Figures
Abstract
Sarcoidosis is a multisystem disease with an unpredictable and sometimes fatal course while the underlying pathomechanism is still unclear. Reasons of the increasing hospitalization rate and mortality in the United States remain in dispute but incriminated are a number of distinct comorbidities and risk factors as well as the application of more aggressive therapeutic agents. Studies reflecting the recent development in central Europe are lacking. Our aim was to investigate the recent mortality and hospitalization rates as well as the underlying comorbidities of hospitalized sarcoidosis patients in Switzerland. In this longitudinal, nested case-control study, a nation-wide database provided by the Swiss Federal Office for Statistics enclosing every hospital entry covering the years 2002–2012 (n = 15,627,573) was analyzed. There were 8,385 cases with a diagnosis of sarcoidosis representing 0.054% (8,385 / 15,627,573) of all hospitalizations in Switzerland. These cases were compared with age- and sex-matched controls without the diagnosis of sarcoidosis. Hospitalization and mortality rates in Switzerland remained stable over the observed time period. Comorbidity analysis revealed that sarcoidosis patients had significantly higher medication-related comorbidities compared to matched controls, probably due to systemic corticosteroids and immunosuppressive therapy. Sarcoidosis patients were also more frequently re-hospitalized (median annual hospitalization rate 0.28 [IQR 0.15-0.65] vs. 0.19 [IQR 0.13-0.36] per year; p < 0.001), had a longer hospital stay (6 [IQR 2-13] vs. 4 [IQR 1-8] days; p < 0.001), had more comorbidities (4 [IQR 2-7] vs. 2 [IQR 1-5]; p < 0.001), and had a significantly higher in-hospital mortality (2.6% [95% CI 2.3%-2.9%] vs. 1.8% [95% CI 1.5%-2.1%] (p < 0.001). A worse outcome was observed among sarcoidosis patients having co-occurrence of associated respiratory diseases. Moreover, age was an important risk factor for re-hospitalization.
Citation: Pohle S, Baty F, Brutsche M (2016) In-Hospital Disease Burden of Sarcoidosis in Switzerland from 2002 to 2012. PLoS ONE 11(3): e0151940. https://doi.org/10.1371/journal.pone.0151940
Editor: Elliott D. Crouser, Wexner Medical Center at The Ohio State University, UNITED STATES
Received: September 17, 2015; Accepted: March 7, 2016; Published: March 17, 2016
Copyright: © 2016 Pohle et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Data analyzed from the current study were extracted from a nation-wide hospitalization database from the Swiss Federal Office for Statistics whose access is regulated by this federal institution. These data can only be requested by directly contacting this institution: Bundesamt für Statistik BFS, Sektion Gesundheit, Espace de l’Europe 10, 2010 Neuchâtel, Switzerland (tel: +41 58 463 6700 / e-mail: gesundheit@bfs.admin.ch).
Funding: The study was supported by an unconditional research grant by the Lungenliga St. Gallen and an institutional grant by the Kantonsspital St. Gallen. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Sarcoidosis is a systemic disease of unknown cause resolving in complete remission in many cases but sometimes leading to substantial and even fatal organic dysfunctions [1]. Nowadays, novel and more aggressive therapeutic agents are used to optimize the treatment of sarcoidosis patients, although their benefit is discussed controversially [1]. To date, there are only few studies reflecting the recent development of the mortality and hospitalization rates of sarcoidosis in Europe. In the United States, the number of hospitalizations and the rate of mortality among sarcoidosis patients have substantially increased over the last decades [2, 3]. The cause of this trend is unknown and not associated with an increasing incidence of sarcoidosis. Incriminated are a number of distinct complications namely pulmonary hypertension, comorbidities (e.g. diabetes mellitus, heart failure) as well as important risk factors like race, age and sex [3–7]. The frequent use of systemic corticosteroids and immunosuppressive therapy might also lead to higher infection rates and associated complications. Furthermore, sarcoidosis has been associated with a higher risk for cancer and pulmonary embolism [8, 9].
The aim of this study was to investigate the hospitalization rate and in-hospital mortality among sarcoidosis patients over the last years in Switzerland, as well as the associated comorbidities and risk factors of this disease.
Materials and Methods
Hospitalization Database
All our observations were derived from a dataset provided by the Swiss Federal Office for Statistics, which offers a nation-wide coverage of all hospitalized patients since 1998. In this database, the patient information is fully anonymized. No written informed consent was given to the patients who were unidentifiable due to the anonymization. The data belong to the Swiss Federal Office for Statistics (Bundesamt für Statistik, Neuchâtel, Switzerland) who provides regulated access to the data for research purpose. In analogy to the study of Deubelbeiss and colleagues [10] or Baty and collaborators [11], data since 2002 were extracted for the analysis, since satisfactory coding quality was reached by then. These data were imported into a relational SQL database (MySQL, version 5.6.24). The database was interfaced with the R statistical software (version 3.1.2) [12], using the dedicated package RMySQL. It included 15,627,573 entries (6,337,575 unique patients) corresponding to all hospitalized cases in Switzerland between 2002 and 2012. Each patient in this database was identified uniquely so that it was possible to track re-hospitalizations of every patient. The database included geographical and temporal information (patient’s residency, hospitalization by canton, year and month of hospitalization, length of hospital stay) as well as age at admission and reason/type of discharge (including death). The patients’ diagnosis list included one main diagnosis as well as up to 50 additional diagnoses coded using the International Classification of Disease version 10 (ICD-10) codes (http://www.who.int/classifications/icd/en/). The coding version ICD-10 was uniformly used all along the study time period.
Nested Case-control Study Design
For the current study, sarcoidosis was defined by the ICD-10 code D86*. The cases were defined as any hospitalizations with the diagnosis of sarcoidosis either as main or concomitant diagnosis. Geriatric, obstetric, psychiatric hospitalization and cases of rehabilitation were excluded from the study. The controls consisted of hospitalizations not including sarcoidosis in the main or concomitant diagnoses. Controls were extracted from the hospitalization database using a random procedure and matching the cases 1:1 for age, gender, and month of hospitalization.
Comorbidities and Mortality
Sarcoidosis-comorbidities were defined as any additional diagnosis coded by the ICD-10 nomenclature. The Charlson’s comorbidity index [13] was used to assign a predicted mortality score to patients.
Statistical Analysis
Descriptive statistics were reported as median and inter-quartile-range (IQR). The enrichment of sarcoidosis-associated comorbidities was tested using conditional logistic regression in order to account for the way the controls were sampled in our nested case-control design. Results were reported as odds-ratio (OR) with 95% confidence intervals (CI) together with p-values from likelihood ratio tests. Time to first re-hospitalization was defined as the time elapsed between the first and second hospitalization record of a given sarcoidosis patient in the database. Data from patients without re-hospitalization, i.e. recorded only once in the database, were censored (lost to follow-up censoring). Time—to—event data were analyzed using Kaplan-Meier estimates, log-rank tests and Cox-proportional hazards regression. Results were reported as hazard ratio (HR) with 95% confidence intervals (CI) together with the Wald test statistic. Hazard ratios provide the percentage of increase in the hazard rates for an increase of one unit in the explanatory variable. Generalized estimating equations (using working independence correlation structure) was used to test associations with one or several covariates while accounting for within-patient variability.
Principal component analysis (PCA) was used to explore correlations among sarcoidosis-associated comorbidities. PCA depicts how comorbidities are related to each other and how sarcoidosis cases are grouped according to comorbidities. A data-matrix summarizing the presence/absence (1/0) of comorbidities in sarcoidosis cases was created. Scaled PCA was applied to the comorbidity matrix. The role played by external explanatory variables, including age, length-of-hospital stay, gender, in-hospital death, hospitalization rate, and number of comorbidities, was further explored using vector fitting procedure to identify the directions of maximal correlation with the external variables in the PCA space [14]. PCA results were directly displayed using biplot representation [15]. PCA biplots are organized as follows: due to the very large number of hospitalization cases, smoothed color density image was used to depict the density of cases in the biplot representation (PCA scores); comorbidities were represented using framed labels (PCA loadings); external variables were superimposed to the biplot using vector representations (arrows) pointing to the direction of most rapid change. The length of the arrows is proportional to the correlation between the PCA configuration and the explanatory variable.
All analyses were done at the hospitalization level, except for the re-hospitalization analysis which was done at the patient level. All statistical computations were done using the R statistical software (version 3.1.2) [12]. The following extension packages were used: survival, geepack, Epi, multipleNCC, ade4 and vegan.
Results
Characteristics of Sarcoidosis Hospitalizations in Switzerland
The number of hospitalized cases with a diagnosis of sarcoidosis either as main or secondary diagnosis between 2002 and 2012 was 8,385 corresponding to 0.054% (8385 / 15,627,573) of all hospitalizations in Switzerland, and representing 4,664 unique patients (Table 1). Overall, 50.2% were male, and the median age was 58 years (IQR: 45–70).
The proportion of hospitalizations due to sarcoidosis was age-dependent. A higher percentage of hospitalization was found among patients aged 40–50 years (1%) and elderly patients defined as more than 70 year old (6.3%). Re-hospitalizations occurred 1.80 times per patient on average (8,385 hospitalizations in 4,664 patients).
In 30% of cases, sarcoidosis was the main reason for hospitalization (primary diagnosis).
A seasonal pattern was found over the observational period with the highest rates of hospitalizations in June and the lowest in October and December (p < 0.001) (Fig 1).
Fig 2 displays the rate of hospitalization in sarcoidosis patients (upper panel) together with the rate of in-hospital mortality (lower panel) over the time period 2002–2012. Overall, both the rate of hospitalization and the in-hospital mortality remained stable. However, there was an increased rate of hospitalization in the age categories (45–58] (p < 0.001) and (70–96] (p < 0.001). The in-hospital mortality rate (lower panel) did not increase significantly in any age category over the observed time period (2002–2012), although there was a trend towards an increase of cases in the older category (age >70) (p = 0.085) (Fig 2).
The upper panel shows the hospitalization rate of sarcoidosis patients between 2002 and 2012 (per 100,000 hospitalizations) displayed for 4 age categories. The lower panel depicts the in-hospital mortality of sarcoidosis patients between 2002 and 2012 subdivided into 4 age categories. The black dashed line in the upper and lower panels represents the overall evolution of the hospitalization rate and in-hospital mortality rate over time, respectively.
Comparison of Sarcoidosis Cases vs. Controls
The characteristics of sarcoidosis cases and matched controls are summarized in Table 1. Overall, sarcoidosis cases had a significantly worse outcome compared to controls: they were more frequently rehospitalized (median annual hospitalization rate 0.28 [IQR 0.15–0.65] vs. 0.19 [IQR 0.13–0.36] per year; p < 0.001), had a longer hospital stay (6 [IQR 2–13] vs. 4 [IQR 1–8] days; p < 0.001), had more comorbidities (4 [IQR 2–7] vs. 2 [IQR 1–5]; p < 0.001), and had a significantly higher in-hospital mortality (2.6% [95% CI 2.3%-2.9%] vs. 1.8% [95% CI 1.5%-2.1%] (p < 0.001).
Comorbidities among Sarcoidosis Cases
Among 4,096 observed comorbidities, 529 were significantly either enriched or under-represented in sarcoidosis cases (p < 0.05). Of these, 57 with the lowest p-values (p < 1 × 10−10) were kept in the analysis (Table 2). Besides well-known sarcoidosis-related manifestations and complications including respiratory failure, interstitial pulmonary disease, pulmonary hypertension and kidney diseases (J960, J961, J841, J969, I270, I279, I2728, L52, N162, N189, N19, R590, R591, G532, H221, G473, E559, E835), we also found a significantly higher declaration of medication and possible medication-related comorbidities (I10, K219, E139, J189, J151, B441, E274, M819, M814, D90, Z922, Z512, Y579, Y420). Furthermore, sarcoidosis patients were associated with other pulmonary diagnosis such as asthma and COPD (J459, J450, J47, J449, J448).
Sarcoidosis cases had a higher rate of pulmonary embolism (1.4% vs. 0.63%; p < 0.001) and a lower rate of cancer disease (12% vs. 13%; p = 0.002). Under-represented comorbidities among sarcoidosis cases consisted of orthopedical, gestational, geriatric, and mental disorders.
Sarcoidosis and Pulmonary Embolism
Overall, there were 119 hospitalized cases with sarcoidosis and pulmonary embolism (ICD-10 code: I26*), corresponding to 1.4% of all sarcoidosis hospitalizations. Compared to the control group, sarcoidosis cases had a higher rate of pulmonary embolism (1.4% vs. 0.63%, p < 0.001). Cases with the combined diagnosis were older (median age: 71 [IQR 58.5–89.0], p < 0.001) and tended to be more frequent among females (sex ratio: 0.42). They had a worse outcome with more comorbidities (median number of comorbidity: 5 [IQR 3–7], p < 0.001), a longer in-hospital stay (median length-of-stay: 12 days [IQR 7–18.5], p < 0.001) and a higher chance of in-hospital mortality (9.2% [95% CI 4.9–16.3], p < 0.001) in comparison to sarcoidosis cases without pulmonary embolism.
Prognostic Relevance and Correlations of Sarcoidosis-associated Comorbidities
Fig 3 displays a PCA biplot showing correlations of comorbidities and the presence of comorbidity clusters among sarcoidosis patients. The comorbidities displayed along the first PCA axis include adverse events related to the use of hormones and medication, as well as diseases including hypertension and diabetes mellitus. Along the second PCA axis, several lung diseases including pneumonia, bronchiectasis, interstitial pulmonary disease, respiratory heart disease, COPD and sleep apnea are lying in the same direction (lower quadrants), thus correlating with each other. In the upper right box, a set of explanatory variables are presented, mostly pointing to the lower direction. These explanatory variables reflecting poorer conditions (higher hospitalization rate, higher number of comorbidities, longer length-of-stay, and older ages) are associated with the presence of concomitant lung diseases.
Hospitalization cases (PCA scores) are represented using smoothed blue colored density, comorbidities are illustrated by framed labels (PCA loadings). Comorbidities lying in the same direction are correlating with each other, the further away from the center of the plot, the stronger is the influence of comorbidities. The number displayed on the upper left corner indicates the size of the grid. The upper right box represents external variables including age, gender, length-of-stay [LOS], in-hospital mortality [Death], hospitalization rate, and number of comorbidities fitted to the PCA and displayed using vector representation (arrows). The longer the arrow, the stronger the association to the corresponding comorbidities.
Risk Factors for Re-hospitalization
Fig 4 shows the Kaplan-Meier curves representing the association between age, sex and number of comorbidities (panels A, B, C, respectively) and the time to first re-hospitalization. The risk of re-hospitalization was significantly higher in the elderly population with an increase of 1.5% in the hazard rates for each increase of 1 year of age (HR = 1.015 [95% CI: 1.012–1.019], p < 0.001). The risk of re-hospitalization was also significantly higher in patients having a larger amount of comorbidities with an increase of 5.4% in the hazard rates for each additional comorbidity (HR = 1.054 [95% CI: 1.042–1.067], p < 0.001). No statistically significant difference was found between females and males (HR = 1.104 [95% CI: 0.996–1.223], p = 0.058). Fig 5 displays the comorbidities significantly associated with the time to first re-hospitalization. Comorbidities significantly increasing the risk of re-hospitalization in sarcoidosis patients included kidney diseases (N189, N162), osteoporosis (M819, M814), respiratory failure (J960, J961), pulmonary heart disease (I279), pulmonary hypertension (I270) and essential hypertension (I10).
For this representation, the explanatory variables age and number of comorbidities were categorized using the quartiles of their distribution. Log-rank test p-values are reported in the upper left corner of each panel.
The hazard ratios (and 95% confidence intervals) of the time to first re-hospitalization associated with each comorbidity are displayed. Comorbidities that were significantly associated with the time to first re-hospitalization (p < 0.05) are depicted in red.
Discussion
In this longitudinal, nested case-control study, we investigated the recent mortality and hospitalization rates as well as the underlying comorbidities of hospitalized sarcoidosis patients in Switzerland from 2002 to 2012. Furthermore, we analyzed possible risk factors and predictors of outcome compared to matched controls.
Overall, our analysis showed stable mortality and hospitalization rates among sarcoidosis patients in Switzerland over the observed time period. This is in contrast to recent findings of a significant mortality increase, at least for Caucasians, in the United States [4–6]. Reasons could be the comparatively short time period in our study regarding epidemiological changes as well as national differences. On the other hand, the results of the U.K. National Health Service showed concordant results to our study with stable mortality rates over the time period 1979–2008 [16].
By clearly displaying the burden of the disease, our data reveal that sarcoidosis patients showed a significantly worse outcome compared to matched controls regarding re-hospitalization, in-hospital mortality and length of hospital stay. Similar results were recently found when comparing sarcoidosis patients with pedestrian traffic accident patients [6]. In the Black Women’s Health Study, a follow-up study of 59,000 African American women between 1995 and 2009, Tukey and colleagues described a significantly higher mortality rate in women with sarcoidosis in comparison to women without the disease [17]. About 25% of the deaths were directly attributable to the disease itself.
In our study, we identified possible predictors and risk factors associated with a worse outcome. We found that age was a significant prognostic factor of the disease, which is supported by other previous studies [2]. There was an increased in-hospital mortality in sarcoidosis patients older than 70 years and a significant increase regarding the hospitalization rate in the middle and older age categories. Moreover elderly patients had a higher risk of re-hospitalization.
Specific comorbidities possibly contributing to a worse outcome were also detected. For instance, associated respiratory diseases were significantly correlated to a longer hospital stay and a higher number of comorbidities. This defines a phenotype of sarcoidosis patients with a higher morbidity. Among these associated respiratory diseases, there was a higher risk of pulmonary embolism. A link between sarcoidosis and pulmonary embolism was postulated in previous studies [8] while the underlying patho-mechanisms are unclear. We could identify possible predictors of co-occurrence, since hospitalized cases with the combined diagnosis were older, had more comorbidities and were more frequently females. Some previous studies postulated a higher risk for cancer diseases in sarcoidosis patients [18]. Our data from Switzerland showed a slight decrease of cancer diseases in sarcoidosis patients, which is in agreement with the data from a Danish and UK study [19, 20].
Medication-related events seem to substantially affect sarcoidosis patients leading to a significantly higher hospitalization rate as well as to a shorter time to re-hospitalization. Long-term benefits of the corticosteroidal and immunosuppressive therapy are often discussed controversially [1]. Our findings support that these therapeutic agents are sometimes harmful. More attention should be payed to possible therapeutic adverse events and the focus should be put on a close monitoring together with a more restrictive way of administrating medication.
One of the strength of our study is the comprehensive number of patients registered by the Swiss coding system, providing good data quality. In addition, this study is one of a few European studies investigating hospitalizations and mortality rates among sarcoidosis patients and the only one displaying risk factors for re-hospitalization. As a limitation, our analysis is based on a database which is essentially hospitalization-oriented. Our cohort only included hospitalized sarcoidosis cases while a high fraction of patients is typically treated in an out-patient setting. Further information such as medication or laboratory data were not available nor was it possible to contact patients to complete or confirm data.
Overall, the current study is showing stable mortality and hospitalization rates among sarcoidosis patients over the observed time period. However, the burden of the disease becomes apparent, as patients with sarcoidosis show a significantly worse outcome compared to matched controls. This could be due to the disease itself as well as to a cluster of specific comorbidities and complications namely associated respiratory diseases. On the other hand, there is also evidence that side effects of more aggressive treatment approaches could explain this poorer outcome, probably especially affecting patients at older age.
A better knowledge of these prognostic factors could lead to a better understanding of the disease and guide to a more specific monitoring as well as more individualized therapeutic approaches. Further research should be done to evaluate this and maybe contribute to a better outcome.
Acknowledgments
The study was supported by an unconditional research grant by the Lungenliga St. Gallen and an institutional grant by the Kantonsspital St. Gallen.
Author Contributions
Conceived and designed the experiments: MB SP. Analyzed the data: FB. Wrote the paper: SP FB MB.
References
- 1. Valeyre D, Prasse A, Nunes H, Uzunhan Y, Brillet PY, Muller-Quernheim J. Sarcoidosis. Lancet. 2014 Mar;383(9923):1155–1167. pmid:24090799
- 2. Gerke AK, Yang M, Tang F, Cavanaugh JE, Polgreen PM. Increased hospitalizations among sarcoidosis patients from 1998 to 2008: a population-based cohort study. BMC Pulm Med. 2012;12:19. pmid:22584044
- 3. Foreman MG, Mannino DM, Kamugisha L, Westney GE. Hospitalization for patients with sarcoidosis: 1979–2000. Sarcoidosis Vasc Diffuse Lung Dis. 2006 Jun;23(2):124–129. pmid:17937108
- 4. Swigris JJ, Olson AL, Huie TJ, Fernandez-Perez ER, Solomon J, Sprunger D, et al. Sarcoidosis-related mortality in the United States from 1988 to 2007. Am J Respir Crit Care Med. 2011 Jun;183(11):1524–1530. pmid:21330454
- 5. Gideon NM, Mannino DM. Sarcoidosis mortality in the United States 1979–1991: an analysis of multiple-cause mortality data. Am J Med. 1996 Apr;100(4):423–427. pmid:8610729
- 6. Mirsaeidi M, Machado RF, Schraufnagel D, Sweiss NJ, Baughman RP. Racial difference in sarcoidosis mortality in the United States. Chest. 2015 Feb;147(2):438–449. pmid:25188873
- 7. Nardi A, Brillet PY, Letoumelin P, Girard F, Brauner M, Uzunhan Y, et al. Stage IV sarcoidosis: comparison of survival with the general population and causes of death. Eur Respir J. 2011 Dec;38(6):1368–1373. pmid:22075486
- 8. Swigris JJ, Olson AL, Huie TJ, Fernandez-Perez ER, Solomon JJ, Sprunger D, et al. Increased risk of pulmonary embolism among US decedents with sarcoidosis from 1988 to 2007. Chest. 2011 Nov;140(5):1261–1266. pmid:21565969
- 9. Ji J, Shu X, Li X, Sundquist K, Sundquist J, Hemminki K. Cancer risk in hospitalized sarcoidosis patients: a follow-up study in Sweden. Ann Oncol. 2009 Jun;20(6):1121–1126. pmid:19211624
- 10. Deubelbeiss U, Gemperli A, Schindler C, Baty F, Brutsche MH. Prevalence of sarcoidosis in Switzerland is associated with environmental factors. Eur Respir J. 2010 May;35(5):1088–1097. pmid:19897550
- 11. Baty F, Putora PM, Isenring B, Blum T, Brutsche M. Comorbidities and burden of COPD: a population based case-control study. PLoS One. 2013;8(5):e63285. pmid:23691009
- 12.
R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria; 2014. Available from: http://www.R-project.org/.
- 13. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373–383. pmid:3558716
- 14.
Oksanen J. Multivariate analysis of ecological communities in R: vegan tutorial; 2010. Accessed 2013 Apr 8. Available from: http://cc.oulu.fi/~jarioksa/opetus/metodi/vegantutor.pdf.
- 15.
Greenacre M. Biplots in Practice. Fundación BBVA; 2010. Accessed 2013 Apr 8. Available from: http://books.google.ch/books?id=dv4LrFP7U_EC.
- 16. Duncan ME, Goldacre MJ. Mortality trends for tuberculosis and sarcoidosis in English populations, 1979–2008. Int J Tuberc Lung Dis. 2012 Jan;16(1):38–42. pmid:22236843
- 17. Tukey MH, Berman JS, Boggs DA. Mortality among African American women with sarcoidosis: data from the Black Women’s Health Study. Sarcoidosis Vasc Diffuse Lung Dis. 2013 Jul;30(2):128–133.
- 18. Bonifazi M, Bravi F, Gasparini S, La Vecchia C, Gabrielli A, Wells AU, et al. Sarcoidosis and cancer risk: systematic review and meta-analysis of observational studies. Chest. 2015 Mar;147(3):778–791. pmid:25340385
- 19. Rømer FK, Hommelgaard P, Schou G. Sarcoidosis and cancer revisited: a long-term follow-up study of 555 Danish sarcoidosis patients. Eur Respir J. 1998 Oct;12(4):906–912. pmid:9817167
- 20. Le Jeune I, Gribbin J, West J, Smith C, Cullinan P, Hubbard R. The incidence of cancer in patients with idiopathic pulmonary fibrosis and sarcoidosis in the UK. Respir Med. 2007 Dec;101(12):2534–2540. pmid:17870458