Abstract
Little is known about the prevalence, clinical characteristics and impact of hypothyroidism in patients with sarcoidosis. We aimed to determine the prevalence and clinical features of hypothyroidism and its relation to organ involvement and other clinical manifestations in patients with sarcoidosis.
We conducted a national registry-based study investigating 3835 respondents to the Sarcoidosis Advanced Registry for Cures Questionnaire between June 2014 and August 2019. This registry is based on a self-reported, web-based questionnaire that provides data related to demographics, diagnostics, sarcoidosis manifestations and treatment. We compared sarcoidosis patients with and without self-reported hypothyroidism. We used multivariable logistic regression and adjusted for potential confounders to determine the association of hypothyroidism with nonorgan-specific manifestations.
14% of the sarcoidosis patients self-reported hypothyroidism and were generally middle-aged white women. Hypothyroid patients had more comorbid conditions and were more likely to have multiorgan sarcoidosis involvement, especially with cutaneous, ocular, joints, liver and lacrimal gland involvement. Self-reported hypothyroidism was associated with depression (adjusted odds ratio (aOR) 1.3, 95% CI 1.01–1.6), antidepressant use (aOR 1.3, 95% CI 1.1–1.7), obesity (aOR 1.7, 95% CI 1.4–2.1), sleep apnoea (aOR 1.7, 95% CI 1.3–2.2), chronic fatigue syndrome (aOR 1.5, 95% CI 1.2–2) and was borderline associated with fibromyalgia (aOR 1.3, 95% CI 1–1.8). Physical impairment was more common in patients with hypothyroidism.
Hypothyroidism is a frequent comorbidity in sarcoidosis patients that might be a potentially reversible contributor to fatigue, depression and physical impairment in this population. We recommend considering routine screening for hypothyroidism in sarcoidosis patients especially in those with multiorgan sarcoidosis, fatigue and depression.
Abstract
Hypothyroidism is a frequent comorbidity in sarcoidosis patients, is often associated with sarcoidosis multiorgan involvement, and might be a potentially reversible contributor to fatigue, depression and physical impairment in this population https://bit.ly/2XGdjKg
Introduction
Sarcoidosis can pose a devastating impact on healthcare-related quality of life (HRQoL) and can have significant manifestations that appear to be nonorgan specific, such as fatigue and depression [1]. Due to multiorgan involvement and the frequent comorbidities associated with sarcoidosis and its treatment, it can be challenging to determine whether such manifestations are related to sarcoidosis or another comorbid condition [2].
Autoimmune diseases are known to be associated with sarcoidosis; some have theorised that sarcoidosis itself may be an autoimmune process [3, 4]. This link with autoimmune conditions perhaps explains the association previously described between sarcoidosis and autoimmune thyroid disorders, albeit in small studies or case series [5]. The exact prevalence and characteristics of hypothyroidism in sarcoidosis is unknown [5–7]. A recent nationwide Taiwanese study found that 12% of sarcoidosis patients had an autoimmune thyroid disorder [3].
Hypothyroidism has been linked to depression [8, 9], obesity [10], sleep apnoea [11], chronic fatigue [12] and fibromyalgia [13] in the general population. Many patients with sarcoidosis suffer from these same problems that are thought to be nonorgan specific. Interestingly, in a study evaluating idiopathic pulmonary fibrosis patients, another type of interstitial lung disease, 17% had hypothyroidism and it was found to be a predictor of mortality [14]. However, the contribution of hypothyroidism to the disease manifestations, morbidity and mortality in sarcoidosis is unknown. Understanding the links between hypothyroidism and sarcoidosis may lead to clearer understanding of the pathogenesis of sarcoidosis. It may also help clinicians to understand the factors affecting HRQoL in sarcoidosis.
In this study, we sought to determine the prevalence of hypothyroidism in a self-reported nationwide sarcoidosis registry from the United States. We also aimed to characterise the clinical features and cluster of organ involvement in sarcoidosis patients who also have hypothyroidism and to study the possible association of self-reported hypothyroidism with nonorgan-specific manifestations such as depression and fatigue.
Methods
Study population
We conducted a national registry-based study investigating 3835 respondents to the Foundation for Sarcoidosis Research (FSR) Sarcoidosis Advanced Registry for Cures Questionnaire (SARC) [15]. This registry from FSR was open to all English-speaking patients self-identifying as having sarcoidosis via a web-based questionnaire that includes 72 questions. Respondents were either directly recruited through their treating physicians, by the FSR or via national and international organisations. It provides an observational cohort platform to collect cross-sectional and longitudinal self-reported data related to demographics, diagnostics, organ involvement, treatment and the physical and psychosocial impact of sarcoidosis. We included patient surveys completed between June 2014 and August 2019. Respondents were able to update their surveys longitudinally with time. Data from the most recent survey for each respondent were used for our analysis. We excluded 13 patients who were reported to be deceased or had missing information regarding the question “is the patient living” bringing the final cohort to a total of 3822 respondents. The study was approved by the University of Florida Institutional Review Board (no. 201902211).
Objectives and data management
The primary objective of the study was to define the prevalence of self-reported hypothyroidism in patients with sarcoidosis. Secondary objectives were to determine the demographic and clinical characteristics of the sarcoidosis patients who reported hypothyroidism; and to study the association of hypothyroidism with the physical and nonorgan-specific manifestations of sarcoidosis. We divided the patients into two groups: those who self-reported hypothyroidism and those who did not report hypothyroidism based on a response to a question in the survey regarding hypothyroidism (supplementary table 1). We compared the following data between the two groups and reported it in three categories: 1) patient characteristics, diagnosis and treatment of sarcoidosis; 2) organ involvement and comorbidities developed after the diagnosis of sarcoidosis; and 3) data regarding the physical impairment and nonorgan-specific manifestations of the disease. The responses “unsure” or “prefer not to answer” were labelled as missing information. Only complete cases were analysed for each variable. In each table, we presented the nominator (positive cases) and denominator (complete cases) for each categorical variable. Data missingness was <2% for demographics, <1% for insurance status, 19.8% for education, 22.7% for income, 9.7% for hospitalisation history, <1% for steroid use history, 35.7% for cytotoxic agent use history, 47% for tumour necrosis factor (TNF) inhibitor use history and <2% for comorbid conditions. Missing data regarding organ involvement ranged from 11% for lung involvement to 36.8% for bone involvement. The question regarding endocrine and vitamin D disorders was a tick-box format so cases with empty boxes were regarded as negative (supplementary table 1).
Baseline demographics and clinical characteristics of sarcoidosis patients with hypothyroidism as compared with patients with no hypothyroidism#
Patients reported comorbidities that have newly developed after the diagnosis of sarcoidosis. We divided newly developed comorbidities into those probably related to sarcoidosis itself (cancer, chronic fatigue syndrome, chronic pain syndrome, congestive heart failure, depression, fibromyalgia, lymphoma and sleep disorders) and those probably related to the use of corticosteroids (diabetes mellitus, cataracts, glaucoma, obesity, hypertension, osteoporosis/osteopenia and sleep apnoea) [16]. Multiorgan sarcoidosis was defined by the reported presence of three or more organs affected by sarcoidosis [1, 17]. Systemic medications used to treat sarcoidosis were divided into four categories: 1) steroids (prednisone, methylprednisolone, dexamethasone); 2) cytotoxic agents (hydroxychloroquine, chloroquine, methotrexate, azathioprine, leflunomide, mycophenolate, cyclophosphamide); 3) TNF inhibitors (infliximab, adalimumab, certolizumab, golimumab, etanercept); and 4) other systemic therapies (rituximab, pentoxifylline, intravenous Ig, thalidomide, adrenocorticotropic hormone).
We studied the association between self-reported hypothyroidism and the following comorbidities: self-reported depression, chronic fatigue syndrome, fibromyalgia, sleep apnoea, sleep disorders and obesity. We also looked at whether hypothyroidism was associated with patient reports of feeling depressed always or often, using antidepressant medications and reporting feeling tired always or often. To investigate physical impairment and social impact, questions regarding use of any mobility assistive device, employment-based disability, missing more than seven work days in the past year and impact on job termination were analysed (survey question details in supplementary table 1).
Validation cohort
To validate our results, we queried the University of Florida integrated data repository from January 2011 until September 2020 and identified patients who carry the diagnosis of sarcoidosis with and without hypothyroidism and compared the presence and absence of depression, fatigue, fibromyalgia, obesity and obstructive sleep apnoea based on International Classification of Diseases (ICD) codes. The following ICD codes were used: sarcoidosis (ICD-9=135; ICD-10=D86); hypothyroidism (ICD-9=244; ICD-10=E03); depression (ICD-9=311, 296.2 and 296.3; ICD-10=F32, F33); fatigue (ICD-9=780.7 and ICD-10=R53); fibromyalgia (ICD-9=729.1 and ICD-10=M79.7); obesity (ICD-9=278, ICD-10=E65-E68) and obstructive sleep apnoea ((OSA) ICD-9=327.23 and ICD-10=G47.33).
Data analysis
The distribution of patient characteristics and outcomes were summarised as percentages for categorical variables and mean ±sd for continuous variables. To compare differences between groups, we used an independent sample t-test for continuous variables and Chi-squared test for categorical variables. A clustered bar chart was created to demonstrate the frequencies of other endocrine and vitamin D disorders and different skin manifestations between groups. To study the association between self-reported hypothyroidism and potential nonorgan-related and social manifestations in sarcoidosis patients, multivariable logistic regression was performed adjusted for sex, race, age at diagnosis, use of steroids, multiorgan involvement (as a binary variable) and whether the patient had other sarcoidosis-related comorbidities (as a binary variable). These confounders were selected on an a priori basis. We also adjusted for healthcare insurance status with the noted disparity between the two groups and its potential impact on the outcomes. Based on this multivariable logistic regression analysis, we constructed a forest plot to illustrate the odds ratio with 95% confidence intervals of the likelihood of nonorgan-related manifestations in sarcoidosis patients with and without self-reported hypothyroidism. Statistical analyses were performed using IBM SPSS Statistics for Windows, version 23.0 (released 2015, IBM Corp, Armonk, NY, USA).
Results
Demographics and baseline characteristics
Among the 3822 sarcoidosis patients identified for our analysis, 538 reported hypothyroidism with a prevalence of 14.1%. The patients with self-reported hypothyroidism were generally middle-aged and were slightly older than the nonhypothyroidism patients at the time of sarcoidosis diagnosis (45.1±12.5 versus 43.1±13.7 years, p=0.001). The hypothyroidism patients were predominantly white women (table 1). Subgroup analysis showed that the prevalence of hypothyroidism in all white females enrolled in the study was 20% (supplementary table 2). Patients with hypothyroidism were more frequently seen by endocrinologists and other specialised healthcare providers before the diagnosis of sarcoidosis was made. During the course of sarcoidosis, patients with hypothyroidism received systemic corticosteroids more frequently as well as chloroquines, pentoxifylline and intravenous Igs, when compared to patients without hypothyroidism. No significant difference was noted in the use of TNF inhibitors or other cytotoxic agents (table 1 and supplementary figure 1).
Sarcoidosis involvement of other organs as reported by sarcoidosis patients with hypothyroidism as compared to those with no hypothyroidism#
Clustered bar chart demonstrating the frequencies of other endocrine and vitamin D disorders in sarcoidosis patients with hypothyroidism as compared with those with no hypothyroidism.
Sarcoidosis organ involvement and comorbidities
Sarcoidosis patients with self-reported hypothyroidism also reported more multiorgan involvement (51% versus 44%, p<0.001). Other self-reported endocrine and vitamin D disorders are presented in figure 1. These included disorders related to corticosteroid use and calcium metabolism. Furthermore, hypothyroid patients had more skin, joint, eye, liver and lacrimal gland sarcoidosis-related manifestations (table 2). Although there was no difference in the prevalence of pulmonary sarcoidosis, respondents with self-reported hypothyroidism reported more use of home ventilators and noninvasive positive pressure ventilation machines (31% versus 24%, p=0.001; data not shown). The cutaneous sarcoidosis manifestations that were more common in the hypothyroid group included abnormal skin pigmentation, erythema nodosum, lupus pernio, papules, plaques and s.c. nodules around scars and tattoos (supplementary figure 2). Additionally, the hypothyroid group had more frequent comorbidities either related to sarcoidosis disease or to steroids use (table 2).
Forest plot showing the adjusted odds ratio of the likelihood of different nonorgan-specific conditions in patients with sarcoidosis who also have hypothyroidism versus those with no hypothyroidism. Odds ratio with confidence intervals are based on multivariable logistic regression analysis adjusted for sex, race, age at diagnosis, healthcare insurance status, use of steroids, multiorgan involvement and other sarcoidosis-related comorbidities (cancer, chronic pain syndrome, congestive heart failure and lymphoma).
Physical impairment and nonorgan-specific manifestations
On univariable analysis, when compared to patients with absence of hypothyroidism, patients with self-reported hypothyroidism were more likely to report depression, obesity, sleep disorders, sleep apnoea, chronic fatigue syndrome and fibromyalgia. They also reported using antidepressants, feeling depressed most of the time and feeling tired most of the time more frequently when compared to patients without hypothyroidism (table 3). Adjusting for baseline demographics and potential confounders including corticosteroids use in multivariable logistic regression analysis, self-reported hypothyroidism was associated with depression (adjusted odds ratio (aOR) 1.3, 95% CI 1.01–1.6), antidepressant use (aOR 1.3, 95% CI 1.1–1.7), obesity (aOR 1.7, 95% CI 1.4–2.1), sleep apnoea (aOR 1.7, 95% CI 1.3–2.2), chronic fatigue syndrome (aOR 1.5, 95% CI 1.2–2) and was borderline associated with fibromyalgia (aOR 1.3, 95% CI 1–1.8) (table 3 and figure 2). Furthermore, hypothyroidism was associated with employment-based disability, missing work days and disease-related job termination in univariable analysis but only with missing work days when adjusting for potential confounders (table 3).
Physical impairment and psychosocial impact of hypothyroidism in patients with sarcoidosis
Validation cohort results
Using the University of Florida database, we identified 2189 adult sarcoidosis patients, of which 1324 (60.5%) were female, 834 (38%) were African American and 328 (15%) also carried the diagnosis of hypothyroidism. Hypothyroidism patients were mainly female (75%) and white (65%). Compared to sarcoidosis patients without hypothyroidism, hypothyroidism patients significantly had more depression (OR 2.6, 95% CI 2–3.4), fatigue (OR 2.1, 95% CI 1.7–2.7), fibromyalgia (OR 2.8, 95% CI 2–3.9), obesity (OR 1.7, 95% CI 1.3–2.2) and OSA (OR 1.9, 95% CI 1.4–2.4) (supplementary table 3).
Discussion
To our knowledge, this is the largest study in the published literature to assess the prevalence, characteristics and potential impact of hypothyroidism in sarcoidosis patients [3, 6, 7, 18–24]. Our study shows that 14% of sarcoidosis patients reported hypothyroidism. These patients were mainly middle-aged white females. Hypothyroidism was more likely to be reported in sarcoidosis with multiorgan involvement, especially when involving the skin, joints, eyes, liver and/or the lacrimal glands. Furthermore, self-reported hypothyroidism was associated with depression, fatigue, obesity, sleep apnoea and physical impairment.
In this nationwide sarcoidosis cohort, the reported prevalence of self-reported hypothyroidism is remarkably higher than what is reported in the general population, which ranges between 3% and 7% in the United States [12]. Notably, sarcoidosis patients with hypothyroidism had a similar demographic profile as the general population of hypothyroidism patients [12]. Our prevalence is similar to what has been observed in other smaller sarcoidosis cohorts. Antonelli et al. [6] reported 17% prevalence of subclinical hypothyroidism and 5% prevalence of overt hypothyroidism in sarcoidosis patients from Italy. Malli et al. [7] reported a 15% prevalence in a cohort from Greece. In a nationwide case–control study from Taiwan looking at sarcoidosis comorbid conditions, the prevalence of autoimmune thyroid disease was around 12% [3].
Our study does not answer the question of whether hypothyroidism is caused by autoimmune thyroiditis, direct sarcoidosis infiltration of the thyroid gland, other causes or is multifactorial. Although direct evidence of autoimmunity as the cause of sarcoidosis is lacking, associations with autoimmune conditions, including thyroid disease, have been noted [3, 4, 24]. Enhanced T-cell immunity along with loss of T-cell tolerance to self-antigens and possible exogenous antigenic exposure have all been proposed to play a role in the pathophysiology of sarcoidosis [4, 25]. Additionally, the increase in circulating Igs and the possible response to B-cell-depleting therapy in refractory cases might also suggest a role of the humoral immune system [26]. Recently, sarcoidosis patients were found to have elevated levels of age-associated B-cells in their blood and bronchoalveolar lavage as compared to healthy subjects, supporting the role of B-cell immunity [27]. This autoimmune phenomenon in sarcoidosis might explain the higher prevalence of hypothyroidism noted in our sarcoidosis cohort as compared to the general population. Previous studies have also shown that sarcoidosis patients have higher levels of thyroid peroxidase antibodies [19] and/or thyroglobulin antibodies [19, 20] when compared to control subjects, which might have a pathophysiological role in the association between sarcoidosis and autoimmune thyroid disease [5]. Another possible mechanism of hypothyroidism in sarcoidosis is direct granulomatous thyroid infiltration, which is reported in around 4% in autopsy studies of systemic sarcoidosis patients [5, 28]. Various treatments for sarcoidosis, including systemic use of steroids or other immunosuppression treatments, can alter thyroid function tests [29]. High doses of systemic steroids can reduce thyroid-stimulating hormone levels but this effect is usually transient and reversible and not typically associated with clinical central hypothyroidism [29]. Other immune suppressing agents such as cyclophosphamide [30], infliximab [31] and chloroquine [32] have been reported to cause hypothyroidism. In our analysis, patients with self-reported hypothyroidism were more likely to have received corticosteroids, chloroquines, pentoxifylline and intravenous Ig when compared to sarcoidosis patients without hypothyroidism, which might suggest a potential contribution from medications in hypothyroidism, and or to treat associated organ involvement.
A novel finding in our study is the link that we noted between self-reported hypothyroidism and multiorgan involvement of sarcoidosis. It is unclear from our analysis whether the same mechanism underlying the development of multiorgan sarcoidosis also underlies the development of hypothyroidism; autoimmunity has been postulated as a mechanism for development of nonpulmonary disease although an exuberant antigen-driven immune response is often noted in multiorgan disease [33]. Other than the lungs and lymphatic system, the most common organ systems reported to be involved with sarcoidosis in the self-reported hypothyroid group were the skin, joints and eyes. The association with cutaneous sarcoidosis was previously reported by Anolik et al. [22] who linked the development of thyroid disorders with cutaneous sarcoidosis, although the underlying mechanisms of this association are not clear.
Nonorgan-specific symptoms and social and physical impairment related to sarcoidosis can have a large burden on the patients' HRQoL. Sarcoidosis-associated fatigue is a challenging problem that is common in sarcoidosis [2]. It is likely multifactorial in origin and often very difficult to treat [2]. Several studies evaluated pharmacological and nonpharmacological approaches to treat fatigue in sarcoidosis; however, none of these studies looked at hypothyroidism and fatigue in this patient population [34–39]. In one study of German patients evaluating excessive daytime sleepiness and fatigue predictors in sarcoidosis, Bosse-Henck et al. [40] found an increased prevalence of “thyroid disorders” in patients with fatigue and excessive daytime sleepiness. Interestingly, in a recently published article phenotyping sarcoidosis organ involvement, fatigue was linked to female sex in sarcoidosis and to cardiac, central nervous system and skin sarcoidosis involvement. However, hypothyroidism was not included in that model [41]. In our analysis, sarcoidosis patients with hypothyroidism were 1.5-times more likely to report fatigue when compared to sarcoidosis patients without hypothyroidism. Chronic fatigue was previously reported to be associated with other comorbidities, including depression, obesity, sleep apnoea and fibromyalgia [2, 40]. We found an association between self-reported hypothyroidism and all four conditions even after controlling for corticosteroid use, suggesting that fatigue in sarcoidosis is multifactorial and may also be linked to hypothyroidism.
We acknowledge several limitations. Despite the extensive recruitment efforts and the relatively large sample size, the participation is still limited by access, awareness and ability to take an online survey. It is clear that the data disproportionately represented white patients (80%) and women (73.6%), which limits the generalizability of our results. However, our reported prevalence of hypothyroidism is in the range with what has been reported in other smaller but more defined sarcoidosis cohorts. Furthermore, using the University of Florida database of sarcoidosis patients, including 40% men and 38% African American people, the prevalence of hypothyroidism was 15%, which was also associated with depression, fatigue, fibromyalgia, obesity and obstructive sleep apnoea, suggesting that our results may have generalizability despite these limitations. Another limitation is the recruitment process, given the platform in which the survey is administered, which leads to referral bias. Furthermore, patients who chose to participate in the survey may not necessarily represent the general sarcoidosis population as those with more severe disease were likely over-represented. Also, the self-reported nature of the data opens the door for nondifferential misclassification of exposure and outcomes. This may explain the higher rates of extrathoracic sarcoidosis in almost all organ systems when compared to published literature. Unfortunately, self-reported registry studies are commonly subject to this bias; however, the large participant number, statistical adjustment for demographics and potential confounders and the external validation that we performed may address this limitation to some degree. Finally, we cannot establish a causality relationship between hypothyroidism and the tested outcomes due to the self-reported observational nature of this study. However, our study highlights the importance of considering hypothyroidism as a potentially treatable factor among other factors contributing to the physical and nonorgan-related disability in sarcoidosis.
Conclusion
In conclusion, self-reported hypothyroidism is prevalent in sarcoidosis and may be an important potentially treatable contributor of disabling symptoms. It is often associated with multiorgan involvement. Based on our findings, routine screening for hypothyroidism should be considered in sarcoidosis patients, especially in those with multiorgan involvement, fatigue and depression. Further investigation is needed to determine common pathogenic pathways and to assess the impact of treatment of hypothyroidism has on the nonorgan-specific symptoms in sarcoidosis.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary material 00754-2020.SUPPLEMENT
Acknowledgment
The authors would like to thank the FSR for the use of FSR-SARC Patient Registry data to conduct this study.
Footnotes
This article has supplementary material available from openres.ersjournals.com.
Data availability: Data are available upon request through the FSR-SARC Registry Committee. You can contact the committee at datarequests{at}stopsarcoidosis.org. Restrictions on access to data are to ensure patient privacy for all persons in the FSR-SARC Registry.
Author contributions: B.N. Alzghoul: study conception and design, project administration, IRB protocol preparation, statistical analysis, initial draft writing and manuscript reviewing and editing; F.N. Amer: study conception and design, figures and tables preparation and manuscript reviewing and editing; A. Innabi and B. Alzghoul: manuscript reviewing and editing; M.T. Mardini and C. Bai: statistical analysis and manuscript reviewing and editing; T. Al-Hakim, N. Singh, M. Buchanan and L. Serchuck: registry design and enrollment, manuscript reviewing and editing; D. Barb, D. Gomez Manjarres, W.W. Woodmansee, and L.A. Maier: study conception and design and manuscript reviewing and editing; D.C. Patel: study conception and design, study administration, supervision and manuscript reviewing and editing. B.N. Alzghoul and D.C. Patel are the guarantors of the study. All authors reviewed and approved the final manuscript.
Conflict of interest: B.N. Alzghoul has nothing to disclose.
Conflict of interest: F.N. Amer has nothing to disclose.
Conflict of interest: D. Barb has nothing to disclose.
Conflict of interest: A. Innabi has nothing to disclose.
Conflict of interest: M.T. Mardini has nothing to disclose.
Conflict of interest: C. Bai has nothing to disclose.
Conflict of interest: B. Alzghoul has nothing to disclose.
Conflict of interest: T. Al-Hakim reports helping with the FSR-SARC registry design and enrollment.
Conflict of interest: N. Singh reports helping with the FSR-SARC registry design and enrollment.
Conflict of interest: M. Buchanan reports helping with the FSR-SARC registry design and enrollment.
Conflict of interest: L. Serchuck reports helping with the FSR-SARC registry design and enrollment.
Conflict of interest: D. Gomez Manjarres has nothing to disclose.
Conflict of interest: W.W. Woodmansee has nothing to disclose.
Conflict of interest: L.A. Maier reports being a nonpaid member of the Scientific Advisory Board of the FSR, and receives grant funding for research as a member of the FSR-Clinical Studies Network. She helped with the FSR-SARC registry design and enrolment.
Conflict of interest: D.C. Patel has nothing to disclose.
Support statement: This study received nonmonetary support from the FSR.
- Received October 16, 2020.
- Accepted December 24, 2020.
- Copyright ©The authors 2021
This version is distributed under the terms of the Creative Commons Attribution Non-Commercial Licence 4.0. For commercial reproduction rights and permissions contact permissions{at}ersnet.org
References
