gestation CDH infants and the benefits are still unclear. 42 Ventilator management Pressure limits. Twelve of the 13 reports specifically mention avoiding lung injury by including inspiratory pressure limits as a guideline of therapy. 15 22,24 27 The successful centers range of acceptable peak inspiratory pressures (PIPs) with conventional mechanical ventilation (CMV) was wide, ranging from 20 to 25 cm H 2 O 17,20,26,27 in some centers to 30 35 cm H 2 O (Table 3) in others. 19,25 While peak pressures were emphasized in the reports, acceptable mean airway pressures (MAPs) ranged from 12 to 18 cm H 2 O. 21,22,25,27 Osiovich 27 tolerated PIPs up to 25 cm H 2 O, and used PaCO 2 and OI (MAP/PaO 2 100) as the primary indices for decision making. Some centers set PIP limits and transitioned to alternate forms of cardiorespiratory support, either high-frequency ventilation or ECMO, when inspiratory pressure limits were reached with an inadequate response. 16 18,26,27 Finer et al. 18 suggested that limiting PIPs meant accepting lower PaO 2 levels, particularly in the first few hours of life, and that this required a change in mindset on the part of the clinician. Finer s group attributed a decreased incidence of pneumothorax in their most recent epoch to this approach to care. Boloker et al. 17 reported that no inborn survivor required a PIP greater than 25 cm H 2 O, and 66% of the outborn deaths used PIP >25 cm H 2 O. Other centers used higher PIPs or MAPs and still demonstrated high overall survival. 19,25,27 These findings suggest that setting consistent limits for PIP and duration of high MAP is a reasonable approach to CDH management, but there appears to be no absolute magic number.

Congenital diaphragmatic hernia 541 Table 5 Pulmonary hypertension/blood pressure Alkalosis NaHCO3 or THAM Pressors used MnBP goal MnBP acceptable INO Notes Bagolan (2000 2001) Bolokor (1992 1999) Finer (1989 1995) Frenckner (1990 1995) Kays (1992 1998) Osiovich (1993 1999) Reyes (1993 1995) Somaschini (1994 1998) Ssemakula (1985 1996) Weber (1988 1997) >7.25 Dopamine 3 10 mcg/kg/min if Poor UOP 5% albumin bolus if refractory to dopamine No ph>7.25 acceptable Only for severe metabolic acidosis Dopamine and dobutamine (rarely) Dopamine starts at 50s (to overcome 5 mg/kg/min if shunts) with good MnBP<50 mm Hg, urine output no improvement, increase to 10 mg/kg/ min, no improvement, 5% albumin 10 cc/kg, no improvement start epinephrine at 0.1 mg/kg/min >45 mm Hg 40 45 mm Hg Yes, PHN Shunt>5% (SaO 2 )or Refractory preductal hypoxemia Yes, for failure to oxygenate 45 50, with good urine output Surgery delayed until hemodynamic stability our experience y infants with CDH and severe hypoxemia can slowly improve without changes in ventilator settings as long as one is paying meticulous attention to the maintenance of blood pressure, tissue perfusion, and sedation. Yes, 5 treated slight hypercarbia and hypoxemia have been accepted in some cases. 4 required ECMO, 1 successful postoperative rescue. No, ph down to 7.20, PaCO 2 up to 65 if otherwise stable Yes, for metabolic acidosis NaHCO 3 only to keep ph>7.20 Dopamine 3 mg/kg/ min for all, increase dopamine and/or volume for hypotension At or above patient s gestational age in weeks Yes, 18 treated, 14 to ECMO. HCO 3 Not reported Trial: 20 PPM for hypoxemia (OI 25) Dopamine and dobutamine 7.4 7.5, PaCO 2 25 35 torr Yes, late change to permissive hypercapnea Induced respiratory alkalosis NaHCO 3 THAM, for mild metabolic alkalosis 27 (87%) of the 31 infants that required ECMO had a trial of ino and/or HFO compared to 14 (50%) of the 28 infants that did not require ECMO. Equal to gestation Yes, 6 treated 5 survived, 3 were treated with ECMO Yes, PHN by echo; OI>30 Not necessarily lung protective ventilator strategy, but significant ECMO use Induced metabolic Inotropes No, n ¼ 24 Used permissive hypercapnea for several

542 Congenital diaphragmatic hernia Table 5 Continued Alkalosis NaHCO3 or THAM Pressors used MnBP goal MnBP acceptable INO Notes Wilson (1991 1994) Downard (2000 2002) (Same center) Wung (1983 1995) alkalosis (do not know which agent) No Avoid NaHCO 3, use to avoid significant metabolic acidosis No, avoid hyperventilation Dopamine Dobutamine Rarely epinephrine, usually bridge to ECMO Dobutamine 10 mg/kg/min years y As long as the ph level is maintained relatively normal, hypercapnia does not result in severe PHN 50s 40s Post 1991: minimize barotrauma by largely ignoring the right-to-left shunt. Aggressively used ino in most recent epoch (2000 2002) Tolazoline for pulmonary hypertension causing hypoxemia, with adequate chest wall excursion (survey carried out before NO) Abbreviations: ECMO, extracorporeal membrane oxygenation; ino, inhaled nitric oxide;, not applicable; NO, nitric oxide; OI, oxygenation index; PaCO 2, partial pressure CO 2 ; SaO 2, oxygen saturations. High-frequency ventilation. Four centers used high-frequency oscillatory ventilation (HFOV) as rescue therapy for severe hypercapnia or hypoxia. 16,18,19,27 Four centers used high-frequency (conventional) ventilation (IMV rate of 100, inspiratory time 0.2 s, PEEP 0 to 2 cm H 2 O) for hypercapnia or hypoxia. 17,20,25,26 In the most recent report, Bagolan et al. 16 used HFOV once PIPs were >24 cm H 2 O and sought a hypercapneic state. Two centers demonstrated success using HFOV as initial therapy. 21,22 Reyes et al. 21 reported HFOV settings that achieved optimal lung volume, defined as eight-rib expansion on chest X-ray. 43 Osiovich 27 mentioned HFOV as therapy for hypercapnia, but did not explicitly mention the use of HFOV as a rescue for poor oxygenation. While duration of high MAPs was not explicitly addressed in these reports, the combined reports seem to support the judicious use of HFOV, particularly in those infants requiring higher PIPs with CMV (Table 3). A previous report demonstrated poor survival despite the use of HFOV; however, since adopting a lung-protective strategy without intentional hyperventilation, this center has noted improved survival with a protocol that includes HFOV. 44,45 Blood gas parameters. Along with pressure limit guidelines, some centers set ideal blood gas parameters. If initial ventilator settings failed to achieve ideal blood gases, another acceptable standard for blood gases was chosen. 16 20,26 If acceptable results were not achieved, patients were switched to high-frequency ventilation (Table 3). 16 20,25,26 Finer et al. 18 suggested that early hypoxemia (preductal O 2 saturations (SaO 2 ) as low as 70% in the first 2 h) can slowly improve without ventilator changes as long as one is paying attention to maintenance of blood pressure, tissue perfusion and sedation. Bagolan et al. 16 allowed preductal SaO 2 as low as 70% in the first two hours, but required improvement from 70 to 85% SaO 2 in the subsequent 2 h. A ph greater than 7.25 and PaCO 2 60 to 65 torr was acceptable and served as indirect measures of acceptable tissue perfusion and ventilation. Similarly, Osiovich 27 considered a PaCO 2 <65 torr acceptable and did not escalate respiratory support for PaCO 2 in this range. Osiovich 27 also used the calculated OI as a decision point for using additional therapy (ino) given the high ventilator cost of oxygenation. Boloker et al. 17 considered the potential toxicity of oxygen and maintained ideal preductal SaO 2 90 to 95%, but weaned FiO 2 to maintain an SaO 2 as low as 80% provided there was evidence of good end-organ function. Similarly, Kays et al. 20 allowed postductal SaO 2 as low as 60% (PaO 2 as low as 30 torr), as long as preductal SaO 2 was >85%, PaCO 2 <65 torr, and there was evidence of adequate postductal tissue perfusion. Most centers defined adequate postductal tissue perfusion as some combination of ph, serum lactate levels and urine output (Table 4). Pulmonary hypertension management Blood pressure management. With the wide spectrum of physiologic states noted in infants with CDH, the underlying pulmonary vascular and cardiac derangements may change significantly during the early postnatal course. PPHN is defined in one report as a peak pulmonary artery (PA) pressure to systemic

Congenital diaphragmatic hernia 543 Table 6 Surgical criteria and timing Clinical criteria for surgery Average age at surgery, latest epoch Surgery on ECMO: timing and survival Notes Bagolan (1996 2001) 3 epochs Boloker (1992 1999) Downard (2000 2002) Finer (1989 1995) Frenckner (1990 1995) Kays (1992 1998) Osiovich (1993 1999) Reyes (1993 1995) Somaschini (1994 1998) Ssemakula (1985 1996) Delayed repair 48 h with: PCO 2 <60, PO 2 >40, SaO 2 >85 (on FiO 2 <50%) Until PPHN resolved (no pre- to postductal gradient, confirmed by ECHO) and minimal ventilatory support. See 3rd epoch of care in Wung et al. Most recent epoch reported at Wilson s center. Delayed repair, aggressive treatment of pulmonary hypertension (with selective pulmonary vasodilators) Since 1993, surgery delayed until FiO 2 <40% PIP<24 cm H 2 O, MAP<12 cm H 2 O (with normal acid-base status and normal BP) See surgery on ECMO/timing Stabilized, then delay of 24 96 h (interval increased over time). Stable : adequate blood gases and no signs of right-to-left shunt (as per preand postductal blood gases and cardiac echo; while maintained on FiO 2 ¼ 0.40 with gentle ventilation Delayed at least 1 day. Continue delay if unstable or require significant ventilator support (same as previous era). If condition plateaus, repair quickly, then try to stabilize for a few hours, and if no success go to ECMO. Preference is to have repair before ECMO Surgery was delayed until infants were stable on minimal ventilation and after ECMO decanulation Neonatal and surgical teams jointly reviewed the patient to determine the optimal time for surgical intervention. Stable BP and HR, adequate peripheral perfusion, acceptable blood gases, adequate lung expansion on CXR, evidence of fall in pulmonary artery pressure Operation performed as an elective procedure after medical stabilization. Optimal blood gas with FiO 2 <0.40 and weaning from mechanical ventilation. Surgery while on HFOV If clinically stable, surgery at 24 48 h of life Preoperative pneumothorax significantly decreased in the third epoch (P<0.03) Numbers not reported. Used infrequently: 6/67 (9%) inborns and 10/53 outborns (19%) No chest tube. Infants undergoing repair on ECMO were treated preoperative with amino-caproic acid Judicious use of ECMO See Wilson et al. for details Non-ECMO: 5.8 days (mean) ECMO: 8.2 days (mean) Non-ECMO: mean 2.1 days (range 0.2 6.2 days) Post- ECMO: FiO 2 <0.40, PIP<25, IMV<40, delay of 4 days As long as patient is improving, delay for 3 5 days. ECMO patients: surgical repair after decanulation Total: (median) 1.4 (range 0.22 13.4) days Non-ECMO survivors: 1 day ECMO: 3.7 days After March 1993, no repairs on ECMO N ¼ 4, all survived to discharge, with one late death. Repair on ECMO when patient could not be weaned after 1 week on ECMO, patient was repaired on ECMO Patients placed on ECMO before surgery underwent repair 1 or 2 days before decanulation (OR) 2 days post-ecmo decanulation Patch repair: 79% survival, primary repair 96% survival (NS) Requirement: stability with loss of most edema, FIO 2 <0.4, PIP<24, MAP<12, normal acid-base balance, stable BP Postoperative pulmonary hypertension may be avoided by preoperative delays in surgery until pulmonary hypertension has subsided Kays comments on surgery: Would rather operate at 24 h of life in a baby destined for ECMO, and manage ECMO postoperatively than to further delay surgery and be forced to manage a newborn with unrepaired CDH on ECMO. 6 patches, 5 survivors. Age at surgery for those who eventually needed ECMO, 89 h (mean) 2.8±1.45 days 9 received Gore-tex patch survival: 9/28 (32%) 24 48 h after going on ECMO. Survival: 15/31 (48%) 17/98 required patch repair, 2 of 17 (12%) survived

544 Congenital diaphragmatic hernia Table 6 Continued Clinical criteria for surgery Average age at surgery, latest epoch Surgery on ECMO: timing and survival Notes Weber (1988 1997) Wilson (1991 1994) 3 epochs Wung (1983 1995) 3 epochs Delayed repair (n ¼ 24), or repair on ECMO (9 immediate repaired babies were transferred in during this epoch), or repair on ECMO (n ¼ 23) Surgery on ECMO, survival: 13/23 (56%) Delayed repair from 1987 to 1994. Yes. Details not reported ECMO utilization peaked from 1987 to 1991, and dropped in the permissive hypercapnia era (1991 1994) Epoch 3, most recent: delayed repair, no chest tube, gentle ventilation. n ¼ 18, no pre- to postductal SpO 2 gradient, no ECHO evidence of right-to-left shunt. 17/18 (94%) survived, only one needed ECMO. 4.2±1.8 days Epoch 1. Emergency surgery 14/17 (82%) survival Epoch 2. Modest delay (24 h) 21/28 (75%) both epochs, prophylactic chest tubes Abbreviations: BP, blood pressure; ECMO, extracorporeal membrane oxygenation; FiO 2, fraction of inspired O 2 ; HFOV, high-frequency oscillatory ventilation; MAP, mean airway pressure; PaCO 2, partial pressure CO 2 ; PIP, peak inspiratory pressure; PPHN, pulmonary hypertension; SaO 2, oxygen saturations. blood pressure ratio of 0.75 or greater. 21 Several centers largely ignored postductal SaO 2, and followed only acceptable preductal SaO 2 and postductal ph, as well as other clinical measures of postductal perfusion. 16,18 20,25,26 While none of the reports described aggressive efforts to override right-to-left shunt by increasing systemic vascular resistance, several centers offer guidance on blood pressure management for CDH patients with PPHN, with use of inotropic support (dopamine, dobutamine or epinephrine) described in six reports. 16,18,20,24 26 Among these, maintaining mean blood pressure >40 torr was common. One center maintained mean blood pressure >50 torr, as long as there was evidence of good tissue perfusion, such as good capillary refill, adequate urine output and normal serum lactic acid levels (Table 5). 18 Inhaled nitric oxide. The successful centers report variable practice and outcome with ino (Table 5). 16,17,19 22 Bagolan et al. 16 used ino to treat PPHN for a pre- and postductal oxygen saturation difference greater than 5% or refractory preductal hypoxemia, while other centers used ino only for refractory hypoxemia ; many of these patients went on to require ECMO. 19 21,27 Wilson et al., 25 in their most recent epoch, aggressively treated intrapulmonary shunt using ino with good results. 15 Osiovich 27 initiated a trial of ino at 20 parts per million for all infants with an OI >25. This center suggested at least some reduction in the need for ECMO, but offered that the response in CDH infants is less predictable than that in infants with classic PPHN associated with meconium aspiration syndrome. 27 Although there may be short-term benefit from the selective use of ino, the routine use of ino for all CDH infants is not supported by results from successful centers. 11,46,47 Furthermore, meta-analysis of randomized clinical trials evaluating the rescue use of ino in neonates with CDH suggests that ino may be associated with worse outcomes (that is, higher rates of ECMO use or death). 48 Sedation, paralysis and environment. Most successful centers reported maintenance of a minimal stimulation environment with a wide range of sedation practice, from deep sedation and paralysis to light sedation and use of pre- and postoperative epidural anesthesia. 16,18,20,22,25 27 The trend noted in these reports is the use of narcotic analgesia and avoidance of paralysis. Timing of repair. In the majority of successful centers, surgeons delayed repair until physiologic stabilization and improvement in PPHN. A few groups reported repair in the first few days, regardless of shunt state, and repaired on ECMO if necessary. 20,23 25,27 None of the centers emergently repaired the defect. Wung et al. 26 described an evolution of care, in which the earliest patients were treated with early repair and prophylactic chest tube, and a more recent epoch with delayed repair and no chest tube, noting concurrent improvement in survival. Bagolan et al. 16 and Boloker et al. specifically noted avoidance of chest tubes, while Kays et al. 20 attributed their center s improved survival in Era 3 for using postoperative chest tubes placed in water seal. The trend of these reports supports surgical repair on stable patients, whether on or off ECMO, with minimal or no PPHN. Questions regarding the use of chest tubes, repair on or off ECMO and benefits of various options for surgical reconstruction for large defects remain unanswered. Criteria and timing for surgical repair are itemized by center report in Table 6.

Congenital diaphragmatic hernia 545 Table 7 ECMO criteria and duration ECMO criteria VA ¼ Venoarterial, VV ¼ Venonenous VA or VV ECMO vent settings ECMO hours, survivors ECMO hours, non-survivors Notes Bagolan (1996 1901) Boloker (1992 1999) Finer (1989 1995) Frenckner (1990 1995) Kays (1992 1998) Osiovich (1993 1999) Reyes (1993 1995) Somaschini (1994 1998) Ssemakula (1985 1996) Weber (1988 1997) Wilson (1991 1994) Wung (1983 1995) Not an ECMO center No lethal anomalies Able to maintain preductal saturation>80% for 1 h Three OIs>0.40 within a 2-h window, with MAPX18 cm H 2 O VA only if internal jugular vein could not be cannulated with a double lumen VV cath Failure of conventional therapy, in combination with either OI>40 or rapid deterioration, BW>2000 g, gestational age>34 weeks, no IVH>grade I, no other major malformations Start ino (or ECMO) if: Cannot maintain preductal saturations>85% or postductal PaO 2 >30 Evidence of inadequate oxygen delivery (rising serum lactate), if ino unsuccessful, start ECMO VV, n ¼ 14, VA, n ¼ 24 VV, n ¼ 1 VA, n ¼ 6 Current preference: VV first, VA if no improvement. SIMV Rate 24 PIP ¼ 22 PEEP ¼ 4 I time ¼ 0.6 s VV ECMO: 98 h VA ECMO: 114 h 241 452 Cycle off ECMO: Postductal SaO 2 100% with PaCO 2 <60 on PIP 22, PEEP<5, rate<40, I time ¼ 0.5 s & FiO 2 <0.50 VV ECMO associated with lower mortality (not significant) An OI>40 ( 3 within 2 h), while on a MAP of 18 or greater Protocol: trial of ino at OI>25 Cardiovascular instability or Refractory hypoxemia Unresponsive to medical therapy and showed N ¼ 3, VA 216 h 348±24 3 ECMO patients. acute deterioration: OI>40, or 1 survivor PaO 2 <40 mm Hg (1) PaO 2 <40 2 h, ph<7.25 2h Repair on ECMO: Repair on 30% of ECMO Intractable hypotension, or 176 h ECMO: patients died with (2) OI>40 2, at least 30 min apart, or Repaired before 392 h PHN and (3) Significant air leak defined as bilateral ECMO: 144 h Repaired before respiratory failure. pneumothoraces with or without ECMO: 264 h Possibly k with pneumomediastinum and less aggressive pneumoperitoneum. ventilation Exclusion: <33-weeks gestation, (see Kays) BW<2000 g, bleeding/coagulopathy, major IVH, mechanical ventilation>10 days, uncorrected cardiac lesions, evidence of irreversible brain damage. Blood gases not used for exclusion Inclusion:, but pre-ecmo data available: VA PaCO 2 64.6±9.60, OI ¼ 35.5±10.10 Exclusion: gestational age<32 weeks, BW<1800 g, ICH>grade II, other medical or additional anomalies that precluded survival OI (OI ¼ FiO 2 mean airway pressure/ VA or VV 92.8±14 156±18 Peak usage PaO 2 )>0.40 1987 1991 Failure of conventional therapy 100% VA, goal to keep Judicious use of Preductal SaO 2 <95% with PCO 2 <50 mm Hg mixed venous O 2 ECMO

546 Congenital diaphragmatic hernia Table 7 Continued ECMO criteria VA ¼ Venoarterial, VV ¼ Venonenous VA or VV ECMO vent settings ECMO hours, survivors ECMO hours, non-survivors Notes observed at any time. Subsequent evidence of inadequate oxygen delivery, progressive metabolic acidosis, or organ failure. Use ECMO before inordinate lung distending pressure. An A-aDO 2 >600 mm Hg>4 h and evidence of adequate lung parenchyma saturation 65 70% Abbreviations: ECMO, extracorporeal membrane oxygenation; FiO 2, fraction of inspired O 2 ; HFCV, high-frequency conventional ventilation; ino, inhaled nitric oxide; OI, oxygenation index; PaCO 2, partial pressure CO 2 ; PaO 2, partial pressure of O 2 ; PEEP, positive end expiratory pressure; PIP, peak inspiratory pressure; MAP, mean airway pressure. ECMO. Indications for ECMO varied in these reports. Several centers used oxygenation index (OI), while others measured alveolar arterial O 2 difference or some combination of an OI and a pre-specified level of respiratory support. 18 20,22,23,25,27 Several centers made ECMO decisions based on evidence of poor systemic perfusion, non-reassuring blood gases, intolerable ventilator parameters or rising serum lactates (Table 7). 17,20,21,26 Some centers only considered ECMO in infants that met well-defined criteria, including evidence of adequate lung parenchyma for postnatal survival, suggested by at least a brief period of adequate preductal oxygenation and/or ventilation (that is, PaCO 2 <50). 17,26 Finer et al. 18 considered ECMO for infants with three OIs >40 (within 2 h) while maintaining MAP>18 cm H 2 O. Osiovich 27 used the same criteria and considered an OI >40 on three consecutive blood gas measurements a failure of conventional therapy (highfrequency ventilation included) necessitating the need for ECMO. Kays et al. s 20 use of ECMO decreased slightly with adoption of lung-sparing strategies, from 50% (7/14) in Era 2 (1988 to 1992) to 39% (23/59) in Era III (1992 to present), the lung-sparing era. However, ECMO survival improved significantly in the same center, from 28% (2/7) in Era 2 to 83% (19/23) in Era 3. Wung et al. 26 reported reduced ECMO use and improved survival with similar practice changes. Their ECMO criteria mandated evidence of adequate ventilation and preductal oxygenation (SaO 2 >85%, PaCO 2 <50 torr). Wilson et al. 25 demonstrated improving survival concurrent with increased ECMO use from 48 to 71%, increased ECMO run times from 132±21 to 220±21 h, and increased survival after instituting a lung-sparing strategy and surgery on ECMO with amino-caproic acid (to minimize hemorrhagic complications). Subsequently, this group has reported decreased ECMO use (14/39, 36%) and a consistently high survival rate (36/39). 15 Discussion The high survival rates reported by these centers suggests that instituting well-defined treatment guidelines, including lungsparing mechanical ventilation strategies and specific criteria for initiating ECMO, may lead to an increase in CDH survival. The details of these reports, however, identify a wide range for each aspect of care. Avoiding hyperventilation and minimizing lung injury by use of gentle ventilation, first advocated by Wung et al., 17,26,49 was common to most of these reports, but the range of PIP limits is wide. 16,18,20 22,27 The reports also support the use of HFV, particularly in those infants requiring higher PIPs with CMV, but they do not provide definitive evidence that this is a life-saving strategy. 16 22,25 27,49 51 Clinicians may be reassured that nonideal blood gases, particularly postductal blood gases, may be tolerated as long as there is some evidence for adequate tissue perfusion. Furthermore, despite previous reports suggesting the futility of ECMO use, 52,53 these reports, consistent with evidence from the CDH Study Group, suggest that some high-risk CDH infants may reach a threshold of supportive care at which ECMO may improve the chances of survival. 17,20,21,25 27,54 These center reports also acknowledge several detrimental strategies in the care of CDH infants in previous eras. Many of these are cataloged in a clinical commentary by Bohn in 2002. 55 Among the most detrimental practices, targeting optimal blood gas values involved aggressive treatment of pulmonary hypertension via hyperventilation-induced alkalosis. This strategy attenuates rightto-left ductal and atrial shunting by decreasing pulmonary vascular resistance, but has been associated with significant ventilator-induced lung injury on clinicopathologic correlation. 56 59 The use of high distending pressures to achieve ideal oxygen saturations, particularly postductal oxygen saturations, is also acknowledged indirectly in these center reports as a strategy to avoid. Another historically important detrimental strategy is the treatment of CDH as a surgical emergency. The literature, including these combined reports, now supports the view that surgical repair of CDH should be performed on an elective basis, when the patient has been stabilized medically. 1,60,61 The improvements in survival noted in serial epochs of care at these centers supports the notion, particularly in infants with moderateto high-risk CDH, that these strategies are best avoided.

Congenital diaphragmatic hernia 547 While there is variation in reporting patterns, some consideration should be given to hidden mortality when evaluating CDH survival statistics, and it is important to acknowledge the potential for bias in reports with historical controls. Some of these center reports reflect a referral patient population exclusively, and population-based studies have demonstrated a hidden mortality associated with outborn CDH infants. 62 That being said, reporting patterns for centers with obstetric wards are also susceptible to biases, and population centers vary significantly with regard to prenatal diagnosis, termination rates and other factors that contribute to hidden mortality. 62 Referral-only centers had a combined survival for isolated CDH of 171/204 (84%), and centers that include both inborn and outborn CDH infants had a combined survival for isolated CDH of 390/457 (85%). In spite of these weaknesses, it is our view that the combined reports reflect a considerable improvement in survival over previous eras, and may provide an important tool for clinicians caring for these high-risk infants. While several aspects of care are common among these centers reports, their success may be due in part to identification of infants at greater risk, perhaps from timely selection of the appropriate support. We hypothesize that the increased survival noted in these centers represents improvement in a subpopulation of high-risk CDH infants, which in previous eras died from an overly aggressive ventilatory strategy or inadequate treatment of potentially treatable aspects of the underlying pathophysiology. Earlier reports suggest that as much as 25% of the observed mortality in CDH is due to potentially treatable aspects of the underlying pathophysiology. 47,56,57,63,64 Future investigations, either observational cohort studies or randomized clinical trials, should stratify analyses based on the categories of illness severity, including anatomic indices and physiologic responses to therapy, rather than simply isolated CDH or CDH with other anomalies. Determination of the appropriate therapies may be aided by the development of a supportive care grid, in which severity of illness and current level of cardio-respiratory support are plotted vs measures of cardiac performance, tissue perfusion and acid-base status. This approach may also aid in development of consistent indications for maintaining, escalating or weaning current therapies. Other aspects of care not explicitly addressed in these center reports still deserve some mention here, as there are certain aspects of the pathophysiology of CDH that have just begun to be explored. Specifically, there appears to be a subpopulation of CDH infants that may be adequately ventilated (removal of excess CO 2 ), but still have impaired oxygen exchange. These infants have varying degrees of pulmonary hypertension that may lead to impaired cardiovascular performance over time. 65 Additionally, there is evidence that in a subset of CDH infants, cardiovascular function is impaired even in the first hours of life. 66 68 Some investigators speculate that, in this subgroup of CDH infants, unloading the myocardium by maintaining patency of the ductus arteriosus may improve acute and chronic cardiopulmonary performance. 66 Future clinical trials may reveal that long-term CDH survival is improved not only by minimizing excessive use of distending pressures, but also by relieving right and left ventricular strain, and other strategies aimed at augmenting cardiac output. 55 Improving survival to consistently greater than 85% may be difficult, as there is still a subpopulation of CDH infants, preterm infants or infants with bilateral severe lung hypoplasia, which are not likely to benefit from any postnatal strategy. Some institutions have used best PaO 2, OI and various ventilation indices for prognostication, and each of these methods have been used to determine the appropriate level of support. 69 In previous eras, these were used as targets to be reached with escalating ventilatory support, to guide decision making. This approach has been largely abandoned because of mounting evidence that even brief periods of hyperventilation may do irreversible damage. 26,49,55 58 The CDH Study Group has published a logistic regression equation that estimates the mortality risk in infants with CDH using parameters available in the first 5 min of life, including birth weight and APGAR scores. 54 Ideally, infants will be classified by both clinical and anatomic criteria, including such tools as the modified McGoon index (MMI). 70 The MMI, reflects (indirectly) the degree of lung hypolasia by indexing the PA diameters to the diameter of the aorta (MMI ¼ [PA1 þ PA2]/Ao). Combining an anatomic index with echocardiographic measures of PA pressure, cardiac function/perfusion and oxygenation (especially OI, which accounts for the ventilator cost of oxygenation) may allow the clinician to better categorize CDH infants based on mortality risk. 46,70 Our interpretation of these reports is that consistency among providers within individual centers may be an essential element of successful CDH management. A consistent approach, with predetermined criteria for use of ECMO and an underlying protect the lung strategy may be as important to the care of these infants as the specific medical interventions chosen. We agree with Finer et al. 18 that a consistent approach is facilitated by development of multidisciplinary standardized treatment guidelines, including input from neonatology, pediatric surgery, ECMO specialists and respiratory therapy. In the absence of much-needed multicenter randomized clinical trials to test specific aspects of the care for CDH infants, development of uniform practices within centers based on successful center reports, and monitoring consistency of practice and physiologic and neurodevelopmental outcomes, remain the best means of optimizing care for infants with CDH. References 1 Clark RH, Hardin Jr WD, Hirschl RB, Jaksic T, Lally KP, Langham Jr MR et al. Current surgical management of congenital diaphragmatic hernia: a report from the congenital diaphragmatic hernia study group. J Pediatr Surg 1998; 33(7): 1004 1009.