Asthma during pregnancy: mechanisms and treatment implications

Historical cohort studies

Most of the earlier studies of asthma and pregnancy were historical cohort studies, which have the advantage of providing a community-wide perspective on asthma 14, 16. While many identify an adverse impact of asthma on pregnancy outcomes, they are limited by little available information about asthma severity, disease progression and medication use during pregnancy 17. Studies based upon medical record review can obtain more detailed clinical information and avoid recall bias, but have several disadvantages, including the possibility that mild asthma may not have been documented 14, 16, 17, with such cases possibly included in the control population 29. This may underestimate any effect of asthma on pregnancy outcomes. Possible confounders, such as maternal smoking or socio-economic status, are not always present in administrative records 16 and coding or data entry errors are possible 14, 16, 34. Furthermore, since these studies rely on retrospective analysis of data, no possibility exists to study the mechanisms involved 35. Despite these drawbacks, the large number of subjects used in these studies give them more power to detect associations between maternal asthma and adverse pregnancy outcomes, which may then be followed up with smaller prospective studies.

The first large study of pregnant asthmatic females was published by Gordon et al. 9 in 1970. Patients with actively treated asthma were included in their analysis (n = 277) and 16 of these had severe asthma characterised by regular attacks during pregnancy. When corrected for ethnic background, there was no increase in the incidence of pre-term delivery or low birth weight in asthmatic mothers. However, there was a relatively large number of maternal (n = 5) or perinatal deaths (n = 16), which were more likely to occur in the severe asthmatics 9.

Bahna and Bjerkedal 10 used the Norwegian medical birth registry (1967–1968) to examine the pregnancies of 381 asthmatics and >112,000 controls who did not suffer from any diseases before or during pregnancy. Pregnancy complications, including hyperemesis gravidarum, haemorrhage and toxaemia, were twice as frequent in asthmatic patients, as were interventions during labour, induced labour and complicated labour. There was a higher rate of neonatal mortality, low birth weight, premature birth and hypoxia at birth in infants from asthmatic mothers. Although information about asthma treatment was not provided, this study was conducted prior to the availability of inhaled corticosteroids (ICS) and at a time when bronchodilators were the main therapy used for asthma 10.

Lao and Huengsburg 11 studied 87 asthmatic patients who delivered between 1984–1987 in Hong Kong. Many of these patients did not require medication for asthma and were considered to be in remission during the study period (n = 33). All other patients were treated with bronchodilators and some with oral or ICS. Mothers with asthma were significantly more likely to have a low birth weight baby, epidural analgesia or caesarean section, compared with the control group, matched for age and parity. When asthma treatment was considered, females who did not use any medication had a higher incidence of low birth weight, and those taking medication had a higher incidence of caesarean section 11.

The effects of asthma and asthma medication on pregnancy outcome were examined in a Californian perinatal database study in 1985–1990, comparing asthmatics (n = 81) to 130 controls selected from the reference population 12. Asthmatics were more at risk of caesarean section, pre-term labour or delivery and pre-term premature rupture of the membranes (PPROM). A significant increase in low birth weight was only observed in the oral-steroid dependent asthmatics (n = 50). However, this was influenced by an unusually high rate of pre-term delivery (54%) in this group. Patients who only used over the counter medications were excluded and thus, this study represented a group of more severe asthmatics 12. However, the effect of severe asthma could not be separated from that of oral steroid use.

A historical cohort of almost 25,000 pregnant females in Canada found a significant association between pregnancy-induced hypertension (PIH) and asthma, which was treated with ICS during pregnancy 13. Of the 1,435 females identified with a history of asthma, only 136 were considered to have asthma during pregnancy, as defined by requiring treatment. The association between asthma history and PIH was not significant after adjusting for confounders. Those females treated for asthma may have had more severe disease, but the role of disease severity and ICS treatment in leading to PIH could not be separated.

Two studies from New Jersey, based upon analysis of hospital records in 1989–1992, have examined neonatal 2 and maternal outcomes 14 in asthmatic females. Data from 2,289 asthmatic females were collected and compared to 9,156 control subjects. Maternal asthma was significantly associated with low birth weight, pre-term delivery, small for gestational age (SGA) neonates, congenital anomalies and prolonged infant hospital stay. After adjustment for potential confounders, including age and education, asthmatic mothers also had an increased risk of pre-term labour, placenta previa, caesarean section, prolonged hospital stay and hypertensive disorders of pregnancy, including pre-eclampsia 14.

Demissie et al. 2 found an increase in transient tachypnoea of the newborn in infants of asthmatic mothers after accounting for confounding risk factors, such as caesarean delivery and premature birth. This association was stronger for male infants than female infants, possibly because male sex is a known risk factor for this condition 36, due to differences in foetal lung maturation between the sexes 37. Schatz et al. 38 earlier described an increased risk of transient tachypnoea, but not respiratory distress syndrome, in a prospective cohort study of 294 asthmatic females compared with 294 controls, but did not find a link with either asthma severity or medication use. The association is of interest, since transient tachypnoea of the newborn is related to a higher prevalence of asthma and atopic symptoms at 5 yrs of age 39. It is possible that there are links between the in utero environment in asthmatic pregnancies and the risk of developing childhood asthma in the offspring, independent of genetic factors. This is demonstrated by the fact that the development of atopy in children is more closely associated with maternal asthma or immunoglobulin (Ig)E levels, rather than paternal asthma or IgE 40–42. Therefore, as well as the immediate implications of poor pregnancy outcomes on the offspring of asthmatic mothers, there may be long-term implications for these children. Appropriately grown neonates of asthmatic mothers had increased nucleated red blood cell counts, haematocrit, leukocytes and lymphocytes within the first day of life, compared with those of nonasthmatic mothers 43. Studies investigating the long-term effects of asthmatic pregnancies on offspring have found an increase in the prevalence of left-handedness 44, wheezing at 15 months of age 38 and childhood respiratory diseases in general 45, while others found no long-term developmental effects 46.

A Canadian historical cohort study of 817 asthmatic females and 13,709 nonasthmatic females in 1991–1993 examined medical records to assess medication use during pregnancy 1. Over 45% of asthmatic females did not use any medication to treat asthma, while 37.1% used β₂-agonists. The use of ICS was unclear, with 17% using “steroids”. Regardless of medication use, asthmatic females were found to be at increased risk of antepartum or post-partum haemorrhage, possibly due to alterations in platelet function in asthmatics 47, 48. In addition, neonates from asthmatic mothers who used steroids were found to be at increased risk of hyperbilirubinemia 1. In a previous prospective cohort study, no alteration in neonatal risk for hyperbilirubinemia in mild or severe asthmatics was found 21.

Kallen et al. 15 examined the effect of asthma on pregnancy outcomes using the medical birth registry and the hospital discharge register (1984–1995) in Sweden. Asthmatic females had an increased risk of pre-term delivery, low birth weight, or prolonged pregnancy (>41 weeks gestation) 15. This study used two approaches to identify females with asthma and would have identified those with very mild asthma (asthma recorded by midwives) as well as those with severe asthma requiring hospitalisation (asthma documented in hospital discharge registers). Despite this, the study did not separate subjects based on disease severity, which may have been a confounder.

Using Canadian administrative data (1989–1996), maternal asthma was associated with all adverse outcomes examined by Wen et al. 16; namely, pre-term labour, PIH, pre-eclampsia, antepartum haemorrhage, membrane disorders (including PPROM), post-partum haemorrhage and caesarean delivery. Many of these associations were stronger in teenage mothers than adult mothers 16, possibly due to the increased risk of complications in adolescent pregnancies.

A study of 2,193 asthmatic and 8,772 nonasthmatic singleton pregnancies in Quebec was carried out in 1991–1996 17. After adjusting for confounders, such as maternal age and pre-existing diabetes or hypertension, maternal asthma was associated with an increased incidence of pre-term labour and delivery, small and large for gestational age neonates, PIH, chorioamnionitis, abruptio placentae and caesarean delivery. Interestingly, this group analysed data separately based upon foetal sex and found that the risks of pre-term birth and pre-eclampsia were higher in asthmatic females pregnant with a female foetus, compared to those pregnant with a male foetus. The prevalence of asthma in this population was very low and similar among those pregnant with a male or female foetus 17. The cause of the increased risk to the female foetus was not examined in this study, but may be related to foetal sex-specific effects on maternal asthma 30, 49.

A Danish population-based study collected data from a birth registry and prescriptions database (1991–1996) to study the use of asthma medications by pregnant females and relate this to perinatal outcome 18. Less than 2% of all Danish females were prescribed asthma medications during pregnancy. Asthma treatment was defined as one of five levels: 1) inhaled β₂-agonist; 2) inhaled steroid; 3) systemic β₂-agonist; 4) systemic steroid; and 5) theophylline. Data was analysed based on whether asthmatic females were prescribed treatment of a higher or lower level than prior to pregnancy. The results showed that asthmatic females who decreased their medication level during pregnancy (78 of 342 asthmatics) had babies with reduced birth weight and length, with a lower mean gestational age compared with nonasthmatic females, or asthmatic females who increased their medication level during pregnancy. This was particularly evident among 22 females who decreased their medication use from inhaled steroid to inhaled β₂-agonist. This study was limited by a lack of information regarding compliance and the reasons for alterations in asthma management 18. These may have included a clinical improvement in lung function and asthma symptoms, or may have been due to a reduction in prescribing as a result of fears of drug use during pregnancy. Furthermore, prescription rates may not necessarily correlate with actual medication used during the pregnancy.

A recent historical cohort study from Sweden found that budesonide use in pregnancy does not affect gestational age, birth weight, birth length or the rate of still-births or multiple births 19. This data came from 2,968 females who used inhaled budesonide during pregnancy and was compared to 7,719 females who used asthma medications other than steroids, and a control population of >293,000 females. The authors acknowledged that a comparison of asthmatic mothers with similar severity would be of benefit, since inadequate asthma control may be a confounder. Another Swedish medical birth registry study previously reported no increase in congenital malformations in females who used budesonide during early pregnancy 50.

Case-control studies

There are only two case-control studies that examine the contribution of maternal asthma to pre-term labour or delivery. In a study of females participating in a pre-natal screening programme in 1994–1995 in the USA, the prevalence of asthma, defined as a lifetime history of asthma diagnosis, was compared in 312 females who delivered pre-term and 424 control subjects who delivered at term 51. Significantly more pre-term cases had a positive asthma history (6.4%) compared with control cases (3.3%), after adjustment for multiple confounders 51. A similar case-control study was previously reported by Kramer et al., 52 with histories of asthma diagnosis or symptoms more commonly associated with idiopathic pre-term labour. However, the risk of pre-term labour was not associated with increased serum IgE or altered response to inhaled methacholine challenge in the mother 52.

Due to their retrospective nature, the main disadvantage of these studies is a lack of information regarding the clinical course of asthma and its treatment. In particular, asthma is often defined as a history of asthma diagnosis, rather than current asthma during the pregnancy.

Cross-sectional surveys and case series

There are numerous cross-sectional surveys and case series examining the relationship between maternal asthma during pregnancy and perinatal outcomes. In many of these studies the lack of a control group is a disadvantage, with most comparing outcomes among females with asthma to a population estimate or institutional rate. In addition, a survey of exposure and outcome at one point in time may be influenced by recall bias and does not address the temporal association between the exposure and outcome that could be better investigated using a longitudinal study design.

A self-report questionnaire administered in Italy in 1987 found that maternal asthma was a risk factor for low birth weight 53. When other variables were considered, this relationship was only found to hold in male infants, mothers who smoked and those who lived in an industrial town, suggesting that other risk factors contribute to the effect of asthma on pregnancy outcome. However, there were only four male infants of low birth weight and three female infants of low birth weight among the 55 mothers with asthma 53. Therefore, these results should be interpreted with caution, as the sample sizes of the subanalyses were very small.

In 1993, >1,000 parents of children aged 5–11 yrs were surveyed in the UK, with regard to the child's history of respiratory disease and pregnancy-related factors, including birth weight and pre-term delivery 45. Pre-term birth was significantly more likely to be reported when the mother was asthmatic, but not when the father was asthmatic, compared with children of nonasthmatic parents. There were several disadvantages to this study, including the self-report of maternal asthma and the potential for recall bias on the part of the parent, given the amount of time that had passed since the pregnancy. In addition, pre-term birth was assessed by asking the question “Was your baby born prematurely?”, which could lead to errors as a result of the participant's misunderstanding of the clinical definition of pre-term birth. Maternal smoking was an additional risk factor for pre-term delivery and no association was found between parental asthma and low birth weight 45.

In a cross-sectional survey of recently post-partum females in Western Australia, there was no significant relationship between doctor-diagnosed asthma, asthma ever or current asthma (defined by attacks of wheezing in the past 12 months) and PIH, low birth weight, pre-term delivery, PROM and threatened abortion 6.

Apter et al. 54 examined a case series of 28 adolescents with severe asthma and found a high rate of exacerbations, hospitalisations and emergency room visits in these patients, which were associated with respiratory tract infections and lack of medication compliance. However, they found no evidence of an increased rate of PIH, pre-term delivery or intrauterine growth restriction (IUGR) in asthmatic adolescents compared with general estimates for adolescent pregnancies 54.

Mabie et al. 55 examined asthma in 200 pregnancies in Tennessee in 1986–1989 by medical record review. There was no increased rate of pre-term delivery or low birth weight among asthmatic females compared with the general population rates, which were very high (17.7 and 6.3%, respectively). However, IUGR was significantly more likely in females with moderate or severe asthma, who required hospitalisation during pregnancy, compared with subjects with mild asthma, who were not hospitalised for asthma during pregnancy. The caesarean section rate and incidence of post-partum exacerbations were also significantly increased in moderate and severe asthmatics compared with mild asthmatics. Asthmatic subjects who had a caesarean delivery were 18 times more likely to have exacerbations of asthma post-partum compared with asthmatics who had a vaginal delivery. The mechanism for this effect is unknown 55. None of the subjects in this study used ICS and additional medical problems, such as hypertension, diabetes and obesity were present in 21% of patients.

A retrospective analysis of medical records by Beckmann 35 for 1992–1997 in the USA assessed outcomes in 782 asthmatic females. Over 90% of these subjects were mild asthmatics (according to hospital records) and almost half did not use any asthma medication during pregnancy. Only 6% of the asthmatics used a β₂-agonist and inhaled steroid for treatment. There was an increased incidence of meconium staining, pre-term labour and oligohydramnios among asthmatic females compared with the general population. This study lacked power to demonstrate a relationship between steroid use and outcomes associated with altered placental function, such as IUGR, PIH and oligohydramnios 35.

Greenberger and Patterson 56 studied 80 pregnancies in females with severe asthma and found that those who had been hospitalised with status asthmaticus delivered neonates of reduced birth weight compared with those who were not hospitalised for asthma, suggesting that an acute attack of asthma may put the foetus at additional risk, particularly of IUGR.

Prospective cohort studies

Recently, a number of large prospective cohort studies of asthma in pregnancy have been published. Studies that prospectively examine pregnant females with asthma alongside a control group of females without asthma have the advantage of being able to assess lung function, treatment or asthma symptoms during pregnancy in relation to pregnancy outcomes, while close follow-up ensures that asthma is well-characterised and effectively managed throughout pregnancy 25. However, significant associations between maternal asthma and adverse outcomes are frequently not observed. This may be due to small sample sizes 30, a high prevalence of ICS use among participants 30, active management of asthmatic subjects 25, 30, 33 or a bias towards mild asthma in some studies 33. While participation in a closely monitored study itself may reduce the risk of an adverse outcome, this remains the most ethical approach to asthma management. One problem when comparing prospective studies is that each population of asthmatics examined varies with regard to steroid use, general treatment and asthma severity, with some studies focussed on mild asthmatics and others on females with severe asthma, making comparison between studies difficult. In addition, standard classification systems are sometimes not employed and the criteria used to assess disease severity differ between studies.

Dombrowski et al. 20 prospectively followed 153 pregnant females with asthma and 116 healthy control females and found a reduced incidence of pre-eclampsia among theophylline users, but only compared with females with asthma who did not use theophylline. They suggested that the ability of theophylline to reduce vascular reactivity and platelet aggregation via increasing cAMP may be responsible for this trend 20. However, another study found that pregnant females using theophylline were more likely to have an asthma exacerbation or develop pre-eclampsia than patients not using theophylline 57. Although asthma severity was not specifically described in these subjects, the authors explain these findings as being possibly due to the higher prevalence of severe asthmatics among the theophylline users and, therefore, the effect on pre-eclampsia may have been independent of theophylline use 57. A recent randomised controlled trial comparing theophylline use with inhaled beclomethasone found no differences between the two medications in maternal or perinatal outcomes, including pre-eclampsia 58.

A study of asthmatic mothers was conducted in Finland in 1978–1982 21. The study prospectively followed 181 asthmatic females during pregnancy, with 17 having two pregnancies in the study. Data on the control population was obtained retrospectively from labour records of subjects matched for age, parity and delivery date. A disadvantage was that only 20% of study subjects were recruited during the first trimester, with 26% of subjects recruited in the third trimester, making it difficult to follow changes in asthma during pregnancy. However, this study did classify females based on asthma severity as very mild, mild, moderately severe or severe. Skin-prick tests and serum IgE were used to assess atopy in these subjects, and although 62% were classified as atopic, this was unrelated to poor pregnancy outcome. There was a significantly higher incidence of pre-eclampsia in asthmatics (15%) compared with control subjects (5%). Mild pre-eclampsia occurred more often in females with severe asthma (29%) compared with females with very mild asthma (9%). The use of systemic steroids may also have contributed to the high frequency of pre-eclampsia, which was 25% compared with 10% in asthmatic females who did not use systemic steroids. Asthmatic subjects had a higher rate of caesarean section, but no differences in perinatal outcome, including birth weight, were found 21.

The same group performed another study of 504 pregnant females with asthma and found that females who had an acute attack during the pregnancy were less likely to have been using ICS prior to the attack 23. However, there was no difference in pregnancy outcomes, including length of gestation and birth weight in females who had an acute attack, compared with both the control group and those who did not have an acute attack during pregnancy. The authors suggest that prompt treatment of these patients with ICS contributed to the positive outcome 23. A disadvantage of these studies was their “mixed cohort” nature, with the asthmatic group being prospectively recruited and followed, and the control group data being retrospectively collected from medical records. This study design may have contributed to an over-estimate of poor outcomes with maternal asthma due to closer monitoring of the asthmatic females compared with control subjects who were not prospectively studied.

Doucette and Bracken 24 performed a prospective cohort study of 32 females with asthma and 3,859 controls. The history of asthma was obtained by self-report during an interview or documentation in medical records and ICS use was not described. There was a two-fold increased risk for pre-term labour and delivery in association with maternal respiratory “problems” during pregnancy, while no effect of asthma on low birth weight was observed 24.

Over many years, Schatz and colleagues 25, 59 have performed the most comprehensive prospective cohort studies of the effects of asthma on pregnancy outcome, as well as the effects of pregnancy on asthma progression. This group actively managed asthmatic females during their pregnancies, measured lung function by spirometry at several time points and related these measurements to pregnancy outcome. In initial studies where 352 asthmatic females had at least three lung function measurements during pregnancy, there was a correlation between mean per cent predicted forced expiratory volume at one second (FEV₁) and birth weight 59. Subjects with an FEV₁ in the lowest quartile (<83% predicted) were significantly more likely to have an infant with a birth weight in the lowest quartile (<3150 g) or a ponderal index <2.2, indicative of asymmetric IUGR 60, 61. There was no relationship between low FEV₁ and pre-term delivery, PIH or pre-eclampsia. A later study by this group on 486 females with actively managed asthma and 486 controls, found no significant differences in the incidences of pre-eclampsia, perinatal mortality, low birth weight, IUGR, pre-term delivery or congenital malformations 25. This study was conducted over a period of 12 yrs and asthmatic subjects were well-characterised and actively managed. Although only 8% of subjects used ICS, the close management of asthmatic females may have contributed to the negative findings. Control subjects were also well-characterised as they also underwent pulmonary function testing, and were matched for maternal age, parity, smoking and delivery date 25.

Jana et al. 26 examined 182 asthmatic pregnancies in India in 1983–1992 and compared outcomes to 364 nonasthmatic pregnancies. Most females had well controlled asthma (91%) and were using medications including oral or inhaled β₂-agonists, theophylline, oral steroids or inhaled beclomethasone. In addition, there was close cooperation between the obstetrician and chest physician in the patient's management. No significant increase in the rate of pre-term labour, low birth weight, caesarean section, perinatal mortality, haemorrhage or foetal distress was found in the asthmatic group compared with the control group. However, 15 of the asthmatics had a severe asthma attack during pregnancy, which required hospitalisation, and in these females there was a significant reduction in birth weight 26, suggesting that poorly controlled asthma may contribute to reduced foetal growth.

A prospective study in Israel comparing asthmatic mothers (n = 101) and control mothers (n = 77), matched for age and ethnicity, collected data by interview at 1 day post-partum and from medical records 27. Altogether, 23% of asthmatics used ICS and asthma was defined as having a history of recurrent episodes of wheeze, chest tightness, shortness of breath and cough. Asthmatic females were classified as mild (no inhaled steroid use), moderate (inhaled steroid use, no hospitalisations for asthma) or severe (inhaled and oral steroid use and possibly hospitalisations for asthma). Significantly more asthmatic females suffered from urinary tract or upper respiratory tract infections (31% of females with mild or moderate asthma and 69% of females with severe asthma) compared with nonasthmatic females (5%). The marked effect of severity on infections may be related to suppression of the immune system following prolonged corticosteroid use. There was no significant effect of asthma on pre-term delivery, gestational age, birth weight and PIH 27.

Sobande et al. 28 studied the pregnancy outcomes of asthmatic patients residing at high altitude in Saudi Arabia, hypothesising that the low oxygen environment may further contribute to pregnancy complications through a worsening of asthma. They studied 88 asthmatic females and 106 nonasthmatic females in 1997–2000. Asthmatic patients were managed by a medical specialist and treated with β₂-agonist alone (n = 57), in combination with oral theophylline (n = 20) or with oral prednisolone (n = 11). Asthmatic pregnancies were more likely to be complicated by pre-eclampsia, congenital malformations, low Apgar score or perinatal mortality and mean birth and placental weights were significantly reduced in asthmatics compared with nonasthmatics. Gestational age at delivery was not different between the groups. It is possible that the hypoxic environment contributed to an amplification of poor outcomes in these asthmatic subjects. However, females were selected for the study because they had visited the emergency room with asthma while they were pregnant, and the outcomes may simply have been observed due to the severity of their asthma. No comparison with a similar group of asthmatic females at low altitude was made 28 and thus the effect of high altitude residence on pregnancy outcomes with maternal asthma could not be properly examined.

As part of the childhood asthma prevention study in Sydney, Australia, pregnant females with physician-diagnosed asthma, and nonasthmatic pregnant females whose partners or other children had asthma, were prospectively studied 29. Because females were recruited at 36 weeks, there was no evaluation of the effect of asthma on pre-term delivery. However, recruitment this late in pregnancy may also have led to an underestimation of the effect of asthma on other outcomes, such as low birth weight and pre-eclampsia, since subjects with these outcomes may also be more likely to deliver early. Of 340 asthmatic females, 31% did not use any medication for asthma during pregnancy, while 35% of females used short acting β₂-agonists alone and 31% used ICS. This study was complicated by the fact that several (21 of 271) nonasthmatics were using β₂-agonists for wheezing during pregnancy, despite no previous doctor diagnosis of asthma. These females may have had mild asthma, but were not assessed during the study or excluded from analysis. Hypertension was significantly increased in the asthmatic group compared with the nonasthmatic group. There was no significant effect of asthma on other outcomes including pre-eclampsia, gestational diabetes, induced labour, caesarean delivery or any neonatal outcomes, including birth weight 29.

Since 1998, the current authors' group in Newcastle, Australia, has prospectively followed asthmatic and nonasthmatic subjects throughout their pregnancies, conducting a detailed examination of the relationships between mother, placenta and foetus 30, 49, 62, 63. Asthmatic females were classified based on both severity and inhaled steroid intake independently, according to the Australian asthma management guidelines 64, which are comparable to the guidelines of the National Heart Lung and Blood Institute 65. Females were assigned an asthma severity rating of mild, moderate or severe according to symptoms, asthma history and other features, including FEV₁ and peak expiratory flow (PEF). Females were assigned to the most severe category that applied for any one of these criteria. These classifications are similar to those used in the multi-centre studies reported recently 33, 66. In addition, females were classified based on their inhaled steroid intake (budesonide, beclomethasone dipropionate or fluticasone propionate), calculated for each trimester and expressed as the mean daily dose of beclomethasone dipropionate (BDP) or equivalent, where 1 µg BDP was considered equivalent to 1 µg budesonide or 0.5 µg fluticasone propionate 67. Classifications were no glucocorticoid (no ICS during pregnancy), low (<400 µg daily), moderate (400–1500 µg daily) or high (>1500 µg ICS daily). Oral steroid medication was used periodically by a small number of subjects. All asthmatic females used the short acting β₂-agonist, salbutamol, for symptom relief when required. Birth weight was examined in asthmatic females (n = 138) and compared with the nonasthmatic control group (n = 44). The current authors demonstrated that growth of the female foetus was significantly reduced when asthmatic females did not use any ICS for treatment 30. This occurred regardless of asthma severity, with most of these females having mild asthma, which was not considered severe enough by their physician to warrant the use of ICS. The use of ICS by asthmatic mothers was associated with female birth weights which were comparable to the nonasthmatic control group, while male birth weights of all asthmatic females were unaffected by asthma or its treatment. The results suggested that a mild inflammatory disease can have significant effects on foetal growth. The current authors propose that this occurs primarily through alterations in placental function 30.

Two large multicentre prospective cohort studies have recently been conducted in different parts of the USA, examining the effects of maternal asthma on pre-term delivery, pre-eclampsia and low birth weight. Bracken et al. 31 studied females in Massachusetts and Connecticut, recruited through obstetric practices and hospital clinics, while Dombrowski et al. 33 conducted a study in 16 university hospital centres across the USA.

The study of Bracken et al. 31 included 832 asthmatic females and 1,266 nonasthmatic controls. Asthma was defined as a lifetime history of doctor-diagnosed asthma, and symptoms and medication use during pregnancy were recorded and each rated as intermittent, mild persistent, moderate persistent or severe persistent, according to the 2002 Global Initiative for Asthma (GINA) guidelines 68. Asthmatic females using theophylline or oral steroids had an increased risk of pre-term delivery. There was no relationship between symptom scores and pre-term delivery risk. However, there was an increased risk of IUGR in females with daily asthma symptoms but no association with treatment. This relationship was strongest in females who had not been diagnosed as asthmatic by a doctor, but who were experiencing symptoms and was not significant when this subgroup was removed from the analysis 31. Assessment by a physician would have been useful to investigate potential cases of asthma in these 449 females 31, 69. Although the assumption was made that most of these females had undiagnosed asthma and would benefit from closer monitoring during pregnancy 69, it seems surprising that one-third of females in a control population of relatively good socio-economic status were under-diagnosed asthmatics. It is possible that the symptoms reported by many of these females (cough or wheeze or chest tightness at least once during pregnancy) may have been general symptoms of dyspnoea due to the pregnancy itself, which are experienced by up to 75% of pregnant females without asthma 70. However, it was alarming that for nearly 100 females, symptoms were mild persistent and for a smaller group of ∼20 females, symptoms occurred daily.

As part of the same prospective cohort study, Triche et al. 32 found that females classified by GINA guidelines as having moderate-to-severe asthma symptoms during pregnancy were at increased risk of pre-eclampsia, suggesting that active asthma symptoms may affect maternal physiology, possibly as a result of increased inflammation.

Dombrowski et al. 33 tested the hypothesis that there would be an increased incidence of pre-term delivery among females with moderate or severe asthma. The classification of mild, moderate and severe asthma was modified from the National Asthma Education Program to include medication use 66. Pregnancy outcomes were examined in 881 nonasthmatic controls, 873 females with mild asthma (FEV₁ ≥80% predicted, not taking daily asthma medications, similar to intermittent asthma by GINA guidelines), 814 females with moderate asthma (FEV₁ 60–79%, using one or more daily asthma medications, similar to mild and moderate persistent) and 52 females with severe asthma (FEV₁<60%, may be using oral corticosteroids, more severe than severe persistent by GINA guidelines as all females used oral steroids). There was an overall increase in the caesarean delivery rate in the moderate/severe group, and neonatal sepsis in the mild asthmatic group, compared with the nonasthmatic control group, but no significant difference in the rates of pre-term delivery (either <32 or <37 weeks gestation) among all asthmatics compared with the control group 33. However, pre-term delivery (<37 weeks gestation) was associated with severe asthma, which may have been due to oral steroid use. Bracken et al. 31 also observed an increased risk for pre-term birth of similar magnitude, in females using oral steroids. These studies from Bracken et al. 31 and Dombrowksi et al. 33 were comparable in terms of gestational age at recruitment (<24 and <26 weeks, respectively), criteria for asthma (doctor diagnosis), the schedule of study visits or telephone interviews, and the data collected. In addition to collecting information about symptoms, medication use and exacerbation history, most patients in the Dombrowski et al. 33 study were actively managed and had spirometry performed as part of their study visits. By contrast, Bracken et al. 31 did not actively manage asthma in their cohort and data was collected by home and telephone interviews. Bracken et al. 31 used a classification system that examined symptoms and treatment as separate characteristics, and performed analyses on mixed groups of females with and without doctor-diagnosed asthma, subdivided based on the reporting of symptoms associated with asthma. Dombrowski et al. 33 gave an overall severity rating that considered symptoms, lung function and medication use. Although the study numbers were larger in the study by Dombrowski et al. 33, their results related to pre-term delivery were similar to those found by Bracken et al. 31.

Summary: effect of asthma on pregnancy

The results of prospective cohort studies into the effects of asthma on pregnancy outcome have not always supported the findings from previous historical cohort studies and case-control studies. However, the prospective cohort study design is superior for examining temporal relationships between maternal asthma during pregnancy and subsequent perinatal outcomes. Associations between asthma and pre-eclampsia, and asthma and low birth weight have most commonly been demonstrated by both historical or prospective cohort and case-control studies. Foetal sex may be a confounder, and studies reporting no adverse perinatal outcomes may have done so due to a lack of separate data analysis for females pregnant with male and female foetuses. In addition, maternal atopy has rarely been examined as a possible risk factor. There are deficiencies in classification approaches in some prospective studies. Classification at enrolment 33, rather than constant monitoring of females 31, and classification that considers asthma throughout pregnancy is one difference between studies. In some studies, control patients were not assessed for the absence of asthma 29, 31, while in other studies, spirometry and close assessment of control subjects confirmed that they did not have asthma 30, 33. The use of standard classification systems and close monitoring of all patients has improved the quality of data obtained from prospective cohort studies of asthma and pregnancy.

A meta-analysis of the association between asthma in pregnancy and low birth weight (<2,500 g) was conducted, with studies grouped by ICS use. In four studies involving 1,453 females with asthma and >156,000 nonasthmatic females, where ICS were not used during pregnancy, there was a significantly increased risk of low birth weight in the asthmatic pregnancies (fig. 1⇓, relative risk (RR) 1.55; 95% confidence interval (CI) 1.28–1.87; p<0.00001). In contrast, there was no significant increase in the risk of low birth weight with asthma when ICS were used by some (range 8–>30%) of the females with asthma (fig. 1⇓, RR 1.19, 95% CI 0.97–1.45, p = 0.10). Females with asthma not using ICS were at significantly increased risk of delivering a low birth weight infant, whereas the use of inhaled corticosteroid medication seemed to protect against this effect, although the difference between the RRs did not reach statistical significance (ratio of RRs 0.63, 95% CI 0.46–0.86 71).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1—

Meta-analysis examining the risk of low birth weight infants from asthmatic and nonasthmatic pregnancies. Studies are grouped based on whether some or all subjects used inhaled corticosteroids (ICS use) or whether none of the subjects used inhaled corticosteroids (no ICS use). RR: relative risk.

Asthma requiring hospitalisation 26 and lack of ICS use in females with mild asthma 30 emerge as risks for low birth weight in prospective cohort studies (table 2⇑). Historical cohort studies (table 1⇑) conducted prior to ICS use 10, or where subjects were either not on any preventer treatment 11 or reduced their use of ICS during pregnancy 18, also support the conclusion that ICS medication may protect against low birth weight in pregnant females with asthma. Pre-term labour or delivery is associated with the use of oral steroids 31, 33 and theophylline 31, or having a respiratory problem during the pregnancy 24, suggesting that this may be an effect of more severe asthma or asthma exacerbations. Data from case-control studies have also demonstrated that an asthma attack in the previous 12 months is a risk factor for pre-term delivery 52. The risk for pre-eclampsia may increase in females with moderate-to-severe persistent asthma symptoms during pregnancy 32, whereas the use of theophylline reduces the risk of pre-eclampsia 20. The effect of acute asthma attacks during pregnancy on pre-eclampsia is unclear, as one study found an effect 28, while another prospective cohort study did not 23. The use of ICS by females in prospective or historical cohort studies did not appear to be related to the development of pre-eclampsia.

Mechanisms for the effect of maternal asthma on pregnancy outcomes

Despite the lack of studies directly addressing possible mechanisms, several authors have proposed that maternal hypoxia, inflammation, medication, smoking [14-–17, 25, 51] and altered placental function 30, 62, 63 may contribute to poor pregnancy outcomes in females with asthma (fig. 2⇓).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2—

The effects of maternal asthma on pregnancy outcome.

Hypoxia may contribute to low birth weight, pre-eclampsia, congenital malformations, spontaneous abortions and placenta previa in asthmatic females 28. Reduced partial pressure of oxygen (PO₂) is a feature of acute severe asthma or status asthmaticus 72–74 and a small decrease in maternal PO₂ can have serious effects on the foetus 75, since the slope of the foetal oxygen dissociation curve is steep in the 50% oxygen saturation range 76. Administration of oxygen to mothers in labour resulted in increased umbilical cord O₂ values at delivery, suggesting that there is a relationship between maternal and foetal oxygen 76. However, maternal hypoxia during asthmatic pregnancies has never been directly investigated in relation to foetal outcome and while it is unlikely to explain the finding of reduced female foetal growth in females with mild asthma 49, it may contribute to reduced growth in subjects who have been hospitalised with an exacerbation of asthma during pregnancy 26, 56.

Various aspects of placental physiology may affect foetal growth in pregnancies complicated by asthma, including placental blood flow, which is important for supply of nutrients to the foetus, and enzyme activity of 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2), which protects the foetus from excess maternal glucocorticoids. Using a perfusion method, placental vascular responses to both dilator and constrictor agonists were significantly reduced in placentae collected from females with moderate and severe asthma, but unaffected in placentae from mild asthmatics compared with nonasthmatic controls 62. These results indicate that altered placental blood flow may contribute to reduced foetal growth in moderate and severe asthmatics, by reducing the supply of nutrients to the foetus. In the placenta of asthmatic females who did not use ICS and were pregnant with a female foetus, a significant reduction in 11β-HSD2 enzyme activity was also observed 63, which was related to the decreased growth observed in female foetuses of mothers with asthma. In addition, there was a significant reduction in cord blood concentrations of oestriol in females, indicating suppression of adrenal function as a result of excess maternal cortisol reaching the foetus 30.

The release of bioactive mediators, such as inflammatory products, from the mother, could also be involved in these mechanisms. Poor pregnancy outcomes, including low birth weight and pre-term delivery, are also features of other inflammatory diseases, including rheumatoid arthritis 77–79, malaria 80–82, systemic lupus erythematosus 83, inflammatory bowel disease 84–86 and periodontal disease 87, 88. Moreover, elevated maternal serum levels or placental gene expression of inflammatory cytokines has also been associated with IUGR 81, 89–92. Bowden et al. 78 found that females with active inflammatory arthritis during pregnancy had smaller babies at birth compared with healthy control subjects or subjects whose disease was in remission. These data indicate that active inflammation during pregnancy may contribute to low birth weight. Similarly Jana et al. 26 found reduced birth weight only when subjects had experienced an acute episode of asthma during pregnancy and Greenberger and Patterson 56 found that females with severe asthma who had experienced at least one episode of status asthmaticus had babies of reduced birth weight compared with those who did not.

In the study by Murphy et al. 30, maternal asthma severity, inflammation, lung function and treatment with ICS were examined in relation to foetal growth. It was proposed that maternal inflammation may be related to reduced female birth weight, since the use of ICS was protective 30. Cytokine expression in the placenta was also examined and an increased ratio of T-helper cell (Th)2:Th1 cytokines in placentae from asthmatic females who did not use ICS and were pregnant with a female foetus was found. Inflammatory pathways in the placenta may be altered as a result of decreased 11β-HSD2 activity and the associated increase in local cortisol concentrations 49. In addition, maternal white blood cell counts were examined and a significant increase in the number of circulating monocytes with advancing gestation was found only in asthmatic females who did not use ICS and were pregnant with a female foetus 30. Changes in circulating inflammatory cells may contribute to altered maternal asthma during pregnancy, which possibly leads to changes in placental function and ultimately, altered foetal development.

Asthma treatment has been widely investigated as a possible mediator of the adverse pregnancy outcomes. However, studies from the current authors' group 30, 63 and others 93 have not found any significant adverse effect of inhaled steroid use on foetal growth in females with asthma. In fact, asthma treatment appeared to protect against this adverse foetal outcome 30. Studies in which asthma was actively managed 25, 30, 33 and where the majority of subjects used ICS, were more likely not to report adverse outcomes in association with maternal asthma, than studies conducted prior to the introduction of ICS, or historical cohort studies where inhaled steroid use was unknown and could not be taken into account into the analysis.

Smoking, a contributor to low birth weight, has consistently been reported to be more common among asthmatics than nonasthmatics 1, 2, 6, 15, 27, 29, 35. However, most studies have found that maternal smoking does not fully explain the association between asthma and adverse pregnancy outcomes 6. Further research into the effect of smoking during asthmatic pregnancies in relation to both adverse pregnancy outcomes and the efficacy of inhaled or oral steroid treatment is warranted.

A common pathway leading to hyperactivity of the smooth muscle in both the bronchioles and the myometrium has been proposed to explain the increased incidence of pre-term labour in asthmatics 24, 52, 94. Bertrand et al. 94 initially suggested this mechanism after finding evidence of airway hyperreactivity in the mothers of premature infants, but this was not confirmed by another group examining airway responsiveness in mothers of premature or low birth weight children 95. At least one study has suggested an additional risk of prolonged pregnancy in asthmatic females, which could not be explained by this same mechanism 15. In addition, data from recent prospective cohort studies suggests that the risk of pre-term delivery may be largely explained by oral steroid use 31, 33.

The large number of adverse outcomes associated with asthma suggests there is a complex interaction of factors associated with the disease and possibly its treatment, which may alter normal maternal physiology during pregnancy. The current authors have identified maternal inflammation leading to altered placental function as a relevant mechanism that importantly can be modified by ICS therapy. In addition, the prevention of severe exacerbations of asthma during pregnancy is likely to lead to improved perinatal outcomes.

Changes in asthma symptoms during pregnancy

Studies examining changes in asthma symptoms during pregnancy date back to 1967 when Williams 8 examined hospital records of 210 asthmatic females and found that overall, 24% worsened and 42% improved during pregnancy. Subjects with severe asthma were more likely to worsen, a finding supported by more recent studies 66, 96.

Schatz et al. 97 examined the progression of asthma in 330 females during pregnancy and up to 12 weeks post-partum. Subjects subjectively rated their asthma as having improved, remained the same or worsened overall during pregnancy, compared to their pre-pregnancy state. In subjects whose asthma worsened, there was a significant increase in the number of days of wheezing and interference with sleep and activity between 25–32 weeks gestation. In asthmatic subjects who felt their asthma improved overall, there was a decrease in wheezing and little change in interference with sleep or activity between 25–32 weeks gestation. In all subjects, there was a fall in wheezing and interference with sleep and activity between 37–40 weeks. These changes highlight the importance of conducting trimester by trimester analyses of changes in asthma, as alterations at one time may not be the same as during another part of the pregnancy. Most subjects who felt their asthma worsened during pregnancy improved post-partum, with significantly fewer days of wheezing at 5–12 weeks post-partum compared with 29–32 weeks. Conversely, most subjects who felt their asthma improved during pregnancy worsened after pregnancy, with significantly more days of interference of activity in this period compared with 29–36 weeks gestation 97. Of concern is the fact that 8% of subjects who stated that their asthma improved still had an emergency department presentation for asthma during pregnancy, which may have put their foetus at risk of poor outcomes. It is therefore not only females who have a subjective worsening of asthma during pregnancy that are clinically important and may benefit from improved asthma management strategies during pregnancy.

Some subjects were assessed in two successive pregnancies and 60% followed the same course of asthma in the second pregnancy as the first 97. However, a substantial minority of patients did not follow the same course of asthma in subsequent pregnancies, suggesting that there must be a determinant of asthma that differed in the different pregnancies 97. A further study from this group examined characteristics such as smoking, maternal body weight, foetal sex, season of delivery and nasal symptoms in pregnancy, to determine whether any of these factors may be causing the pregnancy-associated changes in asthma 100. Season of pregnancy or delivery was not found to effect asthma progression 8, 100, suggesting that allergen exposure may not play a role in asthma alterations with pregnancy. Only the course of rhinitis during pregnancy correlated with the course of asthma during pregnancy 100. Rhinitis worsened or improved in >50% of patients whose asthma had also worsened or improved, respectively 100. The authors suggest that factors, such as IgE, which affect both the upper and lower airways, may be important in changes that occur in asthma during pregnancy 100. Another group found that cockroach-specific IgE levels in serum were linked to clinical asthma severity during pregnancy, and may be useful as a predictive measure 107. Early studies from Gluck and Gluck 96 also found a correlation between increased serum IgE and worsening asthma during pregnancy.

Lao and Huengsburg 11 reported that amongst treated asthmatics, 39% had no change, 30% had an increase and 31% had a decrease in the frequency and severity of symptoms or attacks during pregnancy. When compared with the group whose asthma did not change, females who reported an improvement in asthma during pregnancy had significantly higher per cent predicted PEF 11.

Data obtained from a cross-sectional survey of post-partum females in Western Australia indicated that during pregnancy, 16.4% of subjects experienced improved asthma and 35.4% experienced worsening asthma 6. Wheezing or asthma attacks were experienced by 62% of females during pregnancy. A large number of asthmatics were also smokers in this study, although smoking was not related to the changes in asthma during pregnancy 6.

Recent work suggests that maternal asthma symptoms may be influenced by the foetus. In a blind prospective study, Beecroft et al. 99 studied 34 pregnant mothers with moderate or severe asthma who were using regular treatments. Significantly more mothers pregnant with a female foetus reported shortness of breath, nocturnal waking and a worsening of cough and asthma in general, while mothers pregnant with a male foetus were more likely to report an improvement in asthma 99. Dodds et al. 103 reported that re-analysis of their Canadian population-based study indicated that fewer asthmatic subjects pregnant with a male foetus required steroids for treatment (14%), compared with asthmatic subjects pregnant with a female foetus (20%) 1. Although equal proportions of subjects pregnant with males or females used no drug treatment for asthma, there was a trend towards more subjects pregnant with a male using β₂-agonists alone (40% of subjects) compared with subjects pregnant with a female (35%), suggesting better managed asthma in the subjects pregnant with a male foetus 103. It would be interesting to investigate whether foetal sex influences maternal asthma in a study with larger numbers.

Changes in asthma treatment requirements during pregnancy

The current authors have found evidence that maternal asthma worsens in the presence of a female foetus 30, 49. Since their study found that the female foetus was particularly susceptible to reduced foetal growth in the presence of maternal asthma, it was questioned whether foetal sex may also be altering asthma. In females who were prescribed ICS, there was a significant increase in dose requirements from the first to third trimester only in those pregnant with a female foetus (n = 41), suggesting an increase in maternal inflammatory pathways associated with asthma in the presence of a female foetus 30. The current authors propose that factors derived from the male or female foetus may alter maternal asthma and could result from differences between male and female foetuses in protein or steroid expression, foetal gonadal or lung maturation, or differences in trafficking of foetal or placental cells or DNA from foetus to mother. The current authors also observed a significant rise in maternal circulating monocytes as pregnancy progressed in females who did not use ICS and were pregnant with a female foetus. This may be part of the mechanism explaining worsening maternal asthma during pregnancy with female foetal sex 30.

Changes in lung function and airway hyperresponsiveness during pregnancy

In 1976, Sims et al. 104 performed lung function tests on asthmatic females during pregnancy and post-partum. They found that there were no pregnancy-related changes in FEV₁:vital capacity (VC) ratio in 12 nonasthmatic or 27 asthmatic females either at rest or during exercise 104. Beckmann 35 also reported no changes in peak flow measurements made in each trimester in 22 pregnant females with asthma.

Juniper et al. 105 demonstrated an overall improvement in airway responsiveness to methacholine challenge in the second trimester compared with pre-conception in 16 subjects, but no significant changes in FEV₁ or FEV₁:VC were observed. When individual data was examined, 11 subjects improved and five worsened during pregnancy and there was no relationship between airway responsiveness and serum progesterone or oestriol concentrations 105, 110.

White et al. 106 examined changes in asthma during pregnancy using a symptom questionnaire and daily peak flow measurements. The results from subjective and objective measurements did not always agree, with 71% reporting a perception of improved asthma during pregnancy, and only 34% demonstrating an objective improvement in peak flow during the third trimester 106. This study demonstrates the difficulty of classifying changes in asthma during pregnancy based on patient perception.

Exacerbations of asthma during pregnancy

Hospitalisation during pregnancy has been reported to occur in 1.6% and emergency room visits in 12.6% of asthmatic patients 97. Wendel et al. 111 reported that, with the use of objective pulmonary function tests, 62% of exacerbations during pregnancy required hospitalisation of the asthmatic patient. Stenius-Aarniala et al. 23 found that 9.3% of subjects had an acute asthma attack during pregnancy and this was more common in females who did not use ICS. Acute attacks were normally distributed around 21–24 weeks gestation. They concluded that a mild attack of asthma, if promptly treated does not affect pregnancy or perinatal outcome 23.

Henderson et al. 107 analysed data collected during the 1960s, where 2% of females with asthma were hospitalised and 15% experienced an acute exacerbation without hospitalisation. They found that strongly positive cockroach-specific IgE in serum was associated with an increased risk of exacerbations with status asthmaticus.

Several studies indicate that females with severe asthma are more likely to show signs of worsening asthma during pregnancy than females with milder asthma 8, 66, 96. In a recent study, the relationship between asthma severity classification and subsequent changes in asthma during pregnancy was assessed in >1,700 pregnant asthmatics 66. Exacerbations of asthma occurred in over half of all severe asthmatics, while only 12% of patients with mild asthma had exacerbations during pregnancy. Re-classification of asthma from mild to either moderate or severe occurred in 30% of patients, while only 23% of patients who were initially moderate or severe were later re-classified as mild. Asthma morbidity, encompassing hospitalisations, symptoms, steroid requirements and unscheduled doctor visits, was found to be closely related to the pregnancy classification of asthma severity and overall occurred in ∼20% of subjects during pregnancy 66.

There is little evidence that labour and delivery themselves have any major effect on maternal asthma. If an acute attack occurs at this time, normal medication use is recommended 112. The prospective study of 198 asthmatic females by Stenius-Aarniala et al. 21 found that 14% of patients with atopic asthma and 22% of patients with nonatopic asthma experienced asthma symptoms during labour. They reported that in all subjects, symptoms during labour were mild and well controlled by inhaled β₂-agonists 21. Similar data has been reported by other groups 26, 55 including Schatz et al. 97 where 10% of females experienced mild symptoms during labour and delivery. A larger multi-centre study found that asthma symptoms were present during labour in 17.9% of all patients, with 46% of subjects with severe asthma experiencing symptoms during labour 66.

Mechanisms for the effect of pregnancy on maternal asthma

The mechanisms that contribute to changes in asthma during pregnancy are not well understood, although increases in maternal circulating hormones, altered β₂-adrenoreceptor responsiveness or foetal sex may be involved (fig. 3⇓).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3—

Changes in asthma during pregnancy.

The pregnancy-associated rise in serum-free cortisol may contribute to improvements in asthma during pregnancy 8, 113, since cortisol has anti-inflammatory properties. In addition, oestradiol and progesterone concentrations increase significantly during pregnancy 75. Progesterone is known to contribute to increased minute ventilation during normal pregnancy 114 and is also a potent smooth muscle relaxant 115 and may, therefore, be expected to contribute to improved asthma during pregnancy. Alternatively, changes in β₂-adrenoreceptor responsiveness and airway inflammation as a result of circulating progesterone may contribute to worsening asthma during pregnancy 116. Tan et al. 117 found that in female asthmatics, there was a desensitisation and down-regulation of lymphocyte β₂-adrenoceptors following administration of medroxyprogesterone. Alterations in asthma associated with changes in sex steroid production during the menstrual cycle have previously been observed 98, 118, with up to 40% of females experiencing an exacerbation around the time of menstruation when progesterone and oestradiol levels are low 119. No correlation has been found between the occurrence of pre-menstrual asthma and the progression of asthma during pregnancy 8, 98.

During pregnancy, exposure to foetal antigens, or alterations in immune function, may predispose some females to worsening asthma. Successful pregnancy has previously been described as a Th2 phenomenon 120–123, and asthma itself is primarily a Th2 mediated disease 124. Although in both asthma and pregnancy, the distinction between Th2 and Th1 immune deviation is not definitive 125, 126, in this sense, asthma may be expected to become worse during pregnancy. Rheumatoid arthritis, a Th1 mediated inflammatory disease, is known to go into remission during pregnancy in 75% of patients 127, 128.

The fact that some females experience an improvement in asthma during pregnancy, while others experience a deterioration of asthma, and that different patterns are observed in different pregnancies in the same mother 97, 129, casts doubt on the contribution of these major common hormonal or immune changes of pregnancy. However, studies in nonpregnant females have shown that a high proportion of asthmatics have an abnormal concentration of either progesterone or oestradiol compared with nonasthmatics, and these changes are not consistent across the entire group 130. Such individual abnormalities may explain why the progression of asthma during pregnancy differs between females.

Effect of foetal sex

The possible influence of foetal sex on maternal asthma during pregnancy may not be a novel concept. In 1961, the following comment was made by Schaefer and Silverman 7 in a discussion of his publication on seven cases of asthma in pregnancy: “There have been reports that asthma becomes worse only when the patient is pregnant with a female child and shows no change or gets better when she is pregnant with a male child”. These authors and others 8 did not find any data to support this statement in their own patients. However, reference to this older literature was also made by Green 131 in 1934 and Derbes and Sodeman 132 in 1946, who reviewed the non-English language literature dating back to the 1920s. In several studies, sex of the foetus had an effect on maternal asthma during pregnancy 132. No consistent patterns were observed in these case series. Green suggested that where asthma attacks during pregnancy were associated with a particular foetal sex, the factor responsible came from the sexual organs of the foetus 131. In 1930, Williamson 133 reported case histories of 13 females with asthma and 14 females with hay fever. Surprisingly, subjects had differing histories of urticaria during their pregnancies, which were related to the sex of the child 133. A case series, from New Zealand in 1964, found no consistent overall change in asthma, dependent on the sex of the foetus in subjects with at least three previous pregnancies, with some subjects reporting worsening of asthma in several pregnancies with a male foetus, and others reporting worsening of asthma in several pregnancies with a female foetus 134. The mechanisms leading to changes in asthma during pregnancy in the presence of a male or female foetus require further investigation.

There is likely to be a link between changes in maternal asthma and the increased risk of poor pregnancy outcomes. This has been suggested by several studies where worsening of asthma was associated with reduced birth weight 26, 30, 56. Studies by the current authors have examined possible relationships between the mother's asthma, placental function and foetal development (fig. 4⇓). In the presence of a female foetus, the current authors found that maternal asthma worsens during pregnancy, as demonstrated by an increased requirement for ICS and a significant rise in circulating monocytes 30, 49. These alterations in maternal asthma in the absence of corticosteroid therapy are associated with significantly reduced female birth weight and changes in placental function. Placental 11β-HSD2 activity is significantly reduced, which allows more maternally derived cortisol to reach the female foetus. Further changes in placental function, which may be due to the decrease in 11β-HSD2 activity, include a rise in the local Th2:Th1 cytokine mRNA ratio and decreased glucocorticoid and mineralocorticoid receptor expression. The changes in placental cortisol metabolism contribute to changes in the foetus, reducing growth in late gestation, and suppressing foetal hypothalamic-pituitary-adrenal axis function, as demonstrated by significantly reduced oestriol concentrations in female cord blood 30, 49. Improvements in maternal asthma control during pregnancy may contribute to better foetal outcomes.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4—

The interactions between mother, placenta and foetus in pregnancies complicated by asthma. 11β-HSD2: 11β-hydroxysteroid dehydrogenase type 2; GR: glucocorticoid receptor; Th2: T-helper cell type 2; Th1: T-helper cell type 1; MR: mineralocorticoid receptor.