Endogenous and exogenous sex steroid hormones in asthma and allergy in females: A systematic review and meta-analysis.

To the Editor:

Asthma and allergy are more common in males than in females during early childhood, but the incidence, severity, and impact on quality of life are greater in postpubertal females than in males.1, 2 Female sex steroid hormones may partly explain these differences.1, 2 In 2 previous systematic reviews, early menarche (<12 years) was associated with an increased asthma risk, whereas no significant association was found between menopause and asthma, although subgroup analyses indicated an increased risk in postmenopausal women using hormone replacement therapy (HRT). Consideration of other hormonal factors, along with the full spectrum of relevant outcomes, is necessary for a comprehensive appreciation of the underlying evidence base. We therefore undertook a systematic review investigating the role of endogenous and exogenous hormonal factors in the development and clinical expression of asthma and allergy in females.

Our methods were published a priori (PROSPERO: 2015:CRD42015026762). Further details are available in this article's Online Repository at www.jacionline.org. We included experimental and analytical epidemiological studies of females from puberty to adulthood (<75 years). Exposures were puberty, menarche, menstruation, menopause, hormonal contraceptives, and HRT. Primary outcomes were self-reported or objectively defined incidence or prevalence of asthma, asthma exacerbations, asthma hospitalizations, and asthma medication use.

We searched 11 bibliographic databases, databases of ongoing studies, and conference abstracts, and contacted experts for articles published between January 1990 and November 2015 with no language restrictions. N.M. and B.I.N. independently screened titles, abstracts, and full-text articles; extracted study data; and assessed risk of bias using the Cochrane Risk of Bias Tool (experimental studies) and the Effective Public Health Practice Project tool (observational studies). Discrepancies were resolved by discussion, or arbitration by A.S.

Adjusted effect estimates were combined in random-effects meta-analyses, performed using Stata release 14 (StataCorp, College Station, Tex). Meta-analyses were possible for studies on menarche, menstruation, menopause, hormonal contraceptives, and HRT. Stratified analyses were performed by body mass index and smoking for HRT studies.

Of 22,488 articles retrieved, 64 (reporting 57 studies; observational: 51; experimental: 6) were included with 554,293 participants analyzed (see references E5 and E10-E72 and Fig E1 in this article's Online Repository at www.jacionline.org). Study characteristics are available on request.

Fig E1

Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) flow diagram summarizing the identification, screening, eligibility, and inclusion of studies that investigated associations between endogenous and exogenous sex steroid hormonal factors and asthma and allergy in females. For Google Scholar, the first 400 hits were selected for screening.

Detailed results are given in this article's Online Repository at www.jacionline.org; here, we present key findings. Compared with typical menarche (11-13 years), early menarche (<11 years) was associated with increased risk of new-onset (odds ratio [OR], 1.49; 95% CI, 1.14-1.94) and ever asthma (OR, 1.06; 95% CI, 1.03-1.10), whereas late menarche (>13 years) was associated with increased risk of ever (OR, 1.11; 95% CI, 1.07-1.15), but not new-onset asthma (OR, 1.13; 95% CI, 0.82-1.56) (Fig 1).

Fig 1

Meta-analyses of studies that investigated associations between onset of menarche and asthma and allergy in females. N.Europe, Northern Europe; RR, risk ratio. All effect estimates are adjusted. Weights are from random-effects analysis. Early menarche: <11 years; late menarche: >13 years; the comparator group in each analysis is typical menarche: 11 to 13 years.

Compared with regular menstruation, irregular menstruation was associated with increased risk of current asthma (past 12 months) (OR, 1.59; 95% CI, 1.23-2.05) (see Fig E2, A, in this article's Online Repository at www.jacionline.org), specifically for atopic (OR, 2.57; 95% CI, 1.66-3.98), but not nonatopic asthma (OR, 0.95; 95% CI, 0.54-1.65) (Fig E2, B).

Fig E2

A, Meta-analyses of studies that investigated associations between irregular menstruation and asthma and allergy in females. B, Meta-analyses of studies that investigated associations between irregular menstruation and asthma in females, stratified by atopy. N.Europe, Northern Europe. All effect estimates are adjusted. The comparator group in each analysis is regular menstruation.

Compared with premenopause, menopause onset was associated with increased risk of current asthma (OR, 1.25; 95% CI, 1.04-1.51) and current wheeze (OR, 1.16; 95% CI, 1.05-1.30), but not current allergic rhinitis (OR, 0.94; 95% CI, 0.81-1.10) (see Fig E3 in this article's Online Repository at www.jacionline.org).

Fig E3

Meta-analyses of studies that investigated associations between onset of menopause and asthma and allergy in females (A) and between OCP use and asthma and allergy in females (B). N.Europe, Northern Europe. All effect estimates are adjusted. The comparator group is premenopause (A)/never use (B).

Results for hormonal contraceptives were mixed, with both increased and decreased risks reported (Fig E3).

Compared with never use, ever HRT use (hazard ratio [HR], 1.37; 95% CI, 1.22-1.54), past use (HR, 1.41; 95% CI, 1.22-1.63), current use (HR, 1.48; 95% CI, 1.22-1.78), and current use of estrogen-only HRT (HR, 1.85; 95% CI, 1.50-2.28) were associated with increased risk of new-onset asthma (Fig 2). Current use was also associated with increased risk of current asthma (OR, 1.42; 95% CI, 1.18-1.70), and current wheeze (OR, 1.40; 95% CI, 1.22-1.61), but not current allergic rhinitis (OR, 1.27; 95% CI, 0.97-1.68) (Fig 2). The risk was higher in nonoverweight/nonobese and nonsmoking women than in overweight/obese and smoking women, respectively (see Fig E4 in this article's Online Repository at www.jacionline.org).

Fig 2

Meta-analyses of studies that investigated associations between the use of HRT and asthma and allergy in females. N.Europe, Northern Europe. All effect estimates are adjusted. Weights are from random-effects analysis. The comparator group in each analysis is never use.

Fig E4

Meta-analyses of studies that investigated associations between the use of HRT and asthma and allergy in females, stratified by overweight/obesity (A) and smoking status (B). N.Europe, Northern Europe. All effect estimates are adjusted. The comparator group in each analysis is never use.

Forty-one of the 51 observational studies had moderate risk of bias, whereas the rest had high risk; all 6 experimental studies had high risk of bias (data available on request).

This is the most comprehensive synthesis to date linking sex steroids to the development and expression of asthma and allergy in females. We followed recommended steps for undertaking a high-quality synthesis. However, the lack of high-quality experimental studies limited assessment of causality and relevance to patient care and policy. Although most epidemiological studies adjusted for key confounders, this was often not comprehensive. Many outcomes (eg, medication use, exacerbations, and hospitalizations) were not assessed, and so there is little evidence in relation to these.

Questions that remain to be addressed center on the influence of different types of sex steroids, dose and route of administration of exogenous sex steroids, and the underlying biologic mechanisms through which hormones may influence asthma and allergy. Early menarche and irregular menstruation are often signs of anovulation, indicating episodes of unopposed estrogen exposure to target organs and absences of progesterone exposure. Estrogen-only HRT was associated with asthma, whereas smoking, which may influence estrogen metabolism, had a protective effect. Although higher body mass index is generally associated with a more estrogenic state and also an increased risk of asthma, we found increased risk in both overweight/obese and nonoverweight/nonobese HRT users. At the cellular level, estrogen can have proinflammatory or anti-inflammatory effects, depending on cell type and location. Explanations for the various associations found are undoubtedly complex, and it is unlikely that all can be explained by one biological mechanism. Our results also suggest that atopy may be a contributing factor in some instances (irregular menstruation) but not others (menopause). The differences in findings for early and late menarche may be due to the different asthma outcomes investigated: alternatively, the effect observed in cross-sectional studies may reflect reverse causation. Further mechanistic work is required to elucidate any relationships, as are further longitudinal observational studies with detailed phenotyping of participants.