Effects of maternal dexamethasone administration on fetal Doppler flow velocity waveforms

Authors

  • Yvon Chitrit,

    Consultant, Corresponding author
    1. Department of Obstetrics and Gynaecology, Robert Ballanger Hospital, Aulnay-sous-Bois, France
      Correspondence: Dr Y. Chitrit, Department of Obstetrics and Gynaecology, Centre Hospitalier Général R. Ballanger, Boulevard Robert Ballanger, F-93602 Aulnay-sous-Bois, France.
    Search for more papers by this author
  • Patrick Caubel,

    Consultant
    1. Department of Obstetrics and Gynaecology, Robert Ballanger Hospital, Aulnay-sous-Bois, France
    Search for more papers by this author
  • Rafael Herrero,

    Consultant
    1. Department of Obstetrics and Gynaecology, Robert Ballanger Hospital, Aulnay-sous-Bois, France
    Search for more papers by this author
  • Anne Lyse Schwinte,

    Midwife
    1. Department of Obstetrics and Gynaecology, Robert Ballanger Hospital, Aulnay-sous-Bois, France
    Search for more papers by this author
  • Denis Guillaumin,

    Consultant
    1. Department of Obstetrics and Gynaecology, Robert Ballanger Hospital, Aulnay-sous-Bois, France
    Search for more papers by this author
  • Marie-Christine Boulanger

    Consultant
    1. Department of Obstetrics and Gynaecology, Robert Ballanger Hospital, Aulnay-sous-Bois, France
    Search for more papers by this author

Correspondence: Dr Y. Chitrit, Department of Obstetrics and Gynaecology, Centre Hospitalier Général R. Ballanger, Boulevard Robert Ballanger, F-93602 Aulnay-sous-Bois, France.

Abstract

Objective To investigate the effects of maternal dexamethasone administration on umbilical and fetal cerebral artery flow velocity waveforms.

Design Cross-sectional study.

Setting Department of Obstetrics and Gynaecology, Robert Ballanger Hospital, Aulnay-sous-Bois, France.

Sample Twenty-six pregnant women with singleton pregnancies considered at risk for preterm delivery. At baseline, all pregnancies had normal fetoplacental vascular resistance.

Methods These women were given weekly six intravenous doses of 4 mg of dexamethasone eight hours apart.

Main outcome measures Doppler studies were performed from both umbilical artery (UA) and fetal middle cerebral artery (MCA) before (day 0), during (day 2), immediately after (day 4) and shortly after (day 7) every steroid course.

Results No significant variation was noted in both umbilical artery pulsatility index (PI) and fetal heart rate through dexamethasone therapy. Compared with mean initial values, we found on day 4 a significant decrease in MCA PI of 0.28 (F = 7.17, P < 0.001) and a significant increase in UA:MCA PI ratio of 0.08 (F = 3.85, P= 0.013); in contrast no significant change was documented on days 2 and 7 in both MCA pulsatility index and UA:MCA PI ratio. After multiple regression analysis, only the decrease in fetal middle cerebral artery pulsatility index on day 4 remained significant (F=5.84, P= 0.001).

Conclusions The current study finds in healthy fetuses a transient, significant and unexplained decrease in fetal middle cerebral artery impedance on the fourth day following maternal dexamethasone administration. Further basic research and clinical studies including larger sample sizes or pregnancies with fetoplacental dysfunction are needed.

INTRODUCTION

Maternal administration of synthetic corticosteroids (betamethasone or dexamethasone), by accelerating the maturity of the fetal lung, reduces neonatal mortality, respiratory distress syndrome, intraventricular haemorrhage and necrotising enterocolitis1–4 in preterm infants. Previous studies have shown that steroids have an effect on fetal behaviour and fetal heart rate variability5–13. In fact, conflicting results concerning the effects of betamethasone and dexamethasone on fetal heart pattern have been reported. Recently, Magee et al.5 found that betamethasone and dexamethasone were associated with an increase in long term and short term variability and decreased fetal movements on the first day after steroid administration followed by a decline in fetal heart rate variability on the second day. Similarly, the trial of Mulder et al.6 reported a transient decrease in fetal heart rate variability and in fetal body and breathing movements after betamethasone; conversely, a rise in fetal heart rate variability was noted on the first day after dexamethasone. These effects are more obvious with betamethasone than dexamethasone5,6,13.

Evaluation of fetal wellbeing with Doppler waveform studies after maternal corticosteroid administration is therefore important. Knowledge of fetal haemodynamic effects of exogenous corticosteroids is limited. Little is known about the impact of antenatal corticosteroid therapy on Doppler waveforms in fetal arteries14–20. Meizner et al.14 reported no effect on blood flow velocimetry in the umbilical artery throughout a course of steroids; Cohlen et al.15 found, following antenatal betamethasone therapy, no significant change in the pulsatility index of any of the fetal blood vessels; Wallace et al.19 noted, in pregnancies where umbilical artery end-diastolic flow was absent, an association between betamethasone treatment and decreased placental vascular resistance; Dubiel et al.16 suggested that resistance to blood flow in the intrapulmonary arteries decreased after maternal betamethasone administration; and Wasser-strum et al.17 recorded no significant effect of betamethasone on the ductus arteriosus. The purpose of the present study was to evaluate the effects of maternal dexamethasone administration on both umbilical artery and fetal middle cerebral artery Doppler velocities in pregnancies with normal fetoplacental vascular resistance.

METHODS

Twenty-six women with singleton pregnancies who were expected to deliver preterm were included in the study between January 1997 and November 1998. This investigation was approved by the authors' institutional review board, and oral informed consent was obtained from all participating women. At the time of initial scanning, all pregnancies had umbilical artery flow-velocity waveforms values above the fifth centile according to the reference limits as published by Arduini et al.21. There were no infants with major structural malformations or abnormal karyotype. Gestational age was calculated according to the date of the last menstrual period and confirmed by first trimester ultrasound. If there was a discrepancy (more than five days), ultrasound was used to determine gestational age. Gestational age at initial presentation ranged from 26 to 33 weeks (median 30.0). The median age of the 26 women was 29 years (range 19 to 39). Gravidity ranged from one to nine (median two) and parity from zero to six (median one). Preterm birth was anticipated on the basis of: 1. preterm contractions of the uterus (n= 15); 2. mild or moderate pre-eclampsia (n= 2); 3. suspected small for gestational age (n= 3); 4. placenta praevia (n= 2); 5. intrahepatic cholestasis (n= 2); 6. preterm premature rupture of the membranes (n= 1) and polyhydramnios (n= 1).

These 26 women were given a course of corticosteroid therapy every week. For those who received several courses of steroid, only the first course with complete data was considered. Each course consisted of six 4 mg doses of dexamethasone administered intravenously eight hours apart. The first injection was administered on day 1 around 8:00 a.m. All women except one received additional medication: salbutamol (n= 17), antihypertensive drugs (n= 1), aspirin (n= 4), nitric oxide (n= 1), spasmolytic drugs (n= 1) and cholestyramine (n= 1). In all cases, the same medication was maintained throughout the study, although the route of administration could differ, especially for salbutamol (intravenous or oral).

Doppler studies were performed before (day 0), during (day 2), immediately after (day 4) and shortly after (day 7) the course of dexamethasone therapy. A combined real-time pulsed Doppler system (50 Hz highpass filter, Doppler gate 2 mm) fitted with a 3.75 MHz curvilinear probe (Toshiba SSA, Tokyo, Japan) was used. The spatial peak temporal average power did not exceed 87 mW/cm2. The Doppler angle of insonation was less than 46°, the sweep speed was 2.5 cm/s and the pulse repetition frequency ranged from 3.5 Khz to 5.5 Khz. The women rested in a semi-recumbent position during the Doppler examination. All measurements were performed by the same physician during fetal quiescence and apnoea. Blood flow velocity waveforms were obtained from both the umbilical artery and the fetal middle cerebral arteries. The umbilical artery was insonated close to its placental insertion and the middle cerebral artery about 1 cm distal to its origin from the internal carotid artery. For each artery, the image was frozen when consecutive similar waveforms of good quality were obtained. The outer envelope of each flow velocity-profile was traced and velocity indices were derived from the built-in software. The pulsatility index was measured, a measurement of downstream vascular resistance22 which is independent of the angle of insonation. We calculated the pulsatility index for each vessel by averaging the first two good quality pulsatility indexes obtained from two consecutive waveforms. A copy of the recordings was printed and kept. Intra-observer coefficients of variation were 2.1% and 3.2% for measurements in the umbilical artery and middle cerebral artery, respectively. In addition, the fetal heart rate was derived from the built-in software while measuring the velocity waveforms in the umbilical artery and middle cerebral artery. The umbilical artery:middle cerebral artery pulsatility index ratio (UA:MCA PI ratio) was also estimated.

We estimated that at least 22 women would be required to detect, with a power of 80% at a significance level of 0.05, a 20% difference in both umbilical artery and middle cerebral artery pulsatility indices (0.18, SD 0.30), in UA:MCA PI ratio (0.08, SD 0.13) and in fetal heart rate (6, SD 8). Continuous variables are presented as mean (standard deviation or 95% confidence intervals) or median (range), assessed for normality or not, and comparisons were made using analysis of variance, Student's t test or the Mann-Whitney U test as appropriate. Dependent variables were included in multivariate analyses based on multiple logistic-regression models. Categorical variables are given as number (percent) and associations were tested using the χ2 test. A P value < 0.05 was considered to indicate statistical significance.

RESULTS

Out of the 26 pregnant women, 22 were delivered vaginally and four by caesarean section. Gestational age at delivery ranged from 32 to 41 weeks (median 36.5). Five women were delivered before 34 weeks, eight between 34 and 36 weeks and 13 at 37 weeks or later. There were no perinatal deaths. Birthweight ranged from 1430 to 4360 g (median 2760). Three neonates had birthweights below the fifth centile, according to the reference ranges as published by Alexander et al.23. Neonatal care unit admission was necessary for eleven neonates. Median duration in the neonatal care unit was 21 days (range 3 to 33) for these infants.

Completed series of examinations with adequate signals from both umbilical artery and fetal middle cerebral artery were obtained in all 26 women before and after maternal dexamethasone administration (Fig. 1–4). Mean (SD) values measured on days 0, 2, 4 and 7 in both umbilical artery and middle cerebral artery pulsatility indexes, in umbilical artery: MCA PI ratio and in fetal heart rate are listed in Table 1. From baseline, no significant change was observed in both umbilical artery pulsatility index (F = 0.085, P= 0.968) and fetal heart rate (F = 0.736, P= 0.534) through dexamethasone administration (Table 2).

Figure 1.

Umbilical artery (UA) pulsatility index (PI) values before (day 0), during (day 2), immediately after (day 4), and shortly after (day 7) maternal dexamethasone administration according to gestational age at initial presentation.

Figure 2.

Middle cerebral artery (MCA) pulsatility index (PI) values before (day 0), during (day 2), immediately after (day 4) and shortly after (day 7) maternal dexamethasone administration according to gestational age at initial presentation.

Figure 3.

Umbilical artery: middle cerebral artery pulsatility index ratio (UA:MCA PI ratio) values before (day 0), during (day 2), immediately after (day 4) and shortly after (day 7) maternal dexamethasone administration, according to gestational age at initial presentation.

Figure 4.

Fetal heart rate (FHR) values before (day 0), during (day 2), immediately after (day 4), and shortly after (day 7) maternal dexamethasone administration according to gestational age at initial presentation.

Table 1.  Mean (SD) values measured before (day 0), during (day 2) immediately after (day 4) and shortly after (day 7) maternal dexamethasone administration in umbilical artery pulsatility index (UA PI), middle cerebral artery PI (MCA PI), umbilical artery:middle cerebral artery pulsatility index ratio (UA:MCA PI ratio) and fetal heart rate (FHR).
 Day 0Day 2Day 4Day 7
UA PI1.08 (0.17)1.05 (0.22)1.06 (0.18)1.07 (0.18)
MCA PI1.91 (0.40)1.82 (0.26)1.63 (0.31)1.91 (0.31)
UA:MCA PI ratio0.59 (0.16)0.59 (0.12)0.66 (0.14)0.57 (0.13)
FHR148 (9)148 (9)148 (8)146 (8)
Table 2.  From baseline (day 0). mean (95% CI) changes in umbilical artery pulsatility index (UA PI), fetal middle cerebral artey PI (MCA PI), umbilical artery:middle cerebral artery pulsatility index ratio (UA:MCA PI ratio) and fetal heart rate (FHR) during (day 2), immediately after (day 4) and shortly after (day 7) dexamethasone administration.
Change from day 0Day 2Day 4Day 7
  1. *F= 0.085, P= 0.968; **F= 7.17, P <0.001; F= 3.85, P= 0.013; F= 0.736, P= 0.534.

In UAPI*−0.02 (−0.11 to 0.08)−0.01 (−0.01 to 0.06)<−0.01 (−0.08 to 0.08)
In MCA PI**−0.09 (−0.23 to 0.05)−0.28 (−0.46 to-0.10)<−0.01 (−0.15 to 0.15)
In UI:MCA PI ratiot<−0.01 (−0.06 to 0.06)−0.08 (0.01 to 0.15)−0.01 (−0.08 to 0.05)
In FHR−0.23 (−4.11 to 3.65)−0.54 (4.33 to 3.25)−2.46 (−5.93 to 1.01)

Compared with the mean value at initial presentation, there was a significant decrease in middle cerebral artery pulsatility index of 0.28 on day 4 (Table 2); in contrast, no significant change was documented on day 2 and on day 7 (F = 7.17, P < 0.001). Indeed, the mean middle cerebral artery pulsatility index decreased by 11.4% (95% CI −20.6 to −2.09) of the initial value on day 4 and by 1.74% (95% CI −10.1 to 6.58) on day 2, and increased by 3.52% (95% CI −7.00 to 14.0) on day 7. Similarly, compared with the control day, a significant increase of 0.08 in the mean UA:MCA PI ratio was noted on day 4 (Table 2) whereas no significant change was observed on day 2 and on day 7 (F = 3.85, P= 0.013). Thus, from baseline, the mean UA:MCA PI ratio increased by 19.3% (95% CI 5.76 to 32.9) on day 4, by 3.83% (95% CI −5.46 to 13.1) on day 2 and by 2.29% (95% CI −8.63 to 13.2) on day 7. In addition, changes in middle cerebral artery pulsatility index and in UA:MCA PI ratio on day 4 were significantly correlated neither with gestational age at recording (r=−0.155, P= 0.454 and r= 0.035, P= 0.867, respectively) or at delivery (r= 0.312, P= 0.121 and r=−0.111, P= 0.595, respectively), nor with changes in fetal heart rate on day 4 (r=−0.304, P= 0.132 and r=−0.16, P= 0.939, respectively). After fitting multiple logistic regression models including the dependent variables gestational age at baseline and at delivery and fetal heart rate, the decrease in fetal MCA PI at day 4 following dexamethasone administration remained significant (F = 5.84, P= 0.001), but changes in UA:MCA PI ratio did not (F = 0.810, P= 0.491).

From the initial values, there was a greater than 10% decrease in middle cerebral artery pulsatility index in six (23.1%) fetuses on day 2, in 14 (53.9%) on day 4 and in five (19.2%) on day 7, whereas a greater than 10% increase was noted in four, four and five fetuses on days 2, 4 and 7, respectively. Similarly, compared with the control day, there was a greater than 10% increase in UA:MCA PI ratio in 11 (42.3%) pregnancies on day 2, in 16 (61.5%) on day 4 and in nine (34.6%) on day 7; a greater than 10% decline was showed in seven, four and seven pregnancies on days 2, 4 and 7, respectively. Middle cerebral artery pulsatility index ranged on day 4 from 1.19 to 2.27 (median 1.60); of the eleven (42.3%) fetuses of whom the pulsatility index on day 4 was below the 5th centile21, 10 had shown normal pulsatility index values before therapy. On the other hand, UA:MCA PI ratio ranged on day 4 from 0.42 to 0.88 (median 0.69); all pregnancies had normal ratio values21 on this day. Some neonatal outcomes varied significantly depending on whether the fetuses had a greater than 10% decrease in middle cerebral artery pulsatility index on day 4 or not (Table 3); especially, there were significant differences in duration of gestation at delivery (mean 35.3 weeks, SD 2.43 versus mean 37.3 weeks, SD 2.46) and in birthweight (mean 2447 g, SD 647 versus mean 3014 g, SD 561) between fetuses showing a greater than 10% decrease in MCA pulsatility index and those not showing that change. In contrast, perinatal outcome was not influenced by percentage change in the UA:MCA PI ratio on day 4 (Table 4).

Table 3.  Neonatal outcome depending on whether the fetuses had a ≥ 10% decrease in middle cerebral artery pulsatility index (MCA PI) on day 4 or not. GA = gestational age; NCU = neonatal care unit. Values are given as mean (SD), n [%] or median {range}.
 Decrease in MCA PI 
Outcome≥10%<10%P
GA at baseline (weeks)30.1 (1.51)29.7 (2.02)0.498
 n= 14n= 12 
GA at delivery (weeks)35.3 (2.43)37.3 (2.46)0.044
 n= 14n= 12 
Birthweight (g)2447 (647)3014 (561)0.026
 n= 14n= 12 
APGAR score at 5 min10.0 (9.00 to 10.0)10.0 (7.00 to 10.0)0.404
 n= 14n= 12 
Duration in NCU (days)24.4 (8.98)12.5 (8.43)0.059
 n= 7n= 4 
Deliveryn= 14n= 12 
≤ 37 weeks9 [64.3]4 [33.3]0.238
> 37 weeks5 [35.7]8 [66.7] 
Birthweightn= 14n= 12 
≤ 5th centile2 [14.3]1 [8.33]≥ 0.999
≥ 5th centile12 [85.7]11 [91.7] 
Admission in NCUn= 14n= 12 
Yes7 [50.0]4 [33.3]0.453
No7 [50.0]8 [66.7] 
Table 4.  Neonatal outcome depending on whether the pregnancies had a ≥ 10% increase in umbilical artery: middle cerebral artery pulsatility index ratio (UA:MCA PI ratio) on day 4 or not. GA = gestational age; NCU = neonatal care unit. Values are given as mean (SD), n [%] or median {range}.
 Increase in ratio 
Outcome≥10%<10%P
GA at baseline (weeks)29.8 (1.65)30.2 (1.93)0.533
 n= 16n= 10 
GA at delivery (weeks)35.9 (2.74)36.7 (2.45)0.480
 n= 16n= 10 
Birthweight (g)2688 (755)2743 (516)0.840
 n= 16n= 10 
APGAR score at 5 min10.0 (9.00 to 10.0)10.0 (7.00 to 10.0)0.255
 n= 16n= 10 
Duration in NCU (days)23.0 (8.54)15.0 (12.4)0.233
 n= 1n= 4 
Deliveryn= 16n= 10 
 ≤ 37 weeks9 [56.3]4 [40.0]0.688
 > 37 weeks7 [43.7]6 [60.0] 
Birthweightn= 16n= 10 
 ≤ 5th centile2 [12.5]1 [10.0]≥ 0.999
 > 5th centile14 [87.5]9 [90.0] 
Admission in NCUn= 16n= 10 
 Yes7 [43.8]4 [40.0]≥ 0.999
 No9 [56.2]6 [60.0] 

DISCUSSION

In the present study, only healthy fetuses without fetoplacental insufficiency were considered. Our data show no significant change in umbilical artery impedance during the course of steroid administration. These findings are in agreement with all previous studies8,14,15 except one19. In the latter, it is noteworthy that Wallace et al.19 reported, in pregnancies with fetoplacental dysfunction, a decreased placental vascular resistance as reflected by waveforms obtained from umbilical artery within 24 hours after betamethasone administration. Subsequently, the different impact of corticosteroids in umbilical artery flow velocity waveforms, observed by Wallace et al.19, may be attributed to the characteristics of the fetuses.

In pregnancies with normal fetoplacental vascular resistance, middle cerebral artery pulsatility index decreased significantly on day 4 following dexamethasone administration and returned to pretreatment values within a week. Moreover, a transient and dramatic increase in UA:MCA PI ratio is noted on day 4 but this change does not remain significant after multiple regression analysis. These findings are influenced neither by gestational age at initial presentation or at delivery, nor by changes in fetal heart rate. Since the other drugs (e.g. tocolytic or antihypertensive agents) were introduced before the first Doppler examination and discontinued after the last evaluation, they should not have influenced the changes during dexamethasone therapy. It must be noted that like other authors5,6 who observed most changes in both fetal behaviour and fetal heart variation after stopping corticosteroid administration, we found that most effects in Doppler velocity waveforms were also manifest after discontinuing the therapy. Cohlen et al.15 found that the pulsatility indices of all fetal cerebral arteries decreased immediately after stopping maternal betamethasone administration. The lack of statistical significance in their study may be due to the small sample size or to the choice of steroid. It is of interest to note that the choices of betamethasone or dexamethasone, their dosages and their schedules of administration are empirical in the majority of obstetric units.

In the present analysis, a greater than 10% decrease in middle cerebral artery pulsatility index and a greater than 10% increase in UA:MCA PI ratio were noted on day 4, with respect to the control day, in 53.9% and 61.5% of the 26 cases, respectively. Middle cerebral artery pulsatility index values are below the 5th centile on day 4 in 11 (42.3%) fetuses; by contrast, there were no pregnancies with UA:MCA PI ratio values above the reference limits for gestational age on this day. This latter finding may suggest the absence of an association between maternal dexamethasone administration and fetal distress24. Furthermore in our study, a significant decrease in both duration of gestation at delivery and birthweight is reported in fetuses exhibiting a greater than 10% decrease in MCA PI; however, preterm births and neonates with birthweight below the fifth centile are not more frequent in these infants (Table 4). Thus, these findings should be interpreted with caution.

The presumed mechanisms that explain a transient reduction in middle cerebral artery impedance after maternal dexamethasone administration remain unclear. It is unlikely that the changes in fetal cerebral blood flow after dexamethasone therapy are due to constriction of the ductus arteriosus17,18,20. Steroids may act through circulatory mechanisms. Accordingly, experimental studies25,26 on fetal sheep have shown that maternal corticosteroid administration may cause considerable changes in fetal blood pressure, heart rate and blood volume. Another hypothesis is that corticosteroids exert a direct effect on the fetal brain: glucocorticoid receptors have been located throughout the fetal brain, especially in the hippocampus27,28. Studies on vascular tone after corticosteroid administration have shown varying results. While the predominant effect is increased vascular tone, glucocorticoids have potential biochemical actions for both vasoconstriction (e.g. reduced production of prostacyclin, inhibition of nitric oxide synthase) and vasodilatation (e.g. increased cyclic adenosine monophosphate responses) and may act by nongenomic mechanisms29. Furthermore, the responses may be site-specific and the results of opposing effects may vary with the local milieu. Another possible mechanism for the observed changes in fetal middle cerebral artery flow velocity waveforms is related to placental secretion of corticotrophin-releasing hormone. Robinson et al.30in vitro and Korebrits et al.31in vivo have demonstrated that glucocorticoids stimulate placental synthesis and secretion of corticotrophin-releasing hormone. Furthermore, recent investigations have shown that corticotrophin-releasing hormone causes vasodilatation via induction of nitric oxide synthase32 in the human fetal and placental circulations33. Finally, Dörr et al.34 reported after antenatal betamethasone therapy a dramatic decrease of glucocorticoid plasma levels in the umbilical artery and vein at birth and a complete recovery two hours after. A hypothetical suppression of fetal glucocorticoid secretion by negative-feedback inhibition35 of fetal adrenocorticotrophic hormone after maternal steroid administration may account for the transient change in fetal cerebral artery impedance.

CONCLUSION

This study finds in healthy fetuses a transient, significant and unexplained decrease in middle cerebral artery impedance on the fourth day following maternal dexamethasone administration. There is no evidence that this effect is harmful or beneficial to the fetus. In addition, there may be other possible effects of dexamethasone on the fetal brain. Further basic research and clinical studies including larger sample sizes or pregnancies with fetoplacental dysfunction are needed.

Ancillary