|Home | About | Journals | Submit | Contact Us | Français|
Abbreviations: ACCESS, A Case Control Etiologic Study of Sarcoidosis; BWHS, Black Women's Health Study; CI, confidence interval; IRR, incidence rate ratio.
The authors assessed the relation of hormonal and pregnancy-related factors to the incidence of sarcoidosis in the Black Women's Health Study. On biennial questionnaires, participants (US black women aged 21–69 years at baseline) reported data on diagnoses of sarcoidosis, reproductive history, and medication use. Cox regression models, adjusted for age, education, geographic region, smoking, and body mass index, were used to estimate incidence rate ratios and 95% confidence intervals. During 694,818 person-years of follow-up from 1995 through 2009, 452 incident cases of sarcoidosis were identified. The incidence of sarcoidosis decreased as age at menopause increased (P-trend = 0.03). Both later age at first full-term birth and having a more recent birth were associated with a reduced incidence of sarcoidosis. In models that included both factors, the incidence rate ratios were 0.60 (95% confidence interval: 0.37, 0.97) for age at first birth ≥30 years versus <20 years (P-trend = 0.05) and 0.73 (95% confidence interval: 0.43, 1.24) for <5 years since last birth versus ≥15 years (P-trend = 0.15). No significant associations were observed with age at menarche, parity, lactation, oral contraceptive use, or female hormone use. These results suggest that later full-term pregnancy and longer exposure to endogenous female hormones may be related to a reduced risk of sarcoidosis.
Sarcoidosis is a systemic, granulomatous disorder of unknown etiology for which pulmonary disease is the most common manifestation (1, 2). In the United States, the incidence is greater in African Americans than in whites (3, 4). There is some evidence that the incidence is greater among women than among men (1, 5–8), but the literature is not consistent (7–9). For both men and women, the peak incidence occurs before age 50 years (1, 2, 4, 7). In some populations (8–10), there is a second peak of incidence after age 50 years that is higher in women than in men, suggesting that endogenous hormones may influence risk. Data from animal studies are compatible with this hypothesis. Shirai et al. (11) observed that ovariectomized rats developed more lung granulomas than did nonovariectomized rats in response to Bacillus Calmette-Guérin. A possible protective effect of exogenous female hormones is suggested by the additional findings of Shirai et al. (11) that estrogen supplements diminished the extent of the lung granulomas. In addition, symptoms in sarcoidosis patients sometimes improve during pregnancy, suggesting a possible benefit of pregnancy hormones (12–21). However, these observations stem from case reports and case series, where the lack of a clear comparison group limits interpretation.
Noncaseating granuloma formation, the pathologic hallmark of sarcoidosis (1, 22, 23), is dominated by type 1 helper T cells, which evoke phagocyte-dependent inflammation (2, 24–29). This differs from type 2 helper T-cell-dominated inflammation, which inhibits phagocyte function. High-estrogen states (e.g., pregnancy) favor a type 2 helper T-cell response, while an androgenic or low-estrogen state favors the development of a type 1 helper T-cell response and granuloma formation. The importance of the immune pathway has recently been supported by findings that variants in the BTNL2 gene (30), which are involved in T-cell activation, are associated with sarcoidosis risk (30, 31). While there is ample evidence that sex hormones can affect the immune system (32, 33), opposing effects make it difficult to predict what the ultimate relation with disease risk might be.
We hypothesized that hormonal and pregnancy-related factors influence the risk of sarcoidosis. We assessed the relation of exogenous estrogen exposure (oral contraceptive use, supplemental female hormone use) and markers of endogenous estrogen exposure (age at menarche, age at menopause, pregnancy-related variables) to the incidence of sarcoidosis in African-American women. The data were obtained from the Black Women's Health Study (BWHS), a follow-up study of 59,000 US black women.
The human subjects' protocol for this study was approved by the Boston University Medical Center Institutional Review Board. The BWHS began in 1995 when 59,001 US black women enrolled through postal health questionnaires which were sent to subscribers of Essence magazine, members of selected black women's professional organizations, and friends and relatives of early respondents (34); 93.6% of BWHS participants recruited were Essence subscribers. Participants indicated their informed consent by completing the questionnaires. At baseline, subjects were 21–69 years of age (median, 38 years), 97% had completed high school, and 44% had completed college. Over 80% were from California, Georgia, Illinois, Indiana, Louisiana, Maryland, Massachusetts, Michigan, New Jersey, New York, South Carolina, Virginia, and the District of Columbia. Participants complete biennial questionnaires to update their health information. Follow-up has averaged 80% of the original cohort through 7 questionnaire cycles.
On the 1995 baseline questionnaire, BWHS participants provided data on demographic factors, reproductive and medical history, smoking and alcohol use, physical activity, anthropometric measures (e.g., height, weight, waist circumference), use of selected medications such as oral contraceptives and female hormone supplements, diet, and use of medical care.
On the 1995 baseline questionnaire, BWHS participants were asked whether a physician had ever told them that they had any of a list of medical conditions. The list of diagnoses did not specify sarcoidosis, but many women wrote it in under “other conditions.” The 1997 questionnaire and all subsequent follow-up questionnaires asked specifically about sarcoidosis.
Women who report a diagnosis of sarcoidosis are asked for permission to contact their physicians for information on diagnosis and treatment (35). The physicians are asked to complete an assessment questionnaire with questions about the study participant's diagnosis and treatment. Physicians who are unwilling to complete the questionnaire are asked for a copy of the patient's medical records pertaining to sarcoidosis. To date, the diagnosis of sarcoidosis has been confirmed for 96% of the 148 cases for whom physician questionnaires or medical records were obtained. Based on the high level of agreement between self-reports and physician reports/records, all women who reported incident sarcoidosis on a BWHS questionnaire were included as cases of sarcoidosis unless the diagnosis was disconfirmed by medical record.
Data on exposure variables were collected at baseline in 1995 and every 2 years on follow-up questionnaires. Exposure variables included age at menarche, parity, age at first birth, age at last birth, menopausal status, age at menopause, oral contraceptive use, and female hormone use; data on all variables except age at menarche were updated in each questionnaire cycle. Lactation history was obtained on each questionnaire, with the exception of 1997. Attained educational level was assessed in 1995 and again in 2003. Women were classified as premenopausal if they reported that they still had their usual menstrual periods or if they were currently going through menopause (perimenopause). Women were classified as postmenopausal if they reported cessation of menstrual periods due to natural causes or due to surgery with removal of both ovaries. Age at menopause was considered to be unknown for women who had a hysterectomy with retention of one or both ovaries. Years since last birth were calculated as the difference between current age and age at last birth.
Prevalent cases of sarcoidosis (diagnosed before baseline in 1995) were excluded (n = 684), as were prevalent cases of cancer (n = 1,431), leaving a total of 56,886 women in the analytic sample. Follow-up was conducted from 1995 to 2009. Person-years were calculated from baseline to year of diagnosis of sarcoidosis, loss to follow-up, death, or the end of follow-up, whichever occurred first. Incidence rate ratios and 95% confidence intervals were estimated for each reproductive or hormonal factor by means of Cox proportional hazards models, using SAS statistical software, version 9.1 (36). We adjusted for age (years), education (≤12, 13–15, or ≥16 years), geographic region (Northeast, South, Midwest, or West), pack-years of smoking (never smoked, <5, 5–14, 15–24, or ≥25), body mass index (weight (kg)/height (m)2; <25, 25–29, or ≥30), and questionnaire cycle (1995–1997, 1997–1999, 1999–2001, 2001–2003, 2003–2005, 2005–2007, or 2007–2009). In the analyses of use of female hormone supplements, we additionally adjusted for age at menopause (<40, 40–44, 45–49, or ≥50 years). When we adjusted for household income and household size, results were almost identical. Thus, we did not include these variables in the final model. All exposure variables and covariates, with the exception of age at menarche, were treated as time-varying, with use of the Anderson Gill data structure (37). We performed tests for trend by entering individual variables into the model in their ordinal form.
In 1995, the median age of BWHS participants was 38 years (range, 21–69 years). Participants were approximately equally distributed in the Northeast, South, Midwest, and West, with 81% having completed education beyond high school. During follow-up from 1995 to 2009 (694,818 person-years), 452 incident cases of sarcoidosis were identified. The median age of the cases at diagnosis was 44 years (range, 22–77 years).
The relation of markers of endogenous estrogen exposure to sarcoidosis risk is presented in Table 1. There was no apparent association between age at menarche and sarcoidosis risk. Among menopausal women, relative to age at menopause <40 years, the incidence rate ratios (IRRs) for ages at menopause of 40–44, 45–49, and ≥50 years were 0.92 (95% confidence interval (CI): 0.51, 1.67), 0.65 (95% CI: 0.34, 1.22), and 0.60 (95% CI: 0.31, 1.15), respectively (P-trend = 0.03).
There was no association with parity (Table 2). Among parous women, the IRR was reduced for first birth at age 30 years or older (IRR = 0.52, 95% CI: 0.34, 0.82) relative to a first birth before age 20 years (P-trend = 0.02). The IRR decreased as time since last birth decreased: For less than 5 years since the last birth relative to 15 or more years since the last birth, the IRR was 0.56 (95% CI: 0.35, 0.91; P-trend = 0.01). When we considered both age at first birth and years since last birth in the same model, the IRRs were 0.60 (95% CI: 0.37, 0.97) for first birth at age 30 years or older relative to first birth before age 20 years (P-trend = 0.05) and 0.73 (95% CI: 0.43, 1.24) for less than 5 years since the last birth relative to 15 or more years since the last birth (P-trend = 0.15). The IRR for ever lactating relative to never lactating was 0.87 (95% CI: 0.69, 1.09), and there was no pattern according to duration of lactation (P-trend = 0.25). The associations of parity, age at first birth, years since last birth, and lactation with sarcoidosis risk were largely consistent across categories of age (<40 years, ≥40 years) and education (≤12 years, 13–15 years, ≥16 years) (data not shown).
Results for the relation of exogenous estrogen use to sarcoidosis risk are presented in Table 3. Oral contraceptive use was not associated with risk: The IRRs were close to 1.0 for ever use and for categories of duration and recency of use. For female hormone use, the IRR for ever use relative to never use was 1.20 (95% CI: 0.87, 1.64). The IRRs were elevated but not statistically significant for short-duration use and use that had begun at least 5 years previously.
We repeated the analyses using baseline exposure data only. The results were unchanged. We also restricted analyses to cases that were confirmed by physician questionnaire or medical record. The results were closely similar to those presented.
Despite the fact that African-American women experience the highest incidence of sarcoidosis in the United States (2, 3), few studies have specifically focused on disease etiology in this population. In the largest case-control study of sarcoidosis yet conducted, A Case Control Etiologic Study of Sarcoidosis (ACCESS), approximately one-third of 736 cases are black women (38, 39). The ACCESS investigators have explored environmental (40) and genetic (41) risk factors for the disease but have not reported on reproductive factors. To our knowledge, our follow-up study is the first study to assess reproductive and hormonal factors in relation to incidence of sarcoidosis.
The strongest finding to emerge from our study was a reduced incidence of sarcoidosis associated with a later age at first birth; there was also an association, although weaker, of reduced incidence with having had a recent birth. We do not know the explanation for these observations, but our analysis of reproductive and hormonal factors in relation to sarcoidosis was prompted in part by the observation that hormonal changes experienced during pregnancy may influence the onset and course of the disease (12, 16, 42). Since the 1940s, numerous case reports and case series from the United States and Europe have described effects of pregnancy on the course of sarcoidosis (12–18, 21, 42–44). The majority of reports observed that most patients either experienced no clinical change in disease severity (13–18, 43) or experienced an improvement of symptoms (13–17, 42, 44), with a minority of patients undergoing a worsening or progression of the disease during the prenatal period (12–16, 18, 42–44). The favorable effect of pregnancy has been hypothesized to result from elevated levels of circulating corticosteroids during pregnancy (18, 45). While these previous reports concerned the clinical course of disease among women already diagnosed with sarcoidosis, the focus of the current study was on pregnancy-related risk factors in relation to incidence of the disease.
Some studies have suggested that sarcoidosis incidence is greater in women than in men (1, 2, 4), whereas large population-based incidence studies in the United States and Scandinavia have found similar rates in men and women (7–9). Medical surveillance could have contributed to the findings of a higher incidence in women than in men (1, 5–7). Interestingly, studies conducted in Japan (10) and Scandinavia (8, 9) have indicated a second peak of disease incidence in women aged 50 years or older, in addition to the usual peak in incidence before age 40 years. In the present study, one marker of endogenous hormone exposure, age at menarche, was unrelated to risk of sarcoidosis. For another marker, age at menopause, risk decreased as age at menopause increased. The latter association provides some support for a protective effect of endogenous hormones against the occurrence of sarcoidosis. However, the analyses of age at menopause were based on small numbers. Misclassification of this variable is expected to be random and thus drive estimates for extreme categories towards the null.
Animal and clinical data support a possible inverse association of supplemental hormone therapy with sarcoidosis disease activity. In addition to the beneficial effect of exogenous estrogen in rats reported by Shirai et al. (11), Chida et al. (46) described the resolution of disease in a menopausal woman following the initiation of conjugated estrogen and medroxyprogesterone. We observed no significant association of female hormone supplement use with risk of sarcoidosis. Some of the IRRs exceeded 1.0, but none were statistically significant. We also observed no association with use of oral contraceptives.
Because of the prospective design of the present study, markers of estrogen exposure were reported prior to diagnoses of sarcoidosis, minimizing recall bias and establishing the temporal sequence between exposure and outcome. We were able to analyze multiple reproductive and hormone-related variables which served as markers, or proxies, for endogenous and exogenous estrogen exposure; blood hormone levels were not available. These variables have been assessed in the BWHS in relation to risks of breast cancer (34, 47, 48) and uterine fibroids (49), and expected associations were observed. Specifically, parity was inversely associated with breast cancer in older women and in women who had breastfed (34, 48), while higher risk was associated with recent and long-term use of female hormones (47). Age at menarche, parity, age at first birth, and recent birth were all inversely associated with the risk of uterine fibroids (49).
Important potential confounding factors, including age, geographic region, educational attainment, smoking, and body mass index, were taken into account in the analysis. Follow-up rates within the cohort were high, reducing the likelihood of differential losses related to both exposure and outcome. Sarcoidosis cases were identified by self-report, since it is infeasible in large observational studies to examine all participants for disease. However, our validation effort in a subset of women showed a high degree of accuracy of self-reporting (35). There is evidence that sarcoidosis has an infectious etiology (4, 50–53). We had no information on infections in BWHS participants.
BWHS participants are not a random sample of US black women. Participants in the BWHS need to be sufficiently literate to fill out detailed, scannable health questionnaires; thus, study participants underrepresent the 15% of black women nationally of the same ages who did not graduate from high school (54). On the other hand, study participants are representative of most US black women and are residents of all regions of the United States.
In summary, the present study provides suggestive evidence that late age at first birth and a longer cumulative exposure to endogenous female hormones may be associated with a reduced incidence of sarcoidosis. Further studies of reproductive and hormonal factors in relation to risk of sarcoidosis are needed to confirm or negate these results.
Author affiliations: Slone Epidemiology Center, Boston University, Boston, Massachusetts (Yvette C. Cozier, Julie R. Palmer, Deborah A. Boggs, Lauren A. Wise, Lynn Rosenberg); and The Pulmonary Center, Boston University School of Medicine, Boston, Massachusetts (Jeffrey S. Berman).
This work was supported by grant K01HL088709 from the National Heart, Lung, and Blood Institute and grant CA058420 from the Division of Cancer Control and Population Science, National Cancer Institute.
The authors thank Dr. Rie Adser Virkus of Hillerød Hospital (Hillerød, Denmark) for her helpful comments.
The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Conflict of interest: none declared.