Prevalence and Predictors of Vitamin D Deficiency in Healthy Adults

Deborah M. Mitchell, MD; Maria P. Henao, BA; Joel S. Finkelstein, MD; Sherri-Ann M. Burnett-Bowie, MD, MPH


Endocr Pract. 2012;18(6):914-923. 

In This Article


This study examined the rates of vitamin D deficiency in healthy younger adults in the Boston, Massachusetts area. Current recommendations suggest that only patients with a known risk factor should be screened for vitamin D deficiency, excluding a substantial portion of the population. Overall, we found high rates of vitamin D deficiency in this "lower-risk" cohort.

Many previous studies of vitamin D deficiency have been conducted in population-based samples[7,18] or in populations thought to be at unusually high risk.[8,19,20] Fewer studies have examined rates of deficiency in healthy, nonelderly adults. In a previous study of patients admitted to a general medical service, we found that 40% of those 65 and younger, despite having no chronic medical problems, had 25(OH)D levels <15 ng/mL.[21] In another study that sought to select a healthier cohort by excluding those with malabsorption, 30% of the subjects had 25(OH)D levels ≤20 ng/mL at the end of winter.[22] Our study, in which 39% of subjects had 25(OH)D levels ≤20 ng/mL, extends these findings to an even healthier cohort.

The Institute of Medicine recently proposed that a 25(OH)D level ≤20 ng/mL be used to define vitamin D deficiency.[14] Using this definition, vitamin D deficiency was highly prevalent in black subjects (70%), who are known to be a high-risk group (23,24), but also prevalent in white and Asian subjects (27% and 63%, respectively). These data are particularly concerning given the observed increase in PTH levels with decreasing 25(OH)D levels, suggesting that our otherwise healthy subjects were at risk for increased bone turnover, impaired bone mineralization, and bone loss.[25] Interestingly, even at the lowest levels of 25(OH)D, only a minority of subjects had elevated levels of PTH. Conversely, 2 subjects with 25(OH)D levels ≥30 ng/mL, considered adequate by most experts, had elevated PTH levels. These data point to the likely critical role of dietary calcium intake in modulating the effect of low 25(OH)D levels on PTH levels, and thus on skeletal health.[6]

25(OH)D levels >30 ng/mL have been proposed as "vitamin D sufficiency" by many authors.[4,17,26] This classification is based on data regarding the 25(OH)D threshold required for optimal suppression of PTH,[27–29] calcium absorption,[30] and prevention of osteomalacia.[31] Additionally, several nonskeletal health outcomes including decreased risk of autoimmunity, cardiovascular disease, and cancer, have been associated with 25(OH)D levels >30 ng/mL in observational studies.[32] Randomized controlled trials are underway to further evaluate this association.[33] Using this definition, 36% of our cohort would be classified as "vitamin D sufficient", including only 12% of black and 17% of Asian subjects.

As in previous studies, we found higher 25(OH)D levels in white in comparison with black subjects.[34] Seventy percent of our black subjects had 25(OH)D levels ≤20 ng/mL with 25% of the black subjects having 25(OH)D ≤10 ng/mL. Most previous studies (35), though not all (36), have reported that the increase in 25(OH)D level in response to ultraviolet B (UVB) exposure is blunted in people with increased skin pigmentation in comparison with people with less skin pigmentation. Our finding of an attenuated increase in 25(OH)D levels in black subjects during the summer months is consistent with prior reports and supports a primary role for skin pigmentation in mediating the observed differences among racial groups.[37]

The Asian subjects in our study also had significantly lower 25(OH)D levels and higher rates of deficiency than white subjects. While high rates of vitamin D deficiency have been previously reported in Asia (38), few studies have examined 25(OH)D levels in Asian-Americans.[39,40] These studies have been limited by small numbers and by restriction to high-risk groups such as children and older adults.[8,39,40] Our data suggest that young adult Asian-Americans comprise a previously unrecognized group at high risk for vitamin D deficiency.

The observation that female subjects had higher 25(OH)D levels was surprising, as previous studies have found higher levels in male subjects.[34] The 2001 to 2004 National Health and Nutrition Examination survey, unlike older surveys, however, showed no difference in 25(OH)D levels by sex.[18] This altered pattern may be due to trends of decreased outdoor activity among males.[18] Potential explanations for the higher 25(OH)D levels among females in our cohort include increased adherence to multivitamin use, particularly as the females in this study were of child-bearing age. There may also have been differences in outdoor activities, southern travel, or other sources of UVB exposure.

Importantly, our data suggest that even routine multivitamin use, that is, without adhering to a strict protocol, offers protection against the development of vitamin D deficiency. Because subjects were taking these supplements not as part of a study, daily adherence was likely variable, representing "typical" rather than "perfect" use. These subjects are thus similar in their multivitamin use to patients seen in an average clinical practice. When analyzed by season, multivitamin users did not have the post-summer decline in 25(OH)D levels seen in multivitamin nonusers. This finding suggests that increased oral intake of vitamin D compensated for decreased UVB exposure. It is possible, however, that the observed association of multivitamin use with 25(OH)D levels is mediated by unmeasured confounding factors, including higher dietary intake of vitamin D or higher UVB exposure among vitamin users.

Our logistic regression model provides a prediction of the likelihood of vitamin D deficiency using readily available clinical data. In this model, black and Asian race were the strongest predictors of risk of deficiency, followed by absence of multivitamin use. Spring season and male sex had significant but smaller effects. Overall, this model indicates that white individuals taking multivitamins are at relatively low risk of deficiency regardless of age, sex, or season (probability of deficiency ranging from 4% to 18%). All other groups, including those not traditionally thought to be at higher risk, however, were at risk of being vitamin D deficient (with probabilities ranging from 19% to 89%).

Models predictive of 25(OH)D levels have been developed by other investigators from large prospective cohort studies.[41,42] Predictors found in those studies have included race, age, sex, region of residence, vitamin D intake, total energy intake, body mass index, physical activity, smoking status, and season.[41,42] These models predicted 25(OH)D levels and the risk of deficiency similarly to our model. They differed from ours, however, in the homogeneous demographic makeup of the subjects who were, for the most part, older, male,[41] or Caucasian.[42] Additionally, subjects with potentially confounding medical conditions were not systematically excluded, thus, those subjects were not as "healthy" as our cohort. Finally, those models included predictors such as dietary vitamin D intake and physical activity score, which may be challenging to measure in a busy clinical setting.

Our study had some limitations. We did not collect information regarding dietary vitamin D intake, UVB exposure, or measures of adiposity. Regarding multi-vitamin use, we do not have data on vitamin D content or adherence, which may confound the effects of multivitamin use in different subgroups. Finally, our data are cross-sectional; thus, we do not have insight on longitudinal changes in 25(OH)D.