Vitamin D Deficiency in Older Men
The Journal of Clinical Endocrinology & Metabolism April 2009 Vol. 94, No. 4 1214-1222
Eric Orwoll, Carrie M. Nielson, Lynn M. Marshall, Lori Lambert, Kathleen F. Holton, Andrew R. Hoffman, Elizabeth Barrett-Connor, James M. Shikany, Tien Dam, Jane A. Cauley for the Osteoporotic Fractures in Men (MrOS) Study Group
Bone and Mineral Unit (E.O., C.M.N., L.M.M., L.L., K.F.H.), Oregon Health and Science University, Portland, Oregon 97239; Palo Alto VA Medical Center (A.R.H.), Palo Alto, California 94304; University of California, San Diego (E.B.-C., T.D.), San Diego, California 92093; University of Alabama (J.M.S.), Birmingham, Alabama 35294; and Department of Epidemiology (J..A.C.), University of Pittsburgh, Pittsburgh, Pennsylvania 15260
"Deficiency [25(OH)D <20 ng/ml] was present in 26%, and insufficiency (<30 ng/ml) was present in 72%."
"Other studies reported that aging (2, 27), obesity (28, 29, 30), and winter season (31) are associated with lower 25(OH)D levels.... The winter season considerably increased the proportion classified as deficient.....This information suggests that obese, older men may particularly benefit from vitamin D supplementation, especially in the winter and spring, and that trials are warranted to assess the potentially widespread benefits.....Higher activity level (PASE score) and amount of outside activity were associated with higher 25(OH)D levels....Total 25(OH)D levels were lower in African-Americans (18.5 ± 8.8 ng/ml) than in Caucasians (25.5 ± 7.8 ng/ml). A total of 64.8 and 22.2% of African-Americans had 25(OH)D levels below 20 and 10 ng/ml, respectively, whereas 23.2 and 1.8% of Caucasians were below those levels.. Hispanics also had lower 25(OH)D levels....Total 25(OH)D levels tended to be lower in older men (r = -0.09; P = 0.0002) (Table 2). In men aged 80-84 yr, 31% had levels below 20 ng/ml, and in men at least 85 yr old, 40% were deficient (Fig. 1A). Levels of 25(OH)D3 were lower as age increased (r = -0.10; P < 0.0001), but older men tended to have slightly higher 25(OH)D2 levels (r = 0.11; P = 0.03)........Because older men are commonly afflicted with disorders potentially linked to vitamin D deficiency and some studies have suggested benefit from supplementation, there is a need to further examine the usefulness of vitamin D supplements in clinical trials with adequate power to address multiple outcomes....
...In bivariate analyses, several factors were associated with vitamin D deficiency (<20 ng/ml) (Table 3). After adjustment for these factors, older age, obesity, Black or Hispanic race, and winter or spring season remained independently associated with a greater prevalence of deficiency (Table 4). For instance, the risk of vitamin D deficiency approximately doubled with age of at least 85 yr, BMI of at least 35 kg/m2, Black race, and winter season. Dietary intake of vitamin D of at least 400 IU and lawn work/gardening were both associated with a lower prevalence of deficiency. The additive impact of these risk factors was considerable. In Caucasian men who were older than 80 yr, with a BMI greater than 25 kg/m2, who had a vitamin D intake below 400 IU/d, did not engage in lawn/garden work, and who were sampled in winter or spring, the prevalence of vitamin D deficiency was 86%; whereas there was a prevalence of 24% in younger, thinner men who had higher vitamin D intakes, who were sampled in summer or fall, and who engaged in lawn/garden work."
Context: Vitamin D deficiency is not adequately evaluated in older men.
Objective: The aim of the study was to determine the prevalence of vitamin D deficiency and identify risk factors for its occurrence.
Design and Setting: We conducted a cross-sectional evaluation of 1606 older men in the general community who were enrolled in the Osteoporotic Fractures in Men Study.
Participants: A randomly selected subcohort of a large population of men from six U.S. communities participated in the study. Eligible men were at least 65 yr old, without bilateral hip replacements, and able to walk without assistance from another person.
Main Outcome Measures: Serum concentrations of 25-hydroxyvitamin D2 [25(OH)D2] and 25(OH)D3 were measured using mass spectrometry.
Results: Deficiency [25(OH)D <20 ng/ml] was present in 26%, and insufficiency (<30 ng/ml) was present in 72%. Deficiency was particularly common among men during the winter and spring (especially in the northern communities) and in the oldest and more obese men. For instance, in Caucasian men in winter or spring who were >80 yr old, did not engage in lawn/garden work, and had a body mass index greater than 25 kg/m2 and vitamin D intake below 400 IU/d, the prevalence of vitamin D deficiency was 86%. 25(OH)D2 levels were present in a small fraction of men and accounted for a low proportion of total 25(OH)D levels. The use of vitamin D supplements was reported by 58% of men, but supplement use had a small effect on total 25(OH)D levels and, despite supplement use, low levels remained frequent.
Conclusions: Vitamin D deficiency is common in older men and is especially prevalent in obese, sedentary men living at higher latitudes. Use of vitamin D supplements at levels reported here did not result in adequate vitamin D nutrition.
Vitamin D is necessary for a wide variety of essential biological functions such as bone and mineral metabolism, muscle function, and immunity (1). Moreover, vitamin D insufficiency has been noted to be present in a variety of populations (2) and is implicated in the causation of common disorders including osteoporosis and fractures, falls, cancer, psoriasis, and others (3). Vitamin D deficiency is of particular interest because it is easily, safely, and inexpensively corrected with adequate supplementation. Although findings have been somewhat inconsistent, trials of the effects of vitamin D supplementation suggest beneficial effects on important clinical outcomes (2, 4, 5). As a result, increased attention has been directed at improving vitamin D nutrition, including recent recommendations to increase the routine intake of vitamin D (6).
Complicating the assessment of vitamin D status are concerns about measurement technique. Immunoassay and protein binding based methods have yielded discrepant results (7), and some of these assays may yield spurious levels when compared with liquid chromatography/mass spectrometry (8, 9, 10).
Despite concerns about vitamin D adequacy in the elderly, there remain too few large, community-based studies that use accurate and precise 25-hydroxyvitamin D [25(OH)D] assays to determine the extent of deficiency and identify those at risk. For instance, there are inadequate data about vitamin D levels in older men, a group commonly affected by disorders linked to vitamin D insufficiency (e.g. falls, fractures, cancer). To understand more completely the vitamin D status in a large population of elderly men and identify factors with potential for clinical intervention, we measured levels of 25(OH)D2 (a sterol produced in yeast and obtained primarily from nutritional supplements and supplemented foods) and 25(OH)D3 (obtained from nutritional supplements, supplemented foods, food, and UV irradiation of the skin).
Characteristics of study participants are shown in Table 1. The average age was 73.8 ± 5.9 yr, with a range of 65-99. Most were Caucasian; 10.3% were from other racial and ethnic groups. The average serum total 25(OH)D level was 25.1 ± 7.9 ng/ml (range, 3.1-58.3 ng/ml). A total of 71.1% had 25(OH)D levels below 30 ng/ml, 25.7% had levels below 20 ng/ml, and 2.9% had levels below 10 ng/ml.
A total of 352 men had measures of 25(OH)D in serum obtained on two occasions 1.9 ± 0.4 yr apart. After adjusting for season, there were no statistically significant differences in the means of the two measurements, and differences between measures in individuals were generally small. The partial correlation coefficient for baseline and second visit total 25(OH)D, adjusted for season change, was 0.72 (P < 0.0001). All the 151 men with two measures who were classified as deficient at baseline were also deficient at the second visit.
Almost all circulating 25(OH)D was derived from vitamin D3; only 26% of subjects had detectable levels of 25(OH)D2, and mean levels were low (8.4 ± 3.3 ng/ml) in those in whom it was detectable. The level of total 25(OH)D was weakly associated with the amount of vitamin D reported from supplement use (r = 0.28; P < 0.0001). Those who reported supplement use had slightly higher total 25(OH)D levels and more often had detectable levels of 25(OH)D2 (43.6% vs. 1.5%).
Age, obesity, and race
Total 25(OH)D levels tended to be lower in older men (r = -0.09; P = 0.0002) (Table 2). In men aged 80-84 yr, 31% had levels below 20 ng/ml, and in men at least 85 yr old, 40% were deficient (Fig. 1A). Levels of 25(OH)D3 were lower as age increased (r = -0.10; P < 0.0001), but older men tended to have slightly higher 25(OH)D2 levels (r = 0.11; P = 0.03).
Obesity was also associated with lower total 25(OH)D levels (Table 2). Of those with BMI in the normal range (<25 kg/m2), 21% had 25(OH)D levels below 20 ng/ml; in overweight men (BMI = 25-29 kg/m2), 24.5% had vitamin D deficiency; and in obese men (BMI 30 kg/m2), 33.6% had vitamin D deficiency and 4.3% had levels below 10 ng/ml. Of those who were very obese (BMI 35 kg/m2), 53.3% were deficient (Fig. 1B).
Total 25(OH)D levels were lower in African-Americans (18.5 ± 8.8 ng/ml) than in Caucasians (25.5 ± 7.8 ng/ml). Whereas average 25(OH)D2 levels were similar (27.8% detectable, mean = 9.6 ng/ml, vs. 26.6% detectable, mean = 8.3 ng/ml, respectively), 25(OH)D3 levels were lower in African-Americans (15.8 vs. 23.3 ng/ml). A total of 64.8 and 22.2% of African-Americans had 25(OH)D levels below 20 and 10 ng/ml, respectively, whereas 23.2 and 1.8% of Caucasians were below those levels. Hispanics also had lower 25(OH)D levels (Table 2).
Geography and season
Geography and season did not have major effects on vitamin D levels. Slightly higher total 25(OH)D levels were present in San Diego and Palo Alto, whereas levels were slightly lower in Minneapolis. Moreover, mean total 25(OH)D levels were highest in summer and lowest in winter, although the mean differences between seasons were small (27.5 vs. 22.2 ng/ml, respectively) (Table 2 and Fig. 2). In men with two measurements who had the first one in summer and the second one during winter, there were no differences in mean 25(OH)D2 levels, but average 25(OH)D3 levels were 5.6 ng/ml lower during winter.
Despite the modest seasonal and geographical differences, these factors affected the proportion of men identified with vitamin D deficiency. For instance, the proportion with vitamin D deficiency (<20 ng/ml) was higher in winter (30.9%) than in spring or fall (28.2% and 22.8%, respectively), and was lowest in summer (18.1%). Vitamin D deficiency was more common at the more northern clinical sites. In winter, the mean total 25(OH)D level in Minneapolis was 21.9 ng/ml (37.8% deficient) and in summer was 26..3 ng/ml (13.5% deficient). In contrast, for San Diego the winter mean was 25.0 ng/ml (26.8% deficient), and the summer mean was 30.8 ng/ml (5.0% deficient).
Higher activity level (PASE score) and amount of outside activity were associated with higher 25(OH)D levels (Table 2). For instance, for those who reported no lawn work or gardening, mean total 25(OH)D was 23.2 ng/ml, and 35.2% were deficient; whereas men who reported at least one outdoor yard activity had a mean 25(OH)D of 25.6 ng/ml (P < 0.0001), and 22.6% were deficient.
Vitamin D from diet and supplements
Diet supplied small amounts of vitamin D (mean intake, 163 ± 115 IU/d); only 10 subjects (0.6%) received the recommended daily allowance of 600 IU/d from diet alone (Table 1), and dietary intake correlated weakly with total 25(OH)D levels (r = 0.08). The use of vitamin D supplements was reported by 58.5% of subjects, but none reported taking more than 600 IU/d. Supplements were associated with higher total 25(OH)D levels (Table 2). The 15.5% who reported supplement use had 25(OH)D levels below 20 ng/ml, whereas 39.6% who did not report supplement use had concentrations below that level. Nevertheless, a high proportion had vitamin D insufficiency with all supplement doses (percent of subjects with 25(OH)D levels <30 ng/ml: supplement intake <200 IU/d, 83.6%; 200 to <400 IU, 68.3%; 400 to <600 IU, 69.3%, and 600 IU, 61.3%).
In bivariate analyses, several factors were associated with vitamin D deficiency (<20 ng/ml) (Table 3). After adjustment for these factors, older age, obesity, Black or Hispanic race, and winter or spring season remained independently associated with a greater prevalence of deficiency (Table 4). For instance, the risk of vitamin D deficiency approximately doubled with age of at least 85 yr, BMI of at least 35 kg/m2, Black race, and winter season. Dietary intake of vitamin D of at least 400 IU and lawn work/gardening were both associated with a lower prevalence of deficiency. The additive impact of these risk factors was considerable. In Caucasian men who were older than 80 yr, with a BMI greater than 25 kg/m2, who had a vitamin D intake below 400 IU/d, did not engage in lawn/garden work, and who were sampled in winter or spring, the prevalence of vitamin D deficiency was 86%; whereas there was a prevalence of 24% in younger, thinner men who had higher vitamin D intakes, who were sampled in summer or fall, and who engaged in lawn/garden work.
Subjects and Methods
Participants were enrolled in the Osteoporotic Fractures in Men Study (MrOS). From March 2000 through April 2002, a total of 5995 community-dwelling men at six clinical centers in the United States (Birmingham, AL; Minneapolis, MN; Palo Alto, CA; Monongahela Valley near Pittsburgh, PA; Portland, OR; and San Diego, CA) agreed to participate in a study of healthy aging, with a focus on osteoporosis and fractures. Eligible men were at least 65 yr old, without bilateral hip replacements, and able to walk without assistance from another person. Details of the MrOS design and cohort have been described (11, 12). The Institutional Review Board at each center approved the study, and written consent was obtained from all participants.
Baseline serum for 25(OH)D assays were from 1608 participants who were randomly chosen from 5908 participants in the cohort who had at least one vial of serum available. One participant had insufficient serum to complete the assays. Another man had 25(OH)D levels more than 3 SD above the mean (75.6 ng/ml) and was omitted for analyses, leaving results from 1606 participants.
In addition, assays were performed on a second serum sample from men who volunteered for a follow-up visit an average of 1.9 yr after baseline. The results from those men who were also in the random sample selected at baseline (n = 352) were used to examine the stability of 25(OH)D levels over time.
Fasting morning blood was collected, and serum was prepared immediately after phlebotomy and stored at -70 C in vials foil-wrapped to prevent light exposure. Samples remained frozen until assay. Measures for 25(OH)D2 (derived from ergocalciferol) and 25(OH)D3 (derived from cholecalciferol) were performed at the Mayo Clinic using mass spectrometry as previously described (13). Deuterated stable isotope [d3-25(OH)D] was added to a 0.2-ml serum sample as internal standard. 25(OH)D2, 25(OH)D3, and the internal standard were extracted using acetonitrile precipitation. The extracts were further purified online and analyzed by liquid chromatography/tandem mass spectrometry using multiple reaction monitoring. 25(OH)D2 and 25(OH)D3 were reported individually. The minimum detectable limit was 4 ng/ml for 25(OH)D2 and 2 ng/ml for 25(OH)D3. Aliquots of a single serum pool were included in alternate assay runs. Using the pooled serum, the interassay coefficient of variation for 25(OH)D3 was 4.4%, and the intraassay coefficient of variation was 4.9%. In the study of stability of 25(OH)D levels over time, baseline and follow-up samples were run sequentially within an assay run, and the order of the pair in the run was randomly determined.
Total 25(OH)D was the sum of 25(OH)D3 and 25(OH)D2 levels. Baseline total 25(OH)D levels were categorized as normal, insufficient, or deficient according to the definitions proposed by Holick (3). Deficiency was defined as less than 20 ng/ml, insufficiency as 20-29 ng/ml, and sufficiency as 30-149 ng/ml. No participant had "toxic" levels (>150 ng/ml). We also report the proportion of participants with levels below 10 ng/ml.
Demographic factors, smoking, alcohol consumption, and medications were determined by questionnaire. Physical activity was assessed with the Physical Activity Score for the Elderly (PASE) (14). In PASE, walking for exercise was estimated from the questions "Over the past 7 d, how often did you take a walk outside your home or yard for any reason? For example, for fun or exercise, walking to work, walking the dog, etc.?" During the clinic interview, an additional question was asked: "Do you take walks for exercise, daily or almost every day?" Outdoor activity was determined from the question "During the past 7 d, did you engage in any of the following activities: lawn work or yard care, including snow or leaf removal, wood chopping, etc.? Outdoor gardening?" A modified Block food frequency questionnaire was administered to assess usual dietary and supplement intake over the past year (Block Dietary Data Services, Berkeley, CA) (15). Vitamin D intake from food and supplements was determined, but information did not include whether vitamin D supplements included D3 or D2; hence, only total 25(OH)D content was considered. Participants brought in all medications they used within the last 30 d. A computerized dictionary, based on the original Established Populations for Epidemiologic Studies of the Elderly (EPESE) coding system (16) was used to categorize the medications. Each medication was matched to its ingredient(s) based on the Iowa Drug Information Service (IDIS) Drug Vocabulary (College of Pharmacy, University of Iowa, Iowa City, IA). Height (centimeters) was measured on Harpenden stadiometers, and weight (kilograms) was measured on standard balance beam or digital scales. Body mass index (BMI) was calculated as kilograms per meter squared.
The distributions of 25(OH)D, 25(OH)D3, and detectable 25(OH)D2 levels were approximately normal. However, because the majority of 25(OH)D2 results were undetectable, the variable was examined both as dichotomous (detectable or not) and as continuous if in the detectable range (>4 ng/ml), and both results are reported. Analyses using 25(OH)D2 as a continuous variable used only participants with detectable levels. Pearson correlations were calculated for each 25(OH)D measure and continuous variables collected at baseline, including age, body size measures, and reported intake of vitamin D and calcium from diet and supplements. Means and SD values of baseline total 25(OH)D, 25(OH)D2, and 25(OH)D3 were calculated by participant characteristics, including demographic and physical activity variables, collected at the baseline exam. ANOVA and t tests were used to compare means, and 2 tests were used to compare proportions.
Mean 25(OH)D concentrations were evaluated by season and clinic site. Season of baseline visit was coded as winter (January-March), spring (April-June), summer (July-September), and fall (October-December). Clinic sites also differed in their climate and latitude, which may affect participants' exposure to sunlight; therefore, 25(OH)D levels were evaluated by site. Absolute change in 25(OH)D levels between baseline and a subsequent visit (1.9 ± 0.4 yr later) were calculated for 352 participants who had baseline and follow-up measures, and these levels were normalized for season of collection. Seasonal normalization was accomplished as follows. A five-category "season change" variable was created to describe whether a participant's follow-up visit was: 1) in the same season as baseline; 2) in a "lighter" season than baseline (e.g. baseline visit was in winter and follow-up was in spring); 3) in a "much lighter" season (e..g. baseline visit was in winter and follow-up was in summer), and so on for "dark" and "much darker" season changes. Mean differences between baseline and follow-up in 25(OH)D and 25(OH)D3 were calculated for participants in each of the five categories of the season change variable. The mean differences were added to each follow-up value. For example, participants whose follow-up visit was in a much darker season had an average 5.55 ng/ml lower 25(OH)D3 value at follow-up than at baseline. Therefore, 5.55 ng/ml was added to each participant's 25(OH)D3 value in that season change category. Values in other season-change categories were adjusted similarly.
Factors associated with vitamin D deficiency were identified using log-binomial models (PROC GENMOD). This method provides better estimates of prevalence ratios than does logistic regression for common outcomes, such as vitamin D deficiency (17). That is, logistic regression would be expected to overestimate strengths of association reported in this study (18). Prevalence ratio estimates for deficiency were calculated for each of several baseline characteristics alone and after adjustment for age, season, and race. Upon this basic model adjusted for these three variables, a multivariable model was constructed using each of the variables associated with vitamin D deficiency with a P < 0.10 in bivariate analysis. A likelihood ratio test was used to determine whether model fit improved significantly (P < 0.05) with the addition of new variables. Additional variables were added to the multivariate model in the order of lowest to highest P value (among those with P < 0.05) until no additional variable significantly improved the model fit as assessed by the likelihood ratio test.
All analyses were conducted using SAS version 9.1.2 (SAS Institute, Cary, NC).