Main

RDS due to immature lung development is the major cause of early neonatal mortality in preterm infants(1, 2). Both prenatal steroid treatment(3–5) and postnatal surfactant therapy(6–9) have significantly reduced morbidity and mortality due to RDS. There is a well documented male disadvantage to mortality from RDS(1, 10–12). After the introduction of prenatal steroid administration as a means of accelerating lung maturation(13), it was suggested that this male/female discordance could be reduced or even eliminated(14).

The lung maturational effects of prenatal steroids are well documented both clinically(14–17) and in animal studies(18–27). However, rather than reducing the male disadvantage, a number of these studies suggest that premature male infants are less likely to benefit from steroid treatment than age-matched females(14, 16, 17). These findings are not conclusive, however, as Crowley et al.(15) found no gender difference when they performed a meta-analysis of 12 randomized clinical trials.

Interpretation of results from clinical trials is often complicated by variables that are inherently difficult to control, such as total steroid exposure time and dose, delivery procedure, and postnatal management. Furthermore, the inclusion in clinical trials of infants spanning a wide range of gestational ages may further complicate interpretation of results, as responsiveness to steroids may vary with gestational age at time of exposure(20, 27). In the following study we investigated gender differences in baseline lung function and response to prenatal steroid treatment in fetal sheep. All fetuses were delivered at 128-d gestational age. Exposure time and dose, delivery procedure, and postnatal management were stringently controlled, thus minimizing the effect of these variables on outcome measurements.

METHODS

Animal selection. Data reported in this study were compiled retrospectively from a series of earlier studies investigating the effects of prenatal hormone treatment on lung maturation in date bred sheep(20–24, 27). Protocols were approved by the Animal Ethics Committees at the Harbor-UCLA Medical Center and the Western Australian Department of Agriculture. We restricted our post hoc analysis of data to animals of one gestational age, as we have previously found that response to treatment varies with gestational age at the time of treatment(20, 27). The treated group was restricted to animals receiving a single fetal dose over a set exposure time, as response also varies with dose and exposure time(22, 23). A total of 115 fetal sheep delivered at 128 d of gestational age, which were randomly assigned to saline(control) or steroid (treatment) groups, have been included. Of the total, 83 animals received one or more injections of saline 3 wk to 24 h before delivery(Table 1). The remaining 32 animals received a single direct fetal injection of either 0.5 mg/kg betamethasone or 0.5 mg/kg betamethasone plus 15 μg/kg l-thyroxine, 48 h before delivery. We did not exclude those animals which received betamethasone plus l-thyroxine from our analysis, as they did not differ significantly from those animals treated with betamethasone alone. All hormone administration was by direct fetal injection. Saline was administered either directly to the fetus, via the maternal circulation, or via intraamniotic injection (Table 1).

Table 1 Fetal treatment protocols
Full size table

Fetal treatment. Fetuses were injected intramuscularly using a 9-cm 20-gauge needle. Those animals randomized to receive hormone treatment received 0.5 mg/kg betamethasone (Celestone Soluspan, Schering Pharmaceuticals) alone or in conjunction with 15 μg/kg l-thyroxine in 2 mL of sterile isotonic saline. Control animals received 2 mL of saline.

Animal preparation. A detailed account of animal preparation can be found in previous publications(20–24, 27), therefore only a brief description is presented below. At 128 d the ewes were sedated (intramuscular ketamine, 1 g) and received spinal anesthesia (2% lidocaine, 4 mL). The fetal head was exposed through midline abdominal and uterine incisions. A tracheotomy was performed while still in utero, and a 4.0-mm endotracheal tube was secured in place. Lung liquid was removed by suction through the endotracheal tube. Animals were delivered and the umbilical cord cut.

After delivery, lambs were placed on an infant ventilator (Bournes, BP200 pressure-control mode) set to deliver 100% oxygen at a rate of 40 breaths/min and PEEP 3 cm H2O. An I:E ratio of 1:1, giving an inspiratory time of 0.75 s was used to enable more efficient oxygenation of these immature, extremely surfactant-deficient animals. PIP was initially set at 35 cm H2O. Both tidal volume and Paco2 were monitored closely, and PIP was adjusted to maintain adequate ventilation with a maximum setting of 40 cm H2O to avoid pneumothorax. Where possible, Paco2 was maintained within the range of 45-50 mm Hg. Those animals that did not fall within this range at maximum peak pressure were permitted to become hypercarbic. An arterial catheter was advanced to the level of the descending aorta via the umbilical artery, and lambs were anesthetized by slow arterial infusion of pentobarbital sodium (15 mg/kg). Dextrose (5%) in water was infused at a rate of 12 mL/h. Blood gas samples were taken at regular intervals. Blood pressure and heart rate were recorded throughout the course of mechanical ventilation. Temperature was maintained at 39°C. Animals were ventilated for a total of 40 min.

Dynamic respiratory mechanics. A pressure transducer (model 8507C-2, Endevco, San Juan Capistrano, CA) and pneumotachograph (model 35-597 Hans Rudolph, Kansas City, MO) were placed between the tracheotomy tube and the ventilator to measure tracheal pressure (Ptr) and flow(V′), respectively. Volume (V) was obtained by integrating flow. Measurements of dynamic respiratory mechanics were made at 10-min intervals. At each time point a 20-s epoch of data was collected(sampling frequency 200 Hz) and a single mean value for dynamic resistance(RRS) and elastance (ERS) was calculated over the entire ventilatory cycle, using multiple linear regression analysis of pressure, flow, and volume (Equation 1).

where: Equation (2)

and PaEE is the end-expiratory alveolar pressure. A single compartment model which included a volume-dependent(E2V) elastance term was used to describe the data as we have previously found this term to be necessary to fit data from immature animals(23). Total compliance (C) and conductance (G) were derived from Equation 1, and are the inverse of elastance and resistance, respectively. Specific compliance, compliance per unit lung volume, was the ratio of total compliance to total lung volume, which was estimated from excised lung pressure volume curves.

The VEI, an index that integrates ventilation with respiratory support, was calculated according to the formula VEI = 3800/(P·f·Paco2) where 3800 is a carbon dioxide production constant (mm Hg kg-1 min-1), P is ventilatory pressure (PIP-PEEP), f is ventilatory frequency (constant at 40 breaths/min), and Paco2 is the arterial Pco2.

Excised lung pressure volume curves. Animals were killed with a lethal dose of pentobarbital sodium. A deflation volume-pressure curve was constructed by initially injecting the volume of air required to inflate the lung from a degassed condition to a pressure of 40 cm H2O and then systematically removing the necessary volume required to maintain pressures of 20, 10, 5, and 0 cm H2O. The volume at 40 cm H2O(V40) provided an estimate of total lung volume, as the pressure volume curve was approaching a plateau at this point.

Alveolar and lung tissue PC concentration. Alveolar and lung tissue PC concentrations were available for 73 control and 32 hormone-treated animals. Left lungs were lavaged with iced saline as previously described(19), and aliquots of lavaged fluid were frozen. Alveolar lavage surfactant pool size was estimated by purifying SPC(28) and quantifying phosphorus(29). Lung tissue SPC was similarly estimated on a sample of lavaged lung tissue.

Morphometry. Lung tissue samples were available for a total of 13 male and 11 female saline-treated animals. These animals were randomly chosen for morphometric examination, and lungs from other animals were not processed. They were representative of the group as a whole for all outcome variables examined (data not shown). Only four lung tissue samples fitting the inclusion criteria were available from treated animals, thus it was not possible to extend morphometric examination to include steroid-treated animals.

The right cranial lobe was fixed overnight via bronchial instillation of Karnovsky's fixative, at a constant distending pressure of 30 cm H2O. Remaining lobes were processed for biochemical measurements and were unavailable for morphometric examination. Each right upper lobe was cut into 5-mm transverse slices. Three slices per lobe were randomly chosen by the method of Cavalieri(30) and embedded in paraffin wax. One 5-μm section per block was stained with hematoxylin and eosin. Whole sections were photographed at a magnification of 16× and PF was calculated as the number of points falling on parenchymal tissue as a proportion of the total number of points(Pi/Pt) by superimposing a inear point counting grid containing 462 lines and 928 points.

A video camera (Sony 3CCD color) connected to a Leitz Dialux 20 microscope was used to capture gray scale images from the 5-μm sections at a magnification of 40× (final image magnification, 1800×). Images were visualized on a Macintosh Quadra 840AV computer using National Institutes of Health Image (version 1.59; National Institutes of Health, Bethesda, MD). Images from five nonoverlapping parenchymal regions (containing no airways or blood vessels) were captured and stored, giving a total of 15 fields per animal. Stored images were imported into the program Stereology Toolbox™(Morphometrix, version 1.1, Davis, CA). The number of points which fell on airspace and alveolar wall tissue and the number of air/tissue tissue/air intercepts were counted by superimposing a linear point-counting grid containing 21 lines and 42 points. Surface fraction (Sv) an index of alveolar surface area per unit tissue volume was calculated according to the formula Sv = 2I0/Lr whereI0 is the number of intercepts with the air tissue interface and Lr is the length of the test line. The MLI, an index of alveolar size, was calculated according to the formula MLI = 2Lr/I0. Mean TD was calculated according to the formula TD = Vv/Sv, where Vv is the volume fraction of alveolar wall tissue.

Statistical analysis. For all clinical, mechanical, biochemical, and growth indices, differences in control values between males and females were examined by one-way ANOVA tests. Where there was no difference in control values, the effect of hormone treatment was examined by linear modeling, constraining control male and female values to be equal. The inclusion of an interactive term in the model (treatment·gender) examined any differential effect of treatment on males and females. Where control values were different, the effect of treatment was examined separately for males and females by one-way ANOVA tests. Where borderline statistical significance (0.05 < p < 0.1) was found between males and females for both control and treated values of a given index, the overall gender effect was examined using the general linear modeling procedure in Minitab (version 8.2), which fits normal linear regression models. Differences in morphometric indices between control males and females were examined by one-way ANOVA tests.

Pressure volume data were modeled by assuming a linear relationship between lung volume and the natural logarithm of pressure. The data were well described by a log-linear relationship for all groups (r2 = 0.994-0.996). A constant value of 2.6 was added to pressure before taking the logarithm to avoid taking the logarithm of zero. This value was estimated using a nonlinear least squares procedure to indicate how far and in which direction the log-curve should move to best fit the data. A variance components analysis of covariance model was used to take into account the correlation between lung volume measurements made on the same animal. This model included a term to represent the differential effect of treatment on males and females. Statistical significance was accepted at p < 0.05 for all analyses.

RESULTS

Growth indices. Birth weight was approximately 10% lower in control females than in males (p < 0.0005,Table 2). There was no significant effect of hormone treatment on birth weight for either males or females. Wet lung weight was approximately 8% lower in control females than in males (p = 0.04,Table 2). Hormone treatment led to a 20% reduction in wet lung weight in females (p = 0.001). Although a slight reduction was seen in males, this was not statistically significant. The ration of wet lung weight to birth weight was similar in male and female controls and in hormone-treated males, but was approximately 10-15% lower in hormone-treated females (p = 0.001).

Table 2 Birth weight (BW) and wet lung weight (LW) data for control and hormone-treated preterm lambs
Full size table

Clinical outcomes. All reported clinical and mechanical data were recorded 40 min postdelivery. Control females were able to be ventilated at significantly lower peak pressures than males (p = 0.019,Fig. 1A), although achieving equivalent tidal volumes(Fig. 1B). Hormone treatment led to a significant reduction in ventilatory pressures (p < 0.0005 for both genders) and an increase in tidal volumes for both genders (p < 0.0005). The magnitude of these improvements did not differ between males and females.

Figure 1
figure 1

(A) Ventilatory pressure: (B) tidal volume. Group mean ± SEM values for males (open) and females (filled) are shown. *p < 0.05 vs baseline male;#p < 0.0005 vs appropriate baseline.

Full size image

Pao2 was approximately 50% higher in control females than in males(p = 0.004, Fig. 2A). Hormone treatment led to a 60% increase in females, which was statistically significant (p = 0.002) and an average 40% increase in males, which was of borderline significance only (p = 0.07). Pao2 in hormonetreated males was comparable to values in control females.

Figure 2
figure 2

(A) Arterial oxygen partial pressure:(B) arterial carbon dioxide partial pressure; (C) ventilatory efficiency index. Group mean ± SEM values for males (open) and females (filled). **p < 0.005 vs baseline male; ‡‡p < 0.005 vs baseline female;#p < 0.0005 vs appropriate baseline.

Full size image

Paco2 was on average 10 mm Hg lower in control females than in males; however, this was not significant (Fig. 2B). Both males and females exhibited a reduction of approximately 30% in response to hormone treatment (p < 0.0005), although the 10 mm Hg difference between the two groups remained. Average Paco2 was below 50 mm Hg for treated females but not for males. Although the difference between males and females was not significant for control or treated animals when examined separately, linear regression analysis of pooled data indicated a significant overall gender effect (p = 0.047).

The VEI was approximately 15% higher in control females than in males, although this difference was not statistically significant(Fig. 2C). After treatment, VEI was on average 30% higher in females than in males, but again this difference was not statistically significant. When data for control and treated animals were pooled and examined by multiple linear regression analysis, a significant overall gender effect was seen (p = 0.018). VEI increased 70% in males and 95% in females in response to hormone treatment (p < 0.0005).

Dynamic lung mechanics. As hormone treatment led to a reduction in lung weight of female fetuses, it was not considered appropriate to use this measurement to standardize lung mechanics variables for differences in lung size. Compliance, conductance, and excised lung volumes were therefore standardized for differences in birth weight, which was not altered by hormone treatment.

Neither compliance nor conductance were significantly different for control males and females (Fig. 3A). Hormone treatment led to a 65% increase in compliance for males and an 85% increase in females, both of which were significantly different from corresponding control values(p < 0.0005 for both genders). The magnitude of improvement in compliance was significantly greater in females than in males (p = 0.01 for interaction). Females also registered a significant increase in conductance with hormone treatment (p = 0.013), whereas males did not (Fig. 3B). These differences could not be accounted for by differences in birth weights, as a similar pattern of response was observed for absolute values of compliance and conductance (data not shown). There were no gender differences in either specific compliance or specific conductance for control animals. After hormone treatment, specific compliance decreased by 20% (p = 0.013) in males and by 30% in females(p = 0.008, Fig. 3C). Specific conductance was also reduced after hormone treatment, by 30% in males (p = 0.048) and by 50% in females (p = 0.015, Fig. 3D).

Figure 3
figure 3

(A) Weight-corrected compliance;(B) specific compliance; (C) weight-corrected conductance;(D) specific conductance. Group mean ± SEM values for males(open) and females (filled) are shown. ‡p < 0.05vs baseline female; ‡‡p < 0.0005vs baseline female; *p < 0.05 vs appropriate baseline; #p < 0.0005 vs appropriate baseline.

Full size image

Excised lung pressure volume curves. Pressure volume relationships were well described by log-linear relationships for both control and treated groups (Fig. 4, A and B). The pressure-volume relationships for control male and female animals were the same, with no statistically significant difference in the values of intercept(i.e. volume at pressure zero, V0) or gradient of the regression equations. After hormone treatment, V0 increased significantly for both males and females in comparison to control values (p < 0.0005 for both genders). The gradients of the pressure-volume relationships were also significantly increased in response to hormone treatment (p < 0.0005). There was a significant interaction between treatment and gender, the gradient of the pressure volume curve, and hence V40 (volume at pressure = 40 cm H2O), being greater in treated females than in treated males (p = 0.0019).

Figure 4
figure 4

(A) Deflation pressure volume curves;(B) lung volume vs natural logarithm of deflation pressure. Group mean ± SEM values for control males (open squares) and females (open circles) and treated males (filled squares) and females (filled circles). ¶p < 0.0005 vs baseline(V0); #p < 0.0005 vs appropriate baseline (gradient).

Full size image

Surfactant pool size. Control lung tissue SPC was approximately 50-60-fold greater than alveolar SPC. There were no gender differences in either alveolar lavage or lung tissue SPC concentrations for control animals. After hormone treatment, alveolar lavage SPC doubled in both males and females(Table 3, p < 0.0005). By comparison the percentage increase in lung tissue SPC was modest, and was significant only in females (Table 3, p = 0.02). Although lung tissue SPC increased in females and not in males, posttreatment values were not significantly different for males and females.

Table 3 Alveolar and lung tissue SPC concentrations for control and hormone treated preterm lambs
Full size table

Morphometry. Morphometric examination was performed on a subgroup of control animals, comprising 13 males and 11 females. PF, the proportion of lung volume occupied by parenchymal tissue, did not differ between males and females (Table 4). Control values of MLI, SV, and TD also did not differ between males and females (Table 4).

Table 4 Morphologic parameters from control (saline treated) preterm lambs
Full size table

DISCUSSION

The results of this study in fetal sheep support evidence from a number of clinical trials suggesting that there is a sexual dimorphism in both postnatal lung function and response to prenatal hormone treatment. Although the present study did not detect gender differences in mechanical properties in control animals, females were able to be ventilated at significantly lower peak pressures, exhibited significantly better oxygenation, and had a tendency toward decreased CO2 retention and increased ventilation efficiency. Both males and females exhibited significant physiologic improvements in response to hormone treatment; however, the magnitude of improvement was significantly greater in females than in males for all indices of mechanical behavior as well as for several clinical indices.

Interestingly, both absolute compliance and weight corrected compliance increased after hormone treatment, whereas specific compliance decreased. This observation was true for both males and females and is consistent with normal maturational changes. Between 128 and 135 d of gestation, both absolute and weight-corrected compliance increase by approximately 80%(20, 27). Specific compliance, however, decreased by almost 40%, suggesting that the rate of increase in lung volume exceeds the rate of increase in lung compliance. Conductance exhibits a similar pattern of change with gestational age(27). In the present study, only females exhibited increased conductance after hormone treatment. Thus the changes in lung mechanics in female sheep in response to hormone treatment closely resemble changes that occur with normal lung maturation, and suggest structural changes in both the airways and lung parenchyma. In male sheep the effect of hormone treatment appears to be primarily restricted to remodeling of the lung parenchyma.

Despite large improvements in all physiologic parameters after hormone treatment, these animals are still extremely immature relative to term animals. Hormone treatment increased compliance and total lung volume(V40) in these 128-d animals to approximately 50% of values reported in term animals, and increased VEI to one-third of term values(31). Although alveolar SPC doubled after hormone treatment, these animals remained extremely surfactant-deficient, with post-treatment concentrations reaching only 10% of reported term values(31). Lung mechanics values for hormone-treated females from the present study were comparable with values which we have previously reported for saline-treated lambs delivered at 135 d of gestational age(20, 27). Values for hormone-treated males from the present study were slightly lower than 135-d control values. Thus it would appear that hormone treatment advanced the level of lung maturity by approximately 7 d in females, whereas in males the maturational effect was somewhat less than 7 d.

The lack of difference in control compliance and conductance suggests that the lungs of males and females are structurally similar with respect to mechanical properties. However, the ability of females to absorb oxygen and release carbon dioxide more efficiently suggests that those structures that facilitate gas exchange may differ. More efficient gas exchange could result from reduced thickness of the blood gas barrier, a greater surface area for gas exchange, increased vascularization, or an increase in oxygen-carrying capacity in blood. Our results suggest that the female advantage in gas exchange is not due to an increase in surface area, as morphometric measurements revealed no difference in this parameter between males and females. There was also no detectable gender difference in mean alveolar wall thickness, which suggests that thickness of the blood gas barrier was not different for males and females. However, it should be noted that this measurement reflects the average thickness of the interalveolar septa, which includes both attenuated gas exchange regions and thickened cellular and interstitial regions, and therefore may not be sufficiently sensitive for the detection of small differences occurring in the regions of interest. As these lung samples were fixed by bronchial instillation, it was not possible to examine vascularization. However, it would be appropriate to examine this aspect of lung maturation in future studies. There are currently no available data in the literature relating to gender differences in vascularization of the lungs or oxygen uptake during fetal development.

We found no gender difference in either alveolar wash or lung tissue SPC pool size for control animals. Although hormone treatment led to a significant but modest increase in lung tissue surfactant concentration in females, posttreatment values for males and females were not different. Alveolar surfactant pool size was doubled in response to hormone treatment in both males and females. However, there was no posttreatment gender difference, suggesting that differences in lung mechanics after hormone treatment could not be attributed to differences in alveolar surfactant pool size. Furthermore, it is unlikely this increase in alveolar SPC concentration would be sufficient to account for the improvement in lung mechanics, as the dose of exogenous natural surfactant required to significantly improve lung mechanics in preterm lambs(32) and rabbits(33) is in the order of 25-30 μmol/kg.

The improvement in gas exchange and mechanical properties in response to hormone treatment in these animals most likely reflects restructuring of the lung parenchyma. We have recently examined morphometric changes in the lung parenchyma of fetal sheep at d 121, 128, and 135 of gestation in response to glucocorticoid treatment. Group sizes in this study were not sufficiently large to compare males and females. One of the primary structural changes observed after glucocorticoids was a reduction in alveolar wall thickness(34). We also found a reduction in wet lung weight. The reduction in alveolar wall thickness was of a greater magnitude at 121 than at 128 d (1.5 μm versus 0.6 μm) and coincided with a greater reduction in lung weight (27 versus 10 g). This reduction in lung weight may be partly attributable to thinning of alveolar septa, primarily due to loss of interstitium. In the present study, lung weight, both absolute and as a proportion of body weight, was significantly reduced by steroid treatment in females. This observation may be a reflection of marked thinning of alveolar septa. In contrast lung weights were not altered by steroid treatment in male fetuses. A difference in alveolar wall thickness between hormone-treated males and females may contribute to the difference in both oxygenation and mechanical properties of the lungs after hormone treatment. It is important that morphometric comparisons be made between hormonetreated males and females to answer this question.

Our findings are in agreement with a considerable body of literature that suggests a sexual dimorphism in prenatal lung development(35–42). This sexual dimorphism may be species-dependent, as lung development is more advanced in male rhesus and avian fetuses than in females(43, 44). What limited data are available in humans suggest that females mature slightly ahead of males. Two studies have shown that amniotic fluid SPC levels are higher in females than in males during late gestation, males lagging behind females by 1 to 2 wk(35, 36). The female advantage in lung development may be due to differential effects of estrogenic and androgenic hormones during late gestation. Several studies report an inhibitory effect of dehydrotestosterone(45–47) and a stimulatory effect of estradiol(48–52) on maturation of cultured fetal lung, although estradiol infusion in chronically catheterized fetal sheep failed to enhance lung maturation(53).

A number of studies on cultured fetal lung tissue have reported differences in responsiveness of male and female tissue to steroid exposure(39, 40); however, in vivo data in this area are scant. Kotas and Avery(54) performed a study in a small number of fetal rabbits and reported differences in lung stability and in airspace volume fraction after steroid exposure. Freese and Hallman(41) reported a smaller increase in lavage surfactant phospholipids from male rabbits in response to intrauterine betamethasone exposure than in females of the same gestational age. The question of whether or not a sexual dimorphism in response to steroid treatment exists clinically is presently unanswered. Furthermore it is unlikely to be resolved by further clinical trials, as the widespread use of postnatal surfactant replacement therapy would be expected to confound results.

The present study provides a comprehensive assessment of lung function in response to prenatal hormone treatment in 128-d gestational age fetal sheep under stringently controlled experimental conditions. It provides definitive evidence of gender differences in both postnatal (control) lung function and response to steroids. Clearly the conclusions drawn from the present study can relate only to this gestational age and cannot necessarily be extrapolated to earlier or later gestational ages. However, this developmental stage is of particular interest as it is roughly comparable to human fetuses of 28-30-wk gestation(55) who are at a high risk of developing RDS but have been shown to respond well to steroid intervention.