Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Mar 1.
Published in final edited form as: Am J Hum Biol. 2021 Jul 2;34(3):e23636. doi: 10.1002/ajhb.23636

Vitamin D Deficiency and Insufficiency in Hawaii: Levels and Sources of Serum Vitamin D in Older Adults

Caryn E Oshiro a, Teresa A Hillier a,b, Grant Edmonds c, Missy Peterson c, Patrick L Hill d, Sarah Hampson c
PMCID: PMC8720322  NIHMSID: NIHMS1718770  PMID: 34213035

Abstract

Objective:

To examine the major sources of vitamin D [25-hydroxyvitamin D (25(OH)D)] and evaluate their collective role on rates of vitamin D deficiency/insufficiency among older adults.

Methods:

Cross-sectional analysis of the relationship between serum 25(OH)D levels and sources of vitamin D (self-reported and objectively validated sun exposure, supplementation, food including fortified sources). Study subjects were part of the Hawaii Longitudinal Study of Personality and Health who completed a clinic visit between 55 – 65 years (M = 59.6) and food frequency questionnaire, and provided serum to assay 25(OH)D (n = 223).

Results:

Although mean serum 25(OH)D levels were overall sufficient (34.3 ng/mL, [SD = 10.9]), over one-third of participants (38%) had vitamin D deficiency/insufficiency (< 30 ng/ml). Asians were the most likely to be insufficient and Filipinos were the least likely (43% vs. 11%, respectively). Overall, supplement use and sun exposure were both associated with higher 25(OH)D levels and lower risk of vitamin D deficiency/insufficiency. Moreover, Vitamin D sources varied by race/ethnic groups. In multivariate models, higher BMI, being Asian or Native Hawaiian/Pacific Islander, low supplement use, and low sun exposure were associated with higher risk for vitamin D deficiency/insufficiency (< 30 ng/ml).

Conclusions:

Over 1/3 of the older adult sample was vitamin D deficient/insufficient, despite most of the participants living in a tropical climate with year-round access to sun as a vitamin D source. Sun exposure and supplement use, but not food intake, explained differences in vitamin D deficiency/insufficiency in this population.

Keywords: Vitamin D Insufficiency, Sun exposure, Dietary supplements, Skin tone, Aging

INTRODUCTION

Most experts agree that vitamin D [25-hydroxyvitamin D[25(OH)D]] levels below 30 ng/ml are insufficient.(Holick et al., 2011) Vitamin D insufficiency in the US, Canada, and Europe, has been estimated to range from 20 – 100% in older adults.(Holick et al., 2011; Meehan & Penckofer, 2014) The risk for vitamin D insufficiency increases with aging due to a decreased ability of the skin to synthesize vitamin D,(Laird et al., 2018; MacLaughlin & Holick, 1985) and a reduction in intestinal absorption of nutrients.(Brownie, 2006; Laird et al., 2018) Overall dietary quality may be suboptimal due to limited intake of a variety of foods among older adults.(Laird et al., 2018; Zhu et al., 2010) Few studies have objectively examined multiple major sources of vitamin D, including supplementation and sun exposure, for the purpose of understanding vitamin D insufficiency in older adults.(Ginter et al., 2013; M. G. Kimlin et al., 2014; Nagasaka et al., 2018) The current study evaluates the role of sun exposure and vitamin D intake from food and supplements on serum 25(OH)D levels, accounting for total calories consumed, vitamin D fortification, sunscreen use, and other demographic variables.

Vitamin D insufficiency affects more than one billion people worldwide.(Holick, 2017) Vitamin D is integral to optimal health for not only bone metabolism, but also for optimal immune function, mental health function (less depression), and muscle strength including reduced fall risk.(Meehan & Penckofer, 2014; U.S. Preventive Services Task Force, December 2016) Thresholds for ‘optimal’ or ‘deficient’ 25(OH)D levels are still under debate, partly because optimal levels may vary depending on what function is affected. While vitamin D is most easily attained through ultraviolet radiation (UV) from the sun,(Holick, 2007) even young adults in Hawaii have variability in 25(OH)D levels, with a surprisingly high prevalence of deficiency despite access to abundant sun exposure.(Binkley et al., 2007) A better understanding of the factors that contribute to vitamin D insufficiency across the age continuum is critical for addressing this widespread public health problem.

This study, based in Hawaii, uniquely examined the impact of different sources of vitamin D and insufficiency in an ethnically diverse older adult population, the large majority of whom were living in a tropical location near the equator that lacks marked seasonal variation. Our study population also includes Asians and Native Hawaiians/Pacific Islanders, which are populations that have been understudied in the vitamin D literature. The study objective was to evaluate the relative contribution of these sources and factors to prevalence of vitamin D insufficiency (<30 ng/ml) in a multiethnic population in Hawaii.

METHODS

Participants

Participants were members of the ongoing Hawaii Longitudinal Study of Personality and Health.(Edmonds et al., 2017; Hampson, Goldberg, Vogt, & Dubanoski, 2006) Participants were studied initially when they were in elementary school on the Hawaiian Islands of Oahu and Kauai between 1959 and 1967 where their personalities were assessed by their teachers. They were later recruited as adults for follow-up studies starting in 1998. Follow-up included mailed questionnaires and half-day medical and psychological assessments at clinic visits conducted at approximately age 50 and 60. To be included in the present study, participants must have completed an extended food frequency questionnaire (FFQext) mailed in 2016 and also provided blood for measurement of serum 25(OH)D at the age 60 clinic visit (n = 223; 120 women). All clinic visits were conducted in Hawaii (Oahu or Kauai); 185 (83%) participants were still Hawaii residents, and 38 (17%) lived elsewhere at the time of the age 60 clinic visit and arranged a visit while in Hawaii for other reasons.

Measures

Outcome

Serum Vitamin D: 25-Hydroxyvitamin D (25(OH)D).

This serum measure indicates the overall blood levels of vitamin D from all potential sources: cutaneous (skin synthesis from sunlight), diet, and supplements.(Holick, 2009) Serum 1,25-Dihydroxyvitamin D (1,25(OH)2D3) is the biologically active form of vitamin D and is tightly regulated by the parathyroid hormone with a circulating half-life of 4 – 6 hours.(Holick, 2009) For that reason, 1,25(OH)2D3 is not a dependable indicator of vitamin D status and 25(OH)D is a better overall measurement of vitamin D status on deficiency or sufficiency, including vitamin D from intake and sun exposure.(Holick, 2009) According to the U.S. Preventive Services Task Force, a consensus does not exist on defining optimal levels of 25(OH)D levels or insufficiency.(U.S. Preventive Services Task Force, December 2016) However, there is general agreement that levels at least 20 ng/mL (50 nmol/L) or higher are required to prevent markedly increased fracture risk.(Holick et al., 2011; LeBlanc, Zakher, Daeges, Pappas, & Chou, 2015; Ross et al., 2011) Many, but not all experts suggest optimal 25(OH)D levels are greater than 20 ng/mL, and suggest levels of at least 30 ng/mL to be sufficient.(LeBlanc et al., 2015) We thus analyzed our results based on this definition of vitamin D insufficiency (< 30 ng/ml), and define our outcome variables as Vitamin D deficiency < 20 ng/ml and Vitamin D deficiency/insufficiency < 30 ng/ml to assess both commonly used clinical cut points.

Blood samples were collected at the age 60 clinic visit, and were immediately centrifuged, aliquoted, frozen and then stored at −70C. Serum 25(OH)D levels were measured by the QuestAssuredD™ 25(OH)D LC/MS/MS method.

Exposures

FFQext

The FFQext, a 22-page questionnaire about participants’ diet and supplement use and other behaviors over the past year, was mailed to all participants in 2016 when participants ranged in age from 57 to 66. The FFQext was adapted from the validated 3rd Multiethnic Cohort Questionnaire (MECQx3, 2003 – 2008) which was designed and developed for use in the Hawaii-Los Angeles Multiethnic Cohort study of diet and cancer.(Kolonel et al., 2000) Questions related to vitamin D behaviors such as sunlight exposure, fractures, and fortification were added and existing questions already captured at follow-up visits were removed (e.g., medical history).

Daily vitamin D intake from food.

The FFQext included the quantitative food frequency questionnaire(Stram et al., 2000) from the MECQx3 that has been used to assess food and nutrient intakes of five main race/ethnic groups – Japanese Americans, Native Hawaiians, Whites, Latinos, and African Americans.(Kolonel et al., 2000) Participants were asked about their usual eating habits in the last year including the frequency of food consumption and amount for 200 items. For example, for the item ‘miso soup’, participants were asked to select one of the following categories for average use in the last year (never or hardly ever, once a month, 2 to 3 times a month, once a week, 2 to 3 times a week, 4 to 6 times a week, once a day, 2 or more times a day) and usual serving size of ½ cup or less, small bowl (1cup) or large bowl (2 cups or more). Photographs representing three general plate portion sizes were used to enhance estimation of quantitative intakes of foods. Ethnic foods and dishes with familiar descriptions to the Hawaii population are included (e.g., poi, sushi, dark fish such as Ahi or Aku, and pork laulau).

The scored FFQext provided daily vitamin D intakes (IU) for foods consumed based on frequency of consumption, portion size, and nutrient estimates from the University of Hawaii Cancer Center (UHCC) food composition database, which contains food components for over 2,300 foods; including the United States Department of Agriculture standard reference updates and those which are consumed in Hawaii.(University of Hawaii Cancer Center, 2019)

Total daily kilocalories (kcals).

Daily caloric intake was also calculated from the FFQext and used as a covariate in analyses.

Consumption of specific foods fortified with vitamin D.

The FFQext included a section of questions that asked participants how often they consume foods that are usually fortified with vitamin D (orange juice, soy milk, other milk alternatives, yogurt, dry cereals) (1 = “never or hardly ever,” 2 = “some of the time,” 3 = “always”). The frequency of consumption of each kind of fortified food was summed (internal reliability of scale - alpha = .64). Summing across foods indicated the extent of consuming foods that are fortified with vitamin D regardless of the specific kind of food (e.g., a participant could achieve the same intake by frequent use of yogurt only, or moderate use of more than one fortified food).

Daily vitamin D intake from supplements.

The FFQext also included a 1-page questionnaire about supplement intake. They are first asked about frequency of taking three types of multi-vitamins followed by questions on 14 specific vitamins and minerals. If yes, participants indicated how many times a week they had taken each of three types of multi-vitamins and each of the listed single vitamins and minerals, which included vitamin D alone or combined with something else. If participants took vitamin D, participants were asked about the amount of vitamin D per day.

Based on their responses, UHCC computed an estimated daily intake of vitamin D (IU) using the UHCC supplement composition table, which contains up to 217 nutrients and food components of over 3,300 dietary supplements. The vitamin D amounts (IU) assigned to the single vitamin D supplement line item in the questionnaire took into consideration both single-type and combined supplements.

Using these data, UHCC computed an estimated daily vitamin D intake (IU) based on food alone, supplements alone, and food and supplements combined.

Sun exposure
Skin colorimetry.

Sun exposure was assessed by comparing sun-exposed and unexposed skin tone. Skin tone was measured at the clinic visit using a spectrophotometer (IMS SMARTPROBE 400 Color Reader), which provided the colorimetric values for L* (brightness), a*(red/green coordinate), and b* (yellow/blue coordinate). The degree of skin pigmentation was assessed using the Individual Typology Angle (ITA°) obtained from L* and b* values using the following formula: ITA° = [(L* – 50)/b*]180/3.14.(Rockell, Skeaff, Williams, & Green, 2008) Higher values represent lighter skin tone. Three readings were recorded both for exposed (outer arm) and non-exposed (inner arm) skin, and the difference between the mean exposed and non-exposed values was calculated. Greater discrepancies in skin tone between exposed and unexposed skin tone indicated higher levels of sun exposure.

Questionnaire.

Four questions on the FFQext asked about time spent outdoors during the past month on weekdays, weekend days, on the water, and between 10am and 2pm, and a single item asked how much of that time the participant used sunscreen. These questions were part of a sun exposure questionnaire that has been used by other studies in the aging population. Response options for each question ranged from 1 = “less than 15 minutes,” to 5 = “more than 2 hours”. The four sun exposure items were summed to create a sun exposure variable (alpha = .74). This measure was validated by examining its correlation with the ITA° difference between skin tone of exposed versus non-exposed skin measured by colorimetry at the clinic visit. Larger skin tone differences were significantly correlated with the self-reported sun exposure scale scores (r = .30, p = <.001, n = 218).

Body Mass Index (BMI)

Participants’ height and weight were measured at the age 60 clinic visit and converted to BMI (kg/m2), which was used as a covariate in analyses.

Demographics
Educational attainment.

The most recent report of educational attainment was measured on a 9-step scale (1 = “eighth grade or less”, 9 = “postgraduate or professional degree”).

Annual income.

Participants reported on their approximate annual income at each of four 5-year periods during adulthood and the highest reported income was used as an indicator of income (1 = less than $10,000, 5 = $80,000 or more).

Ethnicity.

Participants self-reported the ethnicity they most identified with by answering the following question: “Which group best describes your cultural identity? (We recognize that many people have a diverse cultural background. If you identify with more than one group, please choose the one group with which you most strongly associate yourself.)” The response categories included: African-American, Aleutian/Alaskan/American Indian, Caucasian (European-American), Chinese, Filipino, Hawaiian/Part-Hawaiian, Japanese, Korean, Latino (Latin American or Puerto Rican), Okinawan, Other Pacific Islander (such as Samoan or Tongan, and Other. Ethnic groups were further collapsed into 5 major groups: Asian (Japanese, Korean, Okinawan, Chinese); Native Hawaiian, part-Hawaiian or other Pacific Islander; White; Filipino; and ‘Other’. The Latino group was collapsed into the ‘Other’ group due to small numbers.

Variable Recoding

All variable distributions were inspected for extreme outliers. Inspection of the distribution of scores of total daily kcal consumed indicated that five participants reported consuming more than 5,000 kcals and five participants reported consuming less than 500 kcals. Following Willett (2013), these outliers were re-coded to 5,000 and 500 kcals respectively to reflect more realistic caloric intake.(Willett, 2013) Outliers on other variables were conservatively defined as participants scoring over four standard deviations above or below the mean (there were no scores over four standard deviations below the mean). Outliers were winsorized (i.e., re-scored) to the value of four standard deviations above the sample mean obtained after removing these extreme scores. Four outliers were winsorized on daily estimates of vitamin D intake from food, four on BMI, and two on serum 25(OH)D.

Statistical Analysis

To start, correlations were estimated for all study variables. Next, a multiple regression was conducted predicting continuous serum levels of vitamin D from sources of vitamin D (self-reported and objectively validated sun exposure, food including fortified sources, and supplements), with sunscreen use (dichotomized as < or ≥15 minutes in the past month), age, educational attainment, annual income, sex, and BMI included as control variables. This analysis was repeated with logistic regression predicting serum levels of vitamin D dichotomized at < or ≥ 30 ng/mL (i.e., vitamin D insufficiency). Both models also evaluated the interaction between sun exposure and use of sunscreen. In sensitivity analyses, both regressions were repeated using values of vitamin D from food and from fortified food corrected for total daily caloric intake (values per 1,000 kcals), and including total daily caloric intake as a control variable. Although the study was not powered to test for gender differences, predictors of deficiency/insufficiency were examined separately in males and females and coefficients for sun exposure, vitamin D, and BMI were in the same direction. Gender was then used a dichotomous variable in the regression models. Interactions between time interval between clinic visit measure of serum 25(OH)D and vitamin D sources from the questionnaire were also examined. Statistical analyses were conducted using IBM SPSS Statistics version 24.(IBM Corp, 2016)

This study was approved by the Kaiser Permanente Hawaii Institutional Review Board [MR25_HI-02TVogt-03]. Written informed consent was obtained from all subjects/patients.

RESULTS

Participant Characteristics

The mean age was 59.6 (SD = 2.1; range 59.3 – 60.2 years) and mean BMI was 27.9 (SD = 6.2, Table 1). Participants’ self-identified ethnic identity was 58% Asian (68 women, 61 men), 20% NH/PI (23 women, 21 men), 11% White (12 women, 13 men), 8% Filipino (8 women, 10 men), and 3% Other (1 woman, 6 men). Participants were, on average, fairly well-educated (M = 7.2 on the 9-point scale, SD = 1.6) and average annual earnings were above the middle of the range of the 5-point scale ranging from $0 to $80,000 or more (M = 3.6). The mean time between completion of the FFQext and the clinic visit was .72 years (SD = .73; range = −.72 – 2.72). The sample mean for BMI was 27.9kg/m2 ± 6.15 and would be classified as overweight.(Center for Disease Control and Prevention, 2017) Mean BMI for the Filipinos and Other ethnicity groups fell within the obese category (≥30 kg/m2).

TABLE 1.

Mean Differences on Study Variables for the Five Race/Ethnic Groups

Total (n = 223) Asian (N = 129) NH/PI (N = 44) Whites (N = 25) Filipino (N = 18) Other (n = 7)
M SD M SD M SD M SD M SD M SD
Age at clinic visit (y) 59.6 2.1 59.3 2.1 59.8 2.2 60.2 2.5 60.1 1.7 58.9 2.2
Education† 7.2 1.6 7.8 1.1 6.0 1.9 6.8 2.0 7.1 1.4 5.9 1.7
Income‡ 3.6 1.0 3.7 1.0 3.2 0.9 3.9 0.9 3.5 0.8 3.5 1.3
BMI (kg/m2) 27.9 6.2 26.8 4.9 29.3 6.6 28.8 8.9 30.5 5.8 31.2 8.7
Serum25(OH)D
ng/mL 34.3 10.9 33.7 10.8 33.7 12.1 38.4 12.0 35.3 7.3 31.1 5.6
(%, <30 ng/ml) 37.7 -- 43.4 -- 36.4 -- 28.0 -- 11.1 -- 42.9 --
(%, <20 ng/ml) 5.4 -- 7.0 -- 6.8 -- 0.0 -- 0.0 -- 0.0 --
Vitamin D Intake
D(mcg) from food 3.2 2.9 2.7 2.2 4.2 3.5 4.3 4.3 3.9 3.4 2.0 1.4
Total daily kcals 1974 1058 1883 939 2303 1340 2127 1122 1854 1010 1330 442
D(mcg)/1000 kcals 1.6 1.0 1.4 .8 1.9 1.1 1.8 1.2 2.1 1.3 1.4 .9
D fortified food§ 2.1 2.1 1.9 2.1 2.4 2.0 1.9 2.0 2.7 2.1 2.4 1.8
D (mcg) supplements 6.8 5.7 7.4 5.8 5.9 5.2 5.0 5.5 6.7 5.5 6.0 7.9
Other Sources
Sun exposure 11.6 4.3 10.6 4.0 13.7 3.9 12.0 4.6 13.5 4.5 9.9 3.1
Sunscreen usea 2.5 1.7 2.6 1.7 2.3 1.7 2.2 1.7 2.2 1.6 2.6 1.8
Sunscreen use (%)a 48.4 -- 52.8 -- 42.0 -- 40.0 -- 41.2 -- 57.1 --
Unexposed skin tone 35.7 13.8 39.1 11.8 29.6 14.9 40.6 12.0 19.6 12.3 37.4 10.0
Skin tone difference 35.7 14.5 35.2 14.2 37.8 12.0 41.0 19.2 31.4 10.9 22.5 13.2
Open in a new tab

M = Mean, SD = Standard Deviation

†

Educational Attainment = 9-step scale (1 = “eighth grade or less to 9 = “postgraduate or professional degree”)

‡

Annual Income = 5-point scale (1 = <$10,000 to 5 = $80,000 or more)

§

Fortified foods = Sum of 5 fortified food items (1 = “never or hardly ever” 2 = “some of the time,” 3 = “always”)

Sun Exposure = Sum of 4 sunlight exposure variables (1 = “less than 15 minutes” to 5 = “more than 2 hours”)

a

Sunscreen use = (1 = “less than 15 minutes” to 5 = “more than 2 hours”), Sunscreen use (%) = ≥15 minutes

There were 722 participants that completed the FFQ sample, and we evaluated the representativeness of the subsample in the present study that also had a 25(OH)D assay (n = 223) by comparing demographic variables (age, gender, ethnicity, and education with those who completed the FFQ but did not have a 25(OH)D assay (n = 499). The vitamin D sample was representative in age, gender, and education distributions but had a higher proportion of Asians and fewer Whites.

Serum 25(OH)D levels

In Table 1 we present means and standard deviations for serum 25(OH)D levels, sources of vitamin D for the entire sample and by the five main ethnicity groups (Asians, NH/PI, Whites, Filipino, and Other). Mean serum 25-(OH)D levels were 34.3 ng/ml (SD = 10.9); 37.7% of participants had levels lower than 30 ng/ml indicating vitamin D deficiency/insufficiency. Vitamin D deficiency (< 20 ng/m) and deficiency/insufficiency rates (< 30 ng/ml) varied by ethnicity: Asians had the highest proportion of vitamin D deficiency/insufficiency (43%) and Filipinos the lowest (11%). Around five percent (5.4%) of the sample were deficient; this was observed only in Asian (7%) and NH/PI (7%) groups; other racial groups did not have any vitamin D deficiency (Table 1). Mean serum 25(OH)D values were similar across Hawaii residents (34 ng/ml) and non-residents (35 ng/ml).

Sources of Vitamin D

Serum 25(OH)D was positively correlated with self-reported sun exposure (r(218) = .21, p = <.01), supplements (r(223) = .24, p < .01), and with sunscreen use (r(217) = .16, p = .02) and negatively correlated with BMI (r(223) = −.13, p = 0.05). Less than half (48%) of participants reported using sunscreen for more than 15 minutes over the past month. Sunscreen use correlated with sunlight exposure (r(215) = .25, p < .01), indicating that participants who reported spending more time in the sun also reported using more sunscreen.

Vitamin D from fortified foods correlated with overall vitamin D intake from food (r(223) = .25, p < .01). Fortified yogurt was used by 44% of participants, dry cereals by 37%, orange juice by 35%, other kinds of milk by 21%, and soymilk by 16%. Unexposed skin tone and skin tone difference values obtained from skin colorimetry varied among ethnic groups as would be expected (Table 1). Larger skin tone differences was correlated with self-reported sun exposure measured by questionnaire (r(218) = .30, p <.01), but was not significantly correlated with serum 25(OH)D (r(223) = −.07, p = .31). Sun exposure measured by questionnaire had a higher, significant correlation with serum 25(OH)D (r(218) = .21, p <.01) than skin tone difference and was therefore included as the measure of sun exposure in our regression models.

Regression Analyses

Neither sunscreen use nor the interaction between sunscreen use and sunlight exposure significantly predicted serum 25(OH)D levels, so these variables were removed from the final model by backwards elimination. Sun exposure was kept in the final model. Results from our final multiple regression model predicting continuous serum 25(OH)D levels are shown in Table 2. The only significant predictors of serum 25(OH)D were intake of vitamin D from supplements (p = .003) and sunlight exposure (p < .001). In our logistic regression model predicting whether serum 25(OH)D were sufficient or insufficient (≥ or < 30 ng/mL) supplement use and sun exposure were also significant, as well as BMI (higher BMI was related to higher risk of vitamin D deficiency/insufficiency, see Table 3). Results of the models did not change after including ethnicity in both models and therefore ethnicity was removed. Sensitivity analyses accounting for daily caloric intake had no effect on the pattern of significant results. Interactions between the time interval of clinic measures and questionnaire response for each of the sources of vitamin D measured by the FFQext were examined and none were significant, indicating that time intervals did not affect these associations.

TABLE 2.

Multiple Regression Predicting Level of Serum 25-(OH)D

B Std. Error β t p
Age at Clinic Visit (y) .257 .373 .052 .691 .491
Highest education level .625 .585 .091 1.069 .287
Highest reported income −.572 .849 −.054 −.673 .502
Males −2.965 1.642 −.142 −1.806 .073
BMI (kg/m2) −.165 .125 −.101 −1.317 .190
Vitamin D (IU) food† .563 .844 .051 .667 .506
Vitamin D (IU) supplements .413 .135 .229 3.049 .003
Vitamin D fortified foods† −.265 .603 −.033 −.439 .661
Sun exposure .835 .202 .319 4.133 <.001
Open in a new tab

Note. Model fit: R = .43, Adjusted R2 =.18, df = 9, 165, p = <.01

†

Adjusted for total caloric intake

TABLE 3.

Logistic Regression Predicting Serum 25(OH)D Insufficiency (< 30ng/ml)

β S.E. Wald χ2 p OR 95% CI OR
Age at Clinic Visit −.134 .086 2.395 .122 .875 .739, 1.036
Highest education level .041 .136 .091 .763 1.042 .798, 1.360
Highest reported income .281 .201 1.959 .162 1.325 .894, 1.963
Males .025 .379 .004 .948 1.025 .487, 2.156
BMI (kg/m2) .066 .030 5.006 .025 1.069 1.008, 1.132
Vitamin D (IU) food† −.271 .215 1.597 .206 .762 .501, 1.161
Vitamin D (IU) supplements −.116 .033 12.599 <.001 .890 .835, .949
Vitamin D fortified foods† .012 .142 .007 .931 1.012 .767, 1.337
Sun exposure −.133 .049 7.325 .007 .876 .795, .964
Open in a new tab
†

Adjusted for total caloric intake

DISCUSSION

Given the additional risks for Vitamin D deficiency/insufficiency faced by older adults, the current study addressed the need for comprehensive investigation into the factors influencing Vitamin D; moreover, it added to the literature by considering this aim with an underrepresented sample in the healthy aging literature in a tropical climate with year-round access to sun for a Vitamin D source. In this sample, although deficiency/insufficiency rates were much lower than in the general U.S. population (77%),(Ginde, Liu, & Camargo, 2009) more than 1/3 of participants (38%) still had vitamin D deficiency/insufficiency despite access to a tropical, sunny climate. Sun exposure and supplementation were the main predictors of serum 25(OH)D in this ethnically diverse sample of adults in their 60s who were mostly residents of Hawaii.

Past research has identified numerous factors that may influence sufficient serum 25(OH)D levels, including length of sun exposure, latitude, season, degree of skin pigmentation (increased melanin reduces vitamin D production), use of sunscreen, and vitamin D supplementation.(Holick, 2017; Hollis, 2005; M. G. Kimlin et al., 2014; Macdonald et al., 2011; Slominski & Postlethwaite, 2015) However, few studies have examined all major vitamin D sources (objectively validated sun exposure, vitamin D supplementation, food including fortified sources) to understand how they each contribute to serum 25(OH)D levels (Gozdzik et al., 2010; Hall et al., 2010; M. G. Kimlin et al., 2014; Nagasaka et al., 2018) especially in older adults.(Ginter et al., 2013; M. G. Kimlin et al., 2014; Nagasaka et al., 2018) A study of Australian adults (18 to 75 years old) identified time spent outdoors, clothing cover, vitamin D supplementation, BMI, physical activity, and personal UV radiation exposure as potentially modifiable factors predicting 52% of the variance in 25(OH)D levels. Another study of South Asians and European adults (20 to 79 years old) in Canada reported that ethnic background and vitamin D supplementation were key predictors of 25(OH)D and that this varied by age group. While in another study of adults >60 years of age, there were no differences in serum levels, dietary intake, and supplementation between ancestral groups (South Asians, East Asians, Europeans).

With regard to the currently observed deficiency/insufficiency levels, even though sun exposure UV blue light (UVB) is reported to be the main source of vitamin D levels before diet intake,(Calvo, Whiting, & Barton, 2005) the current findings align with past studies that still found prevalent vitamin D insufficiency and deficiency levels even in geographical areas with accessibility to sun and high ambient ultraviolet radiation.(Daly et al., 2012; M. Kimlin et al., 2007; M. G. Kimlin et al., 2014; Lips, 2010) In Australia, where there is high ambient UV(M. G. Kimlin et al., 2014) relative to other places, 73% (< 30 ng/ml) were insufficient in vitamin D and 31% (< 25 ng/ml) deficient.(Daly et al., 2012) Similarly, in a sample of 93 young adults in Hawaii that had been outside at least 22.4 hours/week with no sunscreen and 28.9 hours/week with and without sunscreen,(Binkley et al., 2007) 51% were insufficient (< 30 ng/ml). Despite prevalent accessibility to the sun, several reasons present as to why insufficient 25(OH)D levels still persist. For instance, insufficiency could be explained by method of capturing sun exposure or a threshold effect; initial sun exposure is very efficient at vitamin D production, which then declines after a certain exposure time. Moreover, sun-protecting behaviors (such as sunscreen use), skin pigmentation (increased melanin production), and aging also contribute to reducing the effect of sun exposure on meeting sufficient serum 25(OH)D levels. In particular, older adults may spend less time outdoors, have less than optimal skin synthesis of vitamin D, reduced absorption of vitamin D from food sources and a limited dietary quality or intake of a variety of foods. (Brownie, 2006; MacLaughlin & Holick, 1985; Nair & Maseeh, 2012; Zhu et al., 2010)

Melanin, a pigment in the human skin, is found in the basal layer of the epidermis of the skin.(Byard, 1981) Among darker-skinned ethnic groups, there is more melanin in the epidermis(Rockell et al., 2008) which decreases the efficiency of vitamin D production due to the sunscreen effect (e.g., UV filter) of melanin.(Matsuoka, Wortsman, Haddad, Kolm, & Hollis, 1991) It is interesting to note that the current study failed to find evidence for expected skin tone differences in this diverse racial/ethnic sample. This is despite the fact that NH/PI and Filipinos showed the lowest rates of vitamin D deficiency/insufficiency and also having darker unexposed skin tone. It is difficult to speculate on these inconsistent findings, particularly given that the current sample sizes were insufficient for testing group differences. However, potential explanations include that the lack of insufficiency may be attributed to other factors such as cultural differences in sun exposure, supplementation, or sunscreen use rather than skin tone. Future research is needed to investigate these explanations using larger samples for these groups, but the current findings potentially point to interesting differences between studies in Hawaii and previous work on vitamin D insufficiency.

Strengths

Our study had several important strengths. We examined major sources of vitamin D (diet, supplementation, and sun exposure) in an ethnically diverse, older adult cohort, the majority of whom were living in a sunny tropical environment with minimal seasonal variation and represent understudied race/ethnic groups. We assessed dietary and supplement sources of vitamin D using an extended FFQ that has been validated within the Hawaii population and included many culturally different foods, recipes, and supplement use.(Stram et al., 2000)

Moreover, the current study again demonstrated the importance of vitamin D supplementation. Consistent with past work with older adults,(Ginter et al., 2013; M. G. Kimlin et al., 2014; Nagasaka et al., 2018) vitamin D supplementation was another predictor of serum 25(OH)D levels. Ginther et al., reported that whether older adults took vitamin D supplements was a major factor in determining serum 25(OH)D levels though most of these adults were healthy. Nagasaka et al., reported that serum 25(OH)D levels varied by supplementation which also varied by ethnic background, age, and season.(Nagasaka et al., 2018) In another study, no differences between race/ethnic groups in supplementation were reported.(Ginter et al., 2013) In our study, all ethnic groups had mean-levels indicating some supplement use, however the degree of contribution of supplements to improving insufficiency among ethnic groups is not known. Existing studies used different methods to assess vitamin D supplementation, including the employment of supplement labels with exact amounts vs. whether the supplement was taken (yes/no). Dietary supplement use may contribute up to 6 – 47% of the average vitamin D intake(Calvo et al., 2005) which indicates that more detailed information should be gathered in future studies to have a better understanding on the contribution of supplements.

Limitations

We assessed serum 25(OH)D using blood drawn at a clinic visit, and the time interval between the blood draw and our questionnaire of dietary sources consumed in the last year varied across participants, potentially affecting our ability to detect effects of dietary vitamin D intake on serum levels. However, sensitivity analysis of time intervals and major vitamin D sources were non-significant indicating this time interval difference probably did not have a large effect on our results. We also did not have parathyroid hormone measures to examine with blood levels of 25(OH)D. This would be a valuable contribution to our future work.

In the FFQext, the extent of consuming fortified versions of these foods (1 = “never or hardly ever,” 2 = “some of the time,” 3 = “always”) is different from the frequency of consuming that particular item in the last year which is assigned IU and converted to mcg amounts and so mcg amounts from fortified foods may be underestimated. Our questionnaire collected data on vitamin D supplementation frequency and amounts, but did not include data on brand or type of vitamin D supplements, which could have yielded much more detailed information about potency that would provide a better understanding of supplemental contribution to serum 25(OH)D levels.

Most participants (83%) were Hawaii residents at the time of the exam, but a substantial minority were not, potentially affecting strength of sun exposure and other factors. Nevertheless, we found that mean serum 25(OH)D values were similar across these two groups.

CONCLUSION

Less sun exposure and lower levels of vitamin D supplementation, but not lower vitamin D intake through food or sunscreen use, predicted serum 25(OH)D insufficiency in an ethnically diverse cohort of older adults primarily living in Hawaii. Despite abundant sun exposure and vitamin D supplementation, 38% of participants in our sample met criteria for vitamin D deficiency/insufficiency (< 30 ng/mL). Rates of vitamin D insufficiency ranged from 11% in Filipino participants to 43.4% in Asian participants. Future studies should explore whether such differences are related to vitamin D sources and behaviors across ethnic groups. Further exploration of sun protective behaviors (cultural, lifestyle, and preventive behaviors) in Hawaii might better inform the variation we see related to sun exposure, a main predictor of vitamin D. While we await this research, our findings suggest that having access to abundant sun exposure alone may not be adequate to prevent vitamin D deficiency/insufficiency, and higher supplementation may be needed to offset the existing limited contribution from sun exposure and limitations in dietary sources of vitamin D in older adults.

Acknowledgements:

This work was supported by the National Institute on Aging of the National Institutes of Health [R01AG020048]; and the Office of Dietary Supplements, National Institutes of Health [R01AG020048-18S1]. We would also like to acknowledge Carol J. Boushey, PHD, RDN, Associate Research Professor and Director, Nutrition Support Shared Resource at the University of Hawaii Cancer Center (National Cancer Institute at the National Institutes of Health grant P30 CA071789) and Ms. Kim M. Murakami, RDN, Nutritionist/Dietitian Supervisor for their expertise in the adaptation of the 3rd Multiethnic Cohort Questionnaire (MECQx3) to produce the FFQext, scanning of the questionnaires, and timely provision of data descriptions and output from the FFQext.

Footnotes

Declaration of Conflicting Interest:

The authors declare that there is no conflict of interest.

Data Availability Statement:

Research data are not shared.

REFERENCES

  1. Binkley N, Novotny R, Krueger D, Kawahara T, Daida YG, Lensmeyer G, … Drezner MK (2007). Low vitamin D status despite abundant sun exposure. J Clin Endocrinol Metab, 92(6), 2130–2135. doi: 10.1210/jc.2006-2250 [DOI] [PubMed] [Google Scholar]
  2. Brownie S (2006). Why are elderly individuals at risk of nutritional deficiency? Int J Nurs Pract, 12(2), 110–118. doi: 10.1111/j.1440-172X.2006.00557.x [DOI] [PubMed] [Google Scholar]
  3. Byard PJ (1981). Quantitative Genetics of Human Skin Color. Am J Phys Anthropol, 24(S2), 123–137. doi: 10.1002/ajpa.1330240506 [DOI] [Google Scholar]
  4. Calvo MS, Whiting SJ, & Barton CN (2005). Vitamin D intake: a global perspective of current status. J Nutr, 135(2), 310–316. doi: 10.1093/jn/135.2.310 [DOI] [PubMed] [Google Scholar]
  5. Center for Disease Control and Prevention. (2017, August 29, 2017). About Adult BMI. Retrieved from https://www.cdc.gov/healthyweight/assessing/bmi/adult_bmi/index.html
  6. Daly RM, Gagnon C, Lu ZX, Magliano DJ, Dunstan DW, Sikaris KA, … Shaw JE (2012). Prevalence of vitamin D deficiency and its determinants in Australian adults aged 25 years and older: a national, population-based study. Clin Endocrinol (Oxf), 77(1), 26–35. doi: 10.1111/j.1365-2265.2011.04320.x [DOI] [PubMed] [Google Scholar]
  7. Edmonds G, Hampson S, Goldbweg L, Digman J, Dubanoski J, & Oshiro C (2017). The Hawaii Personality and Health Cohort, 1959–1967: Childhood Personality Data.
  8. Ginde AA, Liu MC, & Camargo CA Jr. (2009). Demographic differences and trends of vitamin D insufficiency in the US population, 1988–2004. Arch Intern Med, 169(6), 626–632. doi: 10.1001/archinternmed.2008.604 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Ginter JK, Krithika S, Gozdzik A, Hanwell H, Whiting S, & Parra EJ (2013). Vitamin D status of older adults of diverse ancestry living in the Greater Toronto Area. BMC Geriatr, 13, 66. doi: 10.1186/1471-2318-13-66 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Gozdzik A, Barta JL, Weir A, Cole DE, Vieth R, Whiting SJ, & Parra EJ (2010). Serum 25-hydroxyvitamin D concentrations fluctuate seasonally in young adults of diverse ancestry living in Toronto. J Nutr, 140(12), 2213–2220. doi: 10.3945/jn.110.126284 [DOI] [PubMed] [Google Scholar]
  11. Hall LM, Kimlin MG, Aronov PA, Hammock BD, Slusser JR, Woodhouse LR, & Stephensen CB (2010). Vitamin D intake needed to maintain target serum 25-hydroxyvitamin D concentrations in participants with low sun exposure and dark skin pigmentation is substantially higher than current recommendations. J Nutr, 140(3), 542–550. doi: 10.3945/jn.109.115253 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Hampson SE, Goldberg LR, Vogt TM, & Dubanoski JP (2006). Forty years on: teachers’ assessments of children’s personality traits predict self-reported health behaviors and outcomes at midlife. Health Psychol, 25(1), 57–64. doi: 10.1037/0278-6133.25.1.57 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Holick MF (2007). Vitamin D deficiency. N Engl J Med, 357(3), 266–281. doi: 10.1056/NEJMra070553 [DOI] [PubMed] [Google Scholar]
  14. Holick MF (2009). Vitamin D status: measurement, interpretation, and clinical application. Ann Epidemiol, 19(2), 73–78. doi: 10.1016/j.annepidem.2007.12.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Holick MF (2017). The vitamin D deficiency pandemic: Approaches for diagnosis, treatment and prevention. Rev Endocr Metab Disord, 18(2), 153–165. doi: 10.1007/s11154-017-9424-1 [DOI] [PubMed] [Google Scholar]
  16. Holick MF, Binkley NC, Bischoff-Ferrari HA, Gordon CM, Hanley DA, Heaney RP, … Endocrine S (2011). Evaluation, treatment, and prevention of vitamin D deficiency: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab, 96(7), 1911–1930. doi: 10.1210/jc.2011-0385 [DOI] [PubMed] [Google Scholar]
  17. Hollis BW (2005). Circulating 25-hydroxyvitamin D levels indicative of vitamin D sufficiency: implications for establishing a new effective dietary intake recommendation for vitamin D. J Nutr, 135(2), 317–322. doi: 10.1093/jn/135.2.317 [DOI] [PubMed] [Google Scholar]
  18. IBM Corp. (2016). IBM SPSS Statistics for Windows (Version 24.0). Armonk, NY: IBM Corp. [Google Scholar]
  19. Kimlin M, Harrison S, Nowak M, Moore M, Brodie A, & Lang C (2007). Does a high UV environment ensure adequate vitamin D status? J Photochem Photobiol B, 89(2–3), 139–147. doi: 10.1016/j.jphotobiol.2007.09.008 [DOI] [PubMed] [Google Scholar]
  20. Kimlin MG, Lucas RM, Harrison SL, van der Mei I, Armstrong BK, Whiteman DC, … Sun J (2014). The contributions of solar ultraviolet radiation exposure and other determinants to serum 25-hydroxyvitamin D concentrations in Australian adults: the AusD Study. Am J Epidemiol, 179(7), 864–874. doi: 10.1093/aje/kwt446 [DOI] [PubMed] [Google Scholar]
  21. Kolonel LN, Henderson BE, Hankin JH, Nomura AM, Wilkens LR, Pike MC, … Nagamine FS (2000). A multiethnic cohort in Hawaii and Los Angeles: baseline characteristics. Am J Epidemiol, 151(4), 346–357. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Laird E, O’Halloran AM, Carey D, Healy M, O’Connor D, Moore P, … Kenny RA (2018). The Prevalence of Vitamin D Deficiency and the Determinants of 25(OH)D Concentration in Older Irish Adults: Data From The Irish Longitudinal Study on Ageing (TILDA). J Gerontol A Biol Sci Med Sci, 73(4), 519–525. doi: 10.1093/gerona/glx168 [DOI] [PubMed] [Google Scholar]
  23. LeBlanc ES, Zakher B, Daeges M, Pappas M, & Chou R (2015). Screening for vitamin D deficiency: a systematic review for the U.S. Preventive Services Task Force. Ann Intern Med, 162(2), 109–122. doi: 10.7326/M14-1659 [DOI] [PubMed] [Google Scholar]
  24. Lips P (2010). Worldwide status of vitamin D nutrition. J Steroid Biochem Mol Biol, 121(1–2), 297–300. doi: 10.1016/j.jsbmb.2010.02.021 [DOI] [PubMed] [Google Scholar]
  25. Macdonald HM, Mavroeidi A, Aucott LA, Diffey BL, Fraser WD, Ormerod AD, & Reid DM (2011). Skin color change in Caucasian postmenopausal women predicts summer-winter change in 25-hydroxyvitamin D: findings from the ANSAViD cohort study. J Clin Endocrinol Metab, 96(6), 1677–1686. doi: 10.1210/jc.2010-2032 [DOI] [PubMed] [Google Scholar]
  26. MacLaughlin J, & Holick MF (1985). Aging decreases the capacity of human skin to produce vitamin D3. J Clin Invest, 76(4), 1536–1538. doi: 10.1172/JCI112134 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Matsuoka LY, Wortsman J, Haddad JG, Kolm P, & Hollis BW (1991). Racial pigmentation and the cutaneous synthesis of vitamin D. Arch Dermatol, 127(4), 536–538. [PubMed] [Google Scholar]
  28. Meehan M, & Penckofer S (2014). The Role of Vitamin D in the Aging Adult. J Aging Gerontol, 2(2), 60–71. doi: 10.12974/2309-6128.2014.02.02.1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Nagasaka R, Swist E, Sarafin K, Gagnon C, Rondeau I, Massarelli I, … Ratnayake WMN (2018). Low 25-hydroxyvitamin D levels are more prevalent in Canadians of South Asian than European ancestry inhabiting the National Capital Region of Canada. PLoS One, 13(12), e0207429. doi: 10.1371/journal.pone.0207429 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Nair R, & Maseeh A (2012). Vitamin D: The “sunshine” vitamin. J Pharmacol Pharmacother, 3(2), 118–126. doi: 10.4103/0976-500X.95506 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Rockell JE, Skeaff CM, Williams SM, & Green TJ (2008). Association between quantitative measures of skin color and plasma 25-hydroxyvitamin D. Osteoporos Int, 19(11), 1639–1642. doi: 10.1007/s00198-008-0620-4 [DOI] [PubMed] [Google Scholar]
  32. Ross AC, Manson JE, Abrams SA, Aloia JF, Brannon PM, Clinton SK, … Shapses SA (2011). The 2011 report on dietary reference intakes for calcium and vitamin D from the Institute of Medicine: what clinicians need to know. J Clin Endocrinol Metab, 96(1), 53–58. doi: 10.1210/jc.2010-2704 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Slominski A, & Postlethwaite AE (2015). Skin under the sun: when melanin pigment meets vitamin D. Endocrinology, 156(1), 1–4. doi: 10.1210/en.2014-1918 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Stram DO, Hankin JH, Wilkens LR, Pike MC, Monroe KR, Park S, … Kolonel LN (2000). Calibration of the dietary questionnaire for a multiethnic cohort in Hawaii and Los Angeles. Am J Epidemiol, 151(4), 358–370. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. U.S. Preventive Services Task Force. (December 2016). Final Recommendation Statement: Vitamin D Deficiency: Screening. Retrieved from https://www.uspreventiveservicestaskforce.org/Page/Document/RecommendationStatementFinal/vitamin-d-deficiency-screening
  36. University of Hawaii Cancer Center. (2019). Nutrition Support. Retrieved from https://www.uhcancercenter.org/research/shared-resources/nutrition-support
  37. Willett WC (2013). Nutritional Epidemiology (3rd ed.). New York, NY: Oxford University Press. [Google Scholar]
  38. Zhu K, Devine A, Suleska A, Tan CY, Toh CZ, Kerr D, & Prince RL (2010). Adequacy and change in nutrient and food intakes with aging in a seven-year cohort study in elderly women. J Nutr Health Aging, 14(9), 723–729. doi: 10.1007/s12603-010-0324-2 [DOI] [PubMed] [Google Scholar]

RESOURCES