- Email: [email protected]
Prenatal prediction of lethal pulmonary hypoplasia: The hyperoxygenation test for pulmonary artery reactivity
Prenatal prediction of lethal pulmonary hypoplasia: The hyperoxygenation test for pulmonary artery reactivity Richard E. Broth, MD,a Dennis C. Wood, RDMS,a Juha Rasanen, MD,b Juan Carlos Sabogal, MD,c Ratana Komwilaisak, MD,d Stuart Weiner, MD,a and Vincenzo Berghella, MDa Philadelphia, Pa, Oulu, Finland, Bogota, Colombia, and Khon Kaen, Thailand OBJECTIVE: The purpose of this study was to determine the predictive accuracy of a test for neonatal death from pulmonary hypoplasia by measuring changes in fetal pulmonary artery blood flow on room air and during maternal hyperoxygenation. STUDY DESIGN: Women who were carrying fetuses with congenital anomalies that may cause pulmonary hypoplasia were offered participation in the study as part of a comprehensive fetal echocardiogram. Each fetus at ≥30 weeks of gestation underwent Doppler measurement of the blood flow pattern in the first branch of either the right or the left pulmonary artery before and again during exposure to maternal breathing of 60% oxygen by mask. An increase in the fetal pulmonary blood flow with oxygen (a decrease of ≥20% of the pulsatility index) was considered a reactive test. A change of <20% in the flow pattern during maternal hyperoxygenation was a nonreactive test and suggested pulmonary hypoplasia. The primary outcome for this study was neonatal death from pulmonary hypoplasia. RESULTS: Twenty-nine pregnancies met the criteria for inclusion in our study. Of the 14 fetuses who had a nonreactive hyperoxygenation test, 11 fetuses (79%) died of pulmonary hypoplasia. Of the 15 fetuses who had a reactive hyperoxygenation test, only one fetus (7%) died in the neonatal period. Sensitivity, specificity, and positive and negative predictive values were 92%, 82%, 79%, and 93%, respectively, with an odds ratio of 51 (95% CI, 4.6-560). CONCLUSION: Testing fetal pulmonary vascular reactivity with maternal hyperoxygenation is highly predictive of pulmonary hypoplasia. A reactive test predicted 92% of surviving infants; a nonreactive test predicted 79% of fetal deaths from pulmonary hypoplasia. (Am J Obstet Gynecol 2002;187:940-5)
Key words: Pulmonary hypoplasia, pulmonary vascular reactivity, hyperoxygenation test, fetal echocardiography, pulmonary blood flow, maternal hyperoxygenation
Pulmonary hypoplasia is a term that describes lungs that are sufficiently small enough to impede the exchange of respiratory gases.1 Because gas exchange does not occur in the fetal lung, it is difficult to prove that fetal lungs may be functionally hypoplastic. Pulmonary hypoplasia may be defined as incomplete or underdevelopment of lung tissue that is present at autopsy and is determined by the wet lung-to-body weight ratio, by reduced alveoli count, or by reduced lung DNA content.2 It is a diagnosis that can be made by the pathologist at the time of the autopsy with lung-to-body weight ratio, reduced radial alveolar count, or reduced lung DNA content. The incidence is reported as 11 to 14 per 10,000 live births in the general population and as 7% to 14% in au-
From Jefferson Medical College of Thomas Jefferson University,a the University of Oulu,b Universidad Nacional de Colombia,c and Khon Kaen University.d Presented at the Twenty-second Annual Meeting of the Society for Maternal-Fetal Medicine, New Orleans, La, January 14-19, 2002. Reprints not available from the authors. © 2002, Mosby, Inc. All rights reserved. 0002-9378/2002 $35.00 + 0 6/6/127130 doi:10.1067/mob.2002.127130
940
topsies of all premature infants.3 Perinatal mortality rates approached 70% in most studies.4 Fifteen percent to 20% of neonatal deaths can be attributed to this process. Pulmonary hypoplasia occurs when there is a disruption of normal lung growth and maturation during the critical time period. Numerous conditions can lead to thoracic compression, inhibition of fetal breathing movements, and loss of lung fluids that impede pulmonary growth during the crucial pseudoglandular and canalicular phases between 6 and 26 weeks of gestation. If this is not corrected, pulmonary hypoplasia and neonatal death likely will result. Historically, attempts have been made to assess the postnatal function of the fetal pulmonary system by anatomic criteria. Recent studies have shown that fetal pulmonary vasculature reacts to maternal hyperoxygenation.5,6 Our study looks to predict pulmonary function by assessing pulmonary vascular reactivity in utero. We hypothesized that fetuses with severe lung hypoplasia do not have the same reactivity to maternal hyperoxygenation as normal fetuses of similar gestation. We also believe that, because our test attempts to simulate physiologic factors after delivery (which do not differ despite differ-
Broth et al 941
Volume 187, Number 4 Am J Obstet Gynecol
A
Fig 1. Image of the right pulmonary artery (RPA) shows where to measure the fetal pulmonary artery branch for the hyperoxygenation test. DA, Ductus arteriosis; Ao, aorta; Pul valve, pulmonic valve.
B
ent causes of pulmonary hypoplasia), we could use the hyperoxygenation test in all fetuses at risk. Our objective was to determine, in a noninvasive fashion, the predictive accuracy of fetal pulmonary artery Doppler changes during maternal hyperoxygenation for death from pulmonary hypoplasia. Material and methods This is an institutional review board–approved, retrospective review of data that were obtained during indicated fetal cardiovascular ultrasound examination and were compared with newborn outcomes and infant follow-up from medical records. Thirty-one consecutive women underwent maternal hyperoxygenation to test the pulmonary vascular reactivity of their fetuses. All fetuses were at >30 weeks of gestation at the time of the study. Indications for the fetal oxygen test included previous ultrasound findings of persistent oligohydramnios of renal or nonrenal origin, preterm premature rupture of membranes (PPROM), renal dysplasia/agenesis, obstructive uropathy, twin-to-twin transfusion syndrome, spaceoccupying lesions of the thorax (eg, cardiomegaly, pleural effusion, hydrops, congenital diaphragmatic hernia [CDH], congenital cystic adenomatoid malformation [CCAM], and skeletal dysplasia). Each patient underwent a comprehensive fetal ultrasound study that included a fetal echocardiogram and peripheral, systemic and pulmonary Doppler blood flow waveform analysis. All ultrasound studies were performed and all Doppler waveforms were measured by only one operator (D. C. W.); they were tape recorded, read, and reported on by various pediatric cardiologists. As noted in our previous study,5 our intraobserver variability was 4%. All studies were performed with image-directed pulsed and color Doppler equipment
Fig 2. Sampled waveforms before (A) and after (B) maternal hyperoxygenation with normal reactivity. Note the increased diastolic Doppler pattern. RPA, Right pulmonary artery; PI, pulsatility index; RI, resistance index; S/D, systolic/diastolic ratio.
(Acuson 128XP; Acuson, Mountain View, Calif) with a 5-MHz sector transducer. All measurements were made on-line at the time of capture, with at least three cardiac cycle measurements analyzed from a frozen image. At least two series of measurements were obtained in every case, with several minutes between each capture. The lowest high-pass filter level was used (100 Hz), and the spatial peak temporal average power output for color and pulsed Doppler imaging was kept at <100 mW/cm2. The target vessel that was used in these studies was the first branch of either the right or left pulmonary artery (Fig 1). All attempts were made to obtain the Doppler measurements during periods of fetal apnea. To maintain consistency, an angle of ≤10 degrees insonation of the Doppler beam (as assessed by color Doppler imaging) was obtained. One or both first branch pulmonary artery Doppler waveforms were obtained in every case, and the same vessel was analyzed during maternal hyperoxygenation. Only one test was performed on each fetus.
942 Broth et al
October 2002 Am J Obstet Gynecol
Table I. Fetal diagnoses
A No.
Congenital diaphragmatic hernia Renal/oligohydramnios CCAM Skeletal dysplasia Twin-twin transfusion Pleural effusion
10 7 6 3 2 1
*Sample, 29 fetuses.
Each pregnant woman was placed on an enhanced oxygen mixture that was determined to be a fraction of inspired oxygen of 0.60. This was achieved by plastic rebreathing mask at 8 L per minute of 100% oxygen. Each woman was permitted to breathe through the mask for 10 minutes on the examination table before the pulmonary Doppler studies were repeated. The oxygen mask was removed after the repeat study. All data for this study were obtained from the official report of the pediatric cardiologist. The pulsatility index (PI) values before and after hyperoxygenation of the branch pulmonary arteries were compared. Keeping consistent with our previous study in which we found a 20% decrease in PI with normal gestations of >31 weeks,5 a reactive test was defined as an increase in blood flow, as seen by a decrease in PI of >20% (Fig 2). A nonreactive test had a <20% decrease in the PI value on oxygen as compared to room air (Fig 3). These results were then compared with the neonatal outcome. Autopsies that were performed were reviewed for cause of death, and neonatal charts were reviewed to obtain the cause of death. Established criteria7 were used to define clinical pulmonary hypoplasia. Follow-up of the infants after discharge from the hospital occurred through our fetal therapy division. Our primary outcome criterion was lethal pulmonary hypoplasia, which was diagnosed by either clinical or pathologic criteria. Standard Bayesian analyses were performed along with the Fisher exact test for categoric variables. Results Thirty-one fetuses with a wide array of lesions met criteria. Fetuses of white (57%), Hispanic (25%), and African American (18%) mothers were represented in this population. The mean maternal age was 27.7 ± 5.6 (SD) years. The hyperoxygenation test was performed satisfactorily in all 31 fetuses at a mean gestational age of 32.9 ± 1.9 weeks. Two fetuses were lost to follow-up; therefore, 29 fetuses (93%) were included in this analysis (Table I). Cesarean deliveries were performed for obstetric reasons. The remainder of the infants was delivered vaginally. There were 15 fetuses with reactive testing and
B
Fig 3. Sampled waveforms before (A) and after (B) maternal hyperoxygenation with a negative response. LPA, Left pulmonary artery; RI, resistance index; S/D, systolic diastolic ratio; PI, pulsatility index.
14 fetuses with nonreactive testing (Table II). Of the reactive tests, there was only one false-negative result (ie, a fetus who died after a reactive test). This fetus carried a diagnosis of CCAM and was tested at 32 weeks of gestation. The neonate died soon after delivery at 36 weeks. Of the nonreactive tests, there were three false-positive results (ie, fetuses who lived after nonreactive testing). One of these fetuses carried a diagnosis of CDH and was delivered at 36 weeks; the other fetus had a diagnosis of CDH and omphalocele and was delivered at 37 weeks. Both fetuses had a difficult neonatal course; one infant died at 4 months of age from complications related to omphalocele, and the other infant continued to have significant pulmonary problems that required oxygen therapy until 1 year of age, at which time the child was lost to follow-up. The third fetus, who was studied for persistent oligohydramnios, was found at the time of the study to have intrauterine growth retardation with reversal of blood flow in the umbilical artery and increased diastolic blood flow in the middle cerebral artery. That patient was delivered hours after the study. We speculate that the nonreactivity
Broth et al 943
Volume 187, Number 4 Am J Obstet Gynecol
Table II. Fetal test results and neonatal outcomes
Patient
Fetal age at time of study (wk)
Diagnosis
Reactivity (% change)
Outcome
1 2
32 36
Oligohydramnios CDH
No (7%) No (4%)
Lethal pulmonary hypoplasia Lethal pulmonary hypoplasia
3
33
CCAM
No (6%)
Lethal pulmonary hypoplasia
4 5 6
30 31 30
CDH Oligohydramnios Pleural effusions
No (3%) No (0%) No (1%)
Lethal pulmonary hypoplasia Lethal pulmonary hypoplasia Lethal pulmonary hypoplasia
7 8
36 34
Skeletal dysplasia Oligohydramnios
No (4%) No (3%)
Lethal pulmonary hypoplasia Lethal pulmonary hypoplasia
9
31
No (8%)
Lethal pulmonary hypoplasia
10 11 12 13
32 32 34 36
Twin-twin transfusion syndrome Oligohydramnios Oligohydramnios Oligohydramnios CDH/omphalocele
No (8%) No (3%) No (3%) No (5%)
Lethal pulmonary hypoplasia Lethal pulmonary hypoplasia Live born Nonlethal pulmonary hypoplasia
14
35
CDH
No (3%)
Nonlethal pulmonary hypoplasia
15 16 17 18 19 20 21 22 23 24 25
36 32 34 33 33 35 36 31 31 32 31
Yes (33%) Yes (43%) Yes (69%) Yes (52%) Yes (52%) Yes (64%) Yes (55%) Yes (54%) Yes (35%) Yes (61%)
Lethal pulmonary hypoplasia Live born Live born Live born Live born Live born Live born Live born Live born Live born
26 27 28 29
34 36 32 31
CCAM CDH Skeletal dysplasia CCAM CDH CCAM CDH Small chest CDH CDH Twin-twin transfusion syndrome CCAM CDH Oligohydramnios CCAM
Yes (35%) Yes (41%) Yes (56%) Yes (50%) Yes (34%)
Live born Live born Live born Live born Live born
of the pulmonary blood flow in this fetus was based on redistribution of cardiac output caused by hypoxia from placental insufficiency. The sensitivity, specificity, positive predictive value, and negative predictive value were 92%, 82%, 79%, and 93%, respectively. The odds ratio was 51 (95% CI, 4.6-560). Five of the 12 neonates (42%) who died had an autopsy that proved pulmonary hypoplasia; the remaining 7 neonates were diagnosed clinically. Comment Lethal neonatal pulmonary hypoplasia is associated with a wide variety of lesions of both fetal and maternal
Other information Information not available Extracorporeal membrane oxygenation, mechanical ventilation with oxygen requirements ≥40%, died after 85 d Extracorporeal membrane oxygenation, mechanical ventilation with oxygen requirements 50%, died after 18 d Autopsy Autopsy Oscillator with 100% oxygen requirement, died after 2 d Died in delivery room Oscillator with 95% oxygen requirement, died after 2 d Autopsy Autopsy Autopsy Extracorporeal membrane oxygenation, mechanical ventilation, died at 4 mo of complications of surgery, not pulmonary hypoplasia Extracorporeal membrane oxygenation, oscillator, 34% nasal cannula; transferred to rehabilitation facility after 49 d, required oxygen until lost to follow-up at 1 y Autopsy
origin. External thoracic compression or restriction and internal thoracic mass effect may result in the reduced lung size and function. Fetal lung damage caused during the pseudoglandular stage will result in reduced bronchial and vascular branching, cartilage development, and acinar complexity.3 Insults during the canalicular stage also will reduce acinar complexity and maturation. We believe that a method that evaluates fetal pulmonary function may prove to be a more reliable predictor of postnatal pulmonary dysfunction from pulmonary hypoplasia compared with those that evaluate relative anatomy. Maternal hyperoxygenation decreases human fetal pulmonary arterial vascular impedance and in-
944 Broth et al
creases pulmonary blood flow between 31 and 36 weeks of gestation and mimics the changes in the fetal central hemodynamics that occur after birth.8 We have shown that earlier in pregnancy, between 20 and 26 weeks of gestation, maternal hyperoxygenation does not alter human fetal pulmonary circulation. Therefore, the reactivity of the human fetal pulmonary circulation to oxygen develops sometime between 26 and 30 weeks of gestation. During normoxia, when the human fetal pulmonary arterial bed is under acquired vasoconstriction, right ventricular cardiac output is directed away from the pulmonary circulation to the systemic circulation. Increased oxygen tension decreases pulmonary vascular resistance and thus increases pulmonary blood flow in the normal fetus. The reactivity of the pulmonary arterial circulation to oxygen with advancing gestation has been explained by an increasing amount of smooth muscle in small pulmonary arteries.9 The decrease in the pulmonary vascular resistance is caused mainly by the release of endothelium-derived nitric oxide, which leads to vasodilation of the pulmonary arterial bed.10 The ability to accurately predict those fetuses who will succumb to pulmonary hypoplasia is important for parental counseling and subsequent decision-making regarding obstetric and neonatal treatment. Numerous imaging techniques to predict pulmonary hypoplasia have been tested. These include two-dimensional and three-dimensional ultrasound imaging8,11,12 and magnetic resonance imaging13 to assess anatomic parameters and lung biometry, but are poorly predictive and may be ambiguous because of anatomic and positional variances in lung length and lung area. To date, objective studies have included thoracic circumference to gestational age nomograms,14 thoracic circumference to abdominal circumference ratios,15,16 thoracic circumference/femur length ratios, abdominal circumference/femur length, and cardiac circumference to thoracic circumference,17 and thoracic area-to-heart area or thoracic area-to-heart area 100/thoracic area estimates.18 Other studies have attempted to use fetal breathing movements, perinasal flow, and Doppler velocimetry of the ductus arteriosus, pulmonary veins, and pulmonary arteries19-21 to correlate certain absent or abnormal physiologic patterns with pulmonary hypoplasia. In studying the fetal pulmonary vascularity in the second half of gestation of normal pregnancies, we have shown normal changes in pulmonary blood flow over time under several conditions.5,22,23 We have shown that there exists a normal reaction to maternal hyperoxygenation in the fetus after 30 weeks of gestation. In this study, we have shown that a specific outcome, neonatal death, is predictable in a variety of fetal anomalies by testing those fetuses’ pulmonary vascular reaction to maternal hyperoxygenation. We have applied this concept to fetuses with congenital abnormalities that predispose to pulmonary
October 2002 Am J Obstet Gynecol
hypoplasia. We have shown this test to be both sensitive and specific despite the varying causes of fetal pulmonary hypoplasia. Our data, although limited in number, support our hypothesis that fetal pulmonary Doppler studies in this high-risk group of patients are highly predictive of the presence or absence of lethal pulmonary hypoplasia because a reactive test was 93% (NPV) accurate in predicting survival and a nonreactive test was 79% (PPV) accurate in predicting death. This study was limited to those fetuses of >30 weeks of gestation, which may limit the usefulness of the information for decisions about pregnancy termination. Nonetheless, clinicians and parents can use the information that was derived from this test to tailor the treatment of specific situations, including antepartum testing, emergent delivery, and neonatal resuscitation. It potentially also may decrease the stress of uncertainty that all parents certainly have in this situation. Another limitation is the few number of patients with PPROM in our study. Most women whose membranes rupture during the crucial phase of lung development proceed to delivery before the gestation at which our test can be performed. In those women with PPROM whose pregnancies have survived past 30 weeks of gestation, this test has shown to be accurate about the outcome. Those fetuses who showed no response were not retested at a later date during pregnancy. In conclusion, we have designed a test to evaluate the fetal response to maternal hyperoxygenation in those fetuses believed to be at risk for pulmonary hypoplasia. This test is highly sensitive and specific. The combination of our series of studies of fetal pulmonary vascular physiologic factors,5,22 combined with those studies of other authors,20,21,23 has proved this test to be repeatable and reliable. These data permit the application of physiologic information that is useful in the prediction of the outcome of those fetuses who are at risk for pulmonary hypoplasia. However, before this test is implemented into clinical practice, our data need to be replicated, and a larger sample size is needed. With further study, we hope to be able to pinpoint more accurately the gestational age at which the fetal pulmonary vasculature becomes reactive and use this test earlier than 30 weeks of gestation. We anticipate that, with additional studies, we will be able to predict the severity of the impairment of pulmonary development in the fetus and the relative outcome after delivery. We thank Anthony Sciscione, MD, Ronald Librizzi, MD, Munir Nazir, MD, and Marion Kaufmann, RN, BSN, for helping us to complete this project. REFERENCES
1. Harding R, Hooper SB. Regulation of lung expansion and lung growth before birth. J Applied Physiol 1996;81:209-4.
Volume 187, Number 4 Am J Obstet Gynecol
2. Askenazi SS, Perlman M. Pulmonary hypoplasia: lung weight and radial alveolar count as critical diagnosis. Arch Dis Child 1979;54:631-9. 3. Lauria M, Gonik B, Romero R, Pulmonary hypoplasia: pathogenesis, diagnosis and antenatal prediction. Obstet Gynecol 1995;86:466-75. 4. Rotschild A, Ling EW, Puterman ML, Farquharson D. Neonatal outcome after prolonged preterm rupture of the membranes. Am J Obstet Gynecol 1990;162:46-52. 5. Rasanen J, Wood DC, Debbs R, Cohen J, Weiner S, Huhta JC. Reactivity of the human fetal circulation to maternal hyperoxygenation increases in the second half of pregnancy, a randomized study. Circulation 1998;97:257-62. 6. Soregaroli M, Rizzo G, Danti L, Arduini D, Romanini C. Effects of maternal hyperoxygenation on ductus venosus flow velocity waveforms in normal third-trimester fetuses. Ultrasound Obstet Gynecol 1993; 3:115-9. 7. Swischuk LE, Richardson CJ, Nichols MM, Ingman MJ. Primary pulmonary hypoplasia in the neonate. J Pediatr 1979;95:573-7. 8. Laudy JA, Janssen MM, Struyk PC, Stijnen T, Wladimiroff JW. Three-dimensional ultrasonography of normal fetal lung volume: a preliminary study. Ultrasound Obstet Gynecol 1998; 11:13-6. 9. Levin DL, Rudolph AM, Heymann MA, Phibbs RH. Morphological development of the pulmonary vascular bed in fetal lambs. Circulation 1976;53:144-51. 10. Mital S, Konduri GG. Vascular, K+ ATP channels mediate O2 induced pulmonary vasodilation in fetal lambs [abstract]. Pediatrics 1996;98:527. 11. D’Arcy TJ, Hughes SW, Chiu WS, Clark T, Milner AD, Saunders J, et al. Estimation of fetal lung volume using enhanced 3-dimensional ultrasound: a new method and first result. Br J Obstet Gynaecol 1996;103:1015-20. 12. Lee A, Kratochwil A, Stumpflen I, Deutinger J, Bernaschek G. Fetal lung volume determination by three-dimensional ultrasonography. Am J Obstet Gynecol 1996;175:588-92. 13. Kuwashima S, Nishimura G, Iimura F, Kohno T, Watanabe H, Kohno A, et al. Low intensity fetal lungs on MRI may suggest the diagnosis of pulmonary hypoplasia. Pediatr Radiol 2001;31:669-72.
Broth et al 945
14. Songster GS, Gray DL, Crane JP. Prenatal prediction of lethal pulmonary hypoplasia using ultrasonic fetal chest circumference. Obstet Gynecol 1989;73:261-6. 15. Johnson A, Callan NA, Bhutani VK, Colmorgen GHC, Weiner S, Bolognese RJ. Ultrasonic ratio of fetal thoracic to abdominal circumference: An association with fetal pulmonary hypoplasia. Am J Obstet Gynecol 1987; 157:764-9. 16. D’Alton M, Mercer B, Riddick E, Dudley D. Serial thoracic versus abdominal circumference ratios for the prediction of pulmonary hypoplasia in premature rupture of membranes remote from term. Am J Obstet Gynecol 1992;166: 658-63. 17. Nimrod C, Davies D, Iwanicki S, Harder J, Persaus D, Nichilson S. Ultrasound prediction of pulmonary hypoplasia. Obstet Gynecol 1986;68:495-8. 18. Vintzileos AM, Campbell WA, Rodis JF, Nochimson DJ, Pinette MG, Petrikovsky BM. Comparison of six different ultrasonographic methods for predicting lethal fetal pulmonary hypoplasia. Am J Obstet Gynecol 1989;161:606-12. 19. Mitchell JM, Roberts AB, Lee A. Doppler waveforms from the pulmonary arterial system in normal fetuses and those with pulmonary hypoplasia. Ultrasound Obstet Gynecol 1998;11: 167-72. 20. Yoshimura S, Masuzaki H, Miura K, Muta K, Gotoh H, Ishimaru T. Diagnosis of fetal pulmonary hypoplasia by measurement of blood flow velocity waveforms of pulmonary arteries with Doppler ultrasonography. Am J Obstet Gynecol 1999; 180:441-6. 21. Chaoui R, Kalache K, Tennstedt C, Lenz F, Vogel M. Pulmonary arterial Doppler velocimetry in fetuses with lung hypoplasia. Eur J Obstet Gynecol Reprod Biol 1999;84:179-85. 22. Rasanen J, Huhta JC, Weiner S, Wood DC, Ludomirski A. Fetal branch pulmonary arterial vascular impedence during the second half of pregnancy. Am J Obstet Gynecol 1996;174:1441-9. 23. Sylvester KG, Rasanen J, Kitano Y, Flake AW, Crombleholme TM. Tracheal occlusion reverses the high impedence to flow in the fetal pulmonary circulation and normalizes its physiological response to oxygen at full term. J Pediatr Surg 1998;33:1071-5.
Key words: Pulmonary hypoplasia, pulmonary vascular reactivity, hyperoxygenation test, fetal echocardiography, pulmonary blood flow, maternal hyperoxygenation
Pulmonary hypoplasia is a term that describes lungs that are sufficiently small enough to impede the exchange of respiratory gases.1 Because gas exchange does not occur in the fetal lung, it is difficult to prove that fetal lungs may be functionally hypoplastic. Pulmonary hypoplasia may be defined as incomplete or underdevelopment of lung tissue that is present at autopsy and is determined by the wet lung-to-body weight ratio, by reduced alveoli count, or by reduced lung DNA content.2 It is a diagnosis that can be made by the pathologist at the time of the autopsy with lung-to-body weight ratio, reduced radial alveolar count, or reduced lung DNA content. The incidence is reported as 11 to 14 per 10,000 live births in the general population and as 7% to 14% in au-
From Jefferson Medical College of Thomas Jefferson University,a the University of Oulu,b Universidad Nacional de Colombia,c and Khon Kaen University.d Presented at the Twenty-second Annual Meeting of the Society for Maternal-Fetal Medicine, New Orleans, La, January 14-19, 2002. Reprints not available from the authors. © 2002, Mosby, Inc. All rights reserved. 0002-9378/2002 $35.00 + 0 6/6/127130 doi:10.1067/mob.2002.127130
940
topsies of all premature infants.3 Perinatal mortality rates approached 70% in most studies.4 Fifteen percent to 20% of neonatal deaths can be attributed to this process. Pulmonary hypoplasia occurs when there is a disruption of normal lung growth and maturation during the critical time period. Numerous conditions can lead to thoracic compression, inhibition of fetal breathing movements, and loss of lung fluids that impede pulmonary growth during the crucial pseudoglandular and canalicular phases between 6 and 26 weeks of gestation. If this is not corrected, pulmonary hypoplasia and neonatal death likely will result. Historically, attempts have been made to assess the postnatal function of the fetal pulmonary system by anatomic criteria. Recent studies have shown that fetal pulmonary vasculature reacts to maternal hyperoxygenation.5,6 Our study looks to predict pulmonary function by assessing pulmonary vascular reactivity in utero. We hypothesized that fetuses with severe lung hypoplasia do not have the same reactivity to maternal hyperoxygenation as normal fetuses of similar gestation. We also believe that, because our test attempts to simulate physiologic factors after delivery (which do not differ despite differ-
Broth et al 941
Volume 187, Number 4 Am J Obstet Gynecol
A
Fig 1. Image of the right pulmonary artery (RPA) shows where to measure the fetal pulmonary artery branch for the hyperoxygenation test. DA, Ductus arteriosis; Ao, aorta; Pul valve, pulmonic valve.
B
ent causes of pulmonary hypoplasia), we could use the hyperoxygenation test in all fetuses at risk. Our objective was to determine, in a noninvasive fashion, the predictive accuracy of fetal pulmonary artery Doppler changes during maternal hyperoxygenation for death from pulmonary hypoplasia. Material and methods This is an institutional review board–approved, retrospective review of data that were obtained during indicated fetal cardiovascular ultrasound examination and were compared with newborn outcomes and infant follow-up from medical records. Thirty-one consecutive women underwent maternal hyperoxygenation to test the pulmonary vascular reactivity of their fetuses. All fetuses were at >30 weeks of gestation at the time of the study. Indications for the fetal oxygen test included previous ultrasound findings of persistent oligohydramnios of renal or nonrenal origin, preterm premature rupture of membranes (PPROM), renal dysplasia/agenesis, obstructive uropathy, twin-to-twin transfusion syndrome, spaceoccupying lesions of the thorax (eg, cardiomegaly, pleural effusion, hydrops, congenital diaphragmatic hernia [CDH], congenital cystic adenomatoid malformation [CCAM], and skeletal dysplasia). Each patient underwent a comprehensive fetal ultrasound study that included a fetal echocardiogram and peripheral, systemic and pulmonary Doppler blood flow waveform analysis. All ultrasound studies were performed and all Doppler waveforms were measured by only one operator (D. C. W.); they were tape recorded, read, and reported on by various pediatric cardiologists. As noted in our previous study,5 our intraobserver variability was 4%. All studies were performed with image-directed pulsed and color Doppler equipment
Fig 2. Sampled waveforms before (A) and after (B) maternal hyperoxygenation with normal reactivity. Note the increased diastolic Doppler pattern. RPA, Right pulmonary artery; PI, pulsatility index; RI, resistance index; S/D, systolic/diastolic ratio.
(Acuson 128XP; Acuson, Mountain View, Calif) with a 5-MHz sector transducer. All measurements were made on-line at the time of capture, with at least three cardiac cycle measurements analyzed from a frozen image. At least two series of measurements were obtained in every case, with several minutes between each capture. The lowest high-pass filter level was used (100 Hz), and the spatial peak temporal average power output for color and pulsed Doppler imaging was kept at <100 mW/cm2. The target vessel that was used in these studies was the first branch of either the right or left pulmonary artery (Fig 1). All attempts were made to obtain the Doppler measurements during periods of fetal apnea. To maintain consistency, an angle of ≤10 degrees insonation of the Doppler beam (as assessed by color Doppler imaging) was obtained. One or both first branch pulmonary artery Doppler waveforms were obtained in every case, and the same vessel was analyzed during maternal hyperoxygenation. Only one test was performed on each fetus.
942 Broth et al
October 2002 Am J Obstet Gynecol
Table I. Fetal diagnoses
A No.
Congenital diaphragmatic hernia Renal/oligohydramnios CCAM Skeletal dysplasia Twin-twin transfusion Pleural effusion
10 7 6 3 2 1
*Sample, 29 fetuses.
Each pregnant woman was placed on an enhanced oxygen mixture that was determined to be a fraction of inspired oxygen of 0.60. This was achieved by plastic rebreathing mask at 8 L per minute of 100% oxygen. Each woman was permitted to breathe through the mask for 10 minutes on the examination table before the pulmonary Doppler studies were repeated. The oxygen mask was removed after the repeat study. All data for this study were obtained from the official report of the pediatric cardiologist. The pulsatility index (PI) values before and after hyperoxygenation of the branch pulmonary arteries were compared. Keeping consistent with our previous study in which we found a 20% decrease in PI with normal gestations of >31 weeks,5 a reactive test was defined as an increase in blood flow, as seen by a decrease in PI of >20% (Fig 2). A nonreactive test had a <20% decrease in the PI value on oxygen as compared to room air (Fig 3). These results were then compared with the neonatal outcome. Autopsies that were performed were reviewed for cause of death, and neonatal charts were reviewed to obtain the cause of death. Established criteria7 were used to define clinical pulmonary hypoplasia. Follow-up of the infants after discharge from the hospital occurred through our fetal therapy division. Our primary outcome criterion was lethal pulmonary hypoplasia, which was diagnosed by either clinical or pathologic criteria. Standard Bayesian analyses were performed along with the Fisher exact test for categoric variables. Results Thirty-one fetuses with a wide array of lesions met criteria. Fetuses of white (57%), Hispanic (25%), and African American (18%) mothers were represented in this population. The mean maternal age was 27.7 ± 5.6 (SD) years. The hyperoxygenation test was performed satisfactorily in all 31 fetuses at a mean gestational age of 32.9 ± 1.9 weeks. Two fetuses were lost to follow-up; therefore, 29 fetuses (93%) were included in this analysis (Table I). Cesarean deliveries were performed for obstetric reasons. The remainder of the infants was delivered vaginally. There were 15 fetuses with reactive testing and
B
Fig 3. Sampled waveforms before (A) and after (B) maternal hyperoxygenation with a negative response. LPA, Left pulmonary artery; RI, resistance index; S/D, systolic diastolic ratio; PI, pulsatility index.
14 fetuses with nonreactive testing (Table II). Of the reactive tests, there was only one false-negative result (ie, a fetus who died after a reactive test). This fetus carried a diagnosis of CCAM and was tested at 32 weeks of gestation. The neonate died soon after delivery at 36 weeks. Of the nonreactive tests, there were three false-positive results (ie, fetuses who lived after nonreactive testing). One of these fetuses carried a diagnosis of CDH and was delivered at 36 weeks; the other fetus had a diagnosis of CDH and omphalocele and was delivered at 37 weeks. Both fetuses had a difficult neonatal course; one infant died at 4 months of age from complications related to omphalocele, and the other infant continued to have significant pulmonary problems that required oxygen therapy until 1 year of age, at which time the child was lost to follow-up. The third fetus, who was studied for persistent oligohydramnios, was found at the time of the study to have intrauterine growth retardation with reversal of blood flow in the umbilical artery and increased diastolic blood flow in the middle cerebral artery. That patient was delivered hours after the study. We speculate that the nonreactivity
Broth et al 943
Volume 187, Number 4 Am J Obstet Gynecol
Table II. Fetal test results and neonatal outcomes
Patient
Fetal age at time of study (wk)
Diagnosis
Reactivity (% change)
Outcome
1 2
32 36
Oligohydramnios CDH
No (7%) No (4%)
Lethal pulmonary hypoplasia Lethal pulmonary hypoplasia
3
33
CCAM
No (6%)
Lethal pulmonary hypoplasia
4 5 6
30 31 30
CDH Oligohydramnios Pleural effusions
No (3%) No (0%) No (1%)
Lethal pulmonary hypoplasia Lethal pulmonary hypoplasia Lethal pulmonary hypoplasia
7 8
36 34
Skeletal dysplasia Oligohydramnios
No (4%) No (3%)
Lethal pulmonary hypoplasia Lethal pulmonary hypoplasia
9
31
No (8%)
Lethal pulmonary hypoplasia
10 11 12 13
32 32 34 36
Twin-twin transfusion syndrome Oligohydramnios Oligohydramnios Oligohydramnios CDH/omphalocele
No (8%) No (3%) No (3%) No (5%)
Lethal pulmonary hypoplasia Lethal pulmonary hypoplasia Live born Nonlethal pulmonary hypoplasia
14
35
CDH
No (3%)
Nonlethal pulmonary hypoplasia
15 16 17 18 19 20 21 22 23 24 25
36 32 34 33 33 35 36 31 31 32 31
Yes (33%) Yes (43%) Yes (69%) Yes (52%) Yes (52%) Yes (64%) Yes (55%) Yes (54%) Yes (35%) Yes (61%)
Lethal pulmonary hypoplasia Live born Live born Live born Live born Live born Live born Live born Live born Live born
26 27 28 29
34 36 32 31
CCAM CDH Skeletal dysplasia CCAM CDH CCAM CDH Small chest CDH CDH Twin-twin transfusion syndrome CCAM CDH Oligohydramnios CCAM
Yes (35%) Yes (41%) Yes (56%) Yes (50%) Yes (34%)
Live born Live born Live born Live born Live born
of the pulmonary blood flow in this fetus was based on redistribution of cardiac output caused by hypoxia from placental insufficiency. The sensitivity, specificity, positive predictive value, and negative predictive value were 92%, 82%, 79%, and 93%, respectively. The odds ratio was 51 (95% CI, 4.6-560). Five of the 12 neonates (42%) who died had an autopsy that proved pulmonary hypoplasia; the remaining 7 neonates were diagnosed clinically. Comment Lethal neonatal pulmonary hypoplasia is associated with a wide variety of lesions of both fetal and maternal
Other information Information not available Extracorporeal membrane oxygenation, mechanical ventilation with oxygen requirements ≥40%, died after 85 d Extracorporeal membrane oxygenation, mechanical ventilation with oxygen requirements 50%, died after 18 d Autopsy Autopsy Oscillator with 100% oxygen requirement, died after 2 d Died in delivery room Oscillator with 95% oxygen requirement, died after 2 d Autopsy Autopsy Autopsy Extracorporeal membrane oxygenation, mechanical ventilation, died at 4 mo of complications of surgery, not pulmonary hypoplasia Extracorporeal membrane oxygenation, oscillator, 34% nasal cannula; transferred to rehabilitation facility after 49 d, required oxygen until lost to follow-up at 1 y Autopsy
origin. External thoracic compression or restriction and internal thoracic mass effect may result in the reduced lung size and function. Fetal lung damage caused during the pseudoglandular stage will result in reduced bronchial and vascular branching, cartilage development, and acinar complexity.3 Insults during the canalicular stage also will reduce acinar complexity and maturation. We believe that a method that evaluates fetal pulmonary function may prove to be a more reliable predictor of postnatal pulmonary dysfunction from pulmonary hypoplasia compared with those that evaluate relative anatomy. Maternal hyperoxygenation decreases human fetal pulmonary arterial vascular impedance and in-
944 Broth et al
creases pulmonary blood flow between 31 and 36 weeks of gestation and mimics the changes in the fetal central hemodynamics that occur after birth.8 We have shown that earlier in pregnancy, between 20 and 26 weeks of gestation, maternal hyperoxygenation does not alter human fetal pulmonary circulation. Therefore, the reactivity of the human fetal pulmonary circulation to oxygen develops sometime between 26 and 30 weeks of gestation. During normoxia, when the human fetal pulmonary arterial bed is under acquired vasoconstriction, right ventricular cardiac output is directed away from the pulmonary circulation to the systemic circulation. Increased oxygen tension decreases pulmonary vascular resistance and thus increases pulmonary blood flow in the normal fetus. The reactivity of the pulmonary arterial circulation to oxygen with advancing gestation has been explained by an increasing amount of smooth muscle in small pulmonary arteries.9 The decrease in the pulmonary vascular resistance is caused mainly by the release of endothelium-derived nitric oxide, which leads to vasodilation of the pulmonary arterial bed.10 The ability to accurately predict those fetuses who will succumb to pulmonary hypoplasia is important for parental counseling and subsequent decision-making regarding obstetric and neonatal treatment. Numerous imaging techniques to predict pulmonary hypoplasia have been tested. These include two-dimensional and three-dimensional ultrasound imaging8,11,12 and magnetic resonance imaging13 to assess anatomic parameters and lung biometry, but are poorly predictive and may be ambiguous because of anatomic and positional variances in lung length and lung area. To date, objective studies have included thoracic circumference to gestational age nomograms,14 thoracic circumference to abdominal circumference ratios,15,16 thoracic circumference/femur length ratios, abdominal circumference/femur length, and cardiac circumference to thoracic circumference,17 and thoracic area-to-heart area or thoracic area-to-heart area 100/thoracic area estimates.18 Other studies have attempted to use fetal breathing movements, perinasal flow, and Doppler velocimetry of the ductus arteriosus, pulmonary veins, and pulmonary arteries19-21 to correlate certain absent or abnormal physiologic patterns with pulmonary hypoplasia. In studying the fetal pulmonary vascularity in the second half of gestation of normal pregnancies, we have shown normal changes in pulmonary blood flow over time under several conditions.5,22,23 We have shown that there exists a normal reaction to maternal hyperoxygenation in the fetus after 30 weeks of gestation. In this study, we have shown that a specific outcome, neonatal death, is predictable in a variety of fetal anomalies by testing those fetuses’ pulmonary vascular reaction to maternal hyperoxygenation. We have applied this concept to fetuses with congenital abnormalities that predispose to pulmonary
October 2002 Am J Obstet Gynecol
hypoplasia. We have shown this test to be both sensitive and specific despite the varying causes of fetal pulmonary hypoplasia. Our data, although limited in number, support our hypothesis that fetal pulmonary Doppler studies in this high-risk group of patients are highly predictive of the presence or absence of lethal pulmonary hypoplasia because a reactive test was 93% (NPV) accurate in predicting survival and a nonreactive test was 79% (PPV) accurate in predicting death. This study was limited to those fetuses of >30 weeks of gestation, which may limit the usefulness of the information for decisions about pregnancy termination. Nonetheless, clinicians and parents can use the information that was derived from this test to tailor the treatment of specific situations, including antepartum testing, emergent delivery, and neonatal resuscitation. It potentially also may decrease the stress of uncertainty that all parents certainly have in this situation. Another limitation is the few number of patients with PPROM in our study. Most women whose membranes rupture during the crucial phase of lung development proceed to delivery before the gestation at which our test can be performed. In those women with PPROM whose pregnancies have survived past 30 weeks of gestation, this test has shown to be accurate about the outcome. Those fetuses who showed no response were not retested at a later date during pregnancy. In conclusion, we have designed a test to evaluate the fetal response to maternal hyperoxygenation in those fetuses believed to be at risk for pulmonary hypoplasia. This test is highly sensitive and specific. The combination of our series of studies of fetal pulmonary vascular physiologic factors,5,22 combined with those studies of other authors,20,21,23 has proved this test to be repeatable and reliable. These data permit the application of physiologic information that is useful in the prediction of the outcome of those fetuses who are at risk for pulmonary hypoplasia. However, before this test is implemented into clinical practice, our data need to be replicated, and a larger sample size is needed. With further study, we hope to be able to pinpoint more accurately the gestational age at which the fetal pulmonary vasculature becomes reactive and use this test earlier than 30 weeks of gestation. We anticipate that, with additional studies, we will be able to predict the severity of the impairment of pulmonary development in the fetus and the relative outcome after delivery. We thank Anthony Sciscione, MD, Ronald Librizzi, MD, Munir Nazir, MD, and Marion Kaufmann, RN, BSN, for helping us to complete this project. REFERENCES
1. Harding R, Hooper SB. Regulation of lung expansion and lung growth before birth. J Applied Physiol 1996;81:209-4.
Volume 187, Number 4 Am J Obstet Gynecol
2. Askenazi SS, Perlman M. Pulmonary hypoplasia: lung weight and radial alveolar count as critical diagnosis. Arch Dis Child 1979;54:631-9. 3. Lauria M, Gonik B, Romero R, Pulmonary hypoplasia: pathogenesis, diagnosis and antenatal prediction. Obstet Gynecol 1995;86:466-75. 4. Rotschild A, Ling EW, Puterman ML, Farquharson D. Neonatal outcome after prolonged preterm rupture of the membranes. Am J Obstet Gynecol 1990;162:46-52. 5. Rasanen J, Wood DC, Debbs R, Cohen J, Weiner S, Huhta JC. Reactivity of the human fetal circulation to maternal hyperoxygenation increases in the second half of pregnancy, a randomized study. Circulation 1998;97:257-62. 6. Soregaroli M, Rizzo G, Danti L, Arduini D, Romanini C. Effects of maternal hyperoxygenation on ductus venosus flow velocity waveforms in normal third-trimester fetuses. Ultrasound Obstet Gynecol 1993; 3:115-9. 7. Swischuk LE, Richardson CJ, Nichols MM, Ingman MJ. Primary pulmonary hypoplasia in the neonate. J Pediatr 1979;95:573-7. 8. Laudy JA, Janssen MM, Struyk PC, Stijnen T, Wladimiroff JW. Three-dimensional ultrasonography of normal fetal lung volume: a preliminary study. Ultrasound Obstet Gynecol 1998; 11:13-6. 9. Levin DL, Rudolph AM, Heymann MA, Phibbs RH. Morphological development of the pulmonary vascular bed in fetal lambs. Circulation 1976;53:144-51. 10. Mital S, Konduri GG. Vascular, K+ ATP channels mediate O2 induced pulmonary vasodilation in fetal lambs [abstract]. Pediatrics 1996;98:527. 11. D’Arcy TJ, Hughes SW, Chiu WS, Clark T, Milner AD, Saunders J, et al. Estimation of fetal lung volume using enhanced 3-dimensional ultrasound: a new method and first result. Br J Obstet Gynaecol 1996;103:1015-20. 12. Lee A, Kratochwil A, Stumpflen I, Deutinger J, Bernaschek G. Fetal lung volume determination by three-dimensional ultrasonography. Am J Obstet Gynecol 1996;175:588-92. 13. Kuwashima S, Nishimura G, Iimura F, Kohno T, Watanabe H, Kohno A, et al. Low intensity fetal lungs on MRI may suggest the diagnosis of pulmonary hypoplasia. Pediatr Radiol 2001;31:669-72.
Broth et al 945
14. Songster GS, Gray DL, Crane JP. Prenatal prediction of lethal pulmonary hypoplasia using ultrasonic fetal chest circumference. Obstet Gynecol 1989;73:261-6. 15. Johnson A, Callan NA, Bhutani VK, Colmorgen GHC, Weiner S, Bolognese RJ. Ultrasonic ratio of fetal thoracic to abdominal circumference: An association with fetal pulmonary hypoplasia. Am J Obstet Gynecol 1987; 157:764-9. 16. D’Alton M, Mercer B, Riddick E, Dudley D. Serial thoracic versus abdominal circumference ratios for the prediction of pulmonary hypoplasia in premature rupture of membranes remote from term. Am J Obstet Gynecol 1992;166: 658-63. 17. Nimrod C, Davies D, Iwanicki S, Harder J, Persaus D, Nichilson S. Ultrasound prediction of pulmonary hypoplasia. Obstet Gynecol 1986;68:495-8. 18. Vintzileos AM, Campbell WA, Rodis JF, Nochimson DJ, Pinette MG, Petrikovsky BM. Comparison of six different ultrasonographic methods for predicting lethal fetal pulmonary hypoplasia. Am J Obstet Gynecol 1989;161:606-12. 19. Mitchell JM, Roberts AB, Lee A. Doppler waveforms from the pulmonary arterial system in normal fetuses and those with pulmonary hypoplasia. Ultrasound Obstet Gynecol 1998;11: 167-72. 20. Yoshimura S, Masuzaki H, Miura K, Muta K, Gotoh H, Ishimaru T. Diagnosis of fetal pulmonary hypoplasia by measurement of blood flow velocity waveforms of pulmonary arteries with Doppler ultrasonography. Am J Obstet Gynecol 1999; 180:441-6. 21. Chaoui R, Kalache K, Tennstedt C, Lenz F, Vogel M. Pulmonary arterial Doppler velocimetry in fetuses with lung hypoplasia. Eur J Obstet Gynecol Reprod Biol 1999;84:179-85. 22. Rasanen J, Huhta JC, Weiner S, Wood DC, Ludomirski A. Fetal branch pulmonary arterial vascular impedence during the second half of pregnancy. Am J Obstet Gynecol 1996;174:1441-9. 23. Sylvester KG, Rasanen J, Kitano Y, Flake AW, Crombleholme TM. Tracheal occlusion reverses the high impedence to flow in the fetal pulmonary circulation and normalizes its physiological response to oxygen at full term. J Pediatr Surg 1998;33:1071-5.