Is delayed surgery really better for congenital diaphragmatic hernia?: A prospective randomized clinical trial

Is Delayed Surgery Really Better for Congenital Diaphragmatic Hernia?: A Prospective Randomized Clinical Trial By M.N. de la Hunt, N. Madden, R.J. Bray, E. Cusick, L. Gardner,

J.E.S. Scott, J.N.S. Matthews, J. Beck, C. Sadler, A.M. Barrett, S.A. Boddy, S.A. Hargrave, W. Hinton, L. Rangecroft, R. Spicer, M. Stafford, D. Thomas, C.J. Vallis, and J. Wagget Leeds, England and Newcastle Upon Tyne, England

0 Delayed surgery has become widely accepted in the management of congenital diaphragmatic hernia after comparing outcomes only with historical retrospective controls. It was the aim of this study to compare early and delayed hernia repair in a randomized prospective clinical trial. Fiftyfour infants were randomized to receive either early repair (within 4 hours of admission) or delayed repair (more than 24 hours after birth). The survival rate was higher for the delayed group (57% v 46%), but the difference was not -11; 95% confidence limits: -37.5, significant (difference: 15.5). There were no significant differences between the two groups with respect to length of hospital stay, ventilator dependency, or survival time. Recorded preoperative risk factors were similar for the two groups. Eight infants in the delayed repair group died without having undergone surgery. The optimum time for surgery still needs clarification. Copyright o 1996by W.B. Saunders Company INDEX surgery,

WORDS: pulmonary

Congenital diaphragmatic hypertension.

hernia,

delayed

T

HE MORTALITY RATE for posterior congenital diaphragmatic hernia (CDH) is still about 50% with conventional treatment, despite major advances in intensive care. The philosophy behind surgical treatment has changed radically in recent years. Immediate surgery, once considered mandatory, is now deemed harmful.’ Assuming that severe pulmonary hypoplasia and pulmonary hypertension are the main determinants of survival, and that the mechanical effects of the hernia can be controlled with conservative measures, it is rational to avoid additional surgical stress until the unstable pulmonary circulation becomes less reactive. This hypothesis has not been tested in prospective randomized clinical trials, but has been accepted on the basis of published reports using retrospective controls,*-6 of which only one has shown a higher survival rate.6 It may be that with antenatal diagnosis and better general supportive care, an improvement in surgical

From the Departments of Paediatric Surgery, The Royal Victoria Infirmary, Newcastle Upon Tyne, and The General Infirmary at Leeds; and the Department of Medical Statistics, University of Newcastle Upon Tyne, England. Address reprint requests to M.N. de la Hunt, MS, FRCS, Department of Paediatn’c Surgery, Royal ‘Victoria Infirmary, Newcastle Upon Tyne, NE1 4LP, England. Copyright o 1996 by W.B. Saunders Company 0022-3468J96 j3111-0019$03.00/0 1554

results has been offset by the presentation of a higher-risk population. The aim of the present study was to compare early and delayed CDH repair in a prospective randomized clinical trial. MATERIALS

AND

METHODS

Infants with CDH who presented within 6 hours of birth were considered eligible for the study. At the time of referral to the receiving surgical center (Northern and Yorkshire Regional Neonatal Surgical Units), they were randomized to receive either early surgery or delayed surgery. Those with antenatally suspected CDH were randomized as soon as the diagnosis was confirmed after birth. All outborn babies were transferred to the trial centers as soon as was reasonably possible. After randomization, patients were excluded only if subsequently proven not to have CDH. Analysis was based on intended treatment. The early group had CDH repair within 4 hours of arrival at the receiving surgical center, which enabled preoperative resuscitation. Those in the delayed group had repair no sooner than 24 hours after birth, and when considered stable, which generally meant receiving less than 50% inspired oxygen, with minimal vasodilator or inotropic support for 24 hours or more. General supportive and respiratory care, including criteria for use of pulmonary vasodilators and inotropes, were the same for the two groups, both before and after surgery. Ventilation was begun with pressures of 20 cm Hz0 or less, at a rate of 40 per minute. To achieve optimal ventilation, the rate was altered initially, increasing pressure only when this failed to achieve improvement. Apart from the difference in timing, the surgical approach and procedure were the same for both groups. The main measure of outcome was survival at 6 months. Length of hospital stay and ventilator dependence also were considered. Preoperative risk factors were recorded, including antenatal diagnosis, age at presentation, blood gas values, and ventilation requirements. A target was set of 60 patients, in less than 4 years, to exclude with reasonable certainty an error in favor of the inferior treatment. Results were compared using the x2 test with Yates correction, and the unpaired Student’s t test. Some variables, such as hospital stay, were skewed and were log-transformed before analysis, enabling comparison of geometric means. The protocol for this study was approved by the Ethics Committees of both participating hospitals. RESULTS

From the start of the trial in January 1991, 54 infants were recruited. The median birth weight was 3,000 g (range, 1,620 to 4,735) and mean gestational age was 38 weeks (range, 31 to 42). Thirty two cases (59%) had been diagnosed antenatally. Gestational age at the time of antenatal diagnosis was docuJournalofPediatric

Surgery,

Vol31,

No 11 (November),

1996: pp 1554-1556

DELAYEDSURGERYFORCDH

Table

1. Clinical

1555

Data:

Early Group

Versus

Delayed

Early Surgery

Birth weight(g) Gestational age (wk) Antenatal diagnoses

(n = 28) 3,198 (589) 38.5 (1.9)

PH 0, (kpa) COP (kPa)

6(1 to264) 7 (26%)

48 (25 to 240) 5 (18%)

7.38 (0.15) 9.1 (5.3)

7.37 (0.18) 9.7 (7.3)

5.6 (3.0)

5.9 (3.8)

as total wrth percentage,

as mean with SD,

Table

2. Measures

Early Group

NOTE. geometric

(h)

Data are expressed mean.

Delayed

Group

Surgery (n = 26)

12 (46%) 20.7

16 (57%) 32.7

.42 (-37.5, 15.5) .12 (0.35, 1.13)

5.8

10.7

.lO (0.26,

1.14)

82.4

37.1

.21 (0.13,

1.62)

Early

Age at trme of death

Versus

Delayed Surgery (n = 28)

as geometric

P Value

(Cl)

mean with 95% Cl for ratio of

3,059 (703)

.67 (-276,428)

38.5 (1.9) 14 (50%)

18 (69%)

.15 (-44.9,6.4)

Inborn Patch reparr

13 (46%) 2 (7%)

16 (62%) 10 (38%)

.27 (-42.2, .006 (-52.3,

Best blood <4h

37.8 (2.5)

.26 (-0.52,

1.9)

26.2) -10.3)

gas

PH PO, (kPa) Pco, (kPa)

of Outcome:

Survivors Hosprtal stay (d) Ventilatron (d) Nonsurvrvors

3,134 (572)

age IW Antenatal diagnosis

NOTE.

mented in only 13 cases, but it did not differ significantly between the groups (early: mean, 21 [SD, 71 weeks; late: 23 [SD, 71 weeks) or between survivors and nonsurvivors (26 [SD, 91 weeks; 20 [SD, 41 weeks). The early and delayed groups were similar with respect to recorded preoperative or operative risk factors, including age at time of diagnosis, referral, and arrival (Table 1). Measures of outcome are provided in Table 2. The survival rate was higher for the delayed group, but the difference was not significant (57% v 46%; difference: - 11; 95% confidence limits for difference in survival percentage: -37.5, 15.5). There also was no significant difference in the survival time of nonsurvivors or in the length of hospital stay or ventilator use among survivors. Nine patients in the delayed group died before surgery. All in the early group survived until operation. Survivors are compared with nonsurvivors in Table 3. Early preoperative blood gas measurements, taken in the first 4 hours of life (and thus not likely to be influenced by treatment group), were significantly better for survivors. Large defects, particularly those that required patch repair, were associated with a significantly higher mortality rate (difference: 31.3%; 95% confidence limits: 10.3, 52.3) Antenatal diagnosis, including gestational age at that time, was not a useful prognostic indicator in this group of patients.

P Value (95% Cl)

(n = 26)

Brrth werght (9) Gestational

20 (71%) 18 (64%)

and Nonsurvivors

NOllSUWlVCVS

(n = 28)

(n = 26)

(h)

for Survivors

SUWlVOrS

2,987 (672) 37.8 (2.5)

Patch repair Best blood gas < 4 h

NOTE. Data are expressed or as median with range.

Table 3. Risk Factors

Delayed Surgery

12 (46%) 11 (42%)

Inborn Age at time of operation

Group

7.45 (0.12) 11.2 3.9

Data are expressed

rncrdence with percentage. mean wrth Cl for the ratio

7.29 (0.17) 5.2 6.5 as mean

.0005 (0.07, 0.24) < .OOOl (1.6, 2.9) .0003 (0.45, 0.77)

with

PO, and Pcol

standard

are expressed

deviation

or as

as geometric

There was one major deviation from protocol. In the early group, surgery was delayed for 11 days in an infant who had multiple congenital abnormalities (anorectal, urogenital, spinal, and cardiac), while chromosome analysis was performed. This infant died at 21 days of age. All other patients in the early group had CDH repair within 11 hours of birth. There were 10 late deaths (occurring after more than 5 days), six in the early group and four in the delayed group. These cases might have been salvageable. Of the six in the early repair group, three had contralateral pneumothorax, two had pneumonia documented at the time of necropsy, and one had multiple congenital abnormalities including congenital heart disease (see above). In the delayed group, one died without surgery at 192 hours (never having achieved stability), one died at 149 hours (having had a patch repair and contralateral pneumothorax), and the other two were the only infants in this study who had extracorporeal membrane oxygenation (ECMO). Necropsy information (including lung weight) was available for 10 patients. Lung weights were compared with the expected values reported by Gruenwald and Minh.’ In seven, all of whom died on the first day of life, lung weight was less than 50% of the expected value. Lung weight was 81% of the expected value in the patient who had cardiac and other anomalies. It was 94% in one early-group patient who died after sudden unexpected deterioration (at 30 days) and was 79% in the delayed-surgery infant who died at 192 hours with associated contralateral pneumothorax. DISCUSSION

The case for delayed CDH repair on clinical grounds remains unproven. It is a more convenient

1556

DE L4 HUNT

strategy when operating time is tightly controlled, and thus far it has not been shown to compromise a good outcome. Our results support this. To our knowledge, the only other published prospective study, which included 29 patients, most of whom received ECMO, showed a similar, slightly better (but not statistically significant) survival rate for the delayed group.8 A statistically significant difference between the two groups was not expected; this would take considerably more patients. Such an extensive trial would involve many more centers, or take considerably longer, and both these factors would result in greater diversity among the patients and other aspects of management, which would influence outcome. The present study was ended before the target of 60 patients was reached, largely because of difficulty with recruitment after commencement of the UK Collaborative ECMO Trial, with entry criteria of oxygenation index of greater than 40 or Pacoz of greater than 12 kPa for more than 3 hours in infants born after 35 weeks’ gestation with a birth weight of more than 2 kg. Delayed surgery selects a lower risk group for operation. Nine of our patients died before operation. With the wide acceptance of delayed surgery, there is increasing reluctance to transfer the unstable newborn with CDH to a surgical centers, particularly if the prognosis is poor, and this involves separation from the mother. This will tend to exaggerate the selection bias in favor of delayed surgery in nonrandomized or retrospective studies. There were highly significant differences in the early blood gas values of survivors and nonsurvivors.

ET AL

These measurements were taken in the course of clinical management and not under specified preconditions for trial purposes. They represent a mixture of preductal and postductal values, sampled from various sites, and because of the time bands studied, they were often taken before arrival at the trial centers. Despite this crude comparison, the lack of a difference in values between the early and delayed groups, and the large difference between survivors and nonsurvivors, suggests that disease severity had a much greater influence on survival than does surgical strategy. Necropsy data, where available, confirmed severe pulmonary hypoplasia, making the prospect of survival remote for most of these patients, even if ECMO had been used.9 Radical new approaches are still needed for the patients who are not salvageable with conventional therapy. Antenatal diagnosis was a poor predictor of outcome, even when the defect was identified before 20 weeks’ gestation and accompanied by polyhydramnios. For these patients, it would have been an inappropriate means of selection for fetal intervention.lO The longer hospital stay and ventilator dependence in the delayed-group survivors, although not significant, might suggest that the delayed-repair patients were a higher-risk group or that surgery was delayed too long. Considerably more subjects are required to resolve these issues. There is still much to be done to determine the optimum time for repair of “curable” CDH. We recommend that future studies compare different specified physiological criteria for timing of surgery.

REFERENCES 1. Langer JC, Filler RM, Bohn DJ, et al: Timing of surgery for congenital diaphragmatic hernia: Is emergency operation necessaty? J Pediatr Surg 23:731-734,1988 2. Cartlidge PHT, Mann NP, Kapila L: Preoperative stabilisation in congenital diaphragmatic hernia. Arch Dis Child 61:12261228,1986 3. Hazebroek FWJ, Tibboel D, Bos AP, et al: Congenital diaphragmatic hernia: impact of preoperative stabilization. A prospective pilot study in 13 patients. J Pediatr Surg 121139-1146, 1988 4. Shanbhogue LKR, Tam PKH, Ninan G, et al: Preoperative stabilisation in congenital diaphragmatic hernia. Arch Dis Child 65:1043-1044,199O 5. Goh DW, Drake DP, Brereton RJ, et al: Delayed surgery for congenital diaphragmatic hernia. Br J Surg 79:644-646,1992

6. Charlton AJ, Bruce J, Davenport M: Timing of surgery in congenital diaphragmatic hernia. Anaesthesia 46820.823,199l 7. Gruenwald SS, Minh HN: Evaluation of body and organ weights in perinatal pathology. Normal standards derived from autopsies. Am J Clin Path01 34247~253,196O 8. Nio M, Haase G, Kennaugh J, et al: A prospective randomized trial of delayed versus immediate repair of congenital diaphragmatic hernia. J Pediatr Surg 29:618-621,1994 9. Stolar CJH, Snedecor SM, Bartlett RH: Extracorporeal membrane oxygenation and neonatal respiratory failure: Experience from the Extracorporeal Life Support Organization. J Pediatr Surg 26:563-571,199l 10. Harrison MR, Langer JC, Adzick NS, et al: Correction of congenital diaphragmatic hernia in utero, V. Initial clinical experience. J Pediatr Surg 25:47-57,199O