- Email: [email protected]
Amniotic fluid phospholipids after maternal administration of dexamethasone
Amniotic fluid phospholipids after maternal administration of dexamethasone FARRELL,
M.D.,
MICHAEL
PHILIP
J.
ENGLE,
PH.D.
RICHARD
D.
ZACHMAN,
LUIS
M.
B. CURET,
JOHN
VIJAYA
W.
KENNETH
RAO,
COLLABORATIVE Madison,
M.D.,
PH.D.
M.D.
C. MORRISON,
A.
PH.D.
M.D. PH.D.
POOLE,
PH.D.
GROUP
ON
ANTENATAL
STEROID
Wisconsin, Jackson, Mississippi, and Research Triangle
THERAPY*
Park, North Carolina
The administration of corticosteroids to pregnant women in premature labor can accelerate fetal lung development and potentially prevent neonatal respiratory distress syndrome (RDS). Controversy exists, however, as to whether amniotic fluid phospholipid indices of lung maturation are influenced by such treatment. Without a suitable test for evaluating the fetal response to corticosteroids, there is no method of recognizing whether and when lung development has been stimulated. In an attempt to resolve this issue, we carried out a study of amniofc fluid phosphofipids as pat-l of the National Institutes of Health multicenter trial of prenatal corticosteroids. Amniocenteses were performed before the administration of either steroid hormone or placebo and approximately 1 week after a series of four injections was initiated. Analysis of the ratio of lecithin (phosphatidylcholine) to sphingomyelin (L/S ratio) revealed nearly identical values initially and no significant difference in the posttreatment means when 25 steroid-treated pregnancies were compared to 20 control pregnancies. Although there were significant increases in both groups during the interval between amniocenteses, no statistical difference was found in the extent of change in L/S ratios between the two groups, when pretreatment values were compared with those obtained an average of 1 week later. In addition to evaluating L/S ratios, we performed an assessment of phospholipid concentrations in 17 pregnancies before and after administration of dexamethasone. This revealed no detectable phosphatidylglycerol. There were increases in the absolute concentrations of phosphatidylcholine and disaturated phosphatidylchdine, but these changes were relatively modest in magnitude and could be attributable to either advanced gestational age or dexamethasone. Our results demonstrate that current tests of fetal lung maturity do not provide a routine means for prenatal detection of pulmonary maturational responses to corticosteroids. (AM. J. OBSTET. GYNECOL. 145484, 1983.)
From the Universi~ of Wisconsin, the University Mississippi, and the Research Triangle Institute.
oj
The Collaborative Study on Antenatal Steroid Therapy was sponsored by the Division of Lung Disease, National Heart, Lung, and Blood Institute, and um pmformed pursuant to Contracts NOl-HR-6-2948, 2949, 2950, 2951,2952, and 2953. Presented at the Twenty-ninth Annual Meeting of the Society for Gynecologk Investigation, Dallas, Texas, March 24-27, 1982. Reprint requests: Dr. Philip M. Farrell, Department of Pediatrics, University of Wisconsin, Clinical Sciences Center, Room H4l428, 600 Highland Ave., Madison, Wisconsin 53792. *A complete list is given
484
at the end of the article.
RESPIRATORY DISTRESS syndrome (RDS), also referred to as hyaline membrane disease, is the most common underlying cause of neonatal death in the United States.’ For this reason and because of the incalculable morbidity associated with the disease, there has been great interest in devising strategies for@evention of RDS. During the past decade, two clinical approaches concerned with fetal lung development have been established as methods to potentially lower the incidence of RDS. The first involves assessment of fetal lung maturation by amniocentesis and measurement of phospholipids which reflect the pulmonary surfactant system.2 By using this approach in high-risk pregnan000%9378/83/040484+07$00.70/O
0 1983 The C. V. Mosby Co.
Volume Number
145 4
ties and on patients who require elective intervention (e.g., repeat cesarean sections), obstetricians have been able to ascertain the degree of fetal lung maturation before delivery. The most widely used phospholipid index of lung maturation has been the ratio of lecithin (phosphatidylcholine) to sphingomyelin (L/S ratio); however, recent studies have suggested that phospholipids which are more specific for pulmonary surfactant, such as disaturated phosphatidylcholine3 and phosphatidylglycerol,4 may offer greater reliability. The second method of preventing RDS often follows determination of lung immaturity by amniocentesis and involves maternal administration of corticosteroid hormones. In general, maternal intramuscular injections of potent hydrocortisone analogues, such as dexamethasonej and betamethasone,6 have been utilized to accelerate lung development over 24 to 28 hours. This has been established in recent years’, 8 as a treatment which significantly lowers, but does not eliminate, the risk of neonatal RDS. Because prenatal corticosteroid treatment does not invariably stimulate adequate lung maturation, it would be advantageous to have a prenatal test for identification of fetuses who have responded positively. Thus, interest has focused on possible changes in the amniotic fluid phospholipid pattern after steroid treatment. In the initial trial conducted in New Zealand, Liggins and Howie”, g found no significant influence of betamethasone on L/S ratios as compared to control subjects who were treated with a low dose of cortisone acetate. However, other investigators’0s I1 have reported changes in L/S ratios measured approximately 1 week after steroid treatment. Recently, Arias and associates” reported that semiquantitative measurement of saturated phospholipid in amniotic fluid would serve to detect changes in fetal lung after maternal betamethasone therapy. A large, double-blind, randomized trial of prenatal dexamethasone therapy for prevention of RDS was conducted between 1977 and 1980, with funding from the National Heart, Lung, and Blood Institute, National Institutes of Health. The experimental design of this multicenter study and initial results in regard to the incidence of RDS have been reported in detai1.j In general, the data revealed a significant lowering of the incidence of RDS, particularly in female infants, after administration of 5 mg of dexamethasone up to four times (every 12 hours) before delivery. As part of the clinical protocol, an attempt was made to obtain amniotic fluid before the administration of dexamethasone or placebo and approximately 1 week after the intramuscular injections. These samples of amniotic fluid were analyzed by the standard technique for measure-
Amniotic
fluid phospholipids
after
maternal
dexamethasone
485
ment of the L/S ratio.13 Other indices of fetal lung development were also measured, such as the absolute concentrations of phosphatidylcholine, disaturated phosphatidylcholine, and phosphatidylglycerol. This report summarizes the results of these determinations and presents recommendations in regard to repeat amniocenteses after corticosteroid administration.
Methtxls The protocol for the National Institutes of Health collaborative study “to evaluate the efficacy of antenatal steroid therapy in the prevention of neonatal respiratory distress syndrome” was developed in 1976-1977. The design was that of a double-blind randomized trial that involved pregnant patients at five perinatal centers, as described elsewhere.3 Data obtained from evaluation of mothers, newborn infants, and clinical specimens were reported to a coordinating center, the Research Triangle Institute, for uniform data management and statistical analyses. Various quality control procedures were instituted to assure uniformity at each institution. The reliability of the randomization scheme was continuously monitored and found to be satisfactory.” From March 1, 1977, to March 1, 1980, a total of 7,893 patients was screened for eligibility. In general, patients between 26 and 37 weeks’ gestation who were at risk for premature delivery and consented to participate were considered for the study. Patients beyond 34 weeks’ gestation were only randomized into the trial if amniocentesis revealed an “immature” L/S ratio.* It was also recommended that, whenever clinically indicated, an amniocentesis be performed on any patient being screened, and that a second amniocentesis be carried out approximately 1 week after the intramuscular injections. The initial L/S ratio was obtained on the day of randomization. i.e., within 24 hours of treatment with steroid or placeho. The study drug, dexamethasone, was administered intramuscularly as 5 mg of dexamethasone phosphate every 12 hours for a total of four doses (20 mg total). The placebo contained an equal volume of vehicle solution and was dispensed in coded vials the identity of which was known only to the coordinating center. Assessment of amniotic fluid phospholipids before and after treatment was incorporated into the protocol from the onset of the study, although there was recognition that only a fraction of those entered would have two amniocenteses during an appropriate interval. A total of 47 patients with single fetuses did have pretreatment and posttreatment amniocenteses within 3 to 9 days apart, including 20 in the placebo group and 27 in the dexamethasone treatment group; however, because the only two diabetic patients were in the steroid
488
Farrell et al.
February Am. J. Obstet.
Table I. L/S ratio in specimens of amniotic of either
dexamethasone
fluid obtained
before
and 1 to 7 days after administration
or placebo* LIS ratio vahest
Gestation t Treatment
15, 1983 Gynecol.
group
N
(wk)
All mothers: Placebo Steroid
20 25
Mothers of 32 to 34 weeks’ gestation: Placebo Steroid
10 12
Male infants: Placebo Steroid
11 15
32.5 33.3
ft
Female infants: Placebo Steroid
9 10
32.3 33.3
+f 0.9 0.7
"If: 0.5 0.4
32.3 33.3
-0.6 0.5
P
Pretreat7nent
Posttreatment
1.45 + 0.09 1.42 2 0.08
2.02 2.36
1.57 1.40 +* 0.16 0.07
2.32
tt
0.14 0.28
2.11 + f 0.38 0.21
Intet-ual
$
values GroupSO
0.001 0.003
0.32
0.04 0.06
0.87
1.36 1.37
+f 0.08 0.12
1.85 2.45
zt f 0.20 0.46
0.007
0.01
0.24
1.54 1.50
‘-iz 0.08 0.18
2.22 2.23
22
0.04
0.79
0.18 0.14
0.001
*The second amniocentesis was performed 3 to 9 days after the initial procedure on the day of screening for randomization. tMean f SEM. $P values for comparison of the change in L/S ratio between pretreatment and posttreatment values in each group, adjusting for the following variables: pretreatment L/S, gestational age, time interval between amniocenteses. OP values for comparison of the differences in pretreatment and posttreatment L/S ratios between the placebo and steroid groups, i.e., the magnitude of change in L/S ratio after dexamethasone compared to change after placebo injections.
group, these were excluded from the analysis, thereby leaving 45 pairs of specimens of amniotic fluid. In addition, gastric aspirate samples of amniotic fluid from singleton births were available for L/S ratio measurements, including 59 patients who received placebo and 62 in the dexamethasone group; the incidence of prolonged rupture of membranes was 33.9% in control subjects and 30.6% in the steroid group. Although the original plan was to evaluate only the L/S ratio by a standardized procedure, this objective was expanded in October, 1979, to incorporate additional analyses whenever sufficient volumes of amniotic fluid were available. Thus, the absolute concentrations of phosphatidylcholine, disaturated phosphatidylcholine, phosphatidylglycerol, and sphingomyelin were determined on 17 dexamethasone-treated patients by including a sixth perinatal center that employed the same clinical protocol under the direction of an obstetrician who began the clinical trial as a co-investigator at one of the five clinical centers. Although a control group was not available for comprehensive phospholipid analyses, the laboratory measurements were still performed in a blind fashion by coding the pretreatment and posttreatment samples of amniotic fluid and reporting the data for statistical treatment to the coordinating center. The standard procedure for measuring amniotic fluid L/S ratios was similar to that described by Olson and Graven.13 Specimens were kept on ice briefly before extraction of lipids, if analyses were performed the same day, or were otherwise frozen promptly to
maintain phospholipid stability.14 The samples of amniotic fluid were centrifuged once at 1,000 X g for 5 minutes. No bloody or meconium-stained specimens were included. Extraction of lipids was carried out by mixing amniotic fluid vigorously with one volume of methanol and two volumes of chloroform. After centrifugation, the lower chloroform layer was removed, evaporated to dryness under nitrogen, taken up in a small volume of 2 : 1 chloroform : methanol, and chromatographed on a silica gel H thin-layer plate with a solvent system of chloroform : methanol : water (65 : 25 : 4). Standard samples of phospholipids were also applied to the plates. Detection of phospholipids was performed either by charring with sulfuric acid and heat or with bromthymol blue staining. Reflectance densitometry was then performed to calculate the ratio of lecithin to sphingomyelin. Strict quality control measures were employed to verify the reliability of the L/S ratio procedure at the various laboratories. On seven occasions, 12 to 16 standard samples of amniotic fluid adjusted to known L/S ratios were coded and sent to the five perinatal centers. Results of these determinations were then reported to the coordinating center for data analysis. All laboratories were required to meet defined reliability criteria specified by the National Institutes of Health Steering Committee for this study. Analyses of phosphatidylcholine, disaturated phosphatidylcholine, and phosphatidylglycerol were performed in a central laboratory on frozen specimens of amniotic fluid by methods described elsewhere.15z ”
Volun1e 145 Number
Amniotic
fluid phospholipids
after
maternal
dexamethasone
487
4
Table II. Comprehensive assessment of amniotic after dexamethasone treatment*
fluid phospholipids Conce7ltration
Phospholipid
Sphingomyelin Phosphatidylcholine§ Ratio of phosphatidylcholine to sphingomyelin Disaturated phosphatidylcholine Phosphatidylglycerol
in 17 patients
(nmoleslml)
before and
or ratio? Posttreatment
Pretreatment
6.4 + 1.1 10.9 t 2.5 1.5 r 0.2
f valuf?s
7.4 zk 1.6 28.0 -c 6.1 3.8 2 0.8
0.25 0.001 0.005
6.5 + 1.5
15.2 -c 4.3
0.01
Undetectable
Undetectable
NS
*Mean gestational age was 32.5 weeks. tMean -t SEM. $P value for comparison of the change in value in the interval between amniocenteses. IRefers to total phosphatidylcholine.
Briefly, on the day of extraction, an internal standard of [14C]-dipalmitoyl-phosphatidylcholine was added to monitor phospholipid recovery and correct for losses during processing. Lipids were extracted by the addition of four volumes of 2: 1 chloroform: methanol. After centrifugation, the lower chloroform layer was removed, dried under nitrogen, taken up to a 1 ml volume of chloroform, and split into two equal fractions for either two-dimensional, thin-layer chromatography or osmium tetroxide treatment to produce a disaturated phosphatidylcholine fraction according to the method of Mason and associates.r7 The solvent system was chloroform : methanol : water (65 : 45: 5) in the first dimension and tetrahydrofuran : methylal : methanol:2M ammonium hydroxide (10:6:4:1) for the second direction. This system permitted the separation of phosphatidylcholine, sphingomyelin, and phosphatidylglycerol. The lipid spots were scraped from the plate, extracted from the gel by the method of Bligh and Dyer, ‘” and then analyzed for phospholipid phosphorus by the method of Chen and associates.1s Appropriate aliquots were also analyzed for radioactivity in a Beckman LS 7000 scintillation counter. After the second lipid fraction was treated with osmium tetroxide, separation of disaturated phosphatidylcholine was performed with neutral alumina and thin-layer chromatography performed with the first-dimensional solvent described above. Quantitation was again performed by analysis of phospholipid phosphorus. Gas chromatography was performed on the disaturated phosphatidylcholine fraction to ascertain that a high (greater than 95%) content of saturated fatty acids, especially palmitate,‘” was present. Statistical analyses were performed at the coordinating center. The data were examined by means of analysis of covariance and Wilcoxon rank sum procedures, logistic regression techniques, and t tests. Logarithm of
the posttreatment analyses.
L/S ratio
was used in some of the
Results As shown in Table I, the initial L/S ratios, i.e., pretreatment mean values, were nearly identical for the placebo and steroid groups. In addition, mean gestational ages, which were determined at the time of randomization were comparable in the two groups. The average gestational age for the 45 infants was 32.9 weeks. During the 3- to g-day interval between amniocenteses, the mean L/S ratios increased moderately in each group. The mean change in L/S ratio was 0.57 in the placebo group (from 1.42 to 2.02), whereas the delta value was 0.94 in the steroid treatment group (from 1.42 to 2.36). However, statistical comparison of the change in L/S ratio between the two study groups revealed no significant difference, despite a numerically lower incidence of RDS in the 25 infants exposed to dexamethasone in utero (4.0% versus tO.O% in the control group of 20 neonates: p = 0.30). Evaluation of the extent of change in L/S ratios relative t.o predicting RDS revealed that 12 of 20 control patients and 12 of 25 patients in the steroid group demonstrated an increase sufficient to convert their values from immature to mature. The data from pregnancies of 32 to 34 weeks were also analyzed separately in order to minimize the gestational age variable and examine the L/S ratio issue in patients most likely to be responsive to dexamethasone.6 As indicated in Table I, evaluation of change in L/S ratio again revealed no significant differences between the two subgroups, although the delta value was 70% higher in the steroid-treated population. Because of the influence of fetal sex on dexamethasone responsiveness,‘. 50 we also carried out a separate analysis of data for male and female infants. The results pre-
488
Farrell et al.
sented in Table I demonstrate that the changes in L/S ratio during the pretreatment to posttreatment interval were not statistically different for the male and female subgroups. The 121 gastric aspirate samples of amniotic fluid showed an L/S ratio of 2.88 f 0.21 for the placebo group and 2.70 + 0.18 for the dexamethasone-treated infants. Statistical analysis revealed that there were no significant differences between the two groups (P = 0.52). Furthermore, there was no statistical difference when the change in L/S ratio was evaluated in 41 infants who had gastric aspirates taken subsequent to a pretreatment amniocentesis. Nevertheless, the L/S ratios obtained on gastric aspirates were closely associated with the occurrence of RDS (P = 0.0025). The mean 2 SEM value when RDS occurred was 1.71 i: 0.21, whereas single infants without RDS showed an L/S ratio of 2.97 % 0.15. Thus, although gastric aspirate phospholipids did not statistically relate to treatment groups, they did confirm the expectation that the probability of RDS decreases with increasing L/S ratios. The results obtained from determination of absolute phospholipid concentrations in 17 dexamethasonetreated patients are shown in Table II. The mean gestational age was 32.5 weeks, and the average interval between amniocenteses was 6.8 days. Of interest was the finding that phosphatidylglycerol could not be detected in any of the 34 specimens of amniotic fluid. Although sphingomyelin did not change during the interval between amniocenteses, there were differences in the mean values for phosphatidylcholine (P < 0.005) and disaturated phosphatidylcholine (P < 0.01); however, the degree of the increase was relatively small compared to the tenfold rise normally observed after 34 weeks’ gestation (Engle and Farrell, unpublished data). Furthermore, the percentage of the disaturated phosphatidylcholine fraction did not increase significantly as a result of dexamethasone treatment. On the other hand, comparison of the concentration of phospholipids (by means of a logistic regression model) with neonatal pulmonary status suggested an association with the occurrence of RDS in these 17 pregnancies. In particular, the ratio of phosphatidylcholine to sphingomyelin in 12 patients who did not develop RDS (4.56 t 1.05) was significantly higher than that in five pregnancies that led to neonatal RDS (1.83 2 0.27) (P < 0.03, based on t test analysis of data obtained from the second sample of amniotic fluid); the change in ratio during the pretreatment to posttreatment interval was also greater in the non-RDS group (2.92 ? 0.90 versus 0.56 + 0.51 in those who developed RDS; P < 0.04). Additionally, the percentage of disaturated
February Am. J. Obstet.
15, 1983 Gynecol.
phosphatidylcholine relative to total phosphatidylcholine was higher in patients who did not develop RDS (65.3 k 8.27 versus 33.9 & 5.20 in the RDS group; P < 0.04) because of a much greater increase in the disaturated phosphatidylcholine concentration during the interval between amniocenteses that amounted to 11.3 + 4.23 nmoles/ml compared to 2.50 2 1.88 in the RDS group (P = 0.08).
Comment Although prenatal corticosteroid therapy has been established as a means of lowering the incidence of neonatal RDS in appropriate pregnancies,’ there has been continued controversy over the issue of whether amniotic fluid indices of fetal lung development change in response to steroid administration. On the basis of results obtained in the first clinical trial, Liggins and Howie6* ’ concluded that there was no difference in the L/S ratio after betamethasone therapy. Soon thereafter, Spellacy and associatesi reported a greater rate of rise in the L/S ratio after dexamethasone injections than in uninjected control pregnancies (delta values of 0.68 and 0.40, respectively; P < 0.05). Subsequently, there have been at least 14 investigations of this issue, as reviewed by Morrison and associates,” with about half revealing an increase in the L/S ratio after treatment with various steroid hormones and the remainder demonstrating no change. Even when the L/S ratio has been reported to increase significantly, however, the extent of change has been modest. Because several of the investigations did not include a control group, it is difficult to isolate the steroid treatment variable from other factors, such as gestational age. Arias and associatesi2 attempted to increase the sensitivity of amniotic fluid phospholipid analysis by studying both the L/S ratio and a qualitative measure of disaturated phosphatidylcholine. They reported significantly greater L/S values 1 week after betamethasone treatment than in control pregnancies, but the extent of change was only slight (mean + SD, 1.49 2 0.21 versus 1.09 ? 0.29, respectively). Although six control pregnancies did not show a change in the relative amount of disaturated phosphatidylcholine, 12 of 18 betamethasone-injetted patients had an increase in this fraction within 14 days of treatment (P = 0.016). The study reported herein represents a carefully controlled evaluation of the L/S ratio and several other phospholipid indices of fetal lung development after either dexamethasone injections or the administration of placebo (drug vehicle). In order to minimize the impact of variables which can affect amniotic fluid phospholipids,2 we excluded diabetic pregnancies and
Volume Number
145 4
limited the influence of gestational age, as follows. (1) Pregnancies between 26 and 34 weeks’ gestation were identified as the primary target population for entry, which resulted in a mean gestational age in this study of 32.9 weeks. (2) The second specimen of amniotic fluid was obtained within 9 days of the initial amniocentesis (within 1 week after completion of the intramuscular injections). Although 1 week may not provide adequate time for the acceleration of fetal lung development to be reflected in amniotic fluid, various observations in normal pregnancies suggest that phospholipid indices can increase during that intervaL3 It is equally clear that the longer the period of time before a second amniocentesis, the higher the probability that advancing gestation per se will increase the concentration of lung surfactant-derived phospholipids. In order to optimize experimental design, therefore, it is necessary to limit the interval after steroid treatment before the second amniocentesis. The pretreatment L/S ratio values found in 20 control pregnancies and 25 dexamethasone-injected women indicate a high probability of lung immaturity at the time the intramuscular injections were initiated. Although a significant change in L/S ratio occurred in both groups, the major conclusions from our results are the following. (1) The magnitude of differences in each group is small and only slightly greater than the expected coefficient of variation for an L/S determination. (2) The extent of change after dexamethasone treatment is not significantly greater than the L/S difference in the placebo group. Furthermore, gastric aspirate samples did not differentiate the two study groups, even though these measurements were sufficiently sensitive to identify infants who would develop RDS. Our objective of achieving greater sensitivity and specificity by measuring phospholipids, such as disaturated phosphatidylcholine and phosphatidylglycerol, was hampered by the lack of a matched control group for these assessments. Nevertheless, we are able to conclude that phosphatidylglycerol determinations are not helpful in monitoring the effects of steroid treatment or predicting neonatal RDS for pregnancies of this gestational age. Measurement of the absolute concentration of phosphatidylcholine fractions offers more promise, on the basis of the 256% increase in phosphatidylcholine, a 233% increase in disaturated phosphatidylcholine, and the relationship of these values to the occurrence of RDS. Although the rise in total phosphatidylcholine and the disaturated fraction provides evidence of a fetal lung maturational response to dexamethasone, we must emphasize that the increased
Amniotic
fluid
phospholipids
after
maternal
dexamethasone
489
gestational age is also a factor that influences these data. Furthermore, the magnitude of change in these indices of lung development is relatively modest and may not be sufficient for reliably distinguishing a favorable fetal response. In summary, our results indicate that current tests used routinely to assess fetal lung maturity do not provide a means for reliable detection of fetal pulmonary responses to maternally administered cc>rticosteroids. Because an accurate test is currently unavailable, we conclude that, at present, repeat amniocentesis after steroid administration is not of clinical \alue for monitoring corticosteroid therapy per se. Further research on determination of more sensitive indices of lung development will be necessary before use of such assessments can be applied in clinical perinatology.
The Collaborative Study on Antenatal Steroid Therapy was carried out by the following investigators who also served as members of the Steering Committee: Richard Depp, M.D., and John Boehm, M.D. (Northwestern University, Chicago, Illinois); Richard Zachman, Ph.D., M.D., and Luis Curet, M.D. (University of Wisconsin, Madison, Wisconsin); Charles R. Bauer, M.D.; Louis Fernandez-Rocha, M.D., and Gene Burkett, M.D. (University of Miami, Miami, Florida); Sheldon Korones, M.D., John Morrison, M.D.. Jack Schneider, M.D., and Garland Anderson. M.D. (tiniversity of Tennessee, Memphis, Tennessee): Henrique Rigatto, M.D., Leo Peddle, M.D., and Frank Manning, M.D. (University of Manitoba, Winnipeg, Canada): W. Kenneth Poole, Ph.D., and Vijaya Rao, Ph.D. (Research Triangle Institute, Research Triangle Park, North Carolina); David Fukushima, Ph.D., John O’Connor, Ph.D., and Jack Kream, Ph.D. (Mont&ore Hospital, New York, New York). We also thank Betty Hastings for help with data processing and D. Jeannette Brown for performance of amniotic fluid analyses. A Policy-Data Monitoring Panel is also acknowledged, consisting of Brian Little, M.D. (Chairman), Mary Ellen Avery, M.D., David DeMets, Ph.D., Max Halperin, Ph.D., Patricia King, M.D., Arthur Parmelee, M.D., Samuel Solomon, Ph.D., F.R.S.C., David Sylwester, Ph.D., and James Ware, Ph.D. The National Heart, Lung, and Blood Institute Program Office representatives were Bitten Stripp, Ph.D., Project Officer, and Claude Lenfant, M.D., Director, Division of Lung Diseases, Bethesda, Maryland. On request of the Division of Lung Disease, National Heart, Lung, and Blood Institute, Merck Sharp & Dohme provided the drug and placebo preparations used in this study. This acknowledgement of appreciation is in no way an endorsement 01 a particular product.
490
Farrell et al.
REFERENCES
1. Perelman, R. H., and Farrell, P. M.: Decreasing incidence of fatal hyaline membrane disease based on analysis of 19681978 neonatal mortality statistics, Pediatrics 70: 570, 1982. 2. Tsao, F. H. C., and Zachman, R. D.: In Farrell, P. M., editor: Lung Development: Biological and Clinical Perspectives, New York, 1982, vol. 2, Academic Press, Inc., pp. 167-203. 3. Torday, J., Carson, L., and Lawson, E. E.: Saturated phosphatidylcholine in amniotic fluid and prediction of the respiratory distress syndrome, N. Engl. J. Med. 301:1013, 1979. 4. Tsao, F. H. C., and Zachman, R. D.: Determination of phosphatidylglycerol in amniotic fluid by a simple onedimensional thin-layer chromatography method, Clin. Chem. Acta 118:109, 1982. 5. Collaborative Group on Antenatal Steroid Therapy: Effect of antenatal dexamethasone administration on the prevention of respiratory distress syndrome, AM. J. OBSTET. GYNECOL. 141:276, 1981. 6. Liggins, G. C., and Howie, R. N.: A controlled trial of antepartum glucocorticoid treatment for prevention of the respiratory distress syndrome in premature infants, Pediatrics 50:515, 1972. 7. Zachman, R. D.: In Farrell, P. M., editor: Lung Development: Biological and Clinical Perspectives, New York, 1982, vol. 2, Academic Press, Inc., pp. 275-296. 8. Ballard, R. A., Ballard, P. L., Granberg, J. P., and Sniderman, S.: Prenatal administration of betamethasone for prevention of respiratory distress syndrome, J. Pediatr. 94:97, 1979. 9. Liggins, G. C., and Howie, R. N.: In Cluck, L., editor: Modern Perinatal Medicine, Chicago, 1974, Year Book Medical Publishers, Inc., pp. 415-424. W. N., Buhi, W. C., Riggall, F. C., and Hol10. Spellacy, singer, K. L. Human amniotic fluid lecithimsphingomyelin ratio changes with estrogen or glucocorticoid treatment, AM. J. OBSTET. GYNECOL. 115:216, 1973.
February Am. J. Obstet.
15. 1983 Gynecol.
11. Morrison, J. C., Whybrew, W. D., Bucovaz, E. T., Wiser, W. L., and Fish, S. A.: The lecithin/sphingomyelin ratio in cases associated with fetomaternal disease, AM. J. OBSTET. GYNECOL. 127:363, 1977. 12. Arias, F., Pineda, J., and Johnson, I,. W.: Changes in human amniotic fluid lecithinisphingomyelin ratio and dipalmitoyl lecithin associated with maternal betamethasone therapy, AM. J. OBSTET. GYNECOL. 133:894, 1979. 13. Olson, E. B., and Graven, S. N.: Comparison of visualization methods used to measure the lecithin/sphingomyelin ratio in amniotic fluid, Clin. Chem. 20: 1408, 1974. 14. Schwartz, D. B., Engle, M. J.< Brown, J., and Farrell, P. M.: The stability of phospholipids in amniotic fluid, AM.J.OBSTET. GYNECOL. 141:294, 1981. 15. Curbelo, V., Gail, D. B., and Farrell, P. M.: Determination of disaturated lecithin in rhesus monkey amniotic fluid as an index of fetal lung maturity, AM. J. OBSTET. GYNECOL. 131:764, 1978. 16. Perelman, R. H., Engle, M. J., Kemnitz, J. W., Kotas, R. V., and Farrell, P. M.: Biochemical and physiological development of the fetal rhesus lung, J. Appl. Ph$iol. 53:230. 1982. 17. Mason; R. J., Nellenbogen, J., and Clements, J. A.: Isolation of disaturated phosphatidylcholine with osmium tetroxide, J. Lipid Res. 17:281, 1976. 18. Bligh, E. G., and Dyer, W. J.: A rapid method of total lipid extraction and purification, Can. J. Biochem. 37: 911,1959. 19. Chen, P. S., Toribara, T. Y., and Warner, H.: Microdetermination of phosphorus, Anal. Chem. 28: 1756, 1956. 20. Papageorgiou, A. N., Desgranges, M. F., Masson, M., Colle, E., Shatz, R., and Gelfand, M. M.: The antenatal use of betamethasone in the prevention of respiratory distress syndrome. A controlled double-blind study, Pediatrics 63:73, 1979. 21. Epstein, M. F., Farrell, P. M., and Chez, R. A.: Fetal lung lecithin metabolism and the amniotic fluid L/S ratio in rhesus monkey gestations, AM. J. OBSTET. GYNECOL. 125:545, 1976.
M.D.,
MICHAEL
PHILIP
J.
ENGLE,
PH.D.
RICHARD
D.
ZACHMAN,
LUIS
M.
B. CURET,
JOHN
VIJAYA
W.
KENNETH
RAO,
COLLABORATIVE Madison,
M.D.,
PH.D.
M.D.
C. MORRISON,
A.
PH.D.
M.D. PH.D.
POOLE,
PH.D.
GROUP
ON
ANTENATAL
STEROID
Wisconsin, Jackson, Mississippi, and Research Triangle
THERAPY*
Park, North Carolina
The administration of corticosteroids to pregnant women in premature labor can accelerate fetal lung development and potentially prevent neonatal respiratory distress syndrome (RDS). Controversy exists, however, as to whether amniotic fluid phospholipid indices of lung maturation are influenced by such treatment. Without a suitable test for evaluating the fetal response to corticosteroids, there is no method of recognizing whether and when lung development has been stimulated. In an attempt to resolve this issue, we carried out a study of amniofc fluid phosphofipids as pat-l of the National Institutes of Health multicenter trial of prenatal corticosteroids. Amniocenteses were performed before the administration of either steroid hormone or placebo and approximately 1 week after a series of four injections was initiated. Analysis of the ratio of lecithin (phosphatidylcholine) to sphingomyelin (L/S ratio) revealed nearly identical values initially and no significant difference in the posttreatment means when 25 steroid-treated pregnancies were compared to 20 control pregnancies. Although there were significant increases in both groups during the interval between amniocenteses, no statistical difference was found in the extent of change in L/S ratios between the two groups, when pretreatment values were compared with those obtained an average of 1 week later. In addition to evaluating L/S ratios, we performed an assessment of phospholipid concentrations in 17 pregnancies before and after administration of dexamethasone. This revealed no detectable phosphatidylglycerol. There were increases in the absolute concentrations of phosphatidylcholine and disaturated phosphatidylchdine, but these changes were relatively modest in magnitude and could be attributable to either advanced gestational age or dexamethasone. Our results demonstrate that current tests of fetal lung maturity do not provide a routine means for prenatal detection of pulmonary maturational responses to corticosteroids. (AM. J. OBSTET. GYNECOL. 145484, 1983.)
From the Universi~ of Wisconsin, the University Mississippi, and the Research Triangle Institute.
oj
The Collaborative Study on Antenatal Steroid Therapy was sponsored by the Division of Lung Disease, National Heart, Lung, and Blood Institute, and um pmformed pursuant to Contracts NOl-HR-6-2948, 2949, 2950, 2951,2952, and 2953. Presented at the Twenty-ninth Annual Meeting of the Society for Gynecologk Investigation, Dallas, Texas, March 24-27, 1982. Reprint requests: Dr. Philip M. Farrell, Department of Pediatrics, University of Wisconsin, Clinical Sciences Center, Room H4l428, 600 Highland Ave., Madison, Wisconsin 53792. *A complete list is given
484
at the end of the article.
RESPIRATORY DISTRESS syndrome (RDS), also referred to as hyaline membrane disease, is the most common underlying cause of neonatal death in the United States.’ For this reason and because of the incalculable morbidity associated with the disease, there has been great interest in devising strategies for@evention of RDS. During the past decade, two clinical approaches concerned with fetal lung development have been established as methods to potentially lower the incidence of RDS. The first involves assessment of fetal lung maturation by amniocentesis and measurement of phospholipids which reflect the pulmonary surfactant system.2 By using this approach in high-risk pregnan000%9378/83/040484+07$00.70/O
0 1983 The C. V. Mosby Co.
Volume Number
145 4
ties and on patients who require elective intervention (e.g., repeat cesarean sections), obstetricians have been able to ascertain the degree of fetal lung maturation before delivery. The most widely used phospholipid index of lung maturation has been the ratio of lecithin (phosphatidylcholine) to sphingomyelin (L/S ratio); however, recent studies have suggested that phospholipids which are more specific for pulmonary surfactant, such as disaturated phosphatidylcholine3 and phosphatidylglycerol,4 may offer greater reliability. The second method of preventing RDS often follows determination of lung immaturity by amniocentesis and involves maternal administration of corticosteroid hormones. In general, maternal intramuscular injections of potent hydrocortisone analogues, such as dexamethasonej and betamethasone,6 have been utilized to accelerate lung development over 24 to 28 hours. This has been established in recent years’, 8 as a treatment which significantly lowers, but does not eliminate, the risk of neonatal RDS. Because prenatal corticosteroid treatment does not invariably stimulate adequate lung maturation, it would be advantageous to have a prenatal test for identification of fetuses who have responded positively. Thus, interest has focused on possible changes in the amniotic fluid phospholipid pattern after steroid treatment. In the initial trial conducted in New Zealand, Liggins and Howie”, g found no significant influence of betamethasone on L/S ratios as compared to control subjects who were treated with a low dose of cortisone acetate. However, other investigators’0s I1 have reported changes in L/S ratios measured approximately 1 week after steroid treatment. Recently, Arias and associates” reported that semiquantitative measurement of saturated phospholipid in amniotic fluid would serve to detect changes in fetal lung after maternal betamethasone therapy. A large, double-blind, randomized trial of prenatal dexamethasone therapy for prevention of RDS was conducted between 1977 and 1980, with funding from the National Heart, Lung, and Blood Institute, National Institutes of Health. The experimental design of this multicenter study and initial results in regard to the incidence of RDS have been reported in detai1.j In general, the data revealed a significant lowering of the incidence of RDS, particularly in female infants, after administration of 5 mg of dexamethasone up to four times (every 12 hours) before delivery. As part of the clinical protocol, an attempt was made to obtain amniotic fluid before the administration of dexamethasone or placebo and approximately 1 week after the intramuscular injections. These samples of amniotic fluid were analyzed by the standard technique for measure-
Amniotic
fluid phospholipids
after
maternal
dexamethasone
485
ment of the L/S ratio.13 Other indices of fetal lung development were also measured, such as the absolute concentrations of phosphatidylcholine, disaturated phosphatidylcholine, and phosphatidylglycerol. This report summarizes the results of these determinations and presents recommendations in regard to repeat amniocenteses after corticosteroid administration.
Methtxls The protocol for the National Institutes of Health collaborative study “to evaluate the efficacy of antenatal steroid therapy in the prevention of neonatal respiratory distress syndrome” was developed in 1976-1977. The design was that of a double-blind randomized trial that involved pregnant patients at five perinatal centers, as described elsewhere.3 Data obtained from evaluation of mothers, newborn infants, and clinical specimens were reported to a coordinating center, the Research Triangle Institute, for uniform data management and statistical analyses. Various quality control procedures were instituted to assure uniformity at each institution. The reliability of the randomization scheme was continuously monitored and found to be satisfactory.” From March 1, 1977, to March 1, 1980, a total of 7,893 patients was screened for eligibility. In general, patients between 26 and 37 weeks’ gestation who were at risk for premature delivery and consented to participate were considered for the study. Patients beyond 34 weeks’ gestation were only randomized into the trial if amniocentesis revealed an “immature” L/S ratio.* It was also recommended that, whenever clinically indicated, an amniocentesis be performed on any patient being screened, and that a second amniocentesis be carried out approximately 1 week after the intramuscular injections. The initial L/S ratio was obtained on the day of randomization. i.e., within 24 hours of treatment with steroid or placeho. The study drug, dexamethasone, was administered intramuscularly as 5 mg of dexamethasone phosphate every 12 hours for a total of four doses (20 mg total). The placebo contained an equal volume of vehicle solution and was dispensed in coded vials the identity of which was known only to the coordinating center. Assessment of amniotic fluid phospholipids before and after treatment was incorporated into the protocol from the onset of the study, although there was recognition that only a fraction of those entered would have two amniocenteses during an appropriate interval. A total of 47 patients with single fetuses did have pretreatment and posttreatment amniocenteses within 3 to 9 days apart, including 20 in the placebo group and 27 in the dexamethasone treatment group; however, because the only two diabetic patients were in the steroid
488
Farrell et al.
February Am. J. Obstet.
Table I. L/S ratio in specimens of amniotic of either
dexamethasone
fluid obtained
before
and 1 to 7 days after administration
or placebo* LIS ratio vahest
Gestation t Treatment
15, 1983 Gynecol.
group
N
(wk)
All mothers: Placebo Steroid
20 25
Mothers of 32 to 34 weeks’ gestation: Placebo Steroid
10 12
Male infants: Placebo Steroid
11 15
32.5 33.3
ft
Female infants: Placebo Steroid
9 10
32.3 33.3
+f 0.9 0.7
"If: 0.5 0.4
32.3 33.3
-0.6 0.5
P
Pretreat7nent
Posttreatment
1.45 + 0.09 1.42 2 0.08
2.02 2.36
1.57 1.40 +* 0.16 0.07
2.32
tt
0.14 0.28
2.11 + f 0.38 0.21
Intet-ual
$
values GroupSO
0.001 0.003
0.32
0.04 0.06
0.87
1.36 1.37
+f 0.08 0.12
1.85 2.45
zt f 0.20 0.46
0.007
0.01
0.24
1.54 1.50
‘-iz 0.08 0.18
2.22 2.23
22
0.04
0.79
0.18 0.14
0.001
*The second amniocentesis was performed 3 to 9 days after the initial procedure on the day of screening for randomization. tMean f SEM. $P values for comparison of the change in L/S ratio between pretreatment and posttreatment values in each group, adjusting for the following variables: pretreatment L/S, gestational age, time interval between amniocenteses. OP values for comparison of the differences in pretreatment and posttreatment L/S ratios between the placebo and steroid groups, i.e., the magnitude of change in L/S ratio after dexamethasone compared to change after placebo injections.
group, these were excluded from the analysis, thereby leaving 45 pairs of specimens of amniotic fluid. In addition, gastric aspirate samples of amniotic fluid from singleton births were available for L/S ratio measurements, including 59 patients who received placebo and 62 in the dexamethasone group; the incidence of prolonged rupture of membranes was 33.9% in control subjects and 30.6% in the steroid group. Although the original plan was to evaluate only the L/S ratio by a standardized procedure, this objective was expanded in October, 1979, to incorporate additional analyses whenever sufficient volumes of amniotic fluid were available. Thus, the absolute concentrations of phosphatidylcholine, disaturated phosphatidylcholine, phosphatidylglycerol, and sphingomyelin were determined on 17 dexamethasone-treated patients by including a sixth perinatal center that employed the same clinical protocol under the direction of an obstetrician who began the clinical trial as a co-investigator at one of the five clinical centers. Although a control group was not available for comprehensive phospholipid analyses, the laboratory measurements were still performed in a blind fashion by coding the pretreatment and posttreatment samples of amniotic fluid and reporting the data for statistical treatment to the coordinating center. The standard procedure for measuring amniotic fluid L/S ratios was similar to that described by Olson and Graven.13 Specimens were kept on ice briefly before extraction of lipids, if analyses were performed the same day, or were otherwise frozen promptly to
maintain phospholipid stability.14 The samples of amniotic fluid were centrifuged once at 1,000 X g for 5 minutes. No bloody or meconium-stained specimens were included. Extraction of lipids was carried out by mixing amniotic fluid vigorously with one volume of methanol and two volumes of chloroform. After centrifugation, the lower chloroform layer was removed, evaporated to dryness under nitrogen, taken up in a small volume of 2 : 1 chloroform : methanol, and chromatographed on a silica gel H thin-layer plate with a solvent system of chloroform : methanol : water (65 : 25 : 4). Standard samples of phospholipids were also applied to the plates. Detection of phospholipids was performed either by charring with sulfuric acid and heat or with bromthymol blue staining. Reflectance densitometry was then performed to calculate the ratio of lecithin to sphingomyelin. Strict quality control measures were employed to verify the reliability of the L/S ratio procedure at the various laboratories. On seven occasions, 12 to 16 standard samples of amniotic fluid adjusted to known L/S ratios were coded and sent to the five perinatal centers. Results of these determinations were then reported to the coordinating center for data analysis. All laboratories were required to meet defined reliability criteria specified by the National Institutes of Health Steering Committee for this study. Analyses of phosphatidylcholine, disaturated phosphatidylcholine, and phosphatidylglycerol were performed in a central laboratory on frozen specimens of amniotic fluid by methods described elsewhere.15z ”
Volun1e 145 Number
Amniotic
fluid phospholipids
after
maternal
dexamethasone
487
4
Table II. Comprehensive assessment of amniotic after dexamethasone treatment*
fluid phospholipids Conce7ltration
Phospholipid
Sphingomyelin Phosphatidylcholine§ Ratio of phosphatidylcholine to sphingomyelin Disaturated phosphatidylcholine Phosphatidylglycerol
in 17 patients
(nmoleslml)
before and
or ratio? Posttreatment
Pretreatment
6.4 + 1.1 10.9 t 2.5 1.5 r 0.2
f valuf?s
7.4 zk 1.6 28.0 -c 6.1 3.8 2 0.8
0.25 0.001 0.005
6.5 + 1.5
15.2 -c 4.3
0.01
Undetectable
Undetectable
NS
*Mean gestational age was 32.5 weeks. tMean -t SEM. $P value for comparison of the change in value in the interval between amniocenteses. IRefers to total phosphatidylcholine.
Briefly, on the day of extraction, an internal standard of [14C]-dipalmitoyl-phosphatidylcholine was added to monitor phospholipid recovery and correct for losses during processing. Lipids were extracted by the addition of four volumes of 2: 1 chloroform: methanol. After centrifugation, the lower chloroform layer was removed, dried under nitrogen, taken up to a 1 ml volume of chloroform, and split into two equal fractions for either two-dimensional, thin-layer chromatography or osmium tetroxide treatment to produce a disaturated phosphatidylcholine fraction according to the method of Mason and associates.r7 The solvent system was chloroform : methanol : water (65 : 45: 5) in the first dimension and tetrahydrofuran : methylal : methanol:2M ammonium hydroxide (10:6:4:1) for the second direction. This system permitted the separation of phosphatidylcholine, sphingomyelin, and phosphatidylglycerol. The lipid spots were scraped from the plate, extracted from the gel by the method of Bligh and Dyer, ‘” and then analyzed for phospholipid phosphorus by the method of Chen and associates.1s Appropriate aliquots were also analyzed for radioactivity in a Beckman LS 7000 scintillation counter. After the second lipid fraction was treated with osmium tetroxide, separation of disaturated phosphatidylcholine was performed with neutral alumina and thin-layer chromatography performed with the first-dimensional solvent described above. Quantitation was again performed by analysis of phospholipid phosphorus. Gas chromatography was performed on the disaturated phosphatidylcholine fraction to ascertain that a high (greater than 95%) content of saturated fatty acids, especially palmitate,‘” was present. Statistical analyses were performed at the coordinating center. The data were examined by means of analysis of covariance and Wilcoxon rank sum procedures, logistic regression techniques, and t tests. Logarithm of
the posttreatment analyses.
L/S ratio
was used in some of the
Results As shown in Table I, the initial L/S ratios, i.e., pretreatment mean values, were nearly identical for the placebo and steroid groups. In addition, mean gestational ages, which were determined at the time of randomization were comparable in the two groups. The average gestational age for the 45 infants was 32.9 weeks. During the 3- to g-day interval between amniocenteses, the mean L/S ratios increased moderately in each group. The mean change in L/S ratio was 0.57 in the placebo group (from 1.42 to 2.02), whereas the delta value was 0.94 in the steroid treatment group (from 1.42 to 2.36). However, statistical comparison of the change in L/S ratio between the two study groups revealed no significant difference, despite a numerically lower incidence of RDS in the 25 infants exposed to dexamethasone in utero (4.0% versus tO.O% in the control group of 20 neonates: p = 0.30). Evaluation of the extent of change in L/S ratios relative t.o predicting RDS revealed that 12 of 20 control patients and 12 of 25 patients in the steroid group demonstrated an increase sufficient to convert their values from immature to mature. The data from pregnancies of 32 to 34 weeks were also analyzed separately in order to minimize the gestational age variable and examine the L/S ratio issue in patients most likely to be responsive to dexamethasone.6 As indicated in Table I, evaluation of change in L/S ratio again revealed no significant differences between the two subgroups, although the delta value was 70% higher in the steroid-treated population. Because of the influence of fetal sex on dexamethasone responsiveness,‘. 50 we also carried out a separate analysis of data for male and female infants. The results pre-
488
Farrell et al.
sented in Table I demonstrate that the changes in L/S ratio during the pretreatment to posttreatment interval were not statistically different for the male and female subgroups. The 121 gastric aspirate samples of amniotic fluid showed an L/S ratio of 2.88 f 0.21 for the placebo group and 2.70 + 0.18 for the dexamethasone-treated infants. Statistical analysis revealed that there were no significant differences between the two groups (P = 0.52). Furthermore, there was no statistical difference when the change in L/S ratio was evaluated in 41 infants who had gastric aspirates taken subsequent to a pretreatment amniocentesis. Nevertheless, the L/S ratios obtained on gastric aspirates were closely associated with the occurrence of RDS (P = 0.0025). The mean 2 SEM value when RDS occurred was 1.71 i: 0.21, whereas single infants without RDS showed an L/S ratio of 2.97 % 0.15. Thus, although gastric aspirate phospholipids did not statistically relate to treatment groups, they did confirm the expectation that the probability of RDS decreases with increasing L/S ratios. The results obtained from determination of absolute phospholipid concentrations in 17 dexamethasonetreated patients are shown in Table II. The mean gestational age was 32.5 weeks, and the average interval between amniocenteses was 6.8 days. Of interest was the finding that phosphatidylglycerol could not be detected in any of the 34 specimens of amniotic fluid. Although sphingomyelin did not change during the interval between amniocenteses, there were differences in the mean values for phosphatidylcholine (P < 0.005) and disaturated phosphatidylcholine (P < 0.01); however, the degree of the increase was relatively small compared to the tenfold rise normally observed after 34 weeks’ gestation (Engle and Farrell, unpublished data). Furthermore, the percentage of the disaturated phosphatidylcholine fraction did not increase significantly as a result of dexamethasone treatment. On the other hand, comparison of the concentration of phospholipids (by means of a logistic regression model) with neonatal pulmonary status suggested an association with the occurrence of RDS in these 17 pregnancies. In particular, the ratio of phosphatidylcholine to sphingomyelin in 12 patients who did not develop RDS (4.56 t 1.05) was significantly higher than that in five pregnancies that led to neonatal RDS (1.83 2 0.27) (P < 0.03, based on t test analysis of data obtained from the second sample of amniotic fluid); the change in ratio during the pretreatment to posttreatment interval was also greater in the non-RDS group (2.92 ? 0.90 versus 0.56 + 0.51 in those who developed RDS; P < 0.04). Additionally, the percentage of disaturated
February Am. J. Obstet.
15, 1983 Gynecol.
phosphatidylcholine relative to total phosphatidylcholine was higher in patients who did not develop RDS (65.3 k 8.27 versus 33.9 & 5.20 in the RDS group; P < 0.04) because of a much greater increase in the disaturated phosphatidylcholine concentration during the interval between amniocenteses that amounted to 11.3 + 4.23 nmoles/ml compared to 2.50 2 1.88 in the RDS group (P = 0.08).
Comment Although prenatal corticosteroid therapy has been established as a means of lowering the incidence of neonatal RDS in appropriate pregnancies,’ there has been continued controversy over the issue of whether amniotic fluid indices of fetal lung development change in response to steroid administration. On the basis of results obtained in the first clinical trial, Liggins and Howie6* ’ concluded that there was no difference in the L/S ratio after betamethasone therapy. Soon thereafter, Spellacy and associatesi reported a greater rate of rise in the L/S ratio after dexamethasone injections than in uninjected control pregnancies (delta values of 0.68 and 0.40, respectively; P < 0.05). Subsequently, there have been at least 14 investigations of this issue, as reviewed by Morrison and associates,” with about half revealing an increase in the L/S ratio after treatment with various steroid hormones and the remainder demonstrating no change. Even when the L/S ratio has been reported to increase significantly, however, the extent of change has been modest. Because several of the investigations did not include a control group, it is difficult to isolate the steroid treatment variable from other factors, such as gestational age. Arias and associatesi2 attempted to increase the sensitivity of amniotic fluid phospholipid analysis by studying both the L/S ratio and a qualitative measure of disaturated phosphatidylcholine. They reported significantly greater L/S values 1 week after betamethasone treatment than in control pregnancies, but the extent of change was only slight (mean + SD, 1.49 2 0.21 versus 1.09 ? 0.29, respectively). Although six control pregnancies did not show a change in the relative amount of disaturated phosphatidylcholine, 12 of 18 betamethasone-injetted patients had an increase in this fraction within 14 days of treatment (P = 0.016). The study reported herein represents a carefully controlled evaluation of the L/S ratio and several other phospholipid indices of fetal lung development after either dexamethasone injections or the administration of placebo (drug vehicle). In order to minimize the impact of variables which can affect amniotic fluid phospholipids,2 we excluded diabetic pregnancies and
Volume Number
145 4
limited the influence of gestational age, as follows. (1) Pregnancies between 26 and 34 weeks’ gestation were identified as the primary target population for entry, which resulted in a mean gestational age in this study of 32.9 weeks. (2) The second specimen of amniotic fluid was obtained within 9 days of the initial amniocentesis (within 1 week after completion of the intramuscular injections). Although 1 week may not provide adequate time for the acceleration of fetal lung development to be reflected in amniotic fluid, various observations in normal pregnancies suggest that phospholipid indices can increase during that intervaL3 It is equally clear that the longer the period of time before a second amniocentesis, the higher the probability that advancing gestation per se will increase the concentration of lung surfactant-derived phospholipids. In order to optimize experimental design, therefore, it is necessary to limit the interval after steroid treatment before the second amniocentesis. The pretreatment L/S ratio values found in 20 control pregnancies and 25 dexamethasone-injected women indicate a high probability of lung immaturity at the time the intramuscular injections were initiated. Although a significant change in L/S ratio occurred in both groups, the major conclusions from our results are the following. (1) The magnitude of differences in each group is small and only slightly greater than the expected coefficient of variation for an L/S determination. (2) The extent of change after dexamethasone treatment is not significantly greater than the L/S difference in the placebo group. Furthermore, gastric aspirate samples did not differentiate the two study groups, even though these measurements were sufficiently sensitive to identify infants who would develop RDS. Our objective of achieving greater sensitivity and specificity by measuring phospholipids, such as disaturated phosphatidylcholine and phosphatidylglycerol, was hampered by the lack of a matched control group for these assessments. Nevertheless, we are able to conclude that phosphatidylglycerol determinations are not helpful in monitoring the effects of steroid treatment or predicting neonatal RDS for pregnancies of this gestational age. Measurement of the absolute concentration of phosphatidylcholine fractions offers more promise, on the basis of the 256% increase in phosphatidylcholine, a 233% increase in disaturated phosphatidylcholine, and the relationship of these values to the occurrence of RDS. Although the rise in total phosphatidylcholine and the disaturated fraction provides evidence of a fetal lung maturational response to dexamethasone, we must emphasize that the increased
Amniotic
fluid
phospholipids
after
maternal
dexamethasone
489
gestational age is also a factor that influences these data. Furthermore, the magnitude of change in these indices of lung development is relatively modest and may not be sufficient for reliably distinguishing a favorable fetal response. In summary, our results indicate that current tests used routinely to assess fetal lung maturity do not provide a means for reliable detection of fetal pulmonary responses to maternally administered cc>rticosteroids. Because an accurate test is currently unavailable, we conclude that, at present, repeat amniocentesis after steroid administration is not of clinical \alue for monitoring corticosteroid therapy per se. Further research on determination of more sensitive indices of lung development will be necessary before use of such assessments can be applied in clinical perinatology.
The Collaborative Study on Antenatal Steroid Therapy was carried out by the following investigators who also served as members of the Steering Committee: Richard Depp, M.D., and John Boehm, M.D. (Northwestern University, Chicago, Illinois); Richard Zachman, Ph.D., M.D., and Luis Curet, M.D. (University of Wisconsin, Madison, Wisconsin); Charles R. Bauer, M.D.; Louis Fernandez-Rocha, M.D., and Gene Burkett, M.D. (University of Miami, Miami, Florida); Sheldon Korones, M.D., John Morrison, M.D.. Jack Schneider, M.D., and Garland Anderson. M.D. (tiniversity of Tennessee, Memphis, Tennessee): Henrique Rigatto, M.D., Leo Peddle, M.D., and Frank Manning, M.D. (University of Manitoba, Winnipeg, Canada): W. Kenneth Poole, Ph.D., and Vijaya Rao, Ph.D. (Research Triangle Institute, Research Triangle Park, North Carolina); David Fukushima, Ph.D., John O’Connor, Ph.D., and Jack Kream, Ph.D. (Mont&ore Hospital, New York, New York). We also thank Betty Hastings for help with data processing and D. Jeannette Brown for performance of amniotic fluid analyses. A Policy-Data Monitoring Panel is also acknowledged, consisting of Brian Little, M.D. (Chairman), Mary Ellen Avery, M.D., David DeMets, Ph.D., Max Halperin, Ph.D., Patricia King, M.D., Arthur Parmelee, M.D., Samuel Solomon, Ph.D., F.R.S.C., David Sylwester, Ph.D., and James Ware, Ph.D. The National Heart, Lung, and Blood Institute Program Office representatives were Bitten Stripp, Ph.D., Project Officer, and Claude Lenfant, M.D., Director, Division of Lung Diseases, Bethesda, Maryland. On request of the Division of Lung Disease, National Heart, Lung, and Blood Institute, Merck Sharp & Dohme provided the drug and placebo preparations used in this study. This acknowledgement of appreciation is in no way an endorsement 01 a particular product.
490
Farrell et al.
REFERENCES
1. Perelman, R. H., and Farrell, P. M.: Decreasing incidence of fatal hyaline membrane disease based on analysis of 19681978 neonatal mortality statistics, Pediatrics 70: 570, 1982. 2. Tsao, F. H. C., and Zachman, R. D.: In Farrell, P. M., editor: Lung Development: Biological and Clinical Perspectives, New York, 1982, vol. 2, Academic Press, Inc., pp. 167-203. 3. Torday, J., Carson, L., and Lawson, E. E.: Saturated phosphatidylcholine in amniotic fluid and prediction of the respiratory distress syndrome, N. Engl. J. Med. 301:1013, 1979. 4. Tsao, F. H. C., and Zachman, R. D.: Determination of phosphatidylglycerol in amniotic fluid by a simple onedimensional thin-layer chromatography method, Clin. Chem. Acta 118:109, 1982. 5. Collaborative Group on Antenatal Steroid Therapy: Effect of antenatal dexamethasone administration on the prevention of respiratory distress syndrome, AM. J. OBSTET. GYNECOL. 141:276, 1981. 6. Liggins, G. C., and Howie, R. N.: A controlled trial of antepartum glucocorticoid treatment for prevention of the respiratory distress syndrome in premature infants, Pediatrics 50:515, 1972. 7. Zachman, R. D.: In Farrell, P. M., editor: Lung Development: Biological and Clinical Perspectives, New York, 1982, vol. 2, Academic Press, Inc., pp. 275-296. 8. Ballard, R. A., Ballard, P. L., Granberg, J. P., and Sniderman, S.: Prenatal administration of betamethasone for prevention of respiratory distress syndrome, J. Pediatr. 94:97, 1979. 9. Liggins, G. C., and Howie, R. N.: In Cluck, L., editor: Modern Perinatal Medicine, Chicago, 1974, Year Book Medical Publishers, Inc., pp. 415-424. W. N., Buhi, W. C., Riggall, F. C., and Hol10. Spellacy, singer, K. L. Human amniotic fluid lecithimsphingomyelin ratio changes with estrogen or glucocorticoid treatment, AM. J. OBSTET. GYNECOL. 115:216, 1973.
February Am. J. Obstet.
15. 1983 Gynecol.
11. Morrison, J. C., Whybrew, W. D., Bucovaz, E. T., Wiser, W. L., and Fish, S. A.: The lecithin/sphingomyelin ratio in cases associated with fetomaternal disease, AM. J. OBSTET. GYNECOL. 127:363, 1977. 12. Arias, F., Pineda, J., and Johnson, I,. W.: Changes in human amniotic fluid lecithinisphingomyelin ratio and dipalmitoyl lecithin associated with maternal betamethasone therapy, AM. J. OBSTET. GYNECOL. 133:894, 1979. 13. Olson, E. B., and Graven, S. N.: Comparison of visualization methods used to measure the lecithin/sphingomyelin ratio in amniotic fluid, Clin. Chem. 20: 1408, 1974. 14. Schwartz, D. B., Engle, M. J.< Brown, J., and Farrell, P. M.: The stability of phospholipids in amniotic fluid, AM.J.OBSTET. GYNECOL. 141:294, 1981. 15. Curbelo, V., Gail, D. B., and Farrell, P. M.: Determination of disaturated lecithin in rhesus monkey amniotic fluid as an index of fetal lung maturity, AM. J. OBSTET. GYNECOL. 131:764, 1978. 16. Perelman, R. H., Engle, M. J., Kemnitz, J. W., Kotas, R. V., and Farrell, P. M.: Biochemical and physiological development of the fetal rhesus lung, J. Appl. Ph$iol. 53:230. 1982. 17. Mason; R. J., Nellenbogen, J., and Clements, J. A.: Isolation of disaturated phosphatidylcholine with osmium tetroxide, J. Lipid Res. 17:281, 1976. 18. Bligh, E. G., and Dyer, W. J.: A rapid method of total lipid extraction and purification, Can. J. Biochem. 37: 911,1959. 19. Chen, P. S., Toribara, T. Y., and Warner, H.: Microdetermination of phosphorus, Anal. Chem. 28: 1756, 1956. 20. Papageorgiou, A. N., Desgranges, M. F., Masson, M., Colle, E., Shatz, R., and Gelfand, M. M.: The antenatal use of betamethasone in the prevention of respiratory distress syndrome. A controlled double-blind study, Pediatrics 63:73, 1979. 21. Epstein, M. F., Farrell, P. M., and Chez, R. A.: Fetal lung lecithin metabolism and the amniotic fluid L/S ratio in rhesus monkey gestations, AM. J. OBSTET. GYNECOL. 125:545, 1976.