Healthy lifestyles are associated with better vitamin D status in community‐dwelling older men: The Health In Men Study (HIMS)

Abstract Objective Older people are more prone to vitamin D deficiency than younger populations. Individual lifestyle factors have been associated with vitamin D status. We examined the influence of a combination of lifestyle factors on vitamin D status in older men. Participants and Measurements In a population‐based cohort study of older men (age ≥65 years), a lifestyle score was calculated from eight prudent health‐related behaviours (smoking, exercise, alcohol, fish and meat consumption, adding salt, milk choices and obesity) collected via questionnaire at baseline. Blood samples were collected 5 years afterwards to measure plasma 25‐hydroxyvitamin D (25OHD) levels. Associations between lifestyles and the likelihood of having plasma 25OHD levels of ≥75 versus <75 nmol/L and ≥50 versus <50 nmol/L were tested using logistic regression models. Results Of the 2717 men analysed, mean plasma 25OHD was 69.0 ± 23.5 nmol/L, with 20.7% having plasma 25OHD <50 nmol/L. Men engaging in ≥4 healthy lifestyle behaviours had 20% higher odds of plasma 25OHD ≥75 nmol/L (adjusted OR = 1.20, 95% CI: 1.01−1.45) compared to those with <4 healthy behaviours. No association was found for 25OHD ≥50 nmol/L. Higher physical activity was the only individual component significantly associated with vitamin D sufficiency (highest vs. lowest quintiles of physical activity, adjusted OR = 2.01, 95% CI: 1.47−2.74 for 25OHD ≥50 nmol/L, adjusted OR = 2.35, 95% CI: 1.81−3.06 for 25OHD ≥75 nmol/L). Conclusion Multiple healthy lifestyle behaviours are associated with better vitamin D status in older men. Further work is needed to determine the effects of promoting healthy lifestyle behaviours, including physical activity, on vitamin D sufficiency.


| INTRODUCTION
Bioactive vitamin D is a hormone that stimulates the gut absorption of calcium and phosphate and is essential for bone and muscle health. 1 There are two major forms of vitamin D. Vitamin D 3 is mainly synthesized in the skin after the exposure to sunlight, whereby 7-dehydrocholesterol in the skin is converted to pre-vitamin D 3 during ultraviolet B radiation and is then immediately converted to vitamin D 3 . 1 Latitude, season, skin pigmentation, use of sunscreen and clothing cover, and outdoor activities that affect sunlight exposure can influence dermal synthesis of vitamin D 3 . 1 Small amounts of vitamin D are found in food such as oily fish, meat, eggs, mushrooms (a source of the other form, vitamin D 2 ) and vitamin D fortified food. 2 The blood concentration of the intermediate metabolite, , is used to determine vitamin D status. Severe deficiency of vitamin D, characterized by 25OHD levels below 25 nmol/L, can have detrimental effects on health, including impaired bone mineralization. 3 Endogenous synthesis of vitamin D declines with aging partly due to a reduction in skin 7-dehydrocholesterol, with effects on calcium absorption further exacerbated by reduced renal production of the bioactive form, 1,25-dihydroxyvitamin D (1,25(OH) 2 D). 4 Furthermore, decreased endogenous synthesis of vitamin D is accentuated in older people with sedentary behaviours and limited sunlight exposure, particularly those who are institutionalized, or house bound. 1 The decrease in vitamin D production and calcium absorption together lead to increased risk of negative calcium balance and bone loss in older-aged populations. 5 In recent years, vitamin D deficiency and insufficiency has been recognized as an important public health issue worldwide. 3 Australia is known for its abundance of sunshine, with most of the population obtaining vitamin D through cutaneous synthesis during sun exposure.
However, the national survey indicated a lower prevalence of vitamin D deficiency (25OHD <50 nmol/L) in older age groups compared to younger groups. 6 This difference can be partially attributed to the higher likelihood of older individuals taking vitamin D supplements. 6 However, a significant number of falls-and hip fracture-related hospitalizations in the elderly Australian population remain attributable to vitamin D deficiency. 7 In Australia, serum 25OHD level ≥50 nmol/L by the end of winter is considered adequate, 8 consistent with recommendations from the Institute of Medicine in United States of America. 9 In contrast, the UK Scientific Advisory Committee on Nutrition recommends a lower limit of vitamin D adequacy, described as a 'population protective' concentration with a serum 25OHD ≥25 nmol/L that should be achieved by 97.5% of the population throughout the year. 10 Other health institutes and organizations advocate a higher threshold of 75 nmol/L for the general population. 11 However, there is limited evidence to support higher vitamin D targets. 8,12 The European Society for Clinical and Economic Aspects of Osteoporosis and Osteoarthritis suggested the oldest old (aged >80 years) and frail older people with increased falls and fractures risk should have serum 25OHD levels of at least 75 nmol/L. 13 Healthy lifestyle behaviours, such as regular exercise, and oily fish consumption, are beneficial for bone and muscle health; and these lifestyle factors have been associated with increased blood levels of 25OHD. 1 However, these factors have been studied individually rather than as part of a multi-lifestyle behavioural pattern. Thus, the aim of this study was to investigate the association between healthy lifestyles (a combination of behaviours) and circulating 25OHD levels in a large population of older, communitydwelling men.

| Lifestyle behaviour and prudent lifestyle score
Lifestyle information was collected via a questionnaire in Wave 1 (1996−1999). Physical activity levels were assessed by asking study participants how much time they spent performing non-vigorous (moderate) and vigorous exercise separately. A metabolic equivalent (MET) value of three was assigned to non-vigorous activity and a MET value of five was assigned to vigorous activity. As such, total MET hours per week were then calculated as the sum of the hours of non-vigorous activity multiplied by three and the hours of vigorous activity multiplied by five. 15  Lifestyle scores were then calculated from eight behaviours, with a score allocated for each 'healthy' behaviour: (1) having never smoked or having stopped smoking more than a year ago; (2) doing minimum of 3 h a week of at least non-vigorous (moderate) physical activity; (3) having less than 6 alcoholic drinks/day and no more than 28/week; (4) eating fish at least three times weekly; (5) eating meat less than six times weekly; (6) never or rarely adding salt to food; (7) always using reduced fat or skim milk; (8) having a measured BMI of <25 kg/m 2 . The resulting scores ranged from 0 to 8, with a higher score representing a healthier lifestyle. The prudent lifestyle score has been validated in previous studies that have shown these lifestyle scores predict survival and vascular diseases in the same population. 16,17 In the present study, scores of 7 and 8 were combined to provide adequate group sizes for each score category. ≥50 nmol/L at the end of winter is the current definition of vitamin D adequacy in Australia. 8 Considering the older age of our study subjects with substantially higher requirements for vitamin D and increased risk of fracture and osteoporosis, plasma 25OHD ≥75 nmol was also investigated in the present analysis.

| Statistical analysis
Descriptive characteristics of the study population were expressed as mean ± SD for continuous variables and n (%) for categorical variables, stratified by plasma 25OHD levels <50, 50−74.9 and ≥75 nmol/L. One-way analysis of variance was used to compare continuous variables (age and BMI) between stratification of plasma 25OHD levels <50, 50−74.9 and ≥75 nmol/L, and the Pearson's χ 2 statistic was applied for comparison of categorical variables. Linear regression analyses were performed to determine the relationship between prudent lifestyle score categories and continuous plasma 25OHD levels.
Results were presented as β and corresponding 95% confidence interval (95% CI). Logistic regression analyses were applied to determine the relationship between prudent healthy lifestyle scores and odds of having plasma 25OHD levels greater than 50 and greater than 75 nmol/L. Results were presented as odds ratio (OR) and 95% CI. Adjustment was made for age, season of blood  Table 1 presents the demographic, anthropometric, clinical and lifestyle characteristics of the study participants. Of the 4248 men who participated and donated a blood sample at Wave 2, excluding men with missing data and those who were not fasted at the time of blood collection, a total of 2717 had complete data for fasted plasma 25OHD levels and lifestyle behaviours and were included in the current analysis. Mean (±SD) age of these men was 76.5 ± 3.5 years.

| Population characteristics
The mean BMI was 26.7 ± 3.2 kg/m 2 at Wave 1 and only 32% of men had a BMI < 25 kg/m 2 .
Sixty-five percent of study participants achieved moderate physical activity of at least 3 h per week. Most men consumed less than 6 alcoholic drinks per day or no more than 28 drinks per week (91.5%) and ate meat less than 6 times per week (68.9%).
Less than half of the men reported salt intake limited to never or rarely adding salt to food (40.2%) and half always used reduced fat milk (53.7%). Only 9.2% men reported fish consumption more than three times per week. and Autumn (12.9%). There were 1163 (42.8%) men with plasma 25OHD levels between 50 and 74.9 nmol/L and 992 (36.5%) with plasma 25OHD levels greater than 75 nmol/L. Men with vitamin D <50 nmol/L were older, had higher BMI, were more likely to have MetS or type 2 diabetes, and were less physically active (Table 1). Table 2 presents the linear and logistic regression analysis between lifestyle score and plasma 25OHD levels and odds of vitamin D sufficiency (≥50 and ≥75 nmol/L) for incremental increases in lifestyle score and when categorized as scores <4 and ≥4. Mean plasma 25OHD levels increased with higher lifestyle scores. In the univariate analysis, men who engaged in six or more healthy lifestyle behaviours had higher plasma 25OHD compared to men with a score of one. This association was no longer significant in the multivariable model. When lifestyle scores were grouped into two categories, men who engaged in four or more healthy lifestyle behaviours had a significantly higher plasma 25OHD compared to men with lifestyle score <4. After adjustment for age, season of blood draw, time difference between baseline survey and blood collection, MetS and vitamin D supplementation, the association was no longer significant.

| Association between lifestyles scores and subsequent vitamin D status
In both univariate and multivariable analysis, higher categories of lifestyle score did not predict thresholds of plasma 25OHD ≥50 or ≥75 nmol/L. When lifestyle scores were grouped into two categories, engagement in four or more healthy lifestyle behaviours was not associated with odds of having subsequent total plasma 25OHD T A B L E 1 Demographic, anthropometric, clinical and lifestyle characteristics of the study population, stratified by vitamin D status.

| Individual lifestyles behaviours and subsequent vitamin D status
Smoking and dietary components of healthy lifestyles, including nonexcessive alcohol consumption, eating more fish, less meat and lower salt intake and use of reduced fat/skim milk, as well as obesity status, were not associated with plasma 25OHD levels, nor odds of plasma 25OHD being ≥50 or ≥75 nmol/L (Table 3).
Physical activity level was the only individual lifestyle behaviour that was associated with subsequent higher plasma 25OHD levels.
When physical activity was expressed as total MET hours per week, higher total MET hours was associated with higher plasma 25OHD concentrations and increased odds of plasma 25OHD ≥50 or ≥75 nmol/L (  MS assay, 19 which may partly explain the higher prevalence of 25OHD <50 nmol/L in our study compared to the nationwide Australian Health Survey. In addition, our study site (Perth: 31°S) was located at a higher latitude and thus is exposed to stronger UV radiation than the CHAMP study site (Sydney: 33.9°S). 6  In our population-based study in a sunny location, 21% of the participants in our study had plasma 25OHD <50 nmol/L, and this proportion was 27% in winter. This is a public health concern, given the higher risk of falls and fractures associated with vitamin D deficiency in older populations. 5 Several vitamin D-specific sun exposure guidelines/recommendations are currently available in Australia. 20 However, Australia has the highest global prevalence of melanoma due to the large proportion of fair-skinned populations residing in high UV exposure T A B L E 3 Association between individual lifestyle behaviour and plasma 25OHD levels and odds of vitamin D deficiency (plasma 25OHD ≥50 or ≥75 nmol/L). areas. Therefore, vitamin D supplementation (on its own or combined with calcium) is also recommended for a high-risk population.
However, evidence from randomized control trials (RCTs) that have considered vitamin D supplementation for falls and fracture prevention has been conflicting. 21  A previous study in the same cohort found that plasma 25OHD <50 nmol/L was associated with increased risk of frailty and all-cause mortality. 23 However, evidence of a beneficial effect of maintaining circulating 25OHD levels higher than 75 nmol/L in this population is lacking, thus optimal vitamin D status remains uncertain and use of vitamin D supplementation in this population should be carefully assessed for benefit versus risk. Our current findings potentially strengthen the 'reverse causality' hypothesis popularized by Naveed Sattar, whereby higher 25OHD levels may be a consequence of, rather than a contributor to, generally better health. 24 We found an association between high physical activity levels Surprisingly, we found no association between frequent fish consumption and vitamin D status in this population. This may be explained by the low overall fish consumption in this population, with only 11.3% men reporting fish intake at least three times per week.
The low consumption of fish was consistent with the reports in the National Nutrition Survey in Australia conducted in 1995. 29 Moreover, the vitamin D contents can vary by fish species and cooking methods, 30,31 and that information was not collected in our questionnaire. Given the positive relationship between oily fish consumption and improved vitamin D status 32 35 Similarly, categorization of milk consumption prioritized avoidance of condensed or full cream milk and did not dissect out avoidance of milk consumption.
Lastly, as only men aged 65 years and older were invited to the HIMS, findings from the present analysis may not apply to women and younger men.