Am J Respiratory Crit Care Med 2002 (Sep 15) ; 166 (6): 827—832 ~ FULL TEXT
Tricia M. McKeever, Sarah A. Lewis, Chris Smith and Richard Hubbard
This study of British children suggests that women who have an infection or take antibiotics during pregnancy are more likely to have a child with an allergy-related condition such as asthma, hay fever or eczema. Researchers at the University of Nottingham evaluated the medical records of nearly 25,000 British children and their mothers. The study found that children exposed to antibiotics in the womb had a higher risk of developing asthma, hay fever and eczema than did children whose mothers did not take the medication during pregnancy. Because a person's immune system develops while he or she is still in the womb, some experts speculate that factors that modify microbial exposure at this time may have a long-term effect on the risk of developing allergic disease.
The Full-Text Article:
The prevalence of allergic disease has increased dramatically in the developed world during the second half of the 20th century, and it has been suggested that this increase is in part due to reductions in early microbial exposure . The main evidence for this hypothesis comes from research demonstrating strong protective birth order effects on the risk of having or developing an allergic disease [2–9] and also that exposure to antibiotics early in life increases the risk of developing allergic disease [8, 10–13]. Because the immune system develops in utero, factors that modify microbial exposure at this time may have a long-term impact on the risk of developing allergic disease, but research in this area has been limited.
In this study, we have used a birth cohort analysis using data from the West Midlands General Practice Research Database to investigate the relationship between a variety of exposures that alter microbial load during pregnancy, including diagnosis of infections and prescriptions for antibiotics and the incidence of asthma, eczema, and hay fever in the child. We have also taken the opportunity to study the impact of a number of other perinatal exposures, including maternal depression, contraceptive use, and pregnancy complications and have investigated whether pregnancies that do not produce a live child decrease the risk of allergic disease in subsequent children.
Our results suggest that exposure to antibiotics in utero is associated with a dose-related increase in the child's risk of allergic disease, and we found similar though smaller effects for exposure to infections. We have already reported that there are strong birth order effects within this cohort , but previous pregnancies that did not lead to a live birth did not protect against the development of allergic disease in the child. Perinatal maternal depression was associated with an increased risk of having a diagnosis of all three allergic diseases, particularly asthma. Finally, complications in pregnancy were associated with an increased risk of asthma, but not for eczema or hay fever.
There are a number of advantages and disadvantages to using general practice data such as the General Practice Research Database for research. Our case definition is based on doctor-diagnosed disease. Although this approach has been used for many other research studies, our case definition was dependent on the child going to the doctor and receiving a diagnosis. A major factor to consider, therefore, is ascertainment bias; that is, families who tend to consult the doctor more frequently may be more likely to have both exposures recorded and diagnoses made, thereby producing a positive association. To address this problem, we have adjusted our analyses for consulting behavior, but as children must consult to receive a diagnosis, it is possible that by adjusting for consulting behavior we have “overadjusted” our analyses making our results conservative. There are no specific markers of social class with the General Practice Research Database, but we have investigated whether maternal smoking, which is related to social class, is a confounding factor and found no evidence for it. In addition, our marker of infection is based on doctor-diagnosed disease and therefore will not contain all exposure to infection, although we suspect that during pregnancy a mother is more likely to consult over infections because of concerns for her baby's health.
Our findings of a relationship between exposure to infections in general during pregnancy and an increased risk of allergic disease are consistent with previous research, which has demonstrated that infections in the first trimester of pregnancy are associated with an increased risk of allergic disease [14, 15]. These data are clearly not in keeping with the hygiene hypothesis, which suggests that exposure to early infections may be protective.
Antibiotics can cross the placenta and enter the fetal circulation , and there are data that suggests that exposure to antibiotics early in life may increase the child's risk of developing allergic disease [8, 10–13]. We found a dose–response relationship between antibiotics given during pregnancy and the child's risk of developing an allergic disease, and to our knowledge, this is the first time that this association has been reported. Although the dose–response evidence points toward a causal relationship, it is possible that some ascertainment bias is present, although adjusting for consulting behavior had only a minor effect. To try to examine the impact of family consulting behavior further, we have looked at paternal antibiotics in the pregnancy period, and here we found no associations.
The presence of older siblings has consistently been shown to decrease the incidence of allergic disease, and this is true in this cohort; however, we and others have not been able to explain this effect in terms of early life exposure such as infection , and similarly, antibiotics or infections in pregnancy are not the reasons [6, 10, 17]. An alternative explanation for the birth order effects is that they are mediated by pregnancies changing the mother. A cross-sectional study has found that women who had a greater number of children were less likely to be atopic , and a cohort study found that cord blood immunoglobulin E levels were higher in children with older siblings and that cord blood immunoglobulin E was a better predictor than number of older siblings for atopic sensitization at the age of 4 years . In this study, we found no evidence that previous pregnancies terminated through a therapeutic or spontaneous abortion reduced the risk of allergic disease in subsequent children. It is possible that a protective effect of pregnancy occurs only if the baby is carried to term, but against this is the fact that we were unable to detect a protective effect with stillbirths, although there was limited statistical power for this analysis. However, our data suggest that it is exposure to living older siblings rather than the number of pregnancies themselves that is important.
We found a small association between use of contraception in the 12 months before conception and developing of allergic disease and in keeping with one previous study , although our effect disappeared after adjusting for consulting behavior, suggesting that this may be due to ascertainment bias. It has been shown that cord blood immunoglobulin E was significantly higher in mothers that were taking progesterone therapy during pregnancy . Given the higher incidence of allergic disease in boys compared with girls and the impact of estrogens and progesterone levels during pregnancy, this is an area that requires more research.
A number of studies have found evidence of a relationship between atopic disease and affective disorders [21–27]. Our results suggested that maternal depression may increase the incidence of the child's allergic diseases, particularly asthma. Clearly, ascertainment bias may explain some of this relationship, but our findings are supported by others who have shown that parental stress early in life increases the child's risk of asthma [28–31] and raises both maternal and cord blood immunoglobulin E and invokes a Th2-dominated environment in utero .
In this study, we found evidence that complications recorded during pregnancy increased the incidence of all three allergic diseases, particularly asthma. Because maternal allergic disease is an important risk factor for child allergy  and mothers with asthma may have more complications during pregnancy [33–35], maternal allergy will be an important confounder in the analysis. The size of our effects was unchanged by adjusting for maternal allergic status, however, and similar impacts of pregnancy complications were present when mothers with and without asthma were analyzed separately. The data therefore suggest that pregnancy complications are an independent risk factor for allergy, and these findings are consistent with previous research, which found a relationship between complications in pregnancy and a child's risk of developing allergic disease [15, 36, 37].
In summary, our research suggests that exposures, which modify microbial load during pregnancy, may increase a child's risk of developing allergic disease. Some of these exposures, particularly the use of antibiotics, are potentially avoidable, and thus, further research into the impact of these exposures on the incidence of allergic disease is required. The presence of pregnancies that do not lead to a live birth do not protect against the risk of developing allergic disease, suggesting that the birth order effects seen in this and other cohorts are mediated by exposure to live older siblings, rather than changes in maternal physiology.