INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Several lines of evidence in young women and women with asthma suggest that estrogen and progesterone may play a role in modifying pulmonary function. From age 20 to 50 yr a higher prevalence of asthma among women compared with men has been consistently reported (1, 2). After age 50 yr, the prevalence of asthma has been shown to be higher among women in some studies, and higher among men in others (2- 5). Hospital admissions for asthma are more common among women after early adulthood (1). Premenopausal women have been reported to experience drops in peak flow and worsening of asthma symptoms before and during menses, often experiencing relief after the onset of progesterone and/or estrogen therapy (6). Subsets of pregnant women were reported to experience improvements in asthma symptoms during their pregnancy (15). Oral contraceptive users have been found to have significantly higher total lung capacities when compared with nonusers during the follicular phase of the menstrual cycle (18). Oral contraceptives also appear to reduce premenstrual declines in pulmonary function among women with asthma (19).

Studies examining the relationship between hormones and pulmonary function among postmenopausal women are limited. Two small studies have examined pulmonary function in postmenopausal women while on estrogen replacement therapy: one in a group of women without asthma and one in a group of women with asthma. Women without asthma did not have significantly different forced expiratory volume in 1 s (FEV₁) after the initiation of estrogen therapy (20). However, estrogen therapy was associated with an inhibitory effect on airway reactivity of bronchial smooth muscle (20). Women with asthma experienced significantly decreased peak expiratory flow and increased bronchodilator requirements after estrogen use, but there was no significant difference in spirometric values (21). Both of these studies were small (n = 15 and n = 36, respectively), did not control for potential confounders, and considered only unopposed estrogen therapy. The findings are not consistent with case study findings in postmenopausal women (8) and are also contrary to the findings reported in other cohorts of premenopausal women (6).

Because of the observed benefits of endogenous and exogenous hormones in premenopausal women and women with asthma, we hypothesized that HRT, as estrogen or estrogen/ progestin, would be independently associated with better pulmonary function in a large cohort of postmenopausal women. Specifically, we sought to determine the associations of HRT use with FEV₁, forced vital capacity (FVC), and pulmonary obstruction (FEV₁/FVC below the lowest quintile for the population). We hypothesized that use of HRT would be associated with higher pulmonary function and examined whether this association differed by asthma and smoking status.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Population

The Cardiovascular Health Study (CHS) is a population-based, longitudinal study of adults aged 65 yr and older designed to identify factors related to coronary heart disease and stroke (22). In 1989 and 1990, 5,201 men and women were recruited from a random sample of the Medicare-eligibility list provided by the U.S. Health Care Financing Administration for four U.S. communities: Forsyth County, North Carolina; Sacramento County, California; Washington County, Maryland; and Pittsburgh, Pennsylvania (22). A supplemental cohort of 685 African American adults were recruited from the original CHS communities, except Washington County in 1992-1993. Spirometry was done on both cohorts in 1994-1995. Of the 3,393 women participants in that examination year, 2,353 had information on HRT use and spirometry.

Interview and Examination

Psychosocial factors, medical history, and smoking history were assessed by questionnaire. Examination included blood pressure, anthropometry, electrocardiography, spirometry, laboratory testing, ultrasonography, and echocardiography.

Pulmonary Function

Pulmonary symptom data were collected using a subset of the American Thoracic Society ATS DLD-78 Respiratory Questionnaire (23). A water-sealed, Collins Survey II spirometer (WE Collins, Braintree, MA) was used to measure FEV₁ and FVC. The quality and reproducibility of spirometric data were assessed centrally (24). Obstruction was defined as an FEV₁/FVC value in the lowest quintile (< 0.65). "Definite asthma" was defined as self-report of a current physician diagnosis of asthma (25). "Probable asthma" was defined as self-report of asthma symptoms (25), excluding those with definite asthma or congestive heart failure (26). Chronic bronchitis and chronic emphysema were based on self-report of a current physician diagnosis.

Hormone Replacement Therapy Use

Drug name, strength, and dosing instructions were transcribed from labels of prescription medication containers brought to the clinic by participants (27). Medication data were obtained in 2,807 women. Of these women, 61.3% reported never using estrogen or estrogen/ progestin, 22.6% reported past use, and 16.0% reported current use. Current HRT use was defined as estrogen or estrogen/progestin use. Of the women using HRT, 81.4% were using unopposed estrogen, and 18.6% were using combined estrogen/progestin.

Statistical Analyses

Univariate analyses showed no difference in pulmonary function between past versus never users of HRT. Thus, past and never users were collapsed into a group of noncurrent users.

The prevalence of current HRT use was compared in subgroups using the chi-square test of proportions to determine if a "healthy-user effect" was possible. Factors known to be associated with HRT use (28) or pulmonary function (24) were evaluated as potential confounders.

Differences in unadjusted and adjusted mean FEV₁ and FVC between current HRT users and noncurrent users in the whole population of women, and in subsets defined by asthma and smoking status, were compared. Chi-square tests and logistic regression were used to determine the unadjusted and adjusted relationship between current HRT use and pulmonary obstruction in the entire population of women, and in subsets defined by asthma and smoking status.

Pulmonary function was also compared among noncurrent HRT users, unopposed estrogen users, and combined estrogen/progestin users. The significance level was set at 0.05. Analyses were performed using SPSS for Windows Version 9.0.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Of the total 2,353 women CHS participants who were in the 1993/1994 examination, there were a total of 204 women with asthma and 2,149 women without asthma; 209 current smokers, 795 former smokers, and 1,339 never smokers. The mean age of the cohort was 75.2 yr and 83.1% of this cohort was white.

Among HRT users, 9.4% of women had asthma; similar to 8.5% of the nonusers (see Table 1). Overall, HRT use was most prevalent among never smokers, followed by former smokers and current smokers, respectively. However, former smokers were more likely to be current HRT users than nonusers. HRT use was inversely associated with age and body mass index (BMI). White women were more likely than nonwhite women to use HRT. Individuals with a high school or greater education were more likely to use HRT. Women with better self-reported health status were more likely to use HRT than women with fair and poor self-reported health status. Correlates of HRT use in the CHS cohort have been previously reported (28).

Mean values of FEV₁ and FVC were significantly higher among current HRT users than noncurrent users in the entire population of women and among women without asthma, former smokers, and never smokers (see Table 2). FEV₁ was consistently 9% to 10% higher and FVC was about 8% higher among current HRT users than noncurrent HRT users in the entire population and among women without asthma, former smokers, and never smokers. Women without asthma demonstrated a similar nonsignificant trend of more than a 9% higher FEV₁ among current HRT users than among noncurrent users, and a smaller nonsignificant trend of a 4% higher FVC among HRT users than among noncurrent users. Current smokers had a nonsignificant trend of a smaller magnitude: about a 3% higher FEV₁ among HRT users compared with nonusers. After adjustment for factors known to influence HRT use and pulmonary function, the difference in mean FEV₁ was attenuated: about 3% to 4% higher in HRT users in the entire population (p = 0.031), among never smokers (p = 0.039), among women without asthma (p = 0.086), and among former smokers (p = 0.013). The difference in the adjusted mean FEV₁ between HRT users and nonusers among women with asthma was larger (about 8% higher) but nonsignificant. There were minimal differences in adjusted mean FVC.

There were significant differences in both FEV₁ and FVC based on HRT type. Combined estrogen/progesterone users had the highest levels of FEV₁ (1.92 L), followed by unopposed estrogen users (1.80 L), and then nonusers of HRT (1.66 L) (one-way ANOVA p value < 0.0001, linear test of trend p value 0.0001). There was an 8% difference in FEV₁ between nonusers and unopposed estrogen users, and an additional 7% difference in FEV₁ between unopposed estrogen users and combined estrogen/progesterone users. Combined estrogen/progesterone users also had the highest levels of FVC (2.65 L), followed by estrogen-only users (2.50 L), and then nonusers of HRT (2.34 L) (one-way ANOVA p value < 0.0001, linear test of trend p value 0.0001). This translates to a 7% difference in FVC between estrogen-only users and nonusers, and an additional 6% difference in FVC between estrogen-only users and estrogen plus progesterone users. In summary, estrogen plus progesterone users had higher levels of pulmonary function than both nonusers and unopposed estrogen users. The type of HRT remained a significant predictor of FEV₁ after adjustment for all confounders. The type of HRT was no longer a significant predictor of FVC after multivariate adjustment, although there was a trend toward higher FVC.

For the whole group of women, current HRT users were about 25% less likely to have obstruction than women who did not currently use HRT (OR 0.75, 95% CI [0.56, 0.99]see Table 3). The magnitude of the association was similar in all groups, though the difference between current HRT users and nonusers did not reach significance among all groups. Subgroup analyses had smaller sample sizes that limited the power to detect differences between current HRT and noncurrent HRT users.

After controlling for a number of factors known to influence FEV₁ and HRT use, current HRT use was significantly associated with higher FEV₁ (p = 0.031see Table 2) and less obstruction (OR = 0.67, 95% CI [0.48, 0.95]see Table 4). In analyses limited to nonsmokers, results were similar (OR = 0.66, 95% CI [0.46, 0.96]see Table 4). HRT was also associated with higher FVC, although results approached but did not reach significance.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

To our knowledge, this is the first study to specifically examine the relationship between postmenopausal estrogen and progesterone use and pulmonary function in a large cohort of older women. Women who currently used HRT were less likely to have airway obstruction, even after controlling for variables known to influence pulmonary function. The results remained significant among nonsmokers, decreasing the possible confounding effect of smoking. Women who currently used HRT also had higher FEV₁, though these differences were small. Small differences might be expected in a cohort such as CHS with healthy, normal ranges of pulmonary function. Larger differences in adjusted FEV₁ were found among women with asthma, a subset with a broader range of pulmonary function. Overall and in all subgroups except never smokers, there was no difference in FVC in women on HRT compared with those not on HRT after adjusting for factors known to influence pulmonary function and HRT use. This may suggest HRT affects pulmonary function by reducing airway obstruction rather than by reducing restriction of total lung capacity.

It is possible that the results may be explained by intractable residual confounding. However, we adjusted for many factors that could have influenced our results. Results remained significant and clinically detectable, and are supported by in vitro models. Estrogen and progesterone have been associated with relaxation of airway smooth muscle. Animal and human models have found that both progesterone and estrogen induce a relaxation of bronchial muscle (29). Progesterone and estrogen also decrease muscle contraction of airway smooth muscle (30).

An alternative explanation of the positive association could be that HRT improves the skeletal and muscular integrity of the thoracic spine, improving pulmonary function. Estrogen and progesterone have been found to increase both muscle strength and function and induce skeletal myoblast growth (31). HRT is also associated with increased bone mineral density of the spine (34), which decreases the risk of thoracic spine compression. However, if HRT improves musculoskeletal integrity, which increases total lung capacity, then FVC should have been significantly higher among HRT users after adjustment for confounders as well. Therefore, our findings are better explained by the former biological mechanism.

All women and subgroups of women without asthma, former smokers, and never smokers who were using HRT were more likely to have higher FEV₁ and FVC than women who were not using HRT. Similarly, women with asthma who were currently users of HRT had a trend toward higher FEV₁ and FVC values than women with asthma who did not currently use HRT, which were not significant. Stratified analyses showed similar trends toward higher FEV₁ values in smokers who were currently using HRT compared with smokers who were not currently using HRT, which were not significant. Thus the association was in a positive direction in most groups with consistently higher FEV₁ and FVC among current HRT users in all groups (except for FVC in smokers). The nonsignificant findings among women with asthma and current smokers are likely due to the small sample sizes in these groups, resulting in low power to detect the associations. Further examination of the data also suggests that former smokers who use HRT have pulmonary function values similar to never smokers who do not use HRT (FEV₁ 1.76, FVC 2.51 and FEV₁ 1.72, FVC 2.37, respectively).

Though causality cannot be determined from the cross-sectional analyses, several inferences can be drawn. The magnitude of an 8% improvement in lung function (FEV₁) in a patient with asthma does not meet the standard criterion of a "significant" response to an inhaled bronchodilator (an increase of 12% or more); however, because 8% is the adjusted mean difference in FEV₁ with HRT use, half of our elderly females with asthma had a greater than 8% difference, which may become clinically important in some patients. This supports further examination of the potential effects of HRT administration on pulmonary function in women with asthma.

This study is consistent with prior findings that reduced pulmonary function was associated with periods of reduced estrogen and/or progesterone levels in premenopausal women. We were unable to determine if these changes in pulmonary function parameters resulted in changes in the number of asthma exacerbations over time, as CHS did not collect longitudinal data on asthma exacerbations.

Our findings are not consistent with those from The Nurses' Health Study, which reported that current and past users of postmenopausal hormones were at an increased risk of developing asthma when compared with postmenopausal women who never used HRT (35). The findings of the Nurses' Health Study were not significant in the postmenopausal women who currently used HRT, but the direction of the association appears to contradict our findings that hormone replacement therapy is associated with higher pulmonary function. The CHS cohort is more representative of an older group of women than the Nurses' Health Study cohort. CHS women who use HRT are generally long-term users (28). The findings of the Nurses' Health Study may differ from the findings of the CHS if the adverse associations with pulmonary function in the Nurses' Health Study cohort represent an adverse association among some women during early use of HRT. Our results could be biased if some former HRT users in CHS are women who quit HRT due to asthma exacerbation, lowering the mean FEV₁ in the nonuser group. However, we found that women with asthma and women without asthma were equally likely to be current long-term users of HRT. Additionally, HRT was associated with reduced obstruction in women with asthma, suggesting that HRT use may not aggravate asthma. Therefore in this cohort, women with asthma are just as likely to use HRT as women without asthma and do not have lower pulmonary function than women with asthma who do not use HRT.

Our findings are also not consistent with those reported from the two small prospective studies of pulmonary function in postmenopausal women while on estrogen replacement therapy: one in a group of women without asthma and one in a group of women with asthma (20, 21). A worsening of peak flow among women with asthma after estrogen administration was reported (21). These findings are also opposite to those in young women. The reasons for the differences in our findings may not be clear as the literature in this area is too preliminary. However, these prospective studies utilized estrogen replacement therapy only whereas some women in CHS were taking combined estrogen/progesterone therapy. It may be possible that estrogen and progesterone have either similar or opposite effects on the pulmonary system, which could explain the differences between our findings. This has not previously been examined in vivo. Both estrogen and progesterone have been found to reduce contractility and increase relaxation of bronchial muscle in vitro (29, 30). Because so few studies are available, it is difficult to determine the reasons for the discrepancies at this time.

The main limitation of this study is the cross-sectional design; we cannot make causal inferences about the association of HRT use with pulmonary function. It is difficult to completely adjust for the healthy user effect and is possible that factors not accounted for also explain observed associations. Further, the sample size of women with asthma and smokers currently taking HRT limited power to detect associations of HRT use with pulmonary function in these groups (although trends were apparent, and in the hypothesized direction). Prospective studies among both women without asthma and women with asthma should examine baseline spirometry values before HRT or oral contraceptive use to determine significant changes in pulmonary function.

Conclusions: HRT users in CHS are younger, more highly educated, more likely to be white, less obese, and more likely to be past or never smokers (33). These factors may influence both pulmonary function and HRT use. However, CHS postmenopausal women using HRT were found to have higher levels of FEV₁ and FEV₁/FVC, which were not explained by their lower rates of smoking and other health factors associated with HRT use.