Results

A total of 377 women were on OCP (24.5% of premenopausal women) and 458 were on HRT (15.2% of postmenopausal women). The age span of premenopausal women was 21–49 years and of postmenopausal women was 27–90 years. The general characteristics of women taking and not taking exogenous sex hormones are shown in table 1. In both premenopausal and postmenopausal women those taking sex hormones were younger; the remaining characteristics differed very little.

The prevalence of the chosen asthma indicators in premenopausal women according to treatment with OCP is shown in table 2. None of the asthma indicators was more prevalent in the subgroup on hormonal contraception. Stratification of the women according to smoking habits did not change the results. Similarly, no significant association between any asthma indicator and OCP was seen when using multiple logistic regression with age, length of school education, BMI, and smoking as covariates (not shown).

In postmenopausal women preliminary analyses suggested differences in smoking subgroups and therefore the results are presented separately for never smokers, ex-smokers, and smokers (table 3). In most of the comparisons the asthma indicators were not significantly more prevalent in the women on HRT. However, in never smokers there was a higher prevalence of self-reported asthma and asthma medication in those on HRT (table 3).

The results of multiple logistic regression of the effects of HRT and smoking on asthma indicators in postmenopausal women together with various covariates are shown in table 4. The odds ratios (ORs) for all four asthma indicators were higher in women on HRT with p values of approximately 0.05 and confidence intervals close to 1 (table 4). As expected, the ORs were higher in women with familial predisposition to asthma and short school education. A high BMI was significantly related to wheezing but not to the remaining asthma indicators. Smoking was strongly related to all asthma indicators.

Because of the trends observed in table 3, additional analyses were performed to explore the possible interaction between HRT and smoking on asthma indicators. A borderline interaction (p=0.03) was found which showed that HRT was associated with a much higher risk of asthma in never smokers (OR 3.3 (95% CI 1.6 to 7.1)) than in current smokers (OR 1.2 (95% CI 0.7 to 2.1). A similar interaction (p=0.05) was observed for asthma medication but not for wheezing and cough at exertion.

Discussion

In this study of women selected from the general population, we could not demonstrate any significant association between OCP and asthma, use of asthma medication, and asthma related symptoms. A relatively weak association was detected with HRT, suggesting that postmenopausal women on HRT have a slightly higher risk of asthma and asthma-like symptoms. The size of the population makes it unlikely that important negative effects of exogenous female sex hormones on asthma were undetected.

With regard to OCP, our results are clear and reassuring. Oral contraceptive pills reduce the dynamic fluctuations in oestrogen and progesterone levels over the menstrual cycle and, given that these cyclic changes cause changes in asthma severity, OCP might reduce their amplitude and thus make asthma more stable. This assumption is in line with most previous reports which have shown that OCP may reduce airway responsiveness, peak flow variability, and perceived severity of asthma.11 ,12 These findings are, however, not uniform and observations from a few small studies have even suggested that administration of exogenous sex hormones to premenopausal women may make asthma worse.13-15 However, a recent general practice based study of asthmatic women by Forbeset al found no significant differences in asthma quality of life scores between women on OCP and those not taking exogenous hormones.8 Furthermore, only 4% of the participants reported worsening of their asthma due to the hormonal contraceptives, and these women had significantly worse asthma than the remaining group of participants. This study therefore suggests that, at least in women with relatively mild asthma, oral hormonal contraceptives do not have a pronounced influence on symptom severity. Although our approach differs by looking at a sample of the general population, a lack of association between OCP and asthma and asthma-like symptoms is in keeping with the findings of Forbeset al.8

Postmenopausal women have very limited fluctuations in endogenous sex hormones which, at least theoretically, makes it less likely that HRT would have a substantial impact on asthma symptoms.6However, oestrogen enhances the production of corticosteroid binding globulin and progesterone competes with cortisol for binding sites. Thus, both oestrogen and progesterone may influence the level of free cortisol, leading to decreased levels. Furthermore, it has previously been hypothesised that oestrogen may induce resistance to cortisol in both peripheral tissues and the central nervous system. However, in line with these contradictory theoretical considerations, both smaller clinical studies and case studies of the influence of HRT on asthma severity have reported conflicting observations, some showing remarkable improvement in symptoms and others a pronounced worsening.16-19 Sex steroid hormones may play a role in the expression of asthma in susceptible individuals, but the only available epidemiological evidence for an association between HRT and adult onset asthma comes from the Nurses' Health Study7which showed not only that oestrogen is likely to play a role in the pathophysiology of asthma, but also that long term use and/or high doses of postmenopausal HRT significantly increased the risk for subsequent development of asthma. Although our cross sectional findings of a weak association between HRT and asthma indicators support the hypothesis of a deleterious effect of exogenous female sex hormones on asthma, the association was weak. After stratification for smoking habits, the strongest association was seen between asthma/asthma medication and HRT in never smokers. However, we did not observe a clear increase in asthma symptoms (wheeze and cough) in this smoking subgroup (table 3). The most likely explanation for this discrepancy is that postmenopausal women on HRT are more likely to visit their GPs than women not on HRT and, if they are never smokers, they more likely to have their respiratory symptoms diagnosed as asthma and to receive asthma medication. Another possible explanation is that, because of the strong influence of smoking on asthma-like symptoms, a potential association between HRT and asthma is only detectable in non-smokers. A third possible explanation for the interaction between HRT and smoking in asthma is that the deleterious effect of HRT is counteracted by the anti-oestrogenic effect of smoking. Although these explanations are speculative at present, previous studies have suggested that smoking may exert an anti-oestrogenic effect by altering the metabolism of endogenous and exogenous sex hormones.20 ,21 Finally, although self-reported asthma has been shown to be a valid indicator of asthma in epidemiological settings,22 some degree of misclassification in postmenopausal women has probably occurred in our study with some subjects with mild asthma remaining undiagnosed and others with chronic obstructive pulmonary disease being diagnosed as asthmatics.

We conclude that there is no relationship between the use of oral hormonal contraceptives and asthma and asthma-like symptoms in the general population, while HRT after the menopause seems to be associated with a slightly higher risk of asthma. However, since this association was relatively weak, we think that further studies are needed and that present prescription policies in postmenopausal women do not yet need to be changed.