Endocrine DOI 10.1007/s12020-015-0641-7

ORIGINAL ARTICLE

Longitudinal associations between lifestyle and vitamin D: A general population study with repeated vitamin D measurements Tea Skaaby1 • Lise Lotte Nystrup Husemoen1 • Betina Heinsbæk Thuesen1 • Charlotta Pisinger1 • Anke Hannemann2 • Torben Jørgensen1,3,4 • Allan Linneberg1,5,3

Received: 30 March 2015 / Accepted: 22 May 2015 Ó Springer Science+Business Media New York 2015

Abstract Several lifestyle factors have been found to be associated with vitamin D status in cross-sectional studies, but it is not clear whether a change in these factors can actually affect the vitamin D level. We investigated the association between repeated measurements of physical activity, body mass index (BMI), diet, alcohol consumption, and smoking habits, and corresponding levels of vitamin D during 5 years of follow-up of a large general population sample. We included 4185 persons who participated and had vitamin D (serum-25-hydroxyvitamin D, 25-OH-D) measurements in the Inter99 study at baseline (1999–2001) and 5-year follow-up. In a subsample, 25-OH-D was also measured at 1- and 3-year follow-ups. We used mixed models to examine the association between repeated measurements of lifestyle factors and 25-OH-D levels. In multivariable analyses of repeated measurements, the difference in 25-OH-D was -0.32 ng/ml (95 % CI -0.37, -0.28) per 1 kg/m2 increase in BMI; 4.50 ng/ml (95 % CI 3.84, 5.15) for persons moderately/vigorously physically active versus sedentary; 1.82 ng/ml (95 % CI

& Tea Skaaby [email protected] 1

Research Centre for Prevention and Health, Capital Region of Denmark, Copenhagen, Denmark

2

Institute of Clinical Chemistry and Laboratory Medicine, University Medicine Greifswald, Greifswald, Germany

3

Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark

4

Faculty of Medicine, Alborg University, Aalborg, Denmark

5

Department of Clinical Experimental Research, Rigshospitalet, Glostrup, Denmark

1.09, 2.56) for persons with healthy versus unhealthy dietary habits; 0.05 ng/ml (95 % CI 0.03, 0.07) per 1 standard drink/weak increase in alcohol consumption; and 0.86 ng/ml (95 % CI 0.36, 1.35) for never smokers versus daily smokers. Our study shows that lower BMI, a higher level of physical activity, a healthier diet and possibly a higher alcohol intake, and not smoking, are associated with higher 25-OH-D levels. Keywords Vitamin D  Obesity  Physical activity  Smoking  Alcohol  Diet

Introduction Vitamin D deficiency is common worldwide [1]. Vitamin D is produced in sun exposed skin and retrieved from diet and dietary supplements [2]. Regardless of the source, the two major vitamin D metabolites in the blood are 1,25-dihydroxyvitamin D and 25-hydroxyvitamin D (25-OH-D). The former is the biologically active form, whereas the latter is usually measured to determine vitamin D status [2]. Vitamin D has many physiologic effects, and vitamin D deficiency has been linked with a range of disorders [3–7]. Apart from its key role in growth and maintenance of the skeleton, vitamin D affects a great number of genes including genes involved in angiogenesis and cellular proliferation, differentiation, and apoptosis [2]. Circulating vitamin D may play a role in all-cause mortality and diseases like cardiovascular disease (CVD) and diabetes [2, 5, 8–12], chronic obstructive pulmonary disease [13], cancer [14], and autoimmune disease [6]. Apart from the well-known association between season and vitamin D status [15], several lifestyle factors such as

123

Endocrine

obesity, smoking, physical activity, and diet have been found to be predictors of vitamin D status in cross-sectional analyses [16–22]. Hence, Thuesen et al. found that low vitamin D status was significantly associated with obesity, daily smoking, and a sedentary lifestyle [23]. Hintzpeter et al. found vitamin D intake from both diet and supplements and physical activity to be independent positive determinants of vitamin D status in a German survey [18]. Likewise, Hirani et al. found dietary intake of vitamin D and intake of dietary supplements to be positive predictors of vitamin D status in mutually adjusted regression models in two British surveys [19]. Broch et al. found the major modifiable predictors of low vitamin D levels to be low vitamin D dietary and supplement intake, body mass index (BMI) [30 kg/m2, physical inactivity, and low milk and calcium supplement intake in a US survey [16]. In a Canadian study, Greene-Finestone et al. found that BMI C30 kg/m2 and lower vitamin D supplementation were strong cross-sectional predictors of low vitamin D levels for both men and women [17]. In an Australian survey, Pasco et al. found that physical activity was independently and positively associated with vitamin D status, whereas associations with weight and waist–hip ratio were negative among women aged 20–54 years. Among older women, physical activity and vitamin D intake were positively associated and age, weight, and smoking were negatively associated with vitamin D levels [21]. Although some lifestyle factors are cross-sectionally associated with vitamin D status, it is not clear whether a change in these factors can actually affect vitamin D status. The objective of the present study was to investigate the longitudinal associations between lifestyle factors and 5-year longitudinal changes in vitamin D status as assessed by 25-OH-D in a large general population study.

Materials and methods Study population The Inter99 study was a Danish population-based randomized controlled trial (CT00289237, ClinicalTrials.gov) and investigated the effects of lifestyle intervention on CVD. It was carried out in 1999–2001 and included all 61,301 persons aged 30, 35, 40, 45, 50, 55, and 60 years that lived in 11 municipalities in the south-western part of Copenhagen County on 2 December 1998. An age- and sex-stratified random sample of 13,016 persons was drawn from the study population, of whom 12,934 were eligible, and invited for a health examination. A total of 6906 (53 %) persons turned up for the investigation. Out of these, 122 were excluded because of alcoholism, drug abuse, or linguistic barriers leaving and that left 6784

123

(52 %) participants in the study population [24]. The Inter99 study has previously been described in detail [24]. Participants were randomized into a high and a low intensity intervention group (A and B, respectively). In this study, Inter99 data were considered observational, and analyses were adjusted for study group. The health examination included a self-administered questionnaire, various blood tests, and a physical examination that formed the basis of a risk assessment. Accordingly, participants were divided into high or low risk of ischemic heart disease (IHD). Only those at high risk were invited to follow-up visits 1 and 3 years after baseline, whereas all participants were invited to the 5-year follow-up. Measurements followed identical procedures at baseline, 1-, 3-, and 5-year follow-ups. In the current study, we included 4185 persons who participated and had vitamin D measurements at both baseline and 5-year follow-up. Of those, 1636 and 1819 persons participated and had in addition vitamin D measurements at the 1- and 3-year follow-up, respectively (Fig. 1). Lifestyle factors The questionnaires provided self-reported data on education, diet, leisure time physical activity, smoking habits, and alcohol consumption. Height and weight were measured without shoes and with light clothes. BMI was calculated as weight (kg) divided by height (m) squared. The following classifications were used (at each time point): smoking habits (daily smokers, occasional smokers, former smokers, and never smokers), BMI (kg/m2), alcohol consumption (standard drinks per week), and dietary habits according to intake of vegetables, fruit, fish, and saturated fat (healthy, average, or unhealthy) by a validated 198 food frequency questionnaire (FFQ) [25]. The FFQ is a semiquantitative questionnaire with 198 food items and beverages, and the participants were asked to report average intake of different foods and beverages during the last month with 7–11 possible response in the range never to eight or more times per day. The questionnaire had the following 16 sections: number of meals, breakfast, bread and fat spread on bread, cheese, meat and fish, etc., laid on bread, hot meals, accompaniments to hot meals, sauces, fats for cooking, fast food, vegetables, salad dressing, fruits, snacks, candy/ice cream/chocolate, cookies, and beverages. The questionnaire included specific details about the consumption of different types of bread, fats and fish, and the intake frequency of each unit of food was multiplied by standard portion sizes. The dietary score was developed as a crude index of the overall quality of the dietary habits. Participants were asked to categorize their leisure time physical activity as sedentary, light, moderate, or vigorous according to the following categories,

Endocrine

Time point

Baseline

The Inter99 study population (ninvited=12934)

Questionnaire Health examination Counselling (n=4185a)

follow-up: dietary habits were classified as less healthy, unchanged, or healthier; physical activity was classified as less active, unchanged, or more active; smoking habits were classified as smoking quitters, persistent nonsmokers, or persistent smokers (this category also includes new smokers); BMI was classified as lower ([2 kg/m2 decrease), unchanged (B2 kg/m2 change), or higher ([2 kg/m2 increase); and alcohol intake was classified as decreased, unchanged, or increased. Vitamin D

1 year

3 years

5 years

Questionnaire Health examination Counselling (n=1636b)

Questionnaire Health examination Counselling (n=1819c)

Questionnaire Health examination Counselling (n=4185a)

Fig. 1 Flowchart of the study population. aNumber of persons who participated and had vitamin D measurements at both baseline and 5-year follow-up. bNumber of persons with vitamin D measurement at 1-year follow-up among the 4185 included persons. cNumber of persons with vitamin D measurement at 3-year follow-up among the 4185 included persons

respectively. (1) Mainly reading or watching television, going to the cinema, or doing other sedentary activities in your spare time. (2) Walking, biking, or otherwise physically active at least 4 h per week (gardening, domestic work, table tennis, and bowling). (3) Regular physical activity at least 3 times per week (incl. heavy gardening and other types of hard leisure time work). (4) Regular hard physical training for competitive sports or long distance running several times per week. We combined the groups with moderate or vigorous leisure time physical activity, since the group with vigorous leisure time physical activity group was small. In supplementary analyses, the lifestyle factors were classified according to changes from baseline to 5-year

Levels of 25-OH-D were measured on the IDS-iSYS multidiscipline automated analyzer (Immunodiagnostic Systems Limited) with the IDS-iSYS 25-OH vitamin D assay. Serum samples at each time point were kept without thawing until laboratory analyses in 2013–2014. The coefficients of variation were 14.5 % at low, 8.69 % at medium, and 7.5 % at high concentrations, respectively. Values below the detectable level of 5 ng/ml were given the value 5 ng/ml (Nbaseline = 86, N1_year = 33, N3_years = 21, N5_years = 39). Other covariates Other covariates were classified as follows: Baseline age (30–35 years, 40–50 years, and 55–60 years; season at each time point (March–May, June–August, September– November, or December–February); baseline education (only basic education, low, medium, high, or missing value including students); and time point (0, 1, 3, or 5 years after baseline). Statistical analyses We performed the analyses with SAS, version 9.4 (SAS Institute Inc. Cary, NC USA). p values were two-tailed, and we defined statistical significance as a p value below 0.05. Descriptive characteristics of the participants are presented as % (n) and median (p25, p75) and compared with the Kruskal–Wallis test (p values in the results). We used mixed models with unstructured covariance matrices (where the correlation between each time pair is different) to model the association between baseline, 1-, 3-, and 5-year levels of the five different lifestyle factors and repeated vitamin D measurements at the same time points (Table 2). The model allows for an unbalanced design, so we were able to include not only data from the baseline and 5-year follow-up that all included persons attended, but also data from the 1- and 3-year follow-ups that subsamples of the included persons attended. In the mixed model analyses, we used complete case analysis where only participants with complete information on all considered

123

Endocrine

variables at each time point were included. Model 1 is adjusted for time point and season of blood sample at each time point. Model 2 is further adjusted for gender, baseline age, baseline level of education and randomization status (group A or B), and (at each time point) dietary habits, physical activity, smoking habits, BMI, and alcohol consumption. In additional mutually adjusted mixed model regression analyses, we investigated the association between the changes in BMI, physical activity, diet, smoking, and alcohol consumption, respectively, and vitamin D levels at the baseline and 5-year follow-up. The analyses were adjusted for the baseline value of the outcome in question, except for change in smoking habit where the change variable included the baseline information. Hence, data at 1- and 3-year follow-ups were excluded here. Least square means (LSmeans), that are means estimated from a linear model, and corresponding p values for the interaction between time and the exposure in question are shown in Figs. 2 and 3. This interaction was used to compare the distributions of vitamin D status over time for each of the groups in the exposure in question, e.g., for persons with decreased, unchanged, or increased BMI etc. A statistically significant interaction between time and the exposure in question was interpreted as an indication of an association between changes in the exposure and subsequent changes in vitamin D status. The figures were made in Excel (Microsoft Office 2007, United States).

Results Crude analyses (Table 1) showed that baseline vitamin D levels were higher among women compared to men (p = 0.005), among older age groups (p = 0.001), and in persons with higher intake of alcohol (p \ 0.0001), a healthier diet (p \ 0.0001), and among more physically

Fig. 2 Least square means (LSmeans) of vitamin D at baseline and 5-year follow-up according to change in body mass index (BMI)

123

active (p \ 0.0001). Vitamin D levels were also unequally distributed across levels of education (p \ 0.0001), BMI (p \ 0.0001), and smoking habits (p \ 0.0001) with a tendency toward a negative dose–response relation between amount of tobacco and vitamin D levels among daily smokers. At baseline, the calcium (total) and parathyroid hormone mean levels (standard error) were 2.2 mmol/l (0.2 mmol/l) and 26.7 pg/ml (10.0 pg/ml), respectively. In multivariable analyses of repeated measurements, we found that vitamin D decreased with increasing BMI and increased tobacco consumption over time (Table 2). As opposed to this, we found that vitamin D increased with increasing intake of alcohol over time (Table 2). There was a statistically significant trend into a higher vitamin D level with increased physical activity (Table 2). Likewise, there was a statistically significant trend into a higher vitamin D for persons who improved their diet (Table 2). Additional analyses as visualized in Figs. 2 and 3 showed statistically significant interactions with time for changes in BMI and physical activity, but no statistically significant interaction with time for intake of alcohol, dietary, or smoking habits (data not shown).

Discussion In a large prospective general population study, we found statistically significant longitudinal associations between diet, BMI, physical activity, alcohol consumption, and smoking habits, and vitamin D status as assessed by repeated measurements. Thus, improvement in diet, decreased BMI, increased physical activity, quitting smoking, and increased alcohol consumption were independently associated with an increased 25-OH-D during a 5-year period. Except for the association between changes in intake of alcohol, dietary and smoking habits, and vitamin D status, these findings were supported by additional analyses

Endocrine Fig. 3 Least square means (LSmeans) of vitamin D at baseline and 5-year follow-up according to change in physical activity

of interactions between time and exposure regarding vitamin D status. Our results extend the results from previous studies by investigating the longitudinal associations between lifestyle and vitamin D as assessed by repeated measurements. In line with previous studies, vitamin D status was generally higher in women than men [26, 27]. The observed inverse longitudinal association between BMI and vitamin D levels corroborates the findings by Vimaleswaran et al. in a large bidirectional Mendelian randomization study that found a higher BMI to lead to lower 25-OH-D (but not vice versa) [28]. This is supported by Levy et al. who found that BMI was the second largest (and the largest negative) determinant of vitamin D status in persons consuming a vitamin D supplement [20]. Persons with a high BMI generally have a higher content of body fat that may both serve as a storage reservoir but is also extremely slow to release vitamin D [20]. Low circulating vitamin D levels in obese persons may therefore result from a dilution effect of vitamin D within the large body mass [20]. In line with the observed indication of an association between never smoking versus daily smoking and increased vitamin D status, Cutillas-Marco et al. found that smoking was associated with an increased risk of vitamin D deficiency which was also found by Supervı´a et al. although among female smokers only [29, 30]. However, Supervı´a et al. found statistically significant decreases of 25-OH-D levels after smoking cessation in both genders [29]. The apparent contradiction of these findings may partly be explained by the fact that smoking cessation is often accompanied by weight gain which again may decrease vitamin D status (see above), and thus it is important to take weight changes into account when assessing the association between smoking and vitamin D status. Regarding the observed positive association between alcohol intake and vitamin D status, Lee found that increase in three alcohol-related behaviors (drinking frequency, number of alcoholic drinks consumed, and average

daily alcohol intake) was linearly associated with increase in vitamin D in men but not in women [31]. In comparison, although the effect was lower in women, we found statistically significant associations in both genders with vitamin D (ng/ml) beta-estimates of 0.107 (95 % CI 0.064, 0.150) and 0.039 (95 % CI 0.020, 0.058) in men and women, respectively, per increase of 1 standard drink per week. Although excessive alcohol consumption may have a negative effect on vitamin D status due to a disturbed vitamin D metabolism [32, 33], experimental studies in rats show that chronic ethanol exposure increases the serum levels of 25-OH-D thereby suggesting a possible higher vitamin D status after long-term alcohol consumption [34]. However, Turner et al. found that although ethanol treatment increased serum 25-OH-D, serum 1,25-OH-D was actually decreased [35]. As 1,25-OH-D is the active metabolite, this finding calls for caution. Indeed, 25-OH-D may not be an appropriate marker of vitamin D status when assessing an effect of alcohol intake. The observed association between higher levels of physical activity and higher vitamin D status was independent of changes in BMI which suggests that the effect of physical activity on vitamin D status is not only mediated through BMI. The association may be explained by the fact that physically active spend more time outdoors, thereby exposing them to UV radiation that produces vitamin D in the skin [2, 36]. Also, it is possible that the distribution of fat is not well captured by BMI, i.e., that the more physically active contain more lean body mass that does not sequester vitamin D like fat mass compared to the sedentary regardless of BMI [36]. Dietary vitamin D only contributes a small part of the body vitamin D, as the majority is produced in the skin following sun exposure [37, 38]. As illustrated in a study by Jenab et al., fish and shellfish were the major dietary contributors of vitamin D in both men and women, but there were both important gender and country differences

123

Endocrine Table 1 Serum vitamin D at the four different examinations according to baseline characteristics Participants % (n)

Vitamin D (ng/ml), median (p25, p75)

Baseline

Baseline

1 yeara

3 yearsa

5 years

Male

50.9 (2131)

20.1 (13.4, 28.1)

22.0 (14.7, 29.9)

23.5 (17.1, 31.7)

23.3 (16.1, 31.0)

Female

49.1 (2054)

21.5 (14.8, 28.9)

22.5 (15.3, 30.0)

24.7 (17.5, 33.6)

25.7 (17.8, 32.9)

30–35

12.4 (519)

20.8 (12.8, 28.9)

19.6 (11.1, 30.0)

22.6 (13.4, 29.0)

22.7 (15.3, 31.8)

40–50

62.8 (2626)

20.4 (13.8, 28.1)

22.0 (14.8, 29.8)

23.6 (16.9, 31.8)

24.0 (16.2, 31.6)

55–60

24.8 (1040)

22.1 (15.3, 29.2)

22.9 (16.2, 30.6)

26.0 (19.1, 34.0)

26.4 (19.1, 33.0)

Study point Gender

Age (years)

Education Noneb

13.0 (544)

19.3 (12.5, 27.7)

19.9 (13.3, 28.8)

22.1 (14.6, 31.2)

24.2 (15.5, 32.4)

Low Medium

27.6 (1153) 42.9 (1797)

21.6 (14.4, 29.3) 20.9 (14.9, 28.6)

23.4 (16.0, 30.8) 22.1 (15.6, 30.5)

24.2 (18.4, 32.7) 24.7 (18.1, 33.6)

24.9 (17.1, 31.9) 24.8 (17.6, 32.0)

9.8 (412)

20.9 (14.0, 27.7)

20.6 (12.7, 28.3)

24.2 (17.1, 29.5)

24.2 (16.7, 32.1)

High Body mass index (kg/m2) \18.5

0.8 (33)

17.9 (11.5, 31.4)

13.7 (8.3, 25.7)

21.9 (10.8, 31.1)

25.5 (13.9, 33.1)

18.5–24.9

43.6 (1823)

22.1 (15.2, 30.1)

23.0 (15.6, 31.7)

26.2 (19.5, 34.2)

25.7 (17.7, 33.1)

25–29.9

40.4 (1688)

20.8 (14.2, 28.2)

23.1 (15.5, 30.8)

24.0 (17.4, 32.5)

24.2 (16.8, 31.7)

C30

15.3 (639)

17.7 (11.9, 24.9)

20.9 (13.8, 27.7)

22.0 (16.2, 29.3)

22.1 (15.1, 28.8)

Sedentary

19.5 (801)

17.9 (11.3, 25.1)

19.8 (12.3, 27.7)

22.1 (15.1, 30.6)

21.8 (14.7, 29.7)

Light

62.4 (2571)

21.2 (14.6, 28.5)

22.6 (15.4, 29.9)

24.2 (18.0, 32.2)

24.7 (17.2, 31.9)

Moderate/vigorous

18.1 (745)

23.7 (16.2, 31.6)

25.7 (17.1, 34.1)

27.0 (18.1, 35.1)

26.7 (19.6, 34.4)

Physical activity

Diet Unhealthy

14.2 (577)

18.7 (12.1, 27.8)

20.5 (13.5, 28.9)

22.2 (16.0, 31.1)

21.9 (14.6, 29.8)

Average

70.3 (2849)

20.7 (14.1, 28.4)

22.0 (14.8, 29.9)

24.4 (17.6, 32.3)

24.8 (17.1, 31.9)

Healthy

15.5 (628)

23.6 (16.9, 30.4)

25.4 (18.9, 33.0)

27.3 (18.7, 34.2)

25.9 (18.7, 33.9)

Alcohol (drinks/week) 0

18.0 (330)

18.0 (12.0, 25.6)

19.0 (10.7, 25.8)

22.5 (14.4, 28.5)

20.8 (14.4, 28.9)

B7

20.7 (1820)

20.7 (14.1, 27.6)

22.4 (14.7, 29.2)

24.0 (17.3, 32.5)

24.5 (17.4, 31.9)

B14

21.8 (931)

21.8 (15.1, 29.9)

23.0 (16.0, 31.0)

26.1 (18.4, 33.3)

25.6 (18.4, 32.6)

[14

22.4 (948)

22.4 (15.2, 31.2)

23.2 (15.7, 31.7)

24.6 (18.2, 32.3)

24.6 (16.5, 33.2)

Never smoker

40.1 (1665)

20.8 (14.5, 28.2)

21.8 (14.3, 29.1)

23.8 (16.8, 32.1)

24.6 (17.0, 31.8)

Former smoker

27.8 (1152)

22.0 (15.1, 29.6)

24.0 (16.0, 32.4)

24.7 (19.4, 32.8)

25.6 (18.2, 33.1)

3.9 (161)

20.9 (15.4, 30.2)

21.3 (15.5, 30.0)

26.9 (20.0, 32.0)

23.3 (18.5, 31.8)

Current 0–\15

9.4 (391)

21.4 (14.3, 29.4)

23.2 (16.5, 31.5)

24.9 (18.2, 33.1)

25.9 (17.1, 32.8)

Current 15–\25

14.0 (580)

19.2 (12.5, 27.4)

21.9 (14.7, 29.5)

23.3 (16.9, 32.2)

22.1 (14.8, 31.3)

4.8 (199)

18.0 (10.6, 26.6)

18.5 (12.3, 29.3)

21.3 (15.2, 31.8)

19.9 (14.4, 29.1)

Smoking habits (g/day)

Occasional smoker

Current C25

p25 25-percentile, p75 75-percentile a

Observe that these time points include fewer participants (see Fig. 1)

b

Only basic education

in the vitamin intake [38]. Our data suggest that changes to a healthier diet—as assessed by intake of vegetables, fruit, fish, and saturated fat—may increase vitamin D status. The association was still statistically significant after adjustment for changes in BMI which suggests that diet has an effect on vitamin D status that is not mediated by BMI.

123

The strengths of our study include the longitudinal population-based design, inclusion of both genders, a 5-year follow-up; the detailed information on covariates and the multivariable analyses, and the use of repeated measurements of vitamin D, which more clearly depicts the actual trend than only a single measurement. Also, vitamin

Endocrine Table 2 Longitudinal associations between lifestyle factors and vitamin D level (Nobservations = 10,674) Estimate (95 % CI) of vitamin D (ng/ml) Model 1a

Model 2b

Body mass index, per 1 kg/m2 higher BMI

-0.35 (-0.39, -0.31)

-0.32 (-0.37, -0.28)

p value

p \ 0.0001

p \ 0.0001

Physical activity Sedentary

0 (Reference)

0 (Reference)

Light

2.69 (2.17, 3.21)

1.82 (1.29, 2.34)

Moderate/vigorous p valuec Diet

5.44 (4.80, 6.09)

4.50 (3.84, 5.15)

p \ 0.0001

p \ 0.0001

Unhealthy

0 (Reference)

0 (Reference)

Average

1.37 (0.76, 1.98)

0.57 (-0.04, 1.19)

Healthy

3.34 (2.62, 4.05)

1.82 (1.09, 2.56)

p valuec

p \ 0.0001

p \ 0.0001

Alcohol intake, per 1 standard drink/weak higher intake

0.03 (0.01, 0.05)

0.05 (0.03, 0.07)

p value

p = 0.0004

p \ 0.0001

Smoking habits Daily smokers

0 (Reference)

0 (Reference)

Occasional smokers

1.45 (0.39, 2.51), p = 0.007

1.52 (0.49, 2.56), p = 0.039

Former smokers

1.21 (0.71, 1.72), p \ 0.0001

1.34 (0.83, 1.84), p \ 0.0001

Never smokers

0.51 (0.02, 1.00), p = 0.039

0.86 (0.36, 1.35), p = 0.007

p value

p \ 0.0001

p \ 0.0001

BMI body mass index, CI confidence interval a

Adjusted for time point and season

b

Adjusted for time, season, gender, age, educational level, group, intake of alcohol, diet, smoking habits, physical activity, and BMI

c

p value for trend

D status was measured in the same laboratory and by the same method at all time points. The limitations of our study include the fact that—in contrast to the baseline and 5-year follow-up that included all participants—only high-risk participants were examined 1 and 3 years after baseline. In addition, we used the Inter99 data as observational, although it is actually an interventional study, but we adjusted for changes in lifestyle factors and for study group (see Table 2, model 2). The lack of information about vitamin D supplements would most likely have attenuated any true association (participants were not informed of their vitamin D status, since serum was stored, and vitamin D status was measured several years after the examination). A substantial amount of the included persons had vitamin D levels corresponding to deficiency and insufficiency according to the Endocrine Society Guidelines [39]. Another limitation is the risk of residual confounding and reverse causation inherent in an observational study. However, we minimized the risk of reverse causation by examining longitudinal associations of lifestyle factors and vitamin D. The use of self-reported

exposure data may introduce recall bias and social desirability bias, and non-participation/loss to follow-up may have biased the estimates and generalizability of the results. Also, the classification of covariates into categories may lead to misclassification. This will probably attenuate the estimates. Micronutrient deficiencies are a global health concern. In particular, vitamin D insufficiency and deficiency are common worldwide [2] with almost 80 % of the general population exhibiting suboptimal levels [36]. The Institute of Medicine (IOM) has set the dietary requirements of vitamin D at 600 IU/day. Even in the absence of sun exposure, this amount of vitamin D will result in a serum 25-OH-D level [20 ng/ml, which is considered adequate for bone health. Whereas vitamin D supplements and UV B exposure of the skin are known to successfully increase vitamin D levels [40], the possible effect of changes in lifestyle factors is unresolved. We found that a self-reported higher level of physical activity, a healthier diet, a decrease in BMI and possibly not smoking, and a higher intake of alcohol were

123

Endocrine

associated with increases in 25-OH-D level. The positive association between changes in alcohol consumption and vitamin D status is, however, questionable, since a higher alcohol intake may instead cause a decreased level of the active metabolite, 1,25-OH-D, and it needs further investigation. The possibility that beneficial changes in diet, physical activity, BMI, and smoking can improve vitamin D status is of interest for preventive purposes. Acknowledgments The present study was financially supported by the Health Insurance Foundation (Grant No. 2010 B 131). We would like to thank the participants and all members of the Inter99 staff at Research Centre for Prevention and Health. The Inter99 study was initiated by Torben Jørgensen, DMSci (principal investigator); Knut Borch-Johnsen, DMSci, (co-principal investigator); Troels Thomsen, PhD; and Hans Ibsen, DMSci. The Steering Committee comprises Torben Jørgensen and Charlotta Pisinger, PhD, MPH. Conflict of interest of interest

The authors declare that they have no conflict

Ethical standards The study was approved by the local Ethics Committees and the Danish Data Protection Agency, participants gave their informed written consent, and the recommendations of the Declaration of Helsinki were followed.

References 1. J. Hilger, A. Friedel, R. Herr, T. Rausch, F. Roos, D.A. Wahl et al., A systematic review of vitamin D status in populations worldwide. Br. J. Nutr. 111, 23–45 (2013) 2. M.F. Holick, Vitamin D deficiency. N. Engl. J. Med. 357, 266–281 (2007) 3. 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) 4. 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) 5. T. Skaaby, The relationship of vitamin D status to risk of cardiovascular disease and mortality. Dan. Med. J. 61, 2 (2015) 6. T. Skaaby, L. Husemoen, B. Thuesen, A. Linneberg, Prospective population-based study of the association between vitamin D status and incidence of autoimmune disease. Endocrine (2015). doi:10.1007/s12020-015-0547-4 7. E. Theodoratou, I. Tzoulaki, L. Zgaga, J.P. Ioannidis, Vitamin D and multiple health outcomes: umbrella review of systematic reviews and meta-analyses of observational studies and randomised trials. BMJ 348, g2035 (2014) 8. L.L. Husemoen, T. Skaaby, B.H. Thuesen, T. Jorgensen, R.V. Fenger, A. Linneberg, Serum 25(OH)D and incident type 2 diabetes: a cohort study. Eur. J. Clin. Nutr. 66, 1309–1314 (2012) 9. 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) 10. 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)

123

11. 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) 12. 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) 13. 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) 14. L. Yin, J.M. Ordonez-Mena, T. Chen, B. Schottker, V. Arndt, H. Brenner, Circulating 25-hydroxyvitamin D serum concentration and total cancer incidence and mortality: a systematic review and meta-analysis. Prev. Med. 57, 753–764 (2013) 15. E. Klingberg, G. Olerod, J. Konar, M. Petzold, O. Hammarsten, Seasonal variations in serum 25-hydroxy vitamin D levels in a Swedish cohort. Endocrine (2015). doi:10.1007/s12020-015-0548-3 16. K. Brock, W.Y. Huang, D.R. Fraser, L. Ke, M. Tseng, R. Stolzenberg-Solomon et al., Low vitamin D status is associated with physical inactivity, obesity and low vitamin D intake in a large US sample of healthy middle-aged men and women. J. Steroid Biochem. Mol. Biol. 121, 462–466 (2010) 17. L.S. Greene-Finestone, C. Berger, M. de Groh, D.A. Hanley, N. Hidiroglou, K. Sarafin et al., 25-Hydroxyvitamin D in Canadian adults: biological, environmental, and behavioral correlates. Osteoporos. Int. 22, 1389–1399 (2011) 18. B. Hintzpeter, G.B. Mensink, W. Thierfelder, M.J. Muller, C. Scheidt-Nave, Vitamin D status and health correlates among German adults. Eur. J. Clin. Nutr. 62, 1079–1089 (2008) 19. V. Hirani, A. Mosdol, G. Mishra, Predictors of 25-hydroxyvitamin D status among adults in two British national surveys. Br. J. Nutr. 101, 760–764 (2009) 20. M.A. Levy, T. McKinnon, T. Barker, A. Dern, T. Helland, J. Robertson et al., Predictors of vitamin D status in subjects that consume a vitamin D supplement. Eur. J. Clin. Nutr. 69, 84–89 (2015) 21. J.A. Pasco, M.J. Henry, G.C. Nicholson, S.L. Brennan, M.A. Kotowicz, Behavioural and physical characteristics associated with vitamin D status in women. Bone 44, 1085–1091 (2009) 22. B.H. Thuesen, T. Skaaby, L.L. Husemoen, M. Fenger, T. Jorgensen, A. Linneberg, The association of serum 25-OH vitamin D with atopy, asthma, and lung function in a prospective study of Danish adults. Clin. Exp. Allergy 45, 265–272 (2015) 23. B. Thuesen, L. Husemoen, M. Fenger, J. Jakobsen, P. Schwarz, U. Toft et al., Determinants of vitamin D status in a general population of Danish adults. Bone 50, 605–610 (2012) 24. 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) 25. U. Toft, L. Kristoffersen, S. Ladelund, A. Bysted, J. Jakobsen, C. Lau et al., Relative validity of a food frequency questionnaire used in the Inter99 study. Eur. J. Clin. Nutr. 62, 1038–1046 (2008) 26. A. Ramnemark, M. Norberg, U. Pettersson-Kymmer, M. Eliasson, Adequate vitamin D levels in a Swedish population living above latitude 63 degrees N: The 2009 Northern Sweden MONICA study. Int. J. Circumpolar. Health. 74, 27963 (2015) 27. A.J. Voipio, K.A. Pahkala, J.S. Viikari, V. Mikkila, C G. Magnussen, N. Hutri-Kahonen et al., Determinants of serum 25(OH)D concentration in young and middle-aged adults. The Cardiovascular Risk in Young Finns Study. Ann Med. 1–10 (2015). doi:10.3109/07853890.2015.1020860

Endocrine 28. K.S. Vimaleswaran, D.J. Berry, C. Lu, E. Tikkanen, S. Pilz, L.T. Hiraki et al., Causal relationship between obesity and vitamin D status: bi-directional Mendelian randomization analysis of multiple cohorts. PLoS Med. 10, e1001383 (2013) 29. A. Supervia, X. Nogues, A. Enjuanes, J. Vila, L. Mellibovsky, S. Serrano et al., Effect of smoking and smoking cessation on bone mass, bone remodeling, vitamin D, PTH and sex hormones. J. Musculoskelet. Neuronal. Interact. 6, 234–241 (2006) 30. E. Cutillas-Marco, A. Fuertes-Prosper, W.B. Grant, M. MoralesSuarez-Varela, Vitamin D deficiency in South Europe: effect of smoking and aging. Photodermatol. Photoimmunol. Photomed. 28, 159–161 (2012) 31. K. Lee, Sex-specific relationships between alcohol consumption and vitamin D levels: The Korea National Health and Nutrition Examination Survey 2009. Nutr. Res. Pract. 6, 86–90 (2012) 32. 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) 33. T. Skaaby, L.L. Husemoen, A. Linneberg, Does liver damage explain the inverse association between vitamin D status and mortality? Ann. Epidemiol. 23, 812–814 (2013) 34. M. Gascon-Barre, Interrelationships between vitamin D3 and 25-hydroxyvitamin D3 during chronic ethanol administration in the rat. Metabolism. 31, 67–72 (1982) 35. R.T. Turner, R.C. Aloia, L.D. Segel, K.S. Hannon, N.H. Bell, Chronic alcohol treatment results in disturbed vitamin D

36.

37.

38.

39.

40.

metabolism and skeletal abnormalities in rats. Alcohol. Clin. Exp. Res. 12, 159–162 (1988) J.P. McClung, E. Gaffney-Stomberg, J.J. Lee, Female athletes: a population at risk of vitamin and mineral deficiencies affecting health and performance. J. Trace Elem. Med Biol. 28, 388–392 (2014) A.C. Ellis, J.A. Alvarez, B.A. Gower, G.R. Hunter, Cardiorespiratory fitness in older adult women: relationships with serum 25-hydroxyvitamin D. Endocrine 47, 839–844 (2014) M. Jenab, S. Salvini, C.H. van Gils, M. Brustad, S. ShakyaShrestha, B. Buijsse et al., Dietary intakes of retinol, beta-carotene, vitamin D and vitamin E in the European Prospective Investigation into Cancer and Nutrition cohort. Eur. J. Clin. Nutr. 63(Suppl 4), S150–S178 (2009) M.F. Holick, N.C. Binkley, H.A. Bischoff-Ferrari, C.M. Gordon, D.A. Hanley, R.P. Heaney et al., Evaluation, treatment, and prevention of vitamin D deficiency: an Endocrine Society clinical practice guideline. J. Clin. Endocrinol. Metab. 96, 1911–1930 (2011) Z. Lagunova, A.C. Porojnicu, L. Aksnes, M.F. Holick, V. Iani, O.S. Bruland et al., Effect of vitamin D supplementation and ultraviolet B exposure on serum 25-hydroxyvitamin D concentrations in healthy volunteers: a randomized, crossover clinical trial. Br. J. Dermatol. 169, 434–440 (2013)

123

Longitudinal associations between lifestyle and vitamin D: A general population study with repeated vitamin D measurements.

Several lifestyle factors have been found to be associated with vitamin D status in cross-sectional studies, but it is not clear whether a change in t...
567KB Sizes 1 Downloads 10 Views