Endocrine DOI 10.1007/s12020-015-0547-4
ORIGINAL ARTICLE
Prospective population-based study of the association between vitamin D status and incidence of autoimmune disease Tea Skaaby • Lise Lotte Nystrup Husemoen • Betina Heinsbæk Thuesen • Allan Linneberg
Received: 10 December 2014 / Accepted: 2 February 2015 Ó Springer Science+Business Media New York 2015
Abstract Beside its traditional role in skeletal health, vitamin D is believed to have multiple immunosuppressant properties, and low vitamin D status has been suggested to be a risk factor in the development of autoimmune disease. We investigated the association between vitamin D status and development of autoimmune disease. We included a total of 12,555 individuals from three population-based studies with measurements of vitamin D status (25-hydroxy vitamin D). We followed the participants by linkage to the Danish National Patient Register (median follow-up time 10.8 years). Relative risks of autoimmune disease were estimated by Cox regression and expressed as hazard ratios, HRs (95 % confidence intervals CIs). There were 525 cases of incident autoimmune disease. The risk for a 10 nmol/l higher vitamin D was: for any autoimmune disease (HR = 0.94 % CI 0.90, 0.98); thyrotoxicosis (HR = 0.83, 95 % CI 0.72, 0.96); type 1 diabetes (HR =
0.95, 95 % CI 0.88, 1.02), multiple sclerosis (HR = 0.89, 95 % CI 0.74, 1.07), iridocyclitis (HR = 1.00, 95 % CI 0.86, 1.17); Crohn’s disease (HR = 0.95, 95 % CI 0.80, 1.13), ulcerative colitis (HR = 0.88, 95 % CI 0.75, 1.04); psoriasis vulgaris (HR = 0.99, 95 % CI 0.86, 1.13); seropositive rheumatoid arthritis (HR = 0.97, 95 % CI 0.89, 1.07), and polymyalgia rheumatica (HR = 0.94, 95 % CI 0.83, 1.06). We found statistically significant inverse associations between vitamin D status and development of any autoimmune disease and thyrotoxicosis in particular. Our findings suggest a possible protective role of a higher vitamin D status on autoimmune disease but warrant further studies to clarify causality. Keywords Autoimmune disease Thyrotoxicosis Vitamin D Population-based study
Introduction Electronic supplementary material The online version of this article (doi:10.1007/s12020-015-0547-4) contains supplementary material, which is available to authorized users. T. Skaaby (&) L. L. N. Husemoen B. H. Thuesen A. Linneberg Research Centre for Prevention and Health, Glostrup University Hospital, Nordre Ringvej 57, Building 84/85, 2600 Glostrup, Denmark e-mail: [email protected] A. Linneberg Department of Clinical Experimental Research, Glostrup University Hospital, Nordre Ringvej 57, 2600 Glostrup, Denmark A. Linneberg Department of Clinical Medicine, Faculty of Health and Medical Sciences, University of Copenhagen, Blegdamsvej 3B, 2200 Copenhagen, Denmark
Vitamin D is a fat soluble vitamin that can be ingested from diet and dietary supplements but is primarily produced in sun exposed skin [1]. Beside its traditional role in skeletal health, it has been suggested to play a role in a number of other diseases [2–9], and vitamin D is believed to have multiple immunosuppressant properties [10]. Low vitamin D status is frequent particularly at Northern latitudes where the intensity of the sun is insufficient for vitamin D production in the skin during winter. Since autoimmune diseases are also more frequent here and some are in increase [11], low vitamin D status has been suggested as a possible risk factor for autoimmune disease [12–14]. In animal models, vitamin D supplementation has been shown to be therapeutically effective in autoimmune encephalomyelitis, collagen-induced arthritis, type 1
123
Endocrine
diabetes mellitus, inflammatory bowel disease, autoimmune thyroiditis, and systemic lupus erythematosus (SLE), and in some models of SLE it prevented disease development [12]. Several studies have shown that vitamin D inhibits pro-inflammatory processes by suppressing the increased activity of immune cells that take part in the autoimmune reaction, whereas some reports imply that vitamin D may even be preventive in certain disorders such as multiple sclerosis and diabetes type 1 [12]. The vitamin D intake has been found to be inversely related to the risk of developing multiple sclerosis, and the risk for multiple sclerosis was found to be approximately 50 % greater with vitamin D levels less than 20 ng/ml [15, 16]. Among patients with multiple sclerosis mainly treated with interferon beta-1b, low vitamin D levels early in the disease course were a strong risk factor for longterm multiple sclerosis activity and progression [17]. In addition, studies have reported that supplementing children with up to 2,000 IU of vitamin D per day early in life reduced their risk for type I diabetes by approximately 30 % on average [18]. In a meta-analysis, Song et al. found that low vitamin D intake was associated with an elevated risk of development and activity of rheumatoid arthritis and in 7 of 8 studies (6 cross-sectional and 2 case–control studies) with levels of vitamin D, vitamin D and rheumatoid arthritis activity were inversely related [19]. Since previous studies have mainly focused on the association between vitamin D intake rather vitamin D status in relation to autoimmune disease, and since nearly all of the few studies investigating vitamin D status and autoimmune diseases have been cross-sectional, the role of vitamin D in the development of autoimmune disease is still unresolved [13]. We investigated the association between vitamin D status as assessed by 25-hydroxy vitamin D (25-OH-D) and the incidence of a registry-based diagnosis of autoimmune diseases according to the International Classification of Diseases (ICD) in three Danish population-based studies.
participated which gave a participation rate of 52.5 % [20]. The study was a population-based randomized controlled trial (CT00289237, ClinicalTrials.gov) that investigated the effects of lifestyle intervention on cardiovascular disease [20]. In the Health2006 study, 7,931 persons from 18 to 69 years of age were invited to a general health examination [20] and 3,471 (43.8 %) persons were examined from 2006 to 2008. From the three included studies, a total of 12,555 individuals had measurements of vitamin D as previously described [8] (Monica10: 2,649; Inter99: 6,497; Health2006: 3409). Registry-based diagnoses
Materials and methods
It is possible to link data from Danish registers on an individual level since all persons living in Denmark have a unique and permanent personal civil registration number. The Danish Civil Registration System provided information on emigration and vital status [21]. Information on diseases according to the ICD was obtained from the Danish National Patient Register [22] that holds information on admissions to Danish hospitals since 1978. Admissions are registered by one primary diagnosis and optionally one or more secondary diagnoses. The participants were followed until 31 December 2010. The categorization of diagnoses is summarized in Supplementary material 1. Before 1987, no distinction was made between insulin-dependent and non-insulin-dependent diabetes mellitus (ICD-8: 250 for both types of diabetes) [23]. From 1987 to 1998, insulin-dependent diabetes mellitus was coded as 249 (ICD-8; 1987-1993) or E10 (ICD-10; 1994-present). Most cases of type 1 and type 2 diabetes are likely coded as insulin-dependent diabetes mellitus and noninsulin-dependent diabetes mellitus, respectively. However, some patients with type 2 diabetes need insulin treatment and may thus be categorized as having insulin-dependent diabetes which may have caused a false inflation of this group. Persons with a registry-based diagnosis of the event before baseline were excluded from the prospective analyses.
Study populations
Covariates
We used the three population-based studies, Monica10, Inter99, and Health2006 that were recruited from the Danish Central Personal Register as random samples of the population. The studies included questionnaires, blood tests, and physical examinations. An overview of the studies in this area has previously been published [20]. The Monica10 study was conducted in 1993–94, included 2,656 persons (4,130 invited) between 40 and 71 years and thus had a participation rate of 64.3 % [20]. In the Inter99 study from 1999–2001, 12,934 persons aged 30–60 years were invited. A total of 6,784 persons
Body mass index (BMI) was calculated as weight divided by height squared (kg/m2). The questionnaires gave information on the following covariates: smoking habits according to the consumption of gram tobacco per day (never smokers; former smokers; current smokers \15 g/day; current smokers 15–25 g/day; current smokers C25 g/day); education/vocational training (no education beyond basic, education including students); alcohol consumption (drinks per week); and physical activity during leisure time (sedentary, light, or moderate/vigorous). Blood pressure (mmHg) was calculated as the average of two measurements in the sitting position.
123
Endocrine
Serum lipids were measured from fasting blood samples using enzymatic colorimetric methods (Roche, Mannheim, Germany) [24–27]. The number of missing values was: smoking habits (Nmissing = 41); education (Nmissing = 297); alcohol consumption (Nmissing = 632); physical activity during leisure time (Nmissing = 101); BMI (Nmissing = 6); systolic blood pressure (Nmissing = 2); diastolic blood pressure (Nmissing = 4); serum total cholesterol (Nmissing = 3); and serum triglycerides (Nmissing = 3), giving a total of 993 persons (7.9 %) with a missing value in at least one of the covariates. Statistical analyses Analyses were performed with SAS, version 9.4 (SAS Institute Inc. Cary, NC USA). p values were two-sided, and p values \0.05 were defined as statistically significant. Table 1 shows the baseline characteristics of the participants according to study population and expressed as % (number) or mean (standard deviation, SD) and compared with v2-test or Kruskal–Wallis test, respectively. In Table 2, incident autoimmune disease according to study population is displayed, compared with the v2-test and expressed as % (numbers). Multivariable Cox regression analyses were used to determine the association of baseline vitamin D status with the incidence of autoimmune disease (Table 3). Data from the three included study populations were pooled, we used delayed entry where persons contribute risk time from the date of baseline examination, and age was used as underlying time axis (which means that age is automatically adjusted for). Estimates are presented as hazard ratios, HR (95 % confidence intervals CI). Only persons with complete information
on all considered variables were included in the Cox regression analyses. The few participants lost to follow-up or deceased contributed risk time until the time of their last registered activity or date of death. In model 1, we adjusted for gender and study population. In model 2, we further adjusted for education, season of blood sample, physical activity, smoking habits, alcohol intake, body mass index, systolic and diastolic blood pressure, serum total cholesterol, and serum triglycerides. In additional analyses with meta-analyses of the studyspecific estimates from model 2, we used Stata, version 12.1 (StataCorp LP, College Station, Texas, USA). We used the inverse variance method in fixed and random effects models (Supplementary material 2). We assessed the heterogeneity between the study populations with the I2-test. The fixed effect meta-estimates are summarized in Fig. 1.
Results Baseline characteristics according to study population are summarized in Table 1. Mean baseline vitamin D status was highest in the Monica10 and lowest in the Health2006 study. The percentages with a history of autoimmune disease are highest in the Health2006 study and lowest in the Monica10 study. There were 525 cases of incident autoimmune disease (Table 2). The median follow-up was 10.8 years (Monica10: 16.4; Inter99: 11.0; Health2006: 3.5 years). In multivariable Cox regression analyses, we found significantly lower risks of any autoimmune disease and thyrotoxicosis in particular with HR = 0.94 (95 % CI 0.90, 0.98) and HR = 0.83
Table 1 Baseline characteristics according to study population Study population
p valuea
Mean (SD) or % (n) Monica10
Inter99
Health2006
Year of examination Male gender
1993–1994 50.2 (1,329)
1999–2001 49.2 (3,197)
2006–2008 44.9 (1,531)
Age (years)
55.4 (10.8)
46.1 (7.9)
49.4 (13.0)
\0.0001
BMI (kg/m2)
25.9 (4.2)
26.3 (4.6)
25.9 (4.7)
\0.0001
Alcohol (drinks/week)
10.1 (12.2)
10.3 (13.3)
9.7 (10.2)
0.170
\0.0001
No education beyond basic
25.6 (677)
16.6 (1,040)
13.3 (445)
\0.0001
Never smokers
26.2 (693)
35.1 (2,279)
41.7 (1,409)
\0.0001
Low physical activityb
21.2 (550)
21.6 (1,373)
18.1 (611)
\0.0001
Vitamin D status, nmol/lc
64.7 (27.1)
51.2 (26.6)
44.3 (22.6)
\0.0001
History of autoimmune disease
2.2 (57)
2.6 (166)
3.4 (115)
0.032
BMI body mass index, SD standard deviation a
v2 or Kruskal–Wallis test of the distribution across studies
b
Sedentary leisure time physical activity In comparison, the vitamin D median (p 25, p 75) among persons that developed any autoimmune disease during follow-up was for Health2006: 37.7 (28.0, 55.2), for Inter99: 43.0 (29.0, 60.0), and for Monica10: 56.7 (43.6, 77.5) nmol/l
c
123
Endocrine Table 2 Incident autoimmune disease according to study population Inter99 % (n) Any autoimmune diseasea
Health2006 % (n)
Monica10 % (n)
p value v2-test
In total n
1.2 (41)
8.9 (230)
\0.001
525
4.0 (254)
Thyrotoxicosis
0.4 (27)
0.03 (1)
0.6 (15)
\0.001
43
Type 1 diabetes
1.0 (64)
0.2 (6)
3.2 (83)
\0.001
153
Multiple sclerosis
0.2 (13)
0.1 (2)
0.3 (9)
Iridocyclitis
0.2 (14)
0.1 (2)
0.5 (14)
0.044
24
\0.001
30
Crohn’s disease
0.3 (16)
0.1 (3)
0.2 (5)
0.230
24
Ulcerative colitis Psoriasis vulgaris
0.3 (20) 0.4 (28)
0.2 (5) 0.1 (4)
0.3 (9) 0.7 (18)
0.256 0.002
34 50
Seropositive rheumatoid arthritis
0.7 (47)
0.2 (8)
1.6 (43)
\0.001
98
Polymyalgia rheumatica
0.2 (15)
0.2 (7)
1.3 (35)
\0.001
57
a
Including thyrotoxicosis, type 1 diabetes, multiple sclerosis, iridocyclitis, Crohn’s disease, ulcerative colitis, psoriasis vulgaris, seropositive rheumatoid arthritis, polymyalgia rheumatica, pernicious anemia, autoimmune hemolytic anemia, idiopathic thrombocytopenic purpura, autoimmune thyroiditis, primary adrenocortical insufficiency, Guillain–Barre syndrome, autoimmune hepatitis, primary biliary cirrhosis, celiac disease, pemphigus, pemphigoid, alopecia areata, vitiligo, juvenile arthritis, Wegener’s granulomatosis, dermatopolymyositis, myasthenia gravis, systemic sclerosis, systemic lupus erythematosus, Sjo¨gren’s syndrome, and ankylosing spondylitis
Table 3 Hazard ratios and 95 % confidence intervals of pooled analyses of incident autoimmune disease for a 10 nmol/l higher vitamin D status Events (persons included)
Model 1a HR (95 % CI), p value Per 10 nmol/l higher vitamin D status
Model 2b HR (95 % CI), p value Per 10 nmol/l higher vitamin D status
Outcome Any autoimmune diseasec
466 (11,271)
0.93 (0.89, 0.96), p = 0.0001
0.94 (0.90, 0.98), p = 0.002
Thyrotoxicosis
38 (11,543)
0.89 (0.77, 1.02), p = 0.086
0.83 (0.72, 0.96), p = 0.012
Type 1 diabetes
135 (11,521)
0.89 (0.83, 0.96), p = 0.002
0.95 (0.88, 1.02), p = 0.150
Multiple sclerosis
20 (11,535)
0.94 (0.79, 1.12), p = 0.468
0.89 (0.74, 1.07), p = 0.206
Iridocyclitis
29 (11,544)
0.97 (0.84, 1.13), p = 0.724
1.00 (0.86, 1.17), p = 0.999
Crohn’s disease
22 (11,544)
0.93 (0.78, 1.11), p = 0.413
0.95 (0.80, 1.13), p = 0.570
Ulcerative colitis
32 (11,512)
0.86 (0.73, 1.00), p = 0.053
0.88 (0.75, 1.04), p = 0.137
Psoriasis vulgaris
38 (11,546)
1.00 (0.98, 1.01), p = 0.449
0.99 (0.86, 1.13), p = 0.857
Seropositive rheumatoid arthritis
89 (11,523)
0.94 (0.87, 1.03), p = 0.183
0.97 (0.89, 1.07), p = 0.572
Polymyalgia rheumatic
54 (11,544)
0.96 (0.85, 1.07), p = 0.420
0.94 (0.83, 1.06), p = 0.291
Complete case analysis HR hazard ratio, CI confidence interval a
Adjusted for study and gender (age is underlying time axis)
b
Further adjusted for education, season of blood sample, physical activity, smoking habits, alcohol intake, body mass index, systolic and diastolic blood pressure, serum total cholesterol, and serum triglycerides
c Including thyrotoxicosis, type 1 diabetes, multiple sclerosis, iridocyclitis, Crohn’s disease, ulcerative colitis, psoriasis vulgaris, seropositive rheumatoid arthritis, polymyalgia rheumatica, pernicious anemia, autoimmune hemolytic anemia, idiopathic thrombocytopenic purpura, autoimmune thyroiditis, primary adrenocortical insufficiency, Guillain–Barre syndrome, autoimmune hepatitis, primary biliary cirrhosis, celiac disease, pemphigus, pemphigoid, alopecia areata, vitiligo, juvenile arthritis, Wegener’s granulomatosis, dermatopolymyositis, myasthenia gravis, systemic sclerosis, systemic lupus erythematosus, Sjo¨gren’s syndrome, and ankylosing spondylitis
(95 % CI 0.72, 0.96), respectively, per 10 nmol/l higher vitamin D status (Table 3, model 2). In the partially adjusted analyses, we found significantly lower and borderline significantly lower risks of type 1 diabetes and ulcerative colitis, respectively, with higher vitamin D status (Table 3, model 1). However, both associations were not statistically significant in the fully adjusted analyses (Table 3, model 2).
123
The remaining associations were not statistically significant either in partially or fully adjusted analyses (Table 3, model 1 and 2). The results from meta-analyses (Fig. 1 and Supplementary Figs. 1–10) were similar to the results from the pooled analyses (Table 3, model 2). In all but three of the outcomes, there were no signs of heterogeneity between the
Endocrine Fig. 1 Summarization of metaanalyses of the study-specific estimates of the association between a 10 nmol/l higher vitamin D status and risk of autoimmune disease. HR hazard ratio, CI confidence interval
studies, in two studies there were signs of a slight heterogeneity (Ulcerative colitis: I2 = 7.6 %, iridocyclitis: I2 = 6.9 %), whereas at most and in only one outcome, we found signs of moderate heterogeneity between studies (Thyrotoxicosis: I2 = 50.9 %).
Discussion In a large prospective study of 12,555 persons from three study populations including a total of 525 cases of incident autoimmune disease, we found a statistically significant lower incidence of any autoimmune disease and thyrotoxicosis in particular with higher vitamin D status in both individual-based analyses and meta-analyses of studyspecific estimates. We extended the findings from previous studies by investigating the association of an objective marker of vitamin D status and development of any and specific autoimmune diseases including the association with thyrotoxicosis and psoriasis where literature is particularly sparse. Autoimmune diseases are immune-mediated and commonly lead to severe disability and increased mortality [28]. IgE-mediated sensitisation to autoantigens, autosensitisation, characterizes the development of autoimmune diseases and arises from an abnormal immune response against substances and tissues that are normally present in the body. It may be restricted to particular organs or tissues in different places, e.g. type 1 diabetes, an autoimmune disease caused by cell-mediated immunity against pancreatic islet beta cells; multiple sclerosis, a chronic autoimmune disease where the antigenic target is central nervous system myelin
[29]; rheumatoid arthritis that is characterized by erosive arthritis and joint destruction [14]; and Graves’ hyperthyroidism that is the most frequent cause of thyrotoxicosis where autoantibodies to the TSH receptor stimulate the thyroid gland [30]. The observed inverse association between vitamin D status and thyrotoxicosis is in line with a study that found a high prevalence of vitamin D deficiency in Japanese female patients with Graves’ disease [31], vitamin D deficiency was found to be associated with higher thyroid volume in patients with newly onset Graves’ disease [32], and the active form of vitamin D was found to be lower in patients with Graves’ disease [33]. In an animal study, Misharin et al. found that mice on a vitamin D deficient diet more often developed persistent hyperthyroidism than mice receiving adequate vitamin D supply [30]. In an intervention study, 30 Graves’ disease patients were randomly assigned to either methimazole alone or combination treatment with methimazole and vitamin D3 [34]. Patients treated with the combination therapy had more rapid euthyroidism achievement compared with patients receiving methimazole alone, but the study was small and did not include vitamin D measurements [34]. Also, genetic variability in the proteins involved in vitamin D metabolism may contribute to autoimmune disease and in some but not all studies, a vitamin D receptor polymorphism has been associated with Graves’ disease [35–37]. Of interest, however, a case–control study found that early stages of thyroid immunity were not associated with low vitamin D levels [38]. The observed lack of association between vitamin D status and rheumatoid arthritis is somewhat in contrast to a meta-analysis by Song et al. (215,757 participants) who
123
Endocrine
found that low vitamin D intake was associated with an elevated risk of development and activity of rheumatoid arthritis [19]. Eight studies had data on vitamin D levels (six of which were cross-sectional and two were case– control studies), and all of these studies, except for one, found that vitamin D levels were inversely associated with rheumatoid arthritis activity. Of note, the association observed by Song et al. may be confounded by obesity that has been shown to be associated with both rheumatoid arthritis and vitamin D status which the authors were unable to take into account as opposed to our analyses that were adjusted for BMI. However, our estimate only changed little when adjusting for BMI, i.e., the hazard ratio changed from 0.96 to 0.97. The observed lack of association between vitamin D status and psoriasis is in line with the results from a crosssectional population-based study by Wilson that included 5,841 participants of which 148 reported a psoriasis diagnosis and found that vitamin D deficiency was not more common in psoriasis [39]. Also, a study by Merola et al. found no association between vitamin D intake as assessed by a semi-quantitative questionnaire and incident psoriasis among 70,437 US female nurses aged 47–74 enrolled in the Nurses’ Health Study [40]. However, two previous studies have found low vitamin D status to be more common among persons with psoriasis compared with controls [41, 42]. Gisondi et al. found that vitamin D deficiency was associated with psoriasis independently of age, sex, body mass index, calcium, PTH levels, and season of blood sampling among 145 patients with chronic plaque psoriasis, 112 patients with rheumatoid arthritis and 141 healthy controls [41], whereas OrgazMolina et al. found that vitamin D status was significantly lower in psoriatic patients than in controls in a case– control study including 43 patients with psoriasis and 43 age- and sex-matched control subjects [42]. However, both studies reporting an association recruited persons from hospitals, which represent only a subset of patients, and this may explain the conflicting results. In addition, sun exposure that increases vitamin D status is also known to relieve psoriasis symptoms due to a local effect of the UV radiation which is in fact also used in the treatment of the disease [43, 44] and may thus pose as a negative confounder of the association between vitamin D status and psoriasis, e.g., if there is a delay from disease onset until inclusion in the registry or if persons with early symptoms and no diagnosis of psoriasis have used sun exposure as a preventive measure. The strengths of our study include the longitudinal design with long-term follow-up and the large populationbased samples. We included standardized registry-based diagnoses with few lost to follow-up and an accepted and objective marker of vitamin D.
123
The limitations of the study include the low number of events in some of the autoimmune disease categories, and we may have lacked power to show small-moderate effects. E.g., we found non-significantly lower risks of multiple sclerosis and ulcerative colitis that may be due to a shortness of power/too few events. Hence, future meta-analyses should be undertaken, e.g., to investigate a possible link between vitamin D status and multiple sclerosis and ulcerative colitis. The group of any autoimmune disease may include disease entities with different pathogenesis where vitamin D might have a protective effect against some and a promoting effect on other and hence the effects may counteract each other. Of note, persons with autoimmune diseases have difficulty in exercise and moving and spend less time outdoors and are exposed to less sunlight which may reduce vitamin D levels [14]. To reduce this risk of reverse causation, we excluded persons with a registrybased diagnosis of the autoimmune disease in question at baseline. Also, the possible misclassification of type 2 diabetics among our group of type 1 diabetes may explain the observed lack of association between vitamin D status and diabetes (please see the description of the registrybased diagnoses). Autoimmune diseases are the third leading cause of morbidity and mortality in industrialized countries. Even though targeted treatment against the Th1 pathway, TNF-a in particular, has revolutionized the treatment of autoimmune diseases such as inflammatory bowel disease and rheumatoid arthritis [45], it is important to clarify the etiology, pathogenesis and possible risk factors. We found that a higher vitamin D status was associated with a lower risk of autoimmune disease and in particular thyrotoxicosis. Further studies are necessary to investigate whether low vitamin D status is in fact a causal and reversible factor in the development of autoimmune disease.
Ethics standards The studies were approved by the Ethics Committee of Copenhagen and the Danish Data Protection Agency, and the recommendations of the Declaration of Helsinki were followed. Participants gave their informed written consent. Conflict of interest The authors declare that they have no conflicts of interest.
References 1. M.F. Holick, Vitamin D deficiency. N. Engl. J. Med. 357, 266–281 (2007) 2. T. Skaaby, L.L. Husemoen, C. Pisinger, T. Jorgensen, B.H. Thuesen, M. Fenger et al., Vitamin D status and cause-specific mortality: a general population study. PLoS One 7, e52423 (2012)
Endocrine 3. T. Skaaby, L.L. Husemoen, C. Pisinger, T. Jorgensen, B.H. Thuesen, M. Fenger et al., Vitamin D status and changes in cardiovascular risk factors: a prospective study of a general population. Cardiology 123, 62–70 (2012) 4. T. Skaaby, L.L. Husemoen, C. Pisinger, T. Jorgensen, B.H. Thuesen, K. Rasmussen et al., Vitamin D status and 5-year changes in urine albumin creatinine ratio and parathyroid hormone in a general population. Endocrine 44, 473–480 (2013) 5. T. Skaaby, L.L. Husemoen, T. Martinussen, J.P. Thyssen, M. Melgaard, B.H. Thuesen et al., Vitamin D status, filaggrin genotype, and cardiovascular risk factors: a Mendelian randomization approach. PLoS One 8, e57647 (2013) 6. T. Skaaby, L.L. Husemoen, C. Pisinger, T. Jorgensen, B.H. Thuesen, M. Fenger et al., Vitamin D status and incident cardiovascular disease and all-cause mortality: a general population study. Endocrine 43, 618–625 (2013) 7. T. Skaaby, L.L. Husemoen, A. Borglykke, T. Jorgensen, B.H. Thuesen, C. Pisinger et al., Vitamin D status, liver enzymes, and incident liver disease and mortality: a general population study. Endocrine 47, 213–220 (2013) 8. T. Skaaby, L.L. Husemoen, B.H. Thuesen, C. Pisinger, T. Jorgensen, N. Roswall et al., Prospective population-based study of the association between serum 25-hydroxyvitamin-D levels and the incidence of specific types of cancer. Cancer Epidemiol. Biomarkers Prev. 23, 1220–1229 (2014) 9. T. Skaaby, L.L. Husemoen, B.H. Thuesen, C. Pisinger, T. Jorgensen, R.V. Fenger et al., Vitamin D status and chronic obstructive pulmonary disease: a prospective general population study. PLoS One 9, e90654 (2014) 10. M. Hewison, An update on vitamin D and human immunity. Clin. Endocrinol. (Oxf) 76, 315–325 (2012) 11. A. Latina, D. Gullo, F. Trimarchi, S. Benvenga, Hashimoto’s thyroiditis: similar and dissimilar characteristics in neighboring areas. Possible implications for the epidemiology of thyroid cancer. PLoS One 8, e55450 (2013) 12. Y. Arnson, H. Amital, Y. Shoenfeld, Vitamin D and autoimmunity: new aetiological and therapeutic considerations. Ann. Rheum. Dis. 66, 1137–1142 (2007) 13. P. Pludowski, M.F. Holick, S. Pilz, C.L. Wagner, B.W. Hollis, W.B. Grant et al., Vitamin D effects on musculoskeletal health, immunity, autoimmunity, cardiovascular disease, cancer, fertility, pregnancy, dementia and mortality-A review of recent evidence. Autoimmun. Rev. 12, 976–989 (2013) 14. P. Szodoray, B. Nakken, J. Gaal, R. Jonsson, A. Szegedi, E. Zold et al., The complex role of vitamin D in autoimmune diseases. Scand. J. Immunol. 68, 261–269 (2008) 15. K.L. Munger, S.M. Zhang, E. O’Reilly, M.A. Hernan, M.J. Olek, W.C. Willett et al., Vitamin D intake and incidence of multiple sclerosis. Neurology 62, 60–65 (2004) 16. K.L. Munger, L.I. Levin, B.W. Hollis, N.S. Howard, A. Ascherio, Serum 25-hydroxyvitamin D levels and risk of multiple sclerosis. JAMA 296, 2832–2838 (2006) 17. A. Ascherio, K.L. Munger, R. White, K. Kochert, K.C. Simon, C.H. Polman et al., Vitamin D as an early predictor of multiple sclerosis activity and progression. JAMA Neurol. 71, 306–314 (2014) 18. C. Mathieu, C. Gysemans, A. Giulietti, R. Bouillon, Vitamin D and diabetes. Diabetologia 48, 1247–1257 (2005) 19. G.G. Song, S.C. Bae, Y.H. Lee, Association between vitamin D intake and the risk of rheumatoid arthritis: a meta-analysis. Clin. Rheumatol. 31, 1733–1739 (2012) 20. M. Osler, A. Linneberg, C. Glumer, T. Jorgensen, The cohorts at the Research Centre for Prevention and Health, formerly ‘The Glostrup Population Studies’. Int. J. Epidemiol. 40, 602–610 (2011) 21. C.B. Pedersen, The Danish Civil Registration System. Scand. J. Public Health 39, 22–25 (2011)
22. T.F. Andersen, M. Madsen, J. Jorgensen, L. Mellemkjoer, J.H. Olsen, The Danish National Hospital Register. A valuable source of data for modern health sciences. Dan. Med. Bull. 46, 263–268 (1999) 23. N.M. Nielsen, T. Westergaard, M. Frisch, K. Rostgaard, J. Wohlfahrt, N. Koch-Henriksen et al., Type 1 diabetes and multiple sclerosis: a Danish population-based cohort study. Arch. Neurol. 63, 1001–1004 (2006) 24. M. Aadahl, S.L. von Huth, C. Pisinger, U.N. Toft, C. Glumer, K. Borch-Johnsen et al., Five-year change in physical activity is associated with changes in cardiovascular disease risk factors: the Inter99 study. Prev. Med. 48, 326–331 (2009) 25. L. Ambye, S. Rasmussen, M. Fenger, T. Jorgensen, K. BorchJohnsen, S. Madsbad et al., Studies of the Gly482Ser polymorphism of the peroxisome proliferator-activated receptor gamma coactivator 1alpha (PGC-1alpha) gene in Danish subjects with the metabolic syndrome. Diabetes Res. Clin. Pract. 67, 175–179 (2005) 26. T. Jorgensen, K. Borch-Johnsen, T.F. Thomsen, H. Ibsen, C. Glumer, C. Pisinger, A randomized non-pharmacological intervention study for prevention of ischaemic heart disease: baseline results Inter99. Eur. J. Cardiovasc. Prev. Rehabil. 10, 377–386 (2003) 27. J.P. Thyssen, A. Linneberg, B.C. Carlsen, J.D. Johansen, K. Engkilde, T. Hansen et al., A possible association between a dysfunctional skin barrier (filaggrin null-mutation status) and diabetes: a cross-sectional study. BMJ Open 1, e000062 (2011) 28. G.S. Cooper, B.C. Stroehla, The epidemiology of autoimmune diseases. Autoimmun. Rev. 2, 119–125 (2003) 29. R.L. Rabin, A.I. Levinson, The nexus between atopic disease and autoimmunity: a review of the epidemiological and mechanistic literature. Clin. Exp. Immunol. 153, 19–30 (2008) 30. A. Misharin, M. Hewison, C.R. Chen, V. Lagishetty, H.A. Aliesky, Y. Mizutori et al., Vitamin D deficiency modulates Graves’ hyperthyroidism induced in BALB/c mice by thyrotropin receptor immunization. Endocrinology 150, 1051–1060 (2009) 31. H. Yamashita, S. Noguchi, K. Takatsu, E. Koike, T. Murakami, S. Watanabe et al., High prevalence of vitamin D deficiency in Japanese female patients with Graves’ disease. Endocr. J. 48, 63–69 (2001) 32. T. Yasuda, Y. Okamoto, N. Hamada, K. Miyashita, M. Takahara, F. Sakamoto et al., Serum vitamin D levels are decreased and associated with thyroid volume in female patients with newly onset Graves’ disease. Endocrine 42, 739–741 (2012) 33. H. Czernobilsky, S. Scharla, H. Schmidt-Gayk, R. Ziegler, Enhanced suppression of 1,25(OH)2D3 and intact parathyroid hormone in Graves’ disease as compared to toxic nodular goiter. Calcif. Tissue Int. 42, 5–12 (1988) 34. T. Kawakami-Tani, E. Fukawa, H. Tanaka, Y. Abe, I. Makino, Effect of 1 alpha-hydroxy vitamin D3 on serum levels of thyroid hormones in hyperthyroid patients with untreated Graves’ disease. Metabolism 46, 1184–1188 (1997) 35. M.A. Pani, K. Regulla, M. Segni, S. Hofmann, M. Hufner, A.M. Pasquino et al., A polymorphism within the vitamin D-binding protein gene is associated with Graves’ disease but not with Hashimoto’s thyroiditis. J. Clin. Endocrinol. Metab. 87, 2564–2567 (2002) 36. C.E. Jennings, C.J. Owen, V. Wilson, S.H. Pearce, A haplotype of the CYP27B1 promoter is associated with autoimmune Addison’s disease but not with Graves’ disease in a UK population. J. Mol. Endocrinol. 34, 859–863 (2005) 37. E.R. Lopez, O. Zwermann, M. Segni, G. Meyer, M. Reincke, J. Seissler et al., A promoter polymorphism of the CYP27B1 gene is associated with Addison’s disease, Hashimoto’s thyroiditis, Graves’ disease and type 1 diabetes mellitus in Germans. Eur. J. Endocrinol. 151, 193–197 (2004)
123
Endocrine 38. G. Effraimidis, K. Badenhoop, J.G. Tijssen, W.M. Wiersinga, Vitamin D deficiency is not associated with early stages of thyroid autoimmunity. Eur. J. Endocrinol. 167, 43–48 (2012) 39. P.B. Wilson, Serum 25-hydroxy vitamin D status in individuals with psoriasis in the general population. Endocrine 44, 537–539 (2013) 40. J.F. Merola, J. Han, T. Li, A.A. Qureshi, No association between vitamin D intake and incident psoriasis among US women. Arch. Dermatol. Res. 306, 305–307 (2014) 41. P. Gisondi, M. Rossini, C.A. Di, L. Idolazzi, S. Farina, G. Beltrami et al., Vitamin D status in patients with chronic plaque psoriasis. Br. J. Dermatol. 166, 505–510 (2012) 42. J. Orgaz-Molina, A. Buendia-Eisman, M.A. Arrabal-Polo, J.C. Ruiz, S. Arias-Santiago, Deficiency of serum concentration of
123
25-hydroxy vitamin D in psoriatic patients: a case-control study. J. Am. Acad. Dermatol. 67, 931–938 (2012) 43. A.A. Levin, S. Aleissa, N. Dumont, F. Martinez, C. Donovan, S.C. Au et al., A randomized, prospective, sham-controlled study of localized narrow-band UVB phototherapy in the treatment of plaque psoriasis. J. Drugs Dermatol. 13, 922–926 (2014) 44. I. Heier, E. Soyland, A.L. Krogstad, C. Rodriguez-Gallego, M.S. Nenseter, F.L. Jahnsen, Sun exposure rapidly reduces plasmacytoid dendritic cells and inflammatory dermal dendritic cells in psoriatic skin. Br. J. Dermatol. 165, 792–801 (2011) 45. A. Shah, The pathologic and clinical intersection of atopic and autoimmune disease. Curr. Allergy Asthma Rep. 12, 520–529 (2012)
ORIGINAL ARTICLE
Prospective population-based study of the association between vitamin D status and incidence of autoimmune disease Tea Skaaby • Lise Lotte Nystrup Husemoen • Betina Heinsbæk Thuesen • Allan Linneberg
Received: 10 December 2014 / Accepted: 2 February 2015 Ó Springer Science+Business Media New York 2015
Abstract Beside its traditional role in skeletal health, vitamin D is believed to have multiple immunosuppressant properties, and low vitamin D status has been suggested to be a risk factor in the development of autoimmune disease. We investigated the association between vitamin D status and development of autoimmune disease. We included a total of 12,555 individuals from three population-based studies with measurements of vitamin D status (25-hydroxy vitamin D). We followed the participants by linkage to the Danish National Patient Register (median follow-up time 10.8 years). Relative risks of autoimmune disease were estimated by Cox regression and expressed as hazard ratios, HRs (95 % confidence intervals CIs). There were 525 cases of incident autoimmune disease. The risk for a 10 nmol/l higher vitamin D was: for any autoimmune disease (HR = 0.94 % CI 0.90, 0.98); thyrotoxicosis (HR = 0.83, 95 % CI 0.72, 0.96); type 1 diabetes (HR =
0.95, 95 % CI 0.88, 1.02), multiple sclerosis (HR = 0.89, 95 % CI 0.74, 1.07), iridocyclitis (HR = 1.00, 95 % CI 0.86, 1.17); Crohn’s disease (HR = 0.95, 95 % CI 0.80, 1.13), ulcerative colitis (HR = 0.88, 95 % CI 0.75, 1.04); psoriasis vulgaris (HR = 0.99, 95 % CI 0.86, 1.13); seropositive rheumatoid arthritis (HR = 0.97, 95 % CI 0.89, 1.07), and polymyalgia rheumatica (HR = 0.94, 95 % CI 0.83, 1.06). We found statistically significant inverse associations between vitamin D status and development of any autoimmune disease and thyrotoxicosis in particular. Our findings suggest a possible protective role of a higher vitamin D status on autoimmune disease but warrant further studies to clarify causality. Keywords Autoimmune disease Thyrotoxicosis Vitamin D Population-based study
Introduction Electronic supplementary material The online version of this article (doi:10.1007/s12020-015-0547-4) contains supplementary material, which is available to authorized users. T. Skaaby (&) L. L. N. Husemoen B. H. Thuesen A. Linneberg Research Centre for Prevention and Health, Glostrup University Hospital, Nordre Ringvej 57, Building 84/85, 2600 Glostrup, Denmark e-mail: [email protected] A. Linneberg Department of Clinical Experimental Research, Glostrup University Hospital, Nordre Ringvej 57, 2600 Glostrup, Denmark A. Linneberg Department of Clinical Medicine, Faculty of Health and Medical Sciences, University of Copenhagen, Blegdamsvej 3B, 2200 Copenhagen, Denmark
Vitamin D is a fat soluble vitamin that can be ingested from diet and dietary supplements but is primarily produced in sun exposed skin [1]. Beside its traditional role in skeletal health, it has been suggested to play a role in a number of other diseases [2–9], and vitamin D is believed to have multiple immunosuppressant properties [10]. Low vitamin D status is frequent particularly at Northern latitudes where the intensity of the sun is insufficient for vitamin D production in the skin during winter. Since autoimmune diseases are also more frequent here and some are in increase [11], low vitamin D status has been suggested as a possible risk factor for autoimmune disease [12–14]. In animal models, vitamin D supplementation has been shown to be therapeutically effective in autoimmune encephalomyelitis, collagen-induced arthritis, type 1
123
Endocrine
diabetes mellitus, inflammatory bowel disease, autoimmune thyroiditis, and systemic lupus erythematosus (SLE), and in some models of SLE it prevented disease development [12]. Several studies have shown that vitamin D inhibits pro-inflammatory processes by suppressing the increased activity of immune cells that take part in the autoimmune reaction, whereas some reports imply that vitamin D may even be preventive in certain disorders such as multiple sclerosis and diabetes type 1 [12]. The vitamin D intake has been found to be inversely related to the risk of developing multiple sclerosis, and the risk for multiple sclerosis was found to be approximately 50 % greater with vitamin D levels less than 20 ng/ml [15, 16]. Among patients with multiple sclerosis mainly treated with interferon beta-1b, low vitamin D levels early in the disease course were a strong risk factor for longterm multiple sclerosis activity and progression [17]. In addition, studies have reported that supplementing children with up to 2,000 IU of vitamin D per day early in life reduced their risk for type I diabetes by approximately 30 % on average [18]. In a meta-analysis, Song et al. found that low vitamin D intake was associated with an elevated risk of development and activity of rheumatoid arthritis and in 7 of 8 studies (6 cross-sectional and 2 case–control studies) with levels of vitamin D, vitamin D and rheumatoid arthritis activity were inversely related [19]. Since previous studies have mainly focused on the association between vitamin D intake rather vitamin D status in relation to autoimmune disease, and since nearly all of the few studies investigating vitamin D status and autoimmune diseases have been cross-sectional, the role of vitamin D in the development of autoimmune disease is still unresolved [13]. We investigated the association between vitamin D status as assessed by 25-hydroxy vitamin D (25-OH-D) and the incidence of a registry-based diagnosis of autoimmune diseases according to the International Classification of Diseases (ICD) in three Danish population-based studies.
participated which gave a participation rate of 52.5 % [20]. The study was a population-based randomized controlled trial (CT00289237, ClinicalTrials.gov) that investigated the effects of lifestyle intervention on cardiovascular disease [20]. In the Health2006 study, 7,931 persons from 18 to 69 years of age were invited to a general health examination [20] and 3,471 (43.8 %) persons were examined from 2006 to 2008. From the three included studies, a total of 12,555 individuals had measurements of vitamin D as previously described [8] (Monica10: 2,649; Inter99: 6,497; Health2006: 3409). Registry-based diagnoses
Materials and methods
It is possible to link data from Danish registers on an individual level since all persons living in Denmark have a unique and permanent personal civil registration number. The Danish Civil Registration System provided information on emigration and vital status [21]. Information on diseases according to the ICD was obtained from the Danish National Patient Register [22] that holds information on admissions to Danish hospitals since 1978. Admissions are registered by one primary diagnosis and optionally one or more secondary diagnoses. The participants were followed until 31 December 2010. The categorization of diagnoses is summarized in Supplementary material 1. Before 1987, no distinction was made between insulin-dependent and non-insulin-dependent diabetes mellitus (ICD-8: 250 for both types of diabetes) [23]. From 1987 to 1998, insulin-dependent diabetes mellitus was coded as 249 (ICD-8; 1987-1993) or E10 (ICD-10; 1994-present). Most cases of type 1 and type 2 diabetes are likely coded as insulin-dependent diabetes mellitus and noninsulin-dependent diabetes mellitus, respectively. However, some patients with type 2 diabetes need insulin treatment and may thus be categorized as having insulin-dependent diabetes which may have caused a false inflation of this group. Persons with a registry-based diagnosis of the event before baseline were excluded from the prospective analyses.
Study populations
Covariates
We used the three population-based studies, Monica10, Inter99, and Health2006 that were recruited from the Danish Central Personal Register as random samples of the population. The studies included questionnaires, blood tests, and physical examinations. An overview of the studies in this area has previously been published [20]. The Monica10 study was conducted in 1993–94, included 2,656 persons (4,130 invited) between 40 and 71 years and thus had a participation rate of 64.3 % [20]. In the Inter99 study from 1999–2001, 12,934 persons aged 30–60 years were invited. A total of 6,784 persons
Body mass index (BMI) was calculated as weight divided by height squared (kg/m2). The questionnaires gave information on the following covariates: smoking habits according to the consumption of gram tobacco per day (never smokers; former smokers; current smokers \15 g/day; current smokers 15–25 g/day; current smokers C25 g/day); education/vocational training (no education beyond basic, education including students); alcohol consumption (drinks per week); and physical activity during leisure time (sedentary, light, or moderate/vigorous). Blood pressure (mmHg) was calculated as the average of two measurements in the sitting position.
123
Endocrine
Serum lipids were measured from fasting blood samples using enzymatic colorimetric methods (Roche, Mannheim, Germany) [24–27]. The number of missing values was: smoking habits (Nmissing = 41); education (Nmissing = 297); alcohol consumption (Nmissing = 632); physical activity during leisure time (Nmissing = 101); BMI (Nmissing = 6); systolic blood pressure (Nmissing = 2); diastolic blood pressure (Nmissing = 4); serum total cholesterol (Nmissing = 3); and serum triglycerides (Nmissing = 3), giving a total of 993 persons (7.9 %) with a missing value in at least one of the covariates. Statistical analyses Analyses were performed with SAS, version 9.4 (SAS Institute Inc. Cary, NC USA). p values were two-sided, and p values \0.05 were defined as statistically significant. Table 1 shows the baseline characteristics of the participants according to study population and expressed as % (number) or mean (standard deviation, SD) and compared with v2-test or Kruskal–Wallis test, respectively. In Table 2, incident autoimmune disease according to study population is displayed, compared with the v2-test and expressed as % (numbers). Multivariable Cox regression analyses were used to determine the association of baseline vitamin D status with the incidence of autoimmune disease (Table 3). Data from the three included study populations were pooled, we used delayed entry where persons contribute risk time from the date of baseline examination, and age was used as underlying time axis (which means that age is automatically adjusted for). Estimates are presented as hazard ratios, HR (95 % confidence intervals CI). Only persons with complete information
on all considered variables were included in the Cox regression analyses. The few participants lost to follow-up or deceased contributed risk time until the time of their last registered activity or date of death. In model 1, we adjusted for gender and study population. In model 2, we further adjusted for education, season of blood sample, physical activity, smoking habits, alcohol intake, body mass index, systolic and diastolic blood pressure, serum total cholesterol, and serum triglycerides. In additional analyses with meta-analyses of the studyspecific estimates from model 2, we used Stata, version 12.1 (StataCorp LP, College Station, Texas, USA). We used the inverse variance method in fixed and random effects models (Supplementary material 2). We assessed the heterogeneity between the study populations with the I2-test. The fixed effect meta-estimates are summarized in Fig. 1.
Results Baseline characteristics according to study population are summarized in Table 1. Mean baseline vitamin D status was highest in the Monica10 and lowest in the Health2006 study. The percentages with a history of autoimmune disease are highest in the Health2006 study and lowest in the Monica10 study. There were 525 cases of incident autoimmune disease (Table 2). The median follow-up was 10.8 years (Monica10: 16.4; Inter99: 11.0; Health2006: 3.5 years). In multivariable Cox regression analyses, we found significantly lower risks of any autoimmune disease and thyrotoxicosis in particular with HR = 0.94 (95 % CI 0.90, 0.98) and HR = 0.83
Table 1 Baseline characteristics according to study population Study population
p valuea
Mean (SD) or % (n) Monica10
Inter99
Health2006
Year of examination Male gender
1993–1994 50.2 (1,329)
1999–2001 49.2 (3,197)
2006–2008 44.9 (1,531)
Age (years)
55.4 (10.8)
46.1 (7.9)
49.4 (13.0)
\0.0001
BMI (kg/m2)
25.9 (4.2)
26.3 (4.6)
25.9 (4.7)
\0.0001
Alcohol (drinks/week)
10.1 (12.2)
10.3 (13.3)
9.7 (10.2)
0.170
\0.0001
No education beyond basic
25.6 (677)
16.6 (1,040)
13.3 (445)
\0.0001
Never smokers
26.2 (693)
35.1 (2,279)
41.7 (1,409)
\0.0001
Low physical activityb
21.2 (550)
21.6 (1,373)
18.1 (611)
\0.0001
Vitamin D status, nmol/lc
64.7 (27.1)
51.2 (26.6)
44.3 (22.6)
\0.0001
History of autoimmune disease
2.2 (57)
2.6 (166)
3.4 (115)
0.032
BMI body mass index, SD standard deviation a
v2 or Kruskal–Wallis test of the distribution across studies
b
Sedentary leisure time physical activity In comparison, the vitamin D median (p 25, p 75) among persons that developed any autoimmune disease during follow-up was for Health2006: 37.7 (28.0, 55.2), for Inter99: 43.0 (29.0, 60.0), and for Monica10: 56.7 (43.6, 77.5) nmol/l
c
123
Endocrine Table 2 Incident autoimmune disease according to study population Inter99 % (n) Any autoimmune diseasea
Health2006 % (n)
Monica10 % (n)
p value v2-test
In total n
1.2 (41)
8.9 (230)
\0.001
525
4.0 (254)
Thyrotoxicosis
0.4 (27)
0.03 (1)
0.6 (15)
\0.001
43
Type 1 diabetes
1.0 (64)
0.2 (6)
3.2 (83)
\0.001
153
Multiple sclerosis
0.2 (13)
0.1 (2)
0.3 (9)
Iridocyclitis
0.2 (14)
0.1 (2)
0.5 (14)
0.044
24
\0.001
30
Crohn’s disease
0.3 (16)
0.1 (3)
0.2 (5)
0.230
24
Ulcerative colitis Psoriasis vulgaris
0.3 (20) 0.4 (28)
0.2 (5) 0.1 (4)
0.3 (9) 0.7 (18)
0.256 0.002
34 50
Seropositive rheumatoid arthritis
0.7 (47)
0.2 (8)
1.6 (43)
\0.001
98
Polymyalgia rheumatica
0.2 (15)
0.2 (7)
1.3 (35)
\0.001
57
a
Including thyrotoxicosis, type 1 diabetes, multiple sclerosis, iridocyclitis, Crohn’s disease, ulcerative colitis, psoriasis vulgaris, seropositive rheumatoid arthritis, polymyalgia rheumatica, pernicious anemia, autoimmune hemolytic anemia, idiopathic thrombocytopenic purpura, autoimmune thyroiditis, primary adrenocortical insufficiency, Guillain–Barre syndrome, autoimmune hepatitis, primary biliary cirrhosis, celiac disease, pemphigus, pemphigoid, alopecia areata, vitiligo, juvenile arthritis, Wegener’s granulomatosis, dermatopolymyositis, myasthenia gravis, systemic sclerosis, systemic lupus erythematosus, Sjo¨gren’s syndrome, and ankylosing spondylitis
Table 3 Hazard ratios and 95 % confidence intervals of pooled analyses of incident autoimmune disease for a 10 nmol/l higher vitamin D status Events (persons included)
Model 1a HR (95 % CI), p value Per 10 nmol/l higher vitamin D status
Model 2b HR (95 % CI), p value Per 10 nmol/l higher vitamin D status
Outcome Any autoimmune diseasec
466 (11,271)
0.93 (0.89, 0.96), p = 0.0001
0.94 (0.90, 0.98), p = 0.002
Thyrotoxicosis
38 (11,543)
0.89 (0.77, 1.02), p = 0.086
0.83 (0.72, 0.96), p = 0.012
Type 1 diabetes
135 (11,521)
0.89 (0.83, 0.96), p = 0.002
0.95 (0.88, 1.02), p = 0.150
Multiple sclerosis
20 (11,535)
0.94 (0.79, 1.12), p = 0.468
0.89 (0.74, 1.07), p = 0.206
Iridocyclitis
29 (11,544)
0.97 (0.84, 1.13), p = 0.724
1.00 (0.86, 1.17), p = 0.999
Crohn’s disease
22 (11,544)
0.93 (0.78, 1.11), p = 0.413
0.95 (0.80, 1.13), p = 0.570
Ulcerative colitis
32 (11,512)
0.86 (0.73, 1.00), p = 0.053
0.88 (0.75, 1.04), p = 0.137
Psoriasis vulgaris
38 (11,546)
1.00 (0.98, 1.01), p = 0.449
0.99 (0.86, 1.13), p = 0.857
Seropositive rheumatoid arthritis
89 (11,523)
0.94 (0.87, 1.03), p = 0.183
0.97 (0.89, 1.07), p = 0.572
Polymyalgia rheumatic
54 (11,544)
0.96 (0.85, 1.07), p = 0.420
0.94 (0.83, 1.06), p = 0.291
Complete case analysis HR hazard ratio, CI confidence interval a
Adjusted for study and gender (age is underlying time axis)
b
Further adjusted for education, season of blood sample, physical activity, smoking habits, alcohol intake, body mass index, systolic and diastolic blood pressure, serum total cholesterol, and serum triglycerides
c Including thyrotoxicosis, type 1 diabetes, multiple sclerosis, iridocyclitis, Crohn’s disease, ulcerative colitis, psoriasis vulgaris, seropositive rheumatoid arthritis, polymyalgia rheumatica, pernicious anemia, autoimmune hemolytic anemia, idiopathic thrombocytopenic purpura, autoimmune thyroiditis, primary adrenocortical insufficiency, Guillain–Barre syndrome, autoimmune hepatitis, primary biliary cirrhosis, celiac disease, pemphigus, pemphigoid, alopecia areata, vitiligo, juvenile arthritis, Wegener’s granulomatosis, dermatopolymyositis, myasthenia gravis, systemic sclerosis, systemic lupus erythematosus, Sjo¨gren’s syndrome, and ankylosing spondylitis
(95 % CI 0.72, 0.96), respectively, per 10 nmol/l higher vitamin D status (Table 3, model 2). In the partially adjusted analyses, we found significantly lower and borderline significantly lower risks of type 1 diabetes and ulcerative colitis, respectively, with higher vitamin D status (Table 3, model 1). However, both associations were not statistically significant in the fully adjusted analyses (Table 3, model 2).
123
The remaining associations were not statistically significant either in partially or fully adjusted analyses (Table 3, model 1 and 2). The results from meta-analyses (Fig. 1 and Supplementary Figs. 1–10) were similar to the results from the pooled analyses (Table 3, model 2). In all but three of the outcomes, there were no signs of heterogeneity between the
Endocrine Fig. 1 Summarization of metaanalyses of the study-specific estimates of the association between a 10 nmol/l higher vitamin D status and risk of autoimmune disease. HR hazard ratio, CI confidence interval
studies, in two studies there were signs of a slight heterogeneity (Ulcerative colitis: I2 = 7.6 %, iridocyclitis: I2 = 6.9 %), whereas at most and in only one outcome, we found signs of moderate heterogeneity between studies (Thyrotoxicosis: I2 = 50.9 %).
Discussion In a large prospective study of 12,555 persons from three study populations including a total of 525 cases of incident autoimmune disease, we found a statistically significant lower incidence of any autoimmune disease and thyrotoxicosis in particular with higher vitamin D status in both individual-based analyses and meta-analyses of studyspecific estimates. We extended the findings from previous studies by investigating the association of an objective marker of vitamin D status and development of any and specific autoimmune diseases including the association with thyrotoxicosis and psoriasis where literature is particularly sparse. Autoimmune diseases are immune-mediated and commonly lead to severe disability and increased mortality [28]. IgE-mediated sensitisation to autoantigens, autosensitisation, characterizes the development of autoimmune diseases and arises from an abnormal immune response against substances and tissues that are normally present in the body. It may be restricted to particular organs or tissues in different places, e.g. type 1 diabetes, an autoimmune disease caused by cell-mediated immunity against pancreatic islet beta cells; multiple sclerosis, a chronic autoimmune disease where the antigenic target is central nervous system myelin
[29]; rheumatoid arthritis that is characterized by erosive arthritis and joint destruction [14]; and Graves’ hyperthyroidism that is the most frequent cause of thyrotoxicosis where autoantibodies to the TSH receptor stimulate the thyroid gland [30]. The observed inverse association between vitamin D status and thyrotoxicosis is in line with a study that found a high prevalence of vitamin D deficiency in Japanese female patients with Graves’ disease [31], vitamin D deficiency was found to be associated with higher thyroid volume in patients with newly onset Graves’ disease [32], and the active form of vitamin D was found to be lower in patients with Graves’ disease [33]. In an animal study, Misharin et al. found that mice on a vitamin D deficient diet more often developed persistent hyperthyroidism than mice receiving adequate vitamin D supply [30]. In an intervention study, 30 Graves’ disease patients were randomly assigned to either methimazole alone or combination treatment with methimazole and vitamin D3 [34]. Patients treated with the combination therapy had more rapid euthyroidism achievement compared with patients receiving methimazole alone, but the study was small and did not include vitamin D measurements [34]. Also, genetic variability in the proteins involved in vitamin D metabolism may contribute to autoimmune disease and in some but not all studies, a vitamin D receptor polymorphism has been associated with Graves’ disease [35–37]. Of interest, however, a case–control study found that early stages of thyroid immunity were not associated with low vitamin D levels [38]. The observed lack of association between vitamin D status and rheumatoid arthritis is somewhat in contrast to a meta-analysis by Song et al. (215,757 participants) who
123
Endocrine
found that low vitamin D intake was associated with an elevated risk of development and activity of rheumatoid arthritis [19]. Eight studies had data on vitamin D levels (six of which were cross-sectional and two were case– control studies), and all of these studies, except for one, found that vitamin D levels were inversely associated with rheumatoid arthritis activity. Of note, the association observed by Song et al. may be confounded by obesity that has been shown to be associated with both rheumatoid arthritis and vitamin D status which the authors were unable to take into account as opposed to our analyses that were adjusted for BMI. However, our estimate only changed little when adjusting for BMI, i.e., the hazard ratio changed from 0.96 to 0.97. The observed lack of association between vitamin D status and psoriasis is in line with the results from a crosssectional population-based study by Wilson that included 5,841 participants of which 148 reported a psoriasis diagnosis and found that vitamin D deficiency was not more common in psoriasis [39]. Also, a study by Merola et al. found no association between vitamin D intake as assessed by a semi-quantitative questionnaire and incident psoriasis among 70,437 US female nurses aged 47–74 enrolled in the Nurses’ Health Study [40]. However, two previous studies have found low vitamin D status to be more common among persons with psoriasis compared with controls [41, 42]. Gisondi et al. found that vitamin D deficiency was associated with psoriasis independently of age, sex, body mass index, calcium, PTH levels, and season of blood sampling among 145 patients with chronic plaque psoriasis, 112 patients with rheumatoid arthritis and 141 healthy controls [41], whereas OrgazMolina et al. found that vitamin D status was significantly lower in psoriatic patients than in controls in a case– control study including 43 patients with psoriasis and 43 age- and sex-matched control subjects [42]. However, both studies reporting an association recruited persons from hospitals, which represent only a subset of patients, and this may explain the conflicting results. In addition, sun exposure that increases vitamin D status is also known to relieve psoriasis symptoms due to a local effect of the UV radiation which is in fact also used in the treatment of the disease [43, 44] and may thus pose as a negative confounder of the association between vitamin D status and psoriasis, e.g., if there is a delay from disease onset until inclusion in the registry or if persons with early symptoms and no diagnosis of psoriasis have used sun exposure as a preventive measure. The strengths of our study include the longitudinal design with long-term follow-up and the large populationbased samples. We included standardized registry-based diagnoses with few lost to follow-up and an accepted and objective marker of vitamin D.
123
The limitations of the study include the low number of events in some of the autoimmune disease categories, and we may have lacked power to show small-moderate effects. E.g., we found non-significantly lower risks of multiple sclerosis and ulcerative colitis that may be due to a shortness of power/too few events. Hence, future meta-analyses should be undertaken, e.g., to investigate a possible link between vitamin D status and multiple sclerosis and ulcerative colitis. The group of any autoimmune disease may include disease entities with different pathogenesis where vitamin D might have a protective effect against some and a promoting effect on other and hence the effects may counteract each other. Of note, persons with autoimmune diseases have difficulty in exercise and moving and spend less time outdoors and are exposed to less sunlight which may reduce vitamin D levels [14]. To reduce this risk of reverse causation, we excluded persons with a registrybased diagnosis of the autoimmune disease in question at baseline. Also, the possible misclassification of type 2 diabetics among our group of type 1 diabetes may explain the observed lack of association between vitamin D status and diabetes (please see the description of the registrybased diagnoses). Autoimmune diseases are the third leading cause of morbidity and mortality in industrialized countries. Even though targeted treatment against the Th1 pathway, TNF-a in particular, has revolutionized the treatment of autoimmune diseases such as inflammatory bowel disease and rheumatoid arthritis [45], it is important to clarify the etiology, pathogenesis and possible risk factors. We found that a higher vitamin D status was associated with a lower risk of autoimmune disease and in particular thyrotoxicosis. Further studies are necessary to investigate whether low vitamin D status is in fact a causal and reversible factor in the development of autoimmune disease.
Ethics standards The studies were approved by the Ethics Committee of Copenhagen and the Danish Data Protection Agency, and the recommendations of the Declaration of Helsinki were followed. Participants gave their informed written consent. Conflict of interest The authors declare that they have no conflicts of interest.
References 1. M.F. Holick, Vitamin D deficiency. N. Engl. J. Med. 357, 266–281 (2007) 2. T. Skaaby, L.L. Husemoen, C. Pisinger, T. Jorgensen, B.H. Thuesen, M. Fenger et al., Vitamin D status and cause-specific mortality: a general population study. PLoS One 7, e52423 (2012)
Endocrine 3. T. Skaaby, L.L. Husemoen, C. Pisinger, T. Jorgensen, B.H. Thuesen, M. Fenger et al., Vitamin D status and changes in cardiovascular risk factors: a prospective study of a general population. Cardiology 123, 62–70 (2012) 4. T. Skaaby, L.L. Husemoen, C. Pisinger, T. Jorgensen, B.H. Thuesen, K. Rasmussen et al., Vitamin D status and 5-year changes in urine albumin creatinine ratio and parathyroid hormone in a general population. Endocrine 44, 473–480 (2013) 5. T. Skaaby, L.L. Husemoen, T. Martinussen, J.P. Thyssen, M. Melgaard, B.H. Thuesen et al., Vitamin D status, filaggrin genotype, and cardiovascular risk factors: a Mendelian randomization approach. PLoS One 8, e57647 (2013) 6. T. Skaaby, L.L. Husemoen, C. Pisinger, T. Jorgensen, B.H. Thuesen, M. Fenger et al., Vitamin D status and incident cardiovascular disease and all-cause mortality: a general population study. Endocrine 43, 618–625 (2013) 7. T. Skaaby, L.L. Husemoen, A. Borglykke, T. Jorgensen, B.H. Thuesen, C. Pisinger et al., Vitamin D status, liver enzymes, and incident liver disease and mortality: a general population study. Endocrine 47, 213–220 (2013) 8. T. Skaaby, L.L. Husemoen, B.H. Thuesen, C. Pisinger, T. Jorgensen, N. Roswall et al., Prospective population-based study of the association between serum 25-hydroxyvitamin-D levels and the incidence of specific types of cancer. Cancer Epidemiol. Biomarkers Prev. 23, 1220–1229 (2014) 9. T. Skaaby, L.L. Husemoen, B.H. Thuesen, C. Pisinger, T. Jorgensen, R.V. Fenger et al., Vitamin D status and chronic obstructive pulmonary disease: a prospective general population study. PLoS One 9, e90654 (2014) 10. M. Hewison, An update on vitamin D and human immunity. Clin. Endocrinol. (Oxf) 76, 315–325 (2012) 11. A. Latina, D. Gullo, F. Trimarchi, S. Benvenga, Hashimoto’s thyroiditis: similar and dissimilar characteristics in neighboring areas. Possible implications for the epidemiology of thyroid cancer. PLoS One 8, e55450 (2013) 12. Y. Arnson, H. Amital, Y. Shoenfeld, Vitamin D and autoimmunity: new aetiological and therapeutic considerations. Ann. Rheum. Dis. 66, 1137–1142 (2007) 13. P. Pludowski, M.F. Holick, S. Pilz, C.L. Wagner, B.W. Hollis, W.B. Grant et al., Vitamin D effects on musculoskeletal health, immunity, autoimmunity, cardiovascular disease, cancer, fertility, pregnancy, dementia and mortality-A review of recent evidence. Autoimmun. Rev. 12, 976–989 (2013) 14. P. Szodoray, B. Nakken, J. Gaal, R. Jonsson, A. Szegedi, E. Zold et al., The complex role of vitamin D in autoimmune diseases. Scand. J. Immunol. 68, 261–269 (2008) 15. K.L. Munger, S.M. Zhang, E. O’Reilly, M.A. Hernan, M.J. Olek, W.C. Willett et al., Vitamin D intake and incidence of multiple sclerosis. Neurology 62, 60–65 (2004) 16. K.L. Munger, L.I. Levin, B.W. Hollis, N.S. Howard, A. Ascherio, Serum 25-hydroxyvitamin D levels and risk of multiple sclerosis. JAMA 296, 2832–2838 (2006) 17. A. Ascherio, K.L. Munger, R. White, K. Kochert, K.C. Simon, C.H. Polman et al., Vitamin D as an early predictor of multiple sclerosis activity and progression. JAMA Neurol. 71, 306–314 (2014) 18. C. Mathieu, C. Gysemans, A. Giulietti, R. Bouillon, Vitamin D and diabetes. Diabetologia 48, 1247–1257 (2005) 19. G.G. Song, S.C. Bae, Y.H. Lee, Association between vitamin D intake and the risk of rheumatoid arthritis: a meta-analysis. Clin. Rheumatol. 31, 1733–1739 (2012) 20. M. Osler, A. Linneberg, C. Glumer, T. Jorgensen, The cohorts at the Research Centre for Prevention and Health, formerly ‘The Glostrup Population Studies’. Int. J. Epidemiol. 40, 602–610 (2011) 21. C.B. Pedersen, The Danish Civil Registration System. Scand. J. Public Health 39, 22–25 (2011)
22. T.F. Andersen, M. Madsen, J. Jorgensen, L. Mellemkjoer, J.H. Olsen, The Danish National Hospital Register. A valuable source of data for modern health sciences. Dan. Med. Bull. 46, 263–268 (1999) 23. N.M. Nielsen, T. Westergaard, M. Frisch, K. Rostgaard, J. Wohlfahrt, N. Koch-Henriksen et al., Type 1 diabetes and multiple sclerosis: a Danish population-based cohort study. Arch. Neurol. 63, 1001–1004 (2006) 24. M. Aadahl, S.L. von Huth, C. Pisinger, U.N. Toft, C. Glumer, K. Borch-Johnsen et al., Five-year change in physical activity is associated with changes in cardiovascular disease risk factors: the Inter99 study. Prev. Med. 48, 326–331 (2009) 25. L. Ambye, S. Rasmussen, M. Fenger, T. Jorgensen, K. BorchJohnsen, S. Madsbad et al., Studies of the Gly482Ser polymorphism of the peroxisome proliferator-activated receptor gamma coactivator 1alpha (PGC-1alpha) gene in Danish subjects with the metabolic syndrome. Diabetes Res. Clin. Pract. 67, 175–179 (2005) 26. T. Jorgensen, K. Borch-Johnsen, T.F. Thomsen, H. Ibsen, C. Glumer, C. Pisinger, A randomized non-pharmacological intervention study for prevention of ischaemic heart disease: baseline results Inter99. Eur. J. Cardiovasc. Prev. Rehabil. 10, 377–386 (2003) 27. J.P. Thyssen, A. Linneberg, B.C. Carlsen, J.D. Johansen, K. Engkilde, T. Hansen et al., A possible association between a dysfunctional skin barrier (filaggrin null-mutation status) and diabetes: a cross-sectional study. BMJ Open 1, e000062 (2011) 28. G.S. Cooper, B.C. Stroehla, The epidemiology of autoimmune diseases. Autoimmun. Rev. 2, 119–125 (2003) 29. R.L. Rabin, A.I. Levinson, The nexus between atopic disease and autoimmunity: a review of the epidemiological and mechanistic literature. Clin. Exp. Immunol. 153, 19–30 (2008) 30. A. Misharin, M. Hewison, C.R. Chen, V. Lagishetty, H.A. Aliesky, Y. Mizutori et al., Vitamin D deficiency modulates Graves’ hyperthyroidism induced in BALB/c mice by thyrotropin receptor immunization. Endocrinology 150, 1051–1060 (2009) 31. H. Yamashita, S. Noguchi, K. Takatsu, E. Koike, T. Murakami, S. Watanabe et al., High prevalence of vitamin D deficiency in Japanese female patients with Graves’ disease. Endocr. J. 48, 63–69 (2001) 32. T. Yasuda, Y. Okamoto, N. Hamada, K. Miyashita, M. Takahara, F. Sakamoto et al., Serum vitamin D levels are decreased and associated with thyroid volume in female patients with newly onset Graves’ disease. Endocrine 42, 739–741 (2012) 33. H. Czernobilsky, S. Scharla, H. Schmidt-Gayk, R. Ziegler, Enhanced suppression of 1,25(OH)2D3 and intact parathyroid hormone in Graves’ disease as compared to toxic nodular goiter. Calcif. Tissue Int. 42, 5–12 (1988) 34. T. Kawakami-Tani, E. Fukawa, H. Tanaka, Y. Abe, I. Makino, Effect of 1 alpha-hydroxy vitamin D3 on serum levels of thyroid hormones in hyperthyroid patients with untreated Graves’ disease. Metabolism 46, 1184–1188 (1997) 35. M.A. Pani, K. Regulla, M. Segni, S. Hofmann, M. Hufner, A.M. Pasquino et al., A polymorphism within the vitamin D-binding protein gene is associated with Graves’ disease but not with Hashimoto’s thyroiditis. J. Clin. Endocrinol. Metab. 87, 2564–2567 (2002) 36. C.E. Jennings, C.J. Owen, V. Wilson, S.H. Pearce, A haplotype of the CYP27B1 promoter is associated with autoimmune Addison’s disease but not with Graves’ disease in a UK population. J. Mol. Endocrinol. 34, 859–863 (2005) 37. E.R. Lopez, O. Zwermann, M. Segni, G. Meyer, M. Reincke, J. Seissler et al., A promoter polymorphism of the CYP27B1 gene is associated with Addison’s disease, Hashimoto’s thyroiditis, Graves’ disease and type 1 diabetes mellitus in Germans. Eur. J. Endocrinol. 151, 193–197 (2004)
123
Endocrine 38. G. Effraimidis, K. Badenhoop, J.G. Tijssen, W.M. Wiersinga, Vitamin D deficiency is not associated with early stages of thyroid autoimmunity. Eur. J. Endocrinol. 167, 43–48 (2012) 39. P.B. Wilson, Serum 25-hydroxy vitamin D status in individuals with psoriasis in the general population. Endocrine 44, 537–539 (2013) 40. J.F. Merola, J. Han, T. Li, A.A. Qureshi, No association between vitamin D intake and incident psoriasis among US women. Arch. Dermatol. Res. 306, 305–307 (2014) 41. P. Gisondi, M. Rossini, C.A. Di, L. Idolazzi, S. Farina, G. Beltrami et al., Vitamin D status in patients with chronic plaque psoriasis. Br. J. Dermatol. 166, 505–510 (2012) 42. J. Orgaz-Molina, A. Buendia-Eisman, M.A. Arrabal-Polo, J.C. Ruiz, S. Arias-Santiago, Deficiency of serum concentration of
123
25-hydroxy vitamin D in psoriatic patients: a case-control study. J. Am. Acad. Dermatol. 67, 931–938 (2012) 43. A.A. Levin, S. Aleissa, N. Dumont, F. Martinez, C. Donovan, S.C. Au et al., A randomized, prospective, sham-controlled study of localized narrow-band UVB phototherapy in the treatment of plaque psoriasis. J. Drugs Dermatol. 13, 922–926 (2014) 44. I. Heier, E. Soyland, A.L. Krogstad, C. Rodriguez-Gallego, M.S. Nenseter, F.L. Jahnsen, Sun exposure rapidly reduces plasmacytoid dendritic cells and inflammatory dermal dendritic cells in psoriatic skin. Br. J. Dermatol. 165, 792–801 (2011) 45. A. Shah, The pathologic and clinical intersection of atopic and autoimmune disease. Curr. Allergy Asthma Rep. 12, 520–529 (2012)
