Doppler echocardiographic assessment of pulmonary circulation in severe respiratory failure of the neonate: An aid for extracorporeal lung support indications

Doppler Echocardiographic Assessment of Pulmonary Circulation in Severe Respiratory Failure of the Neonate: An Aid for Extracorporeal Lung Support Indications ByJ.F. Germain,

I. Casadevall, L. Desplanques, J.C. Mercier, J.F. Hartmann, and F. Beaufils

Paris, France l Extracorporeal lung support (ECLS) for newborns with acute respiratory failure has achieved increased popularity over the last decade. However, precise criteria for its implementation remain controversial. The aim of this study was to assess the value of Doppler echocardiography (DE) in 31 neonates with Paor of Sbll mmHg, F10r of 1, and optimal ventilation. Treatment included mechanical ventilation, paralysis, vdume loading, vasopressors, and tolazoline. Markers indicative of ECLS (failure of maximal medical therapy, assessed by AaD of more than 610 mm Hg beyond 8 hours and/or an oxygenation index (01 = mean airway pressure x Ro2% / postductal Pao2) of more than 40 beyond 4 to 6 hours) were present in 23 (group 1) and absent in eight (group 2). Shunt direction

and systolic pulmonary

arterial

pressure (sPAP) calculated from tricuspid insufficiency velocity were assessed using DE. At the time of admission, sPAP was significantly

higher in group 1 (62.1 Y 43.7 mm Hg). On

day 1, group 1 differed from group 2 in maximum sPAP value (73.2 Y 44.4 mm Hg), Pacer (56.1 v 46 mm Hg), right-to-left shunting (85% v 25% of the patients), and pulmonary-tosystemic-pressure systolic ratio (sPAP:sSAP) (1.28 v 0.75). Patients with an sPAP:sSAP ratio of more than 1 and patients with high sPAP associated with high Pacer on day 1, all later (average, IO hours later) fulfilled ECLS criteria; this suggests that DE assessment of pulmonary circulation may yield early and predictive markers of impending ECLS indication. Further confirmation of these results would help avoid unecessary delays in ECLS implementation in newborns with severe respiratory failure. Copyright o 1994 by W. B. Saunders Company INDEX WORDS: Pulmonary hypertension, persistent, neonatal; extracorporeal membrane oxygenation; Doppler echocardiography.

A

CUTE NEONATAL respiratory failure is often complicated by severe hypoxemia. Refractory hypoxemia may result from major ventilation/perfusion mismatch owing to parenchymal injury or from right-to-left shunting through the foramen ovale and/or the ductus arteriosus secondary to suprasystemic persistent pulmonary hypertension. Both mechanisms may be present. Precise recognition of either mechanism is important; the former justifies changes in ventilator settings to recruit functional pulmonary units, and the latter requires medical treatment aiming at pulmonary vasodilation and systemic vascular support.’ Since 1975, full-term neonates with potentially reversible lung disease whose conditions do not respond to maximal medical therapy are increasingly being considered for extracorporeal lung

JournalofPediatric

Surgery, Vol29, No 7 (July), 1994: pp 873-877

support (ECLS). Criteria for ECLS have been based mainly on a predicted mortality of 80% or more, and vary somewhat from institution to institution.’ They usually combine persistently abnormal oxygenation and ventilation parameters, such as the alveoloarterial oxygen gradient (AaDO# and the oxygenation index (OI).3 However, these parameters do not directly assess the status of the pulmonary circulation; this can be evaluated more directly by twodimensional and Doppler echocardiography. Thus, the purpose of the present study was to investigate the pulmonary circulation using Doppler echocardiography in neonates with acute respiratory failure complicated by refractory hypoxemia. Specifically, the aims were to assess the degree of pulmonary hypertension, to compare the results with the usual criteria for ECLS, and to evaluate the contribution of Doppler echocardiography in determining ECLS candidates. MATERIALS AND METHODS

Patients The pulmonary circulation was prospectively investigated in 31 nearly consecutively born infants, referred to the pediatric intensive care unit (PICU) at Hopitat Robert Debre in Paris between October 1989 and July 1991. The diagnosis of acute respiratory failure was based on the following criteria: absence of congenital heart disease after echocardiographic and Doppler assessment and hypoxemia (Pa02 < 50 mm Hg) while breathing 100% oxygen, despite optimal positive pressure mechanical ventilation for more than 12 hours.

Treatment All patients were sedated (midazolam or fentanyl) and paralyzed (vecuronium). When profound hypoxemia was associated with left-to-right shunting and there was no evidence of pulmonary hypertension, mechanical ventilation was adjusted to recruit functional pulmonary units, ie, stepwise increase in positive endexpiratory pressure. On the other hand, patients with criteria of

From the Service de Rkanimation Pkdiatrique Polyvalente, H6pital Robert Debrk, Park, France, and the Facultk de Mkdecine Xavier Bichat, Universitk Paris VI, Paris, France. Date accepted: July 19, 1993. Address reprint requests to J.F. Germain, MD, Service de Rkanimation Pkdiatrique Polyvalente, Hbpital Robert Debr6, 48 Bd Stkoier, 75019 Paris, France. Copyright 0 1994 by W B. Saunders Company 0022-3468/94/2907-OOG6$03.006$o3.oo/0

873

874

persistent pulmonary hypertension of the newborn were managed with maximum medical therapy,’ ie, hyperventilated with high rates ( 2 80 per minute) high peak inspiratory pressures (2 35 cm HzO), and I+02 of 100%. Metabolic acidosis was corrected by sodium bicarbonate. Volume loading was used in cases of hypovolemia and/or with the aim to increase cardiac output and systemic pressure. Newborns without heart failure were treated by systemic vasopressors (dopamine or norepinephrine) if hypotension failed to respond to volume loading. Newborns with heart failure were treated with dobutamine or epinephrine. Tolazoline trials for patients with pulmonary hypertension were attempted only when the systemic hemodynamic status had been successfully stabilized. The indication for ECLS was the failure of maximum medical therapy (as described). ECLS was used if AaDO* was greater than 610 mm Hg (AaD = [barometric pressure - 471 x FIOZ- Pacol IF102 + (1 - F102llO.8) - Paoz) for more than 8 hours and/or if the 01 (mean airway pressure x Floz%/postductal Pao2) exceeded 40 for more than 4 to 6 hours. Patients with acute deterioration or untractable hypercarbia, were also treated by ECLS. Doppler echocardiographic findings were purposefully not included in the decision-making process for ECLS indication. The usual contraindications for ECLS were respected, such as gestational age of less than 35 weeks, body weight of less than 2,000 g, caryotypic abnormalities, nonreversible pulmonary injury, and severe neurological abnormalities. Generally, ECLS was performed using a veno-venous approach, with a single-lumen cannula, a nonocclusive pump, and an alternate clamp, in combination with apneic oxygenation of the native lung.5 Venoarterial bypass (ECMO) using the same type of equipment, without an alternate clamp or apneic oxygenation, was preferred in cases of congenital diaphragmatic hernia. The extracorporeal circuit and the details of ECLS patient management have been described previously.5-7

Investigations Systemic arterial pressure (sSAP) was measured from radial and/or umbilical arterial catheters; central venous pressure was measured from umbilical venous catheters connected to transducers and a multimodular monitor (Merlin system; Hewlett-Packard, Biihlingen Germany). Blood gases were monitored noninvasively and continuously, using dual preductalipostductal pulse oxymetry (N 200; Nellcor, Hayward, U.S.A.) and combined transcutaneous PO2iPco~. In addition, arterial blood gases from radial and/or umbilical arterial catheters were measured at least every 6 hours and every time Doppler echocardiography was performed. All infants underwent sequential echographic studies using a combined two-dimensional echo-Doppler system (sono layer SSH160A, Toshiba, Tochigi-ken, Japan). The studies were performed as soon as possible after initial stabilization and repeated daily as needed during the acute phase of the pulmonary illness. Each echocardiographic study involved a qualitative assessment of left-toright or right-to-left shunting at the foramen ovale or ductus arteriosus levels, using color-coded Doppler. Systolic pulmonary arterial pressure (sPAP) was measured by continuous Doppler, using the modified Bernoulli principle8,9; when tricuspid insuficiency (TI) was present, sPAP was calculated (sPAP = 4 [peak velocity TI, m/s]- + central venous pressure). The sPAP:sSAP ratio was calculated.

GERMAIN ET AL

RESULTS

Thirty-one newborns (18 males, 13 females) met the criteria for the diagnosis of acute respiratory failure. Twenty-three of them met ECLS criteria (group 1); the 0th er eight had high values of AaDOz and/or 01 but of insufficient duration to meet ECLS criteria (group 2). Table 1 shows the general characteristics, physiological data, ventilator settings, and echocardiographic and Doppler assessment at the time of admission. There were no significant differences between the two groups with respect to birth weight (3,237 4 483 g v 3,174 2 705 g), time of intubation, and time of admission. Group 1 had lower Apgar scores at 1 (3.7 ? 3.1 v 7.2 2 2.7) and 5 minutes (5.6 + 3.3 v 9.1 -C 1). Although the gestational period was longer for group 1 (39.4 ? 1.6 weeks v 36.8 + 3 weeks), there was no term contraindication to ECLS in either group. Ventilator settings, including mean airway pressure, were the same in the two groups, except for Table 1. General Characteristics,

Ventilator Settings, Blood Gases,

Severity Indexes, and Doppler Echocardiographic Parameters at Time of Admission No. of Patients (n = 31)

Group 1 (n = 23)

Group 2 (n = 8)

4(13)

2 (91

2 (25)

Primary diagnosis: No. (%) Hyaline membrane disease Meconial aspiration syndrome (MAS)

8 (2’3)

8 (35)

0 (0)

Neonatal sepsis (NS)

7 (23)

3 (13)

4 (50)

MAS and NS Congenital diaphragmatic

5 (16)

4 (17)

1 (12)

5 (16)

4 (17)

1 (12)

2 (6)

2 (9)

0 (0)

35 + 8

38 f 6.8

27 + 3.2

15.7 + 4

16 2 4.3

14.8 ? 2.9

hernia Undetermined diagnosis Ventilator settings Peak inspiratory pressure (mean cm H20 + SD)* Mean airway pressure (mean cm Hz0 2 SD) Blood gases and severity indexes Pao, (mean mm Hg * SD)

54 + 24

54 + 27

Pace, (mean mm Hg f SD)

49 + 21

49.5 2 22

AaDO

532

13

47 + 20

(mean mm Hg

+ SD) 01 (mean r SD)

5642121

579?78

511 + 221

34 2 25

36 2 28

28&

14

55.5 + 18

62 + 18

44+

12

Doppler echocardiographic parameters sPAP (mean mm Hg t SD)* sPAP:sSAP ratio (mean * SD)

Statistical Analysis Data are presented as mean 2 standard deviation. The features of the two groups were compared using the unpaired t test, Yates’ corrected x2 test, Fisher’s exact test, or a nonparametric test. The significance level is OL= .05.

0.99 c 0.38

1.08 -t 0.42

0.83 2 0.26

55.5

74

12.5

Foramen ovale and/or ductus arteriosus rightto-left shunting (% of patients)’

*Significant difference between groups 1 and 2.

DOPPLER ECHOCARDIOGRAPHY

AND PULMONARY HYPERTENSION

peak inspiratory pressure, which was higher in group 1 (37.9 v 27 cm H,O; P = .002). There was no difference with respect to blood gas results, AaD02, 01 or sSAP. Doppler echocardiographic results differed significantly with respect to mean sPAP, which was higher in group 1 (62.1 v 43.75 mm Hg; P = .02), and frequency of right-to-left shunting through the foramen ovale and/or the ductus arteriosus (73.7% for group 1 v 12.5% for group 2; P = .Ol). All patients with sPAP of more than 60 mm Hg were in group 1. We studied the poorest values of the above parameters on the first day in the pediatric intensive care unit (PICU). The results are summarized in Table 2. Although volume loading, drug therapy, and urine output were the same in both groups, there were statistically significant differences between the groups with respect to peak inspiratory pressure (38.3 v 29 cm H,O; P = .002), maximal PaCoz (56.1 v 40 mm Hg; P = .03), sPAP (73.2 v 44.4 mm Hg; P = .OOOl), and right-to-left shunting through the ductus arteriosus and/or foramen ovale at the time of the highest sPAP (85% v 25% of the patients; P = .OOl). On day 1, sPAP could be evaluated in 22 of the 31 newborns; 14 were in group 1, and eight were in group 2. For the Table 2. Poorest Values on Day 1 in the PICU: Ventilator

Settings,

Blood Gases, Severity indexes, and Doppler Echoeardiographic

Parameters

No. of Patients

Group 1

(n = 31)

(n = 23)

Group 2 In = 9)

Ventilator settings Maximal peak inspiratory pressure (mean cm H,O 2 SD)*

36 f 8

38.3 2 7

28 ‘-

4.9

Maximal mean airway pressure (mean cm HZ0 2 SD)

16.4 -’ 4.4

16.7 -t 4.7

15.4 + 3.4

Blood gases and severity indexes Minimal Paoz (mean mm Hg 2 SD)

47 f 11.5

45 * 10.5

53 t- 12.5

52 & 18

562

40 t 6

875

Fig 1. Maximum sPAP:sSAP ratios for groups 1 (m) and 2 (0) on day 1 in the PICU. Each bar represents one patient.

other nine patients, there was lack of tricuspid insufficiency (n = 5), or ECLS had been started shortly after admission, as a lifesaving procedure, before any measurements could be obtained (n = 4). In addition, the sPAP:sSAP ratio was higher in group 1 (1.29 v 0.75; p = .OOOl).The extreme values for both groups did not overlap (Fig 1). The value exceeded 1 for all but one patient from group 1. A value above 1 was observed 10 hours (average) (+ 11) before ECLS criteria were met. The maximum sPAP level according to maximum Pace* is shown in Fig 2. The maximum Paco2 and maximum sPAP were independent, and a straight line completely separated the two groups. All group 1 patients reached the right part of the straight line a mean time of 14 hours (+23) before they met ECLS criteria. Of the 147 Doppler echocardiographic studies, sPAP could be calculated 123 times (84%), from a tricuspid regurgitation jet. The mean sPAP was significantly higher in patients with right-to-left shunting through foramen ovale and/or ductus arteriosus than in those with left-to-right shunting (67 v 42.5 mm Hg; P = .OOl). Right-to-left shunting was always present when sPAP reached 80 mm Hg.

Maximal Pace, (mean mm Hg t SD)”

19

Maximal AaDOz (mean mm Hg 2 SD) Maximal 01 (mean t SD)

593 * 114 38 + 24

610 2 66 41 f 28.5

547 r 193 29 r 14

.

Doppler echocardiographic

.

parameters

.

Maximal sPAP (mean mm Hg + SD)*

62.7 + 20

73~

16

. .

44.5 -c 9

Maximal sPAP:sSAP ratio (mean + SD)”

1.09 2 .cr4

1.29 * 0.3

0.75 * 0.15

68

85

25

Foramen ovale and/or ductus arteriosus rightto-left shunting (% of patients)*

*Significant difference behveen groups 1 and 2.

Fig 2. Maximum sPAP level according to maximum groups 1 (H) and 2 (0) on day 1 in the PICU.

Paco2 for

a76

GERMAIN ET AL

The relationship among AaDO*, 01, and sPAP before ECLS was studied. When AaDO* was greater than 610 mm Hg and/or 01 greater than 40, sPAP was significantly higher than in the other cases (60.5 v 45.5 mm Hg; P = .02). The mortality rate was 48% and did not differ significantly between the groups, although it was higher for group 1(52% v 37.5%). DISCUSSION

Doppler echocardiography is ideally suited to the evaluation of infants with acute respiratory failure. Hypoxemia may result from intrapulmonary or cardiac right-to-left shunting or both. Cardiac right-toleft shunting results from the persistence of elevated pulmonary vascular resistance. Through its ability to disclose evidence of persistent pulmonary hypertension, Doppler echocardiography helps to analyze the mechanism of hypoxemia.’ Alterations in systolic time intervals may be measured by M-mode echocardiographyiOJr and suggest the presence of pulmonary hypertension. Doppler findings include demonstration of right-to-left shunting through the foramen ovale and/or through the ductus arteriosus,12 which color-coded Doppler can help to specify.12 The bloodflow pattern within pulmonary artery branches has been studied; in persistent pulmonary hypertension of the newborn (PPHN), systolic flow velocity does not show any acceleration (in comparison to that of normal neonates). l Finally, Doppler evaluation of pulmonary blood flow recently has been applied to study the vasodilatation response to tolazoline infusion.13 In PPHN, right ventricular dilation and tricuspid regurgitation are common.rl Thus, a more direct estimation of pulmonary pressure is possible, from peak velocity of tricuspid regurgitation and/or from peak velocity through the ductus arteriosus.8 Although the degree of correlation with catheterderived cardiac pressures remains controversial, the accuracy of Doppler-estimated pulmonary arterial pressure is commonly admitted.4 Our study shows that pulmonary hypertension is frequent in acute respiratory failure because the mean sPAP at the time of admission was 55 mm Hg, and right-to-left shunting was present in about 50% of the patients. The sPAP and the frequency of right-to-left shunting increased after admission and correlated with the maximum sPAP:sSAP ratio, the mean value of which was above 1 on the first day in the PICU. The degree of respiratory failure in PPHN is usually evaluated by AaDO and 01. AaDO of 2 610 mm Hg for more than 8 hours was associated with a

mortality rate of more than 78%,? and an 01 of 2 40 (on three to five arterial blood gases, at 30-minute intervals) correlated with an 80% mortality rate.3 The results of the present study show a strong correlation in the evolution of sPAP, AaDO*, and 01. Thus, the sPAP of patients who had an AaDO of at least 610 mm Hg or an 01 of at least 40 was significantly higher than that of patients with lower values. Therefore, we suggest that, in addition to AaDO an 01, sPAP is a valuable parameter in assessing the severity of acute respiratory failure of the neonate. The same classical criteria (persistence of abnormal values of AaDO, and/or 01) are used mainly to decide whether to initiate ECLS.’ However, they are used as dynamic criteria because they must be higher than a precise value for several hours. In this study, it was clearly shown that an isolated high value is not significant because there was no difference between patients who met ECLS criteria and those who did not, even with respect to the maximum values of AaDO and 01. In addition, it may be questionable to wait 4 to 8 hours before deciding to begin a lifesaving procedure. In this respect, our study shows the advantages of adding Doppler echocardiography results. With regard to sPAP, there was a significant difference, at the time of admission, between patients who were to meet ECLS criteria and those who were not. This is not explained by differences in the physiological decrease in pulmonary arterial resistances because the admission time was the same for both groups. In addition, all patients whose sPAP was at least 60 mm Hg at the time of admission ultimately met the ECLS criteria. Consequently, the positive predictive value of “sPAP 2 60 mm Hg” for the later appearance of classical ECLS criteria is 100% in this study. On the other hand, lower sPAP values at the time of admission do not exclude the need for ECLS. This limitation may be minimized by adding the study of the poorest values on day 1 in the PICU. All patients who were to meet ECLS criteria had a high sPAP: sSAP value. The highest value, 0.9, was in group 2; this was lower than the lowest value in group 1 (0.96). In addition, only one group 1 patient did not have an sPAP:sSAP value of more than 1. The patient’s sPAP was high (80 mm Hg), but he was treated by high-dose norepinephrine (1 to 1.5 pg/kg/min), which consequently increased his sSAP dramatically by the time of echographic assessment, explaining his 0.96 ratio value. All other group 1 patients had suprasystemic sPAP levels-always before they met the classical ECLS criteria. As noted in cases of congenital diaphragmatic hernia,r4 a second parameter, Pacoz, may help assess

DOPPLER ECHOCARDIOGRAPHY

AND PULMONARY

HYPERTENSION

the extent of pulmonary parenchymal disease. Because the maximum sPAP and Pacoz differed significantly between the two groups during the first day in the PICU, the diagram showed total separation of the two groups (on either side of a straight line). All patients in group 1 reached the right side of the diagram before they met the ECLS criteria. Therefore, use of the sPAP:sSAP ratio and the PacoJsPAP diagram may allow for earlier initiation of ECLS. Indeed, despite the good results obtained with ECLS treatment 3*7~1s mortality and morbidity are also related to the’duration of pre-ECLS hypoxemia and tissular ischemia.7~16-18 This duration could be reduced by using such predictive indexes. In addition, the early determination of nonresponders to conventional therapy may help to buy time for an attempt of nonconventional techniques, such as highfrequency oscillatory ventilation’” or high-frequency jet ventilation.2’J~2’ However, our results should be evaluated by prospective studies because there are limitations to their interpretation. First, we were not able to evaluate

877

sPAP in nine patients. Nevertheless, ECLS was begun shortly after admission for four of these patients as a lifesaving procedure because of acute deterioration, and we did not need predictive criteria. Second, an increase in Pace, does not always reflect the extent of pulmonary parenchymal disease; it also depends on extrapulmonary shunting. In case of right-to-left extrapulmonary shunting, the Pacoz may increase because of decreased pulmonary blood flo~.~’ Assessment of the alveolo-arterial CO2 gradient, by comparing Pace, and end-tidal COz, may be more accurate in determining the degree of parenchymal disease. Nevertheless, we recommend the routine use of Doppler echocardiography, including measurement of sPAP, in infants with acute respiratory failure. When Doppler echocardiographic assessment shows suprasystemic sPAP or high sPAP associated with high Paco2, nonconventional therapy should be started. In such cases, ECLS is, at this time, the treatment of choice, but nitric oxide may, in the future, modify its indications.‘”

REFERENCES 1. Hammerman C, Yousefzadeh D, Choi JH, et a!: Persistent pulmonary hypertension of the newborn. Managing the unmanageable? Clin Perinato! 16:137-156, 1989 2. Short BL, Pearson GD: Neonatal extracorporeal membrane oxygenation: A review. J Intern Care Med 1:47-54. 1986 3. Bartlett RH. Gazzaniga AB. Toomasian AB: Extracorporeal membrane oxygenation (ECMO) in neonatal respiratory failure. Ann Surg 204:236-245, 1986 4. Zellers T, Gutgese!! HP: Noninvasive estimation of pulmonary artery pressure. J Pediatr 114:735-741. 1989 5. Chevallier JY, Durandy Y, Batisse A, et a!: Preliminary report: Extracorporeal lung support for neonatal acute respiratory failure. Lancet 335:1364-1366, 1990 6. Bartlett RH, Gazzaniga AB, Jefferies MR, et a!: Extracorporeal circulation (ECMO) cardiopulmonary support in infancy. Trans Am Sot Artif Intern Organs 22:80-93,1976 7. Bartlett RH: Extracorporeal life support for cardiopulmonary failure. Curr Prob! Surg 27:623-705, 1990 8. Marx GR, Allen HD, Goldberg SJ: Doppler echocardiographic estimation of systolic pulmonary artery pressure in patients with aortic-pulmonary shunts. J Am Co!! Cardio! 7:880-885,1986 9. Currie PJ, Seward JA, Chan KL, et a!: Continuous wave Doppler determination of right ventricular systolic pressure: A simultaneous Doppler-catheterization study in 127 patients. J Am Co!! Cardio! 6:750-756, 1986 10. Riggs T, Hirschfeld S, Fanaroff A, et a!: Persistence of fetal circulation syndrome: An echocardiographic study. J Pediatr 91:626631, 1977 11. Valdes-Cruz L, Dude!! G, Ferrerra A: Utility of M-mode echocardiography for early identification of infants with persistent pulmonary hypertension of the newborn. Pediatrics 68:515-525, 1981 12. Huhta JC, Cohen M, Gutgesse! HP: Patency of the ductus arteriosus in normal neonates: Two-dimensional echocardiography versus Doppler assessment. J Am Co!! Cardio! 4561-564, 1984

13. Morville P, Egreteau L, Mauran P, et a!: Evaluation of tolazoline test in the treatment of persistent pulmonary hypertension of the newborn (PPHN) by Doppler echocardiography. Pediatr Res 30:633, 1991 (abstr) 14. Bohn DJ, James I, Filler RM, et a!: The relationship between Pace? and ventilation parameters in predicting survival in congenital diaphragmatic hernia. J Pediatr Surg 19:666-671, 1984 15. 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, 1991 16. Heiss KF, Bartlett RH: Extracorporeal membrane oxygenation: An experimental protocol becomes a clinical service. Adv Pediatr 36:117-136, 1989 17. Short BL, Miller MK. Anderson KD: Extracorporeal membrane oxygenation in the management of respiratory failure in the newborn. Clin Perinato! 14:737-748,1987 18. Stork E: Extracorporeal membrane oxygenation in the newborn and beyond. Clin Perinato! 15815829, 1988 19. Bryan AC, Froese AB: Reflections on the HiFi trial. Pediatrics 871565-567,1991 20. Spitzer AR, Butler S, Fox WW: Ventilatory response to combined high frequency jet ventilation and conventional mechanical ventilation for the rescue treatment of severe neonatal lung disease. Pediatr Pulmono! 7:244-249, 1989 21. Baumgart S, Hirsch! RB, Butler SZ, et al: Diagnosis-related criteria in the consideration of extracorporeal membrane oxygenation in neonates previously treated with high frequency jet ventilation. Pediatrics 89:491-494, 1992 22. Dehan M, Guenard H, Hernandorena X, et a!: L’exploration des hypoxemies rtfractaires chez le nouveau-m?. Arch Fr Pediatr 38:101-107. 1981 23. Roberts JD, Polaner DM, Lang P, et a!: Inhaled nitric oxide in persistent pulmonary hypertension of the newborn. Lancet 340:818-819, 1992