Introduction

Sarcoidosis is an inflammatory disease of unknown etiology which is characterized by the development of granulomas in any organ, typically in the lungs [1]. Its onset and progression vary; some patients experience acute episodes with spontaneous resolution [2, 3] while others experience a prolonged course resulting in fibrosis and organ function decline [4]. These severe manifestations can result in adverse outcomes including mortality [5], infection [6], heart failure and myocardial infarction [7]. Sarcoidosis can also impact the ability to work, reducing income and quality of life [8]. Treatment primarily involves systemic corticosteroids, typically prescribed for those with debilitating symptoms or signs of disease progression [9].

The incidence of sarcoidosis peaks in women between 50 and 60 years old, estimated to be 15.4 cases per 100,000 per year in Sweden [10]. This age range corresponds to the time of menopause, where there is a significant decrease in estrogen levels. This suggests that certain female hormones, mainly estrogen, may delay sarcoidosis onset in women by improving the aberration of the T-helper (Th)1/Th2 balance in immune response.

The theory that estrogen affects sarcoidosis risk is supported by two previous studies which found that reproductive indicators for endogenous estrogen (produced/synthesized within the body) are protective for sarcoidosis [11, 12]. However, these two studies reported an increased risk associated with exogenous estrogen, in particular menopausal hormone therapy (MHT). MHT is used to mitigate discomfort caused by decreased levels of circulating estrogen and progesterone after menopause [13] and consists of either estrogen alone, or combined estrogen and progestogen (progesterone or synthetic progesterone-like substance called progestin) [14]. The two previous studies on MHT and sarcoidosis, however, were limited due to low power which could have resulted in inadequate statistical precision and reduced ability to detect associations. Moreover, they relied on self-reported MHT use and did not have information on specific types of MHT so the explicit effect of estrogen on sarcoidosis risk was not possible to disentangle. In addition, their results may not be applicable to today since the characteristics of the population receiving MHT and the prescription patterns have changed since 2002 after reports about increased risk of cardiovascular disease and breast cancer [15,16,17,18]. Specifically, there has been a shift towards more cautious prescribing practices and a preference for lower doses and shorter durations of treatment. It thus remains unclear whether exposure to MHT is associated with sarcoidosis and whether the risk varies between estrogen alone and combined estrogen-progestogen.

We performed a nested case–control study using information derived from well-established nationwide Swedish population-based registers. Our aim was to investigate whether MHT is associated with risk of developing sarcoidosis in women and whether this risk varies by MHT type, route of administration and duration of use.

Methods

The data used in this study are part of a larger register linkage which was designed to conduct matched case–control studies to examine risk factors and matched cohort studies to examine outcomes. This register linkage design is an efficient way to conduct multiple investigations of the causes and consequences of sarcoidosis using the same dataset. In this study, we used these data to conduct a case–control study nested within the Swedish population.

Sarcoidosis cases and general population controls

We identified women with at least two inpatient or outpatient International Classification of Diseases (ICD)-coded visits for sarcoidosis (ICD-8/9 135, ICD-10 D86) in the National Patient Register (NPR; inpatient hospitalizations nationwide since 1987 and visits to outpatient clinics since 2001). A validation study showed that two ICD-coded visits for sarcoidosis in the NPR has a high positive predictive value (0.94) [19]. Women were required to have their first ever visit occurring in 2007–2020, allowing for at least 1.5 years of prescription data before sarcoidosis diagnosis since the Prescribed Drug Register (PDR) became available in July 2005. The PDR captures all prescription dispensations in pharmacies across Sweden since July 2005 including the date and route of administration [20].

Women receiving treatment for sarcoidosis at the time of diagnosis were considered having a more severe sarcoidosis (e.g. debilitating symptoms or organ involvement in need of treatment) [1]. In Sweden, the mainstay of sarcoidosis treatment is systemic corticosteroids and second-line treatment is methotrexate or azathioprine [1]. We classified women as receiving treatment at diagnosis if they had ≥ 1 prescription of either systemic corticosteroids [Anatomic Therapeutic Chemical (ATC) classification codes: H02AB01/02/04/06/07], methotrexate (L01BA01/L04AX03) or azathioprine (L04AX01) in the PDR ± 3 months from their first visit in the NPR listing sarcoidosis. For a subset of cases (n = 108) who were registered in a clinical cohort at Karolinska University Hospital in Stockholm, we retrieved information on sarcoidosis phenotype, i.e. Löfgren or non-Löfgren syndrome.

Women from the general population without sarcoidosis served as controls. Up to 10 population controls without any sarcoidosis visits in the NPR were randomly sampled from the Total Population Register (TPR) and were matched to each unique woman with sarcoidosis on year of birth and residential location at time of sarcoidosis, and required to be living in Sweden at the time the matched case was first identified with sarcoidosis (matching date).

The study population was restricted to women 40 years or older to capture women of menopausal age. Women with a hematologic or lung malignancy diagnosis (ICD-7 162, 163, 200‒205) in the Swedish Cancer Register within 6 months before or after the first visit for sarcoidosis/matching were excluded to avoid including cases where cancer may have been misdiagnosed as sarcoidosis. Additionally, women with a diagnosis of breast cancer, endometrial cancer, cardiovascular disease (stroke, acute myocardial infarction, ischemic heart disease), venous thromboembolism or anticoagulant dispensation before the first visit for sarcoidosis/matching were excluded because they may have a contraindication for MHT use (see flow chart of study population in supplementary Fig. 1 and a list of ICD and ATC codes used for exclusions in supplementary Table 1).

Exposure: menopausal hormone therapy

To identify history of MHT use, cases and controls were linked to the PDR using each person’s unique identification number. Information on dispensations of MHT [ATC codes: G03C (estrogens), G03D (progestogen, if prescribed in combination with estrogens), G03F (estrogen combined progestogen), G03CX (tibolone; used as an alternative for continuous combined estrogen-progesterone hormone therapy which has androgenic properties); supplementary Table 2] was obtained from the PDR before sarcoidosis diagnosis/matching. Both systemic and local treatments were identified using the pharmaceutical form variable from PDR. Drugs for systemic MHT were defined as oral and transdermal products (i.e. oral tablets, dermal patches and dermal gel) and drugs for local MHT were defined as vaginal products (i.e. vaginal creams, rings and pessaries). A detailed description of the exposure variables is presented in Table1.

Table 1 Description of variables used to investigate menopausal hormone therapy (MHT) obtained from the Prescribed Drug Register
Full size table

People with undiagnosed/preclinical sarcoidosis might experience symptoms that mimic menopause, and receive MHT treatment for those symptoms. To mitigate this potential reverse causation, women whose first dispensation occurred within one year before the sarcoidosis diagnosis/matching were not considered exposed in main analyses.

Other variables

We retrieved demographic information from the TPR including the date of birth, country of birth (Nordic, non-Nordic, missing), and county of residence at diagnosis/matching (classified into healthcare regions: Stockholm, Uppsala-Örebro, West, South, Southeast, and North). From the Longitudinal Integration Database for Health Insurance and Labour Market Studies, we obtained data on education level at time of sarcoidosis diagnosis/matching (≤ 9, 10‒12, ≥ 13 years, missing), gross income in 2005 adjusted to 2019 inflation rate [21] (< 100, 100–< 300, ≥ 300 thousand Swedish krona, missing), and sick leave/disability pension during the year 2005 (0, 1‒49, 50‒199, 200‒364, ≥ 365 days, missing). Sick leave, disability and income were collected from 2005, to assure they were from before exposure and outcome, and not later in time when they could be mediators. When using data on sick leave, disability and income from the year before diagnosis, ORs were within ± 0.02 of the ORs using 2005 data. Women who use MHT may be systematically different in terms of socioeconomic and/or health status than women who do not use MHT. Therefore, we collected information on education and income as proxies for socioeconomic status, and number of days of sick leave/disability pension as a proxy for health status. From the Medical Birth Register, we retrieved data on number of births before sarcoidosis diagnosis/matching. It has been found that number of childbirths is associated with age at menopause [22] and to also be associated with sarcoidosis [11, 12]. A family history of sarcoidosis is the strongest risk factor for sarcoidosis, and is a proxy for genetic risk [23]. We therefore searched for biological first-degree relatives (parents, full siblings and offspring) of cases and controls in the Multi-Generation Register and identified those with at least two sarcoidosis diagnoses in the NPR (family history of sarcoidosis – yes; no) at any point in time.

Statistical analysis

Characteristics of sarcoidosis cases and general population controls were reported as means with standard deviations, or as proportions. Conditional logistic regression models were used to estimate adjusted odds ratios with 95% confidence intervals (aOR; 95% CI) for the associations between MHT use, type of formulation, route of administration, and duration of use with incident sarcoidosis. MHT type and route of administration were considered together but since some women with systemic estrogen and with combined estrogen-progestogen had a history of local estrogen, we also considered the route of administration alone. We adjusted for age, education, income, sick leave/disability pension, number of births and family history of sarcoidosis.

A subgroup analysis by age at diagnosis/matching (< 60; 60–69; ≥70 years) was conducted to assess whether the association between MHT and sarcoidosis varies by age.

To address the heterogeneity of sarcoidosis, we investigated the association separately for treated and untreated sarcoidosis. Furthermore, to investigate whether misclassification of our register-based definition for sarcoidosis affected our results, we restricted to cases in the Karolinska clinical cohort who have medical record-confirmed diagnoses. Moreover, we examined Löfgren and non-Löfgren syndrome separately using data from the clinical cohort.

We stratified by time from first MHT dispensation to sarcoidosis diagnosis or matching (0 to 7 years) to investigate whether the OR varied by the time since MHT dispensation.

Three separate sensitivity analyses investigating potential misclassification of MHT were conducted. (1) Since the PDR was established in July 2005, we had incomplete information on the first dispensation date (left censoring) so to see if this affected results we included women who had their first ever visit occurring in 2010–2020, allowing for at least 4.5 years of PDR data, (2) we used a stricter definition for MHT, requiring at least two dispensations for MHT in the PDR, and (3) since tibolone has not only estrogenic and progestogenic properties, but also androgenic, we excluded women who received tibolone before sarcoidosis diagnosis/matching.

Since MHT is not used only for menopausal symptoms in a small group of women, we excluded non-menopause indications for MHT to test if those indications affected results (a list of ICD and ATC codes used for exclusions in supplementary Table 3).

We tested the robustness of the results against the potential unmeasured confounding of smoking and obesity using probabilistic bias analysis [24, 25]. Smoking and obesity have been found to be associated with both sarcoidosis [26,27,28,29] and women’s sex hormones [30, 31]. The assumptions for the analysis are described in the supplementary methods.

The relative risk (RR) can be computed from the OR as RR = (OR)/[(1 − P) + (P * OR)], where P is the prevalence in the unexposed [10]. With a sarcoidosis prevalence of 160/100,000, an OR of e.g. 1.25 corresponds to RR = 1.20, i.e. practically the same as the OR. We therefore refer to higher/increased odds as higher/increased risk.

Data management and statistical analyses were performed using SAS software (version 9.4; SAS institute Inc., Cary, NC, USA). Forest plots were performed using STATA software (version 16.1).

Results

We included 2593 newly diagnosed sarcoidosis cases and 20,003 matched general population controls (supplementary Fig. 1). Cases and controls were on average 58 years old (SD ± 11.0), and were comparable with respect to country of birth, education, and number of births (Table 2). Compared to controls, cases were more likely to have income < 100,000 SEK (40.2% vs. 35.0%), have ≥ 365 days on sick leave and disability pension (10.5% vs. 9.0%) and have a family history of sarcoidosis (3.1% vs. 1.1%).

Table 2 Characteristics of sarcoidosis cases (N = 2593) and general population controls (N = 20,003), 2007–2020
Full size table

A larger percentage of cases had a history of MHT use (28.9% vs. 24.2%), and used local estrogen only (15.4% vs. 12.6%) compared to controls (Fig. 1). The average duration of MHT use was 8.1 months (SD ± 19.8) for cases and 6.8 months (SD ± 18.4) for controls.

Compared with never use, ever MHT use was associated with a 25% increased risk of sarcoidosis (aOR 1.25, 95% CI 1.13–1.38; Fig. 1). When type of MHT and route of administration were considered together, systemic estrogen was associated with the highest risk of sarcoidosis (aOR 1.51, 95% CI 1.23–1.85), followed by local estrogen (aOR 1.25, 95% CI 1.11–1.42), and systemic estrogen-progestogen combined (aOR 1.12, 95% CI 0.96–1.31). Women who received both systemic and local treatments had the highest risk of sarcoidosis (aOR 1.47, 95% CI 1.19–1.81), followed by local only (aOR 1.25, 95% CI 1.11–1.42), and systemic only treatment (aOR 1.15, 95% CI 0.99–1.34). The aOR associated with MHT < 12 months, and those exposed ≥ 12 months were 1.31 (95% CI 1.14–1.52) and 1.21 (95% CI 1.07–1.36), respectively (Fig. 1). A 3-month and a 1-year increase in MHT duration was associated with a 1% and a 3% higher risk of sarcoidosis, respectively (aOR 1.01, 95% CI 1.00–1.01 and aOR 1.03, 95% CI 1.00–1.06).

Fig. 1
figure 1

Association between menopausal hormone therapy and sarcoidosis in a nested case–control study in Sweden of 2593 cases and 20,003 controls, 2007–2020. MHT, menopausal hormone therapy; OR, odds ratio; CI, confidence interval. aOdds ratios from conditional logistic regression models adjusted for age, education, income, sick leave/disability pension, number of births and family history of sarcoidosis. bSystemic administration is defined as oral and transdermal products (i.e. oral tablets, dermal patches and dermal gel) and local as vaginal products (i.e. vaginal creams, rings and pessaries)

Full size image

The estimates were similar across age groups (supplementary Table 4). However, the aORs from the main analysis were slightly attenuated for individuals who received sarcoidosis-related treatment around the time of diagnosis (1.14, 95% CI 0.98‒1.33) but not for untreated (1.33, 95% CI 1.16‒1.51; supplementary Table 5). Restricting to cases from the Karolinska clinical cohort yielded a higher OR (aOR 1.56, 95% CI 0.92‒2.63) and this did not differ greatly by Löfgren vs. non-Löfgren syndrome (supplementary Table 6). Stratifying by time from MHT dispensation to sarcoidosis diagnosis or matching, the aOR of sarcoidosis did not materially change (supplementary Table 7‒8). In addition, sensitivity analyses investigating potential misclassification of MHT (among women who had their first ever visit in 2010–2020, defining MHT use as ≥ 2 dispensations, and excluding women who received tibolone; supplementary Table 9) yielded similar results. The estimates did not change considerably after excluding non-menopause indications for MHT (supplementary Table 10). Last, when accounting for unmeasured confounding by smoking, the estimates were slightly higher (OR 1.31, 95% simulation interval 1.27–1.39), and by obesity were almost the same (OR 1.28, 95% simulation interval 1.24–1.39; supplementary Table 11).

Discussion

In this large nationwide register-based study, we found a 25% increased risk of sarcoidosis associated with a history of MHT use. An increased risk was also observed with different MHT types and route of administration, with women receiving estrogen administered systemically having the highest risk (a 51% increased risk).

Two previous studies found similar results [11, 12]. Specifically, a cohort study using data from the Black Women’s Health Study (BWHS) found that ever MHT use was associated with a 20% increased risk of sarcoidosis [11]. Similarly, a nested case–control study using data from the Northern Sweden Health and Disease Study (NSHDS) showed a 40% higher risk with ever MHT use compared to never use [12]. However, these two studies showed that exposure to endogenous female hormones reduces the risk of sarcoidosis. This indicates that endogenous estrogens may protect against the occurrence of the disease but not exogenous estrogens such as MHT. A similar pattern has been observed with cardiovascular diseases where endogenous estrogen has been found to be protective while this might not be true for estrogen therapy [32]. Recent studies have shown the effect of MHT on cardiovascular disease differs according to timing of initiation, with more beneficial effects when initiated in women < 60 years of age and/or < 10 years after menopause [33] and less favourable effects when initiated in older women and/or > 10 years after menopause [34]. Therefore, the timing of MHT initiation may be an important factor in sarcoidosis too, although we did not find any differences in risk by age group. Future studies should investigate both age at MHT initiation and time since menopause. MHT has been also associated with an increased risk of other immune-mediated inflammatory diseases such as systemic lupus erythematosus [35] and ulcerative colitis [36], indicating that MHT may play a role in the dysregulation of the immune response.

The exact mechanism by which MHT influences sarcoidosis is not known, since there is still an unresolved paradox with respect to the immunomodulating role of estrogens. On one side, estrogens have demonstrated anti-inflammatory activity by inhibiting many pro-inflammatory pathways of innate immunity, adaptive immunity, and inflammatory tissue responses (inhibit the production of Th1 pro-inflammatory cytokines [tumor necrosis factor-α, interferon-γ and interleukin (IL)-2], while they stimulate the production of Th2 anti-inflammatory cytokines [IL-4, IL-10] [37, 38]). On the other side, pro-inflammatory responses have also been shown, including anti-apoptotic effects on immune cells, promotion of neoangiogenesis, and stimulation of B cells [39]. Moreover, estrogens have been shown to activate the mechanistic target of rapamycin (mTOR) and phosphatidylinositol 3-kinase (PI3K) pathways [40, 41], and recent research has suggested that these pathways may play a role in the development and progression of sarcoidosis [42, 43].

The exact mechanism behind the lower risk of sarcoidosis with combined estrogen-progesterone MHT than with only estrogen therapy is not fully understood. However, it is thought that progesterone may play a protective role in reducing the risk of sarcoidosis. Progesterone has anti-inflammatory and immunomodulatory effects [44], which may counteract some of the pro-inflammatory effects of estrogen. Additionally, progesterone has been shown to inhibit the mTOR pathway [45], which has been implicated in the development of sarcoidosis [42, 43].

We observed a 25% increased risk of sarcoidosis associated with local estrogen MHT treatment and we theorized that it could be due to reverse causation. That is, sarcoidosis patients are more likely to receive the diagnosis of urinary tract infections (UTI) which might be partly due to preclinical (asymptomatic) sarcoidosis before diagnosis [46]. Local estrogen administration is more likely to be prescribed to women with a history of UTIs because it reduces the risk of recurrent UTIs [47]. Therefore, we conducted a post-hoc analysis excluding women with a history of UTIs but the OR did not differ greatly from the main analysis (aOR 1.26, 95% CI 1.10–1.43), which does not support our theory. Estrogens administered vaginally can be absorbed into the bloodstream and may have systemic effects [48]. The lower risk associated with local estrogen compared to systemic estrogen (OR 1.25 vs. 1.51, respectively) is consistent with a lower potency of local estrogen administration.

There are several limitations to our study. Information on menopause was not available from nationwide registers, however, we restricted to women 40 years or older to capture women of menopausal age. We did not have information if women received MHT due to menopausal symptoms or other indications not related to menopause. Nevertheless, our estimates remained robust when we excluded non-menopause indications for MHT. There may be some sarcoidosis misclassification, since detailed clinical information in the NPR is not available and sarcoidosis was identified using ICD-coded visits. However, ICD codes for sarcoidosis in the NPR have been shown to have a high positive predictive value [19]. In addition, the odds ratios remained the same (slightly higher) when we restricted to cases in the Karolinska clinical cohort who have medical record-confirmed diagnoses. We only have information on the use of MHT starting in 2005, when data on dispensed medications became available from the PDR. We believe, however, that exposure misclassification is minimal since our estimates remained robust when we included women who had their first ever visit occurring in 2010–2020. Furthermore, when a stricter definition for MHT was used requiring at least two dispensations, our results were similar. There is the possibility of reverse causation bias induced by preclinical sarcoidosis, meaning that symptoms of preclinical sarcoidosis might be mistaken for symptoms of menopause. If this were the case, one would expect that the OR would be higher in the years closest to diagnosis, however, the OR did not vary with time since MHT dispensation (0 to 7 years). Last, we cannot entirely preclude the possibility of unmeasured confounding due to smoking and obesity. However, probabilistic bias analysis showed that these factors have a minimal effect on our results and that we may be underestimating the effect due to the negative confounding of smoking.

A major strength of this study is the use of prospectively collected high-quality population-based data. Using ICD codes for sarcoidosis, which have high validity in the NPR, the study was sufficiently powered to obtain robust inferences. All Swedish residents have universal access to healthcare and we addressed the bias due to socioeconomic and health status by adjusting for education, income and days of sick leave and disability pension. Moreover, after a number of sensitivity analyses examining a range of potential biases, our results did not change. Our study also benefited from detailed information on MHT dispensations from the entire population and considered different types of formulation, and route of administration.

While our study identified a higher risk of sarcoidosis associated with a history of MHT use, the overall decision to initiate MHT should be made on a case-by-case basis. MHT offers significant benefits in alleviating menopause-related hormone deficits, such as relief from vasomotor symptoms, bone density preservation [13], and potentially reduced cardiovascular risks [33]. Moreover, it is essential to understand that the proportion of sarcoidosis cases among women using MHT that can be attributed to MHT is low (attributable proportion 20%). Consequently, the health benefits that MHT can provide to many menopausal women should not be overlooked. However, treatment should be individualized considering various factors, such as a woman’s overall health, medical history, and specific menopausal symptoms.

Interpretations from this study may only be generalizable to older onset sarcoidosis, since sarcoidosis diagnosed at a younger age may differ in terms of etiology (e.g. pathogenetic factors). Lastly, the generalizability of our results may be limited to Northern European ancestry women if the effect of MHT on sarcoidosis risk is different from other ethnic groups. For example, some studies have suggested that estrogen levels are higher in black women and lower in Asian women compared to white women [49,50,51,52]. However, given that our results were similar to those reported by the Black Women’s Health Study, we do not think that the external validity of our study is limited.

Conclusions

Our findings suggest that a history of MHT use is associated with higher risk of sarcoidosis, and women receiving estrogen administered systemically have the highest risk.