Abstract
Background: Association studies have suggested that lower circulating 25-hydroxyvitamin D [25(OH)D] in African Americans may partially underlie higher rates of cardiovascular disease and cancer in this population. Nonetheless, the relation between vitamin D supplementation and 25(OH)D concentrations in African Americans remains undefined.
Objective: Our primary objective was to determine the dose-response relation between vitamin D and plasma 25(OH)D.
Design: A total of 328 African Americans in Boston, MA, were enrolled over 3 winters from 2007 to 2010 and randomly assigned to receive a placebo or 1000, 2000, or 4000 IU vitamin D3/d for 3 mo. Subjects completed sociodemographic and dietary questionnaires, and plasma samples were drawn at baseline and 3 and 6 mo.
Results: Median plasma 25(OH)D concentrations at baseline were 15.1, 16.2, 13.9, and 15.7 ng/mL for subjects randomly assigned to receive the placebo or 1000, 2000, or 4000 IU/d, respectively (P = 0.63). The median plasma 25(OH)D concentration at 3 mo differed significantly between supplementation arms at 13.7, 29.7, 34.8, and 45.9 ng/mL, respectively (P < 0.001). An estimated 1640 IU vitamin D3/d was needed to raise the plasma 25(OH)D concentration to ≥20 ng/mL in ≥97.5% of participants, whereas a dose of 4000 IU/d was needed to achieve concentrations ≥33 ng/mL in ≥80% of subjects. No significant hypercalcemia was seen in a subset of participants.
Conclusions: Within African Americans, an estimated 1640 IU vitamin D3/d was required to achieve concentrations of plasma 25(OH)D recommended by the Institute of Medicine, whereas 4000 IU/d was needed to reach concentrations predicted to reduce cancer and cardiovascular disease risk in prospective observational studies. These results may be helpful for informing future trials of disease prevention. This trial was registered at clinicaltrials.gov as NCT00585637.
See corresponding editorial on page 427.
INTRODUCTION
Beyond vitamin D's role in bone and calcium metabolism, prospective observational cohort studies have suggested inverse associations between circulating 25-hydroxyvitamin D [25(OH)D]5, which is the accepted measure of vitamin D status, and the incidence and mortality of cardiovascular disease (CVD) (1, 2), diabetes (3, 4), and several cancers (5–10). In 2011, the Institute of Medicine (IOM) updated Dietary Reference Intakes for vitamin D (11). A target 25(OH)D concentration of 20 ng/mL was recommended on the basis of studies of fracture risk in whites, which led to a Recommended Dietary Allowance (RDA) (defined as the daily amount needed to achieve 20 ng/mL in ≥97.5% of the population) of 600 IU/d for adults aged 19–70 y and 800 IU/d for adults aged >70 y.
A profound disparity in vitamin D status according to skin pigmentation has complicated the determination of an RDA. Compared with whites, African Americans have virtually one-half concentrations of 25(OH)D (12, 13), largely attributable to decreased vitamin D synthesis in skin with a greater melanin content (14). Consequently, the optimal dose of vitamin D to achieve adequate 25(OH)D concentrations in African Americans is unknown. Moreover, CVD and several cancers are more prevalent and have higher mortality in African Americans than whites (15, 16), and differences in plasma 25(OH)D could partly account for these disparities (17, 18).
A critical need exists to quantify the vitamin D intake required by African Americans to achieve predefined targets of 25(OH)D. Therefore, we conducted a randomized clinical trial of 3 doses of vitamin D3 and placebo in African Americans residing in public-housing communities to assess the dose-response relation in this population and identify the amount needed to raise 25(OH)D to specific predefined thresholds.
SUBJECTS AND METHODS
Study population and design
This study was a randomized, double-blind, placebo-controlled trial of vitamin D3 supplementation in community-based African Americans (www.clinicaltrials.gov; NCT00585637). Participants were drawn from the Open Doors to Health, which is a colorectal cancer prevention study in 1554 subjects from 12 public-housing communities and community- and faith-based organizations in Boston (19). Subjects were 30–80 y old, understood written and spoken English, and self-identified as black or African American (20, 21). To minimize the contribution of UVB radiation to 25(OH)D concentrations (22), enrollment occurred from 31 December 2007 to 31 January 2008, 2 October 2008 to 31 March 2009, and 13 October 2009 to 31 March 2010. The derivation of the final cohort is shown in Figure 1. All subjects provided written consent, and the project was approved by institutional review boards of Harvard School of Public Health and Dana-Farber Cancer Institute.
FIGURE 1.
Consolidated Standards of Reporting Trials diagram. 25(OH)D, 25-hydroxyvitamin D.
Primary care physician (PCP) approval, which was sought after informed consent was obtained, was required to participate. Exclusion criteria included pregnancy, parathyroid, thyroid, or calcium disorders, sarcoidosis, a requirement for calcium channel blockers, type I diabetes, and concurrent active malignancies.
Study treatment and assessments
Participants were assigned to 4 arms that consisted of a placebo or 1000, 2000, or 4000 IU vitamin D3/d for 3 mo in a 1:1:1:1 ratio by using block random assignment stratified by age, sex, and enrollment month. Vitamin D3 capsules were indistinguishable and also contained 100 mg Ca (Pharmavite LLC). Actual vitamin D3 concentrations of capsules were 1291, 2557, and 5070 IU for 1000, 2000, and 4000 IU capsules, respectively, and the stability of capsules was confirmed at 3 y. Study statisticians generated the random allocation sequence, and subjects were enrolled by research assistants. All participants, providers, and study staff were blinded. Participants were followed for toxicity and compliance every 2 wk by phone and every 4 wk in person during supplementation. In addition, serum calcium was measured in subjects who were taking hydrochlorothiazide at 1 (n = 79) and 3 (n = 75) mo. Any subject shown to have a calcium concentration >10.5 mg/dL was immediately discontinued from the study, and the PCP notified. An additional subset of control participants, who did not take hydrochlorothiazide, also underwent calcium assays at 3 mo (n = 44). Electronic pill-dispenser systems and pill counts were also used to track compliance with study supplementation. The study was monitored closely by a data and safety monitoring board composed of external advisors.
Plasma samples were collected in lavender-top evacuated tube containing liquid EDTA at baseline and 3 and 6 mo for 25(OH)D determination. Assays were performed in a single batch by using a radioimmunoassay (23) (DiaSorin Inc) in the laboratory of Bruce Hollis (Medical University of South Carolina, Charleston, SC). Masked quality-control samples were interspersed in cases and all laboratory personnel were blinded. The mean CV of the assay was 9%, and National Institute of Standards and Technology reference ranges (±SDs) were met as follows: concentration 1, 23.3 ± 1.8 ng/mL; concentration 2, 14.6 ± 1.3 ng/mL; concentration 3, 38.6 ± 2.4 ng/mL; and concentration 4, 33.1 ± 2.6 ng/mL. Participants were also asked to complete questionnaires at baseline and 3 and 6 mo that addressed socioeconomic and demographic factors, dietary and lifestyle behaviors, and medication use. Specifically, dietary vitamin D and calcium intakes were assessed by using questions modeled after the validated semifrequency food questionnaire (24).
Skin pigmentation was measured by using a tristimulus colorimeter device (Photovolt 577 Reflectometer; Photovolt Instruments Inc) (25–27). Measurements were done on the upper inner sides of both arms. Reflectance readings were converted to the Commission International d'Eclairage L*a*b* system, and the L* variable (range: 0–100 for white to black) was used to represent skin pigmentation (27). Twenty-five participants had outlier values for L* because of calibration errors that were imputed with the median in regression models.
Study endpoints
The predefined primary endpoint was plasma 25(OH)D after 3 mo of supplementation, with the objective of identifying the dose of vitamin D3 required to raise 25(OH)D concentrations to ≥33 ng/mL in ≥80% of participants in the intent-to-treat population, regardless of compliance. In a secondary analysis, the dose required to reach this threshold in compliant subjects (defined as subjects taking ≥75% of required pills) was determined. Although the protocol initially specified thresholds of 32 and 40 ng/mL for analysis, 33 ng/mL was ultimately chosen in consultation with the data and safety monitoring board on the basis of published observational studies that showed that individuals with 25(OH)D greater than this concentration had significantly lower risk of developing and dying from CVD and cancer (7–10, 28). Moreover, in light of the IOM report that recommended a target 25(OH)D concentration of 20 ng/mL in 2011 (11) and ongoing debate about the preferred 25(OH)D concentration, we sought to determine the daily intake of vitamin D3 needed to achieve 20 ng/mL in 97.5% of the study cohort.
Statistical analysis
Baseline characteristics of subjects were compared between supplementation arms by using Fisher's exact test (Monte Carlo method) for categorical variables and the Kruskal-Wallis test for continuous variables. We tested for a linear trend in median plasma 25(OH)D concentrations in each supplementation arm by performing a quantile regression at each time point. We calculated the percentage of participants with a plasma 25(OH)D concentration ≥33 ng/mL and identified the 25(OH)D concentration at the 20th percentile at 3 and 6 mo. To predict the dose needed to meet the IOM target of 20 ng/mL, a mixed model was used to estimate the best linear, unbiased predictor of within-subject mean plasma 25(OH)D concentrations (29). The mixed model used all available doses and time points and included a random patient effect to account for the within-subject correlation and estimate the variance of the mean within-subject concentration. For each time point, the model fit a quadratic effect of dose on the mean plasma 25(OH)D concentration greater than a dose of 1000 IU vitamin D3/d and a single mean at dose 0 IU vitamin D3/d. To calculate the dose required to achieve plasma 25(OH)D concentrations ≥20 ng/mL at 3 mo in ≥97.5% of the study cohort on the basis of this model, 1.96 times the SD of the mean within-subject concentration was subtracted from the estimated quadratic fit. This curve was used to interpolate to the dose expected to include 97.5% of future subjects’ mean plasma 25(OH)D concentration >20 ng/mL. A 95% bootstrap percentile CI for the interpolated dose was calculated from 1000 bootstrap samples. The bootstrapping was stratified by dose, and the subject was the resampled unit.
In addition we explored predictors of plasma 25(OH)D by using linear regression. Variables evaluated in a stepwise selection included age, sex, year of study enrollment, skin pigmentation, BMI, exercise frequency, smoking status, regular multivitamin and vitamin D supplement use at baseline, dietary vitamin D intake, and baseline 25(OH)D. Individuals with missing data for any of these variables were excluded. We calculated the P-trend by using the vitamin D3 dose as a continuous variable in the final multivariable model (8, 30). Statistical interactions between treatment arm and potentially modifying covariates were assessed by using Wald's test of cross-product terms.
The statistical power for this trial was based on the intent-to-treat population of 80 subjects/arm, regardless of compliance with the study treatment. During the study, the sample size was expanded to allow ≤100 participants/arm to account for 1) the time gap between subject consent and PCP approval and 2) a 10% withdrawal and lost-to-follow-up rate; however, this increase was ultimately not necessary. With the use of a 2-sided t test at the 0.05 significance level, the minimum detectable difference in plasma 25(OH)D between treatment arms was 5.3 and 6.2 ng/mL with 80% and 90% power, respectively. For the secondary analysis in compliant participants, we had 80% power to detect a minimum between-arms rate difference of 0.26 [0.25 compared with 0.51 (ie, 25% of subjects who received 1000 IU vitamin D3/d and attained plasma 25(OH)D concentrations ≥33 ng/mL compared with 51% of subjects who received 4000 IU/d vitamin D3/d)] and >90% power to detect a rate difference of 0.30 (0.25 compared with 0.55) by using a 2-sided Fisher's exact test at the 0.05 significance level. All statistical analyses were performed with SAS 9.2 software (SAS Institute).
RESULTS
Baseline characteristics for the 328 randomly assigned participants (222 women and 106 men) are presented in Table 1. More participants in the placebo and 1000-IU/d arms had a past history of cancer than did subjects who received 2000 or 4000 IU/d. Otherwise, there were no significant differences in any of the characteristics. In the 292 participants with available plasma 25(OH)D at 3 mo, relevant baseline characteristics were also similar between arms (data not shown). Compliance was high at 95.7%, 96.6%, 96.5%, and 97.6% for placebo and 1000-, 2000-, and 4000-IU/d arms, respectively, and did not differ significantly between arms (P = 0.81).
TABLE 1.
Subject characteristics by supplementation arm1
| Vitamin D3 dose assignment (for 3 mo) |
|||||
| Characteristic | Placebo (n = 81) | 1000 IU/d (n = 81) | 2000 IU/d (n = 83) | 4000 IU/d (n = 83) | Total (n = 328) |
| Age (y) | 50.7 (44.2, 58.0)2 | 51.1 (43.4, 60.1) | 50.3 (43.4, 58.2) | 51.3 (44.1, 59.7) | 51.0 (43.6, 59.4) |
| Sex [n (%)] | |||||
| M | 27 (33.3) | 22 (27.2) | 28 (33.7) | 29 (34.9) | 106 (32.3) |
| F | 54 (66.7) | 59 (72.8) | 55 (66.3) | 54 (65.1) | 222 (67.7) |
| Born in United States [n (%)] | |||||
| Yes | 72 (88.9) | 69 (85.2) | 70 (84.3) | 67 (80.7) | 278 (84.8) |
| No | 9 (11.1) | 12 (14.8) | 12 (14.5) | 16 (19.3) | 49 (14.9) |
| Missing/unknown | 0 | 0 | 1 (1.2) | 0 | 1 (0.3) |
| Maximum education level [n (%)] | |||||
| Less than high school | 16 (19.8) | 18 (22.2) | 23 (27.7) | 29 (34.9) | 86 (25.3) |
| High school degree | 32 (39.5) | 29 (35.8) | 23 (27.7) | 16 (19.3) | 100 (30.5) |
| Vocational/some college | 18 (22.2) | 19 (23.5) | 19 (22.9) | 25 (30.1) | 81 (24.7) |
| College degree | 12 (14.8) | 11 (13.6) | 14 (16.9) | 10 (12.1) | 47 (14.3) |
| Postgraduate degree | 3 (3.7) | 4 (4.9) | 3 (3.6) | 3 (3.6) | 13 (3.9) |
| Missing/unknown | 0 | 0 | 1 (1.2) | 0 | 1 (0.3) |
| Employment status [n (%)] | |||||
| Working | 28 (34.6) | 31 (38.3) | 32 (38.6) | 31 (37.3) | 122 (37.2) |
| Not working | 53 (65.4) | 50 (61.7) | 51 (61.4) | 52 (62.7) | 206 (62.8) |
| Marital status [n (%)] | |||||
| Married | 23 (28.4) | 30 (37.0) | 23 (27.7) | 24 (28.9) | 100 (30.5) |
| Not married | 58 (71.6) | 51 (63.0) | 59 (71.1) | 58 (69.9) | 226 (68.9) |
| Missing/unknown | 0 | 0 | 1 (1.2) | 1 (1.2) | 2 (0.6) |
| Median household income [n (%)] | |||||
| <$10,000 | 33 (40.8) | 23 (28.4) | 27 (32.5) | 27 (32.5) | 110 (33.6) |
| $10,000–$19,999 | 15 (18.5) | 17 (21.0) | 17 (20.5) | 17 (20.5) | 66 (20.1) |
| $20,000–$29,999 | 11 (13.6) | 10 (12.3) | 16 (19.3) | 10 (12.0) | 47 (14.3) |
| $30,000–$39,999 | 4 (4.9) | 9 (11.1) | 4 (4.8) | 9 (10.8) | 26 (7.9) |
| $40,000–$49,999 | 4 (4.9) | 5 (6.2) | 4 (4.8) | 4 (4.8) | 17 (5.2) |
| ≥$50,000 | 9 (11.1) | 11 (13.6) | 11 (13.3) | 11 (13.3) | 42 (12.8) |
| Missing/unknown | 5 (6.2) | 6 (7.4) | 4 (4.8) | 5 (6.1) | 20 (6.1) |
| Skin tone (n = 303) (L*)3 | 45.1 (38.2, 50.4) | 44.9 (39.5, 50.2) | 44.4 (41.3, 50.7) | 44.6 (41.1, 48.4) | 44.7 (39.7, 50.1) |
| History of cancer [n (%)]4 | |||||
| Yes5 | 6 (7.4) | 6 (7.4) | 0 | 3 (3.6) | 15 (4.6) |
| No | 75 (92.6) | 74 (91.4) | 83 (100) | 79 (95.2) | 311 (94.8) |
| Missing/unknown | 0 | 1 (1.2) | 0 | 1 (1.2) | 2 (0.6) |
| History of hypertension [n (%)] | |||||
| Yes | 35 (43.2) | 35 (43.2) | 36 (43.4) | 35 (42.2) | 141 (43.0) |
| No | 36 (44.4) | 33 (40.7) | 36 (43.4) | 34 (41.0) | 139 (42.4) |
| Missing/unknown | 10 (12.4) | 13 (16.1) | 11 (13.2) | 14 (16.8) | 48 (14.6) |
| No. of visits with doctor or nurse practitioner in past year | 3.0 (2.0, 6.0) | 4.0 (2.0, 5.0) | 3.0 (2.0, 5.0) | 3.0 (2.0, 5.0) | 3.0 (2.0, 5.0) |
| Smoking status [n (%)] | |||||
| Never | 33 (40.7) | 36 (44.4) | 33 (39.8) | 44 (53.0) | 146 (44.5) |
| Past | 20 (24.7) | 16 (19.8) | 27 (32.5) | 20 (24.1) | 83 (25.3) |
| Current | 28 (34.6) | 29 (35.8) | 23 (27.7) | 19 (22.9) | 99 (30.2) |
| Postmenopausal hormone use [n (%)]6 | |||||
| Yes | 1 (1.9) | 0 | 0 | 0 | 1 (0.5) |
| No | 37 (68.5) | 41 (69.5) | 30 (54.5) | 37 (68.5) | 145 (65.3) |
| Missing/unknown | 16 (29.6) | 18 (30.5) | 25 (45.5) | 17 (31.5) | 76 (34.2) |
| Travel to sunny region during supplementation period [n (%)] | |||||
| Yes | 5 (6.2) | 4 (5.0) | 5 (6.0) | 7 (8.4) | 21 (6.4) |
| No | 68 (83.9) | 62 (76.5) | 70 (84.3) | 72 (86.8) | 272 (82.9) |
| Missing/unknown | 8 (9.9) | 15 (18.5) | 8 (9.7) | 4 (4.8) | 35 (10.7) |
| Frequency of exercise (d/wk)7 (n = 325) | 3.0 (0.5, 5.0) | 3.0 (1.0, 5.0) | 3.0 (0, 5.0) | 3.0 (0, 5.0) | 3.0 (0, 5.0) |
| BMI (kg/m2) | |||||
| Baseline (n = 325) | 31.2 (26.5, 35.9) | 30.5 (27.0, 37.5) | 31.9 (26.2, 36.9) | 31.4 (27.4, 35.7) | 31.2 (26.8, 36.3) |
| 3 mo (n = 290) | 30.9 (26.3, 36.4) | 30.4 (26.6, 36.9) | 32.5 (27.2, 37.1) | 32.0 (27.1, 35.8) | 31.6 (26.5, 36.8) |
| 6 mo (n = 291) | 31.8 (27.8, 37.0) | 29.9 (27.3, 37.3) | 32.5 (26.4, 37.1) | 30.9 (27.0, 35.1) | 31.0 (26.6, 36.9) |
| Dietary vitamin D intake (IU)8 | |||||
| Baseline (n = 328) | 147.3 (71.4, 262.8) | 162.5 (92.6, 295.5) | 144.0 (58.0, 265.1) | 198.1 (83.2, 306.4) | 167.5 (72.3, 291.8) |
| 3 mo (n = 293) | 175.8 (93.7, 277.4) | 201.4 (81.0, 394.7) | 191.4 (93.3, 332.5) | 248.4 (134.5, 413.4) | 204.7 (93.7, 341.1) |
| 6 mo (n = 292) | 181.5 (86.5, 307.9) | 209.3 (91.5, 308.2) | 251.8 (67.2, 405.8) | 245.3 (108.9, 398.1) | 227.1 (86.6, 368.6) |
| Dietary calcium intake (mg)8 | |||||
| Baseline (n = 328) | 277.0 (171.7, 632.3) | 422.9 (226.1, 795.9) | 318.8 (172.7, 637.4) | 445.9 (198.6, 780.4) | 356.6 (188.6, 693.8) |
| 3 mo (n = 293) | 361.1 (198.6, 660.2) | 394.4 (169.8, 856.8) | 323.8 (197.8, 878.7) | 527.4 (256.9, 995.0) | 404.2 (197.8, 789.6) |
| 6 mo (n = 292) | 342.7 (185.9, 769.5) | 430.2 (244.2, 886.6) | 420.3 (203.1, 836.4) | 453.3 (239.1, 825.9) | 424.6 (217.2, 799.9) |
| Regular multivitamin use [n (%)]9 | |||||
| Yes | 10 (12.3) | 18 (22.2) | 15 (18.1) | 22 (26.5) | 65 (19.8) |
| No | 71 (87.7) | 63 (77.8) | 67 (80.7) | 61 (73.5) | 262 (79.9) |
| Missing/unknown | 0 | 0 | 1 (1.2) | 0 | 1 (0.3) |
| Regular vitamin D supplement use [n (%)]9 | |||||
| Yes | 8 (9.9) | 6 (7.4) | 2 (2.4) | 8 (9.6) | 24 (7.3) |
| No | 72 (88.9) | 74 (91.4) | 81 (97.6) | 73 (88.0) | 300 (91.5) |
| Missing/unknown | 1 (1.2) | 1 (1.2) | 0 | 2 (2.4) | 4 (1.2) |
| Regular calcium supplement use [n (%)]9 | |||||
| Yes | 7 (8.7) | 9 (11.1) | 7 (8.4) | 9 (10.8) | 32 (9.8) |
| No | 73 (90.1) | 71 (87.7) | 76 (91.6) | 74 (89.2) | 294 (89.6) |
| Missing/unknown | 1 (1.2) | 1 (1.2) | 0 | 0 | 2 (0.6) |
| Calendar month of enrollment [n (%)] | |||||
| October | 14 (17.3) | 12 (14.8) | 14 (16.9) | 8 (9.7) | 48 (14.6) |
| November | 9 (11.1) | 7 (8.6) | 7 (8.4) | 11 (13.2) | 34 (10.4) |
| December | 15 (18.5) | 14 (17.3) | 14 (16.9) | 15 (18.1) | 58 (17.7) |
| January | 18 (22.2) | 24 (29.7) | 21 (25.3) | 20 (24.1) | 83 (25.3) |
| February | 11 (13.6) | 9 (11.1) | 11 (13.2) | 11 (13.2) | 42 (12.8) |
| March | 14 (17.3) | 15 (18.5) | 16 (19.3) | 18 (21.7) | 63 (19.2) |
| Year of study enrollment [n (%)] | |||||
| 2007–2008 | 9 (11.1) | 13 (16.0) | 10 (12.0) | 13 (15.7) | 45 (13.7) |
| 2008–2009 | 45 (55.6) | 40 (49.4) | 42 (50.6) | 39 (47.0) | 166 (50.6) |
| 2009–2010 | 27 (33.3) | 28 (34.6) | 31 (37.4) | 31 (37.3) | 117 (35.7) |
There were no significant differences in subject characteristics across supplementation arms except where indicated.
Median; 25th, 75th percentiles in parentheses (all such values).
L* is the lightness variable from the Commission International d'Eclairage (CIE) L*a*b* system, ranging from 0 to 100. Outlier values were imputed with the median value (see Subjects and Methods).
= 0.03 (Fisher's exact test with the Monte Carlo method).
Reported cancers include breast cancer, cervical cancer, uterine cancer, lung cancer, prostate cancer, and sarcoma.
Percentages calculated from a total of 222 females.
Exercise defined as moderate to vigorous physical activity for at least 30 min, resulting in a faster-than-normal heart rate, sweating, and deep breathing.
Refers to the intake during the preceding month.
Defined as supplement use for 7 d/wk during the preceding month.
Impact of vitamin D3 supplementation on plasma 25(OH)D
Plasma 25(OH)D concentrations at baseline and 3 and 6 mo are shown in Table 2 and Figure 2. In 328 subjects, the median plasma 25(OH)D concentration at baseline was 15.3 ng/mL (25th, 75th percentiles: 10.4, 22.8) and did not differ significantly between arms (P-trend = 0.77). After 3 mo supplementation, 292 subjects (89.0%) provided plasma for follow-up 25(OH)D determination. Circulating plasma 25(OH)D concentrations at 3 mo differed significantly by supplementation arm, with medians of 13.7, 29.7, 34.8, and 45.9 ng/mL for placebo and 1000-, 2000-, and 4000-IU/d arms, respectively (effect estimate ± SE: 8.0 ± 0.65 ng/mL per 1000 IU vitamin D3; P-trend < 0.001). Notably, plasma 25(OH)D decreased at 3 mo in subjects treated with the placebo. In the intent-to-treat population, the 4000-IU/d arm achieved plasma 25(OH)D concentrations ≥33 ng/mL in 88.0% of subjects compared with in 2.5%, 19.8%, and 53.0% of subjects for placebo and 1000- and 2000-IU/d arms, respectively (P < 0.001). Correspondingly, the 20th percentile plasma 25(OH)D concentration in each of the supplementation arms was 6.7, 22.6, 26.8, and 37.2 ng/mL for placebo and 1000-, 2000-, and 4000-IU/d arms, respectively (P < 0.001). In comparison, the proportion of subjects in each arm who reached the IOM target of 20 ng/mL and The Endocrine Society target of 30 ng/mL (31) at 3 mo was 19.8%, 69.1%, 88.0%, and 92.8% and 3.7%, 37.0%, 63.8%, and 90.4%, respectively (P < 0.001 for both) (data not shown).
TABLE 2.
Plasma 25(OH)D concentrations by supplementation arm1
| Vitamin D3 dose assignment (for 3 mo) |
|||||||
| Variable | Placebo | 1000 IU/d | 2000 IU/d | 4000 IU/d | Total | Effect estimate per 1000 IU vitamin D32 | P |
| ng/mL | |||||||
| Baseline plasma 25(OH)D (ng/mL) | 15.1 (10.4, 23.6)3 | 16.2 (11.0, 22.7) | 13.9 (9.5, 22.3) | 15.7 (11.0, 23.3) | 15.3 (10.4, 22.8) | 0.12 ± 0.40 | 0.774 |
| Subjects (n) | 81 | 81 | 83 | 83 | 328 | — | — |
| Plasma 25(OH)D at 3 mo (ng/mL) | 13.7 (7.2, 18.6) | 29.7 (25.6, 32.9) | 34.8 (28.8, 41.0) | 45.9 (39.4, 55.2) | 31.7 (20.5, 41.8) | 8.00 ± 0.65 | <0.0014 |
| Subjects (n) | 71 | 67 | 76 | 78 | 292 | — | — |
| Plasma 25(OH)D at 6 mo (ng/mL) | 18.1 (12.2, 23.3) | 21.2 (16.8, 27.8) | 27.0 (20.6, 31.1) | 31.2 (26.5, 35.9) | 24.6 (18.0, 31.2) | 3.28 ± 0.35 | <0.0014 |
| Subjects (n) | 75 | 68 | 72 | 77 | 292 | — | — |
| Δ 3 mo to baseline plasma 25(OH)D (ng/mL) | −2.3 (−5.4, 1.7) | 10.8 (2.5, 18.9) | 19.3 (11.6, 26.2) | 30.3 (21.5, 37.6) | 15.0 (1.7, 25.1) | 8.60 ± 0.41 | <0.0014 |
| Subjects (n) | 71 | 67 | 76 | 78 | 292 | — | — |
| Δ 6–3 mo plasma 25(OH)D (ng/mL) | 3.8 (0.9, 7.6) | −5.4 (−13.0, 0.3) | −9.4 (−14.4, −4.1) | −15.7 (−20.9, −9.8) | −7.2 (−14.5, 1.0) | −4.92 ± 0.35 | <0.0014 |
| Subjects (n) | 69 | 67 | 71 | 76 | 283 | — | — |
| Plasma 25(OH)D ≥33 ng/mL at 3 mo [n (%)] | — | <0.0015 | |||||
| Yes | 2 (2.5) | 16 (19.8) | 44 (53.0) | 73 (88.0) | 135 (41.1) | — | — |
| No | 69 (85.2) | 51 (63.0) | 32 (38.6) | 5 (6.0) | 157 (47.9) | — | — |
| Missing | 10 (12.3) | 14 (17.2) | 7 (8.4) | 5 (6.0) | 36 (11.0) | — | — |
| Plasma 25(OH)D ≥33 ng/mL at 6 mo [n (%)] | — | <0.0015 | |||||
| Yes | 6 (7.4) | 6 (7.4) | 14 (16.8) | 29 (34.9) | 55 (16.8) | — | — |
| No | 69 (85.2) | 62 (76.5) | 58 (69.9) | 48 (57.8) | 237 (72.2) | — | — |
| Missing | 6 (7.4) | 13 (16.1) | 11 (13.3) | 6 (7.3) | 36 (11.0) | — | — |
| 20th percentile plasma 25(OH)D (ng/mL) | |||||||
| 3 mo | 6.7 | 22.6 | 26.8 | 37.2 | 17.0 | — | <0.0016 |
| 6 mo | 11.2 | 15.9 | 19.1 | 24.9 | 16.8 | — | <0.0016 |
1 ng/mL = 2.496 nmol/L. 25(OH)D, 25-hydroxyvitamin D.
All values are effect estimates ± SEs.
Median; 25th, 75th percentiles in parentheses (all such values).
P-linear trend in median plasma 25(OH)D concentration calculated by performing quantile regression.
Calculated by using Fisher's exact test (Monte Carlo method).
Calculated by using quantile regression.
FIGURE 2.
Plasma 25(OH)D concentrations (ng/mL) at baseline and 3 and 6 mo according to vitamin D3 dose. 25(OH)D, 25-hydroxyvitamin D.
In 289 compliant participants, the 4000-IU/d dose remained the only arm that achieved the primary objective, with 93.4% of participants having a plasma 25(OH)D concentration ≥33 ng/mL. In contrast, only 2.8%, 23.9%, and 56.0% of compliant subjects who received the placebo or 1000 or 2000 IU/d, respectively, achieved 25(OH)D concentrations ≥33 ng/mL (P < 0.001) (data not shown).
At 6 mo, 292 subjects (89.0%) were available for plasma 25(OH)D determination. Plasma 25(OH)D differed significantly by supplementation arm, with medians of 18.1, 21.2, 27.0, and 31.2 ng/mL, respectively (effect estimate ± SE: 3.28 ± 0.35 ng/mL per 1000 IU vitamin D3; P-trend < 0.001). Except for the placebo arm, median concentrations of 25(OH)D decreased compared with those at 3 mo, although not quite to baseline concentrations.
In light of the 2011 IOM report that established a 25(OH)D concentration of 20 ng/mL to be sufficient for bone health (11), we determined the dose of vitamin D required in our African American cohort to reach that threshold, similar to in previously published studies (32). The empirical Bayesian prediction interval to bound 97.5% of future subjects’ mean plasma 25(OH)D concentrations intersected 20 ng/mL at 1640 IU/d (95% CI: 1490, 2020 IU/d) (Figure 3), which indicated that 97.5% of our population would maintain a mean plasma 25(OH)D concentration of 20 ng/mL at an estimated dosage of 1640 IU/d.
FIGURE 3.
Graphical display of plasma 25(OH)D concentrations (ng/mL) at 3 mo and dose of vitamin D3 supplementation (n = 67 for 1000 IU/d; n = 76 for 2000 IU/d; n = 78 for 4000 IU/d). The solid line is a quadratic fit to the observed mean plasma 25(OH)D concentration. The dashed line falls below the mean line by 1.96 SDs of the distribution of the estimated within-subject mean concentration (obtained from the random patient effect in the mixed model) and represents the empirical Bayesian prediction interval to bound 97.5% of future subjects’ mean plasma 25(OH)D concentrations. This prediction interval crosses the 20-ng/mL line at 1640 IU/d (95% CI: 1490, 2020 IU/d), indicating that an estimated dose of 1640 IU vitamin D3/d is required to achieve an individual mean plasma 25(OH)D concentration at ≥20 ng/mL in ≥97.5% of the study population. 25(OH)D, 25-hydroxyvitamin D.
Predictors of plasma 25(OH)D concentration and response to supplementation
We explored the impact of potential confounding variables on circulating 25(OH)D (Table 3). Age, regular multivitamin and vitamin D supplement use, smoking status, and year of study enrollment were significant independent predictors of baseline plasma 25(OH)D. Predictors of plasma 25(OH)D at 3 mo included the supplementation arm, baseline 25(OH)D, age, BMI, and year of enrollment, and predictors at 6 mo included the supplementation arm, baseline 25(OH)D, BMI, and year of enrollment. Of note, skin pigmentation was not an independent predictor of the 25(OH)D concentration at any time point.
TABLE 3.
Predictors of plasma 25(OH)D concentrations at various time points from a multiple linear regression model by using stepwise selection1
| Covariate | Effect estimate2 | P |
| ng/mL | ||
| Baseline (n = 323) | ||
| Age (y) | 0.17 ± 0.04 | <0.0001 |
| Regular vitamin D supplement use at baseline3 | ||
| No | Referent | — |
| Yes | 8.34 ± 1.72 | <0.0001 |
| Regular multivitamin use at baseline3 | ||
| No | Referent | — |
| Yes | 5.57 ± 1.17 | <0.0001 |
| Smoking status | ||
| Current | Referent | — |
| Past | 1.75 ± 1.15 | 0.13 |
| Never | 2.28 ± 1.01 | 0.02 |
| Year of study enrollment | ||
| 2007–2008 | Referent | — |
| 2008–2009 | 3.38 ± 1.29 | 0.009 |
| 2009–2010 | 4.43 ± 1.34 | 0.001 |
| 3 mo (n = 290) | ||
| Supplementation arm4 | ||
| Placebo | Referent | — |
| 1000 IU vitamin D3/d | 13.93 ± 1.43 | <0.0001 |
| 2000 IU vitamin D3/d | 21.39 ± 1.39 | <0.0001 |
| 4000 IU vitamin D3/d | 32.46 ± 1.38 | <0.0001 |
| Baseline plasma 25(OH)D (ng/mL) | 0.44 ± 0.06 | <0.0001 |
| Age (y) | 0.08 ± 0.04 | <0.001 |
| BMI (kg/m2) | −0.14 ± 0.06 | 0.04 |
| Year of study enrollment | ||
| 2007–2008 | Referent | — |
| 2008–2009 | 0.95 ± 1.49 | 0.52 |
| 2009–2010 | 3.62 ± 1.57 | 0.02 |
| 6 mo (n = 290) | ||
| Supplementation arm5 | ||
| Placebo | Referent | — |
| 1000 IU vitamin D3/d | 3.33 ± 1.06 | 0.002 |
| 2000 IU vitamin D3/d | 8.01 ± 1.04 | <0.0001 |
| 4000 IU vitamin D3/d | 11.98 ± 1.02 | <0.0001 |
| Baseline plasma 25(OH)D (ng/mL) | 0.52 ± 0.04 | <0.0001 |
| BMI (kg/m2) | −0.18 ± 0.05 | 0.0003 |
| Year of study enrollment | ||
| 2007–2008 | Referent | — |
| 2008–2009 | −2.44 ± 1.14 | 0.03 |
| 2009–2010 | −0.84 ± 1.18 | 0.48 |
All values are effect estimates ± SEs. 1 ng/mL = 2.496 nmol/L. Variables evaluated by using stepwise selection included age, sex, year of study enrollment, skin pigmentation, BMI at baseline, exercise frequency at baseline, smoking status, regular multivitamin use at baseline, regular vitamin D supplement use at baseline, dietary vitamin D intake at the relevant time point, travel to a sunny region during the preceding 3 mo (3- and 6-mo analyses only), and baseline plasma 25(OH)D concentration (3- and 6-mo analyses only). Variables were selected for inclusion in the model if P ≤ 0.10. The treatment arm was forced into the model in the 3- and 6-mo analyses. 25(OH)D, 25-hydroxyvitamin D.
All values are effect estimates ± SEs.
Defined as supplement use 7 d/wk during the preceding month.
Calculated by using the vitamin D3 dose as a continuous variable in the multiple linear regression model: 4P-trend < 0.001,
Calculated by using the vitamin D3 dose as a continuous variable in the multiple linear regression model: 5P-trend = 0.10.
We assessed whether any baseline characteristics modified the response to vitamin D supplementation. Current smokers, participants with lower BMI, and those who were not taking regular multivitamins or vitamin D supplements at baseline showed greater increases in plasma 25(OH)D per unit dose of vitamin D3 (Table 4). Of note, baseline 25(OH)D was lower in many subgroups with a greater response, which was consistent with findings from other studies (33).
TABLE 4.
Response in plasma 25(OH)D concentrations to vitamin D3 supplementation across strata of selected variables1
| Subgroup | Subjects | Median plasma 25(OH)D at baseline | Median plasma 25(OH)D at 3 mo | Increase (±SE) in Δ plasma 25(OH)D per 1000 IU vitamin D3/d2 | P-trend3 | P-interaction4 |
| n | ng/mL | ng/mL | ng/mL | |||
| Age5 | — | 0.42 | ||||
| <51.0 y | 142 | 12.8 | 31.3 | 7.45 ± 0.50 | <0.0001 | — |
| ≥51.0 y | 150 | 18.5 | 32.2 | 8.01 ± 0.49 | <0.0001 | — |
| Sex | — | 0.97 | ||||
| M | 92 | 13.3 | 32.0 | 7.76 ± 0.62 | <0.0001 | — |
| F | 200 | 16.0 | 31.7 | 7.73 ± 0.42 | <0.0001 | — |
| Skin tone5 | — | 0.82 | ||||
| <44.7 L* | 134 | 15.1 | 32.2 | 7.73 ± 0.51 | <0.0001 | — |
| ≥44.7 L* | 137 | 14.2 | 31.7 | 7.56 ± 0.50 | <0.0001 | — |
| Smoking status | — | 0.006 | ||||
| Current | 82 | 12.8 | 29.0 | 9.37 ± 0.67 | <0.0001 | — |
| Never or past | 210 | 16.7 | 32.2 | 7.20 ± 0.40 | <0.0001 | — |
| Travel to sunny region during preceding 3 mo | — | 0.56 | ||||
| Yes | 21 | 20.3 | 33.6 | 8.45 ± 1.22 | <0.0001 | — |
| No or unknown | 269 | 14.4 | 31.7 | 7.70 ± 0.36 | <0.0001 | — |
| Frequency of exercise at baseline5 | — | 0.22 | ||||
| <3.0 d/wk | 119 | 14.0 | 30.4 | 7.18 ± 0.54 | <0.0001 | — |
| ≥3.0 d/wk | 170 | 16.5 | 32.0 | 8.07 ± 0.46 | <0.0001 | — |
| BMI at baseline | — | 0.03 | ||||
| <30 kg/m2 (nonobese) | 130 | 15.5 | 32.6 | 8.62 ± 0.52 | <0.0001 | — |
| ≥30 kg/m2 (obese) | 160 | 14.5 | 31.6 | 7.08 ± 0.45 | <0.0001 | — |
| Dietary vitamin D intake reported at 3 mo5 | — | 0.56 | ||||
| <204.7 IU | 145 | 13.0 | 30.4 | 7.99 ± 0.50 | <0.0001 | — |
| ≥204.7 IU | 145 | 17.0 | 32.9 | 7.59 ± 0.49 | <0.0001 | — |
| Regular multivitamin use at baseline6 | — | 0.02 | ||||
| Yes | 58 | 23.4 | 34.3 | 6.20 ± 0.77 | <0.0001 | — |
| No | 233 | 13.3 | 31.3 | 8.15 ± 0.39 | <0.0001 | — |
| Regular vitamin D supplement use at baseline6 | — | 0.02 | ||||
| Yes | 21 | 26.8 | 29.7 | 5.20 ± 1.13 | <0.0001 | — |
| No | 267 | 14.0 | 31.7 | 8.06 ± 0.36 | <0.0001 | — |
| Year of study enrollment | — | 0.26 | ||||
| 2007–2008 | 45 | 14.0 | 31.4 | 8.44 ± 0.91 | <0.0001 | — |
| 2008–2009 | 166 | 15.1 | 30.3 | 7.11 ± 0.49 | <0.0001 | — |
| 2009–2010 | 117 | 17.1 | 34.3 | 8.32 ± 0.57 | <0.0001 | — |
1 ng/mL = 2.496 nmol/L. L*, lightness variable from the Commission International d’Eclairage (CIE) L*a*b* system, ranging from 0 to 100; 25(OH)D, 25-hydroxyvitamin D.
Adjusted for age, year of study enrollment, BMI at baseline, baseline 25(OH)D concentration, and the covariate for which subgroups were being explored.
Calculated by using vitamin D3 dose as a continuous variable in the multiple linear regression model.
Calculated by using Wald's test of cross-product terms.
Cutoff defined by the median value.
Defined as supplement use for 7 d/wk during the preceding month.
Adverse events
Two subjects reported symptoms that were potentially attributable to hypercalcemia (pruritis and polydipsia with polyuria), but subsequent calcium assays were normal. At 1 mo, 79 participants who were taking hydrochlorothiazide were required to have calcium concentrations checked. At 3 mo, 119 participants had serum calcium assayed, 75 (63.0%) subjects of whom were taking hydrochlorothiazide. There were no significant differences in calcium concentrations between arms at 1 mo (P = 0.14) and 3 mo (P = 0.52) (see Supplemental Table 1 under “Supplemental data” in the online issue). In 128 participants with available calcium at either time point, 5 subjects (3.9%) were shown to have concentrations >10.5 mg/dL (range: 10.7–11.2 mg/dL), and all of them were asymptomatic and taking hydrochlorothiazide. Four subjects had elevated calcium at 1 mo and were discontinued from the study per protocol (3 subjects were receiving 1000 IU/d, and one subject was receiving 2000 IU/d). The fifth subject was assigned to receive 4000 IU/d and had elevated calcium at 3 mo at the conclusion of supplementation and, therefore, was kept in the final analysis.
DISCUSSION
In community-based African Americans in Boston, supplementation with 4000 IU vitamin D3/d for 3 mo resulted in 88.0% of subjects achieving plasma 25(OH)D concentrations ≥33 ng/mL, whereas an estimated 1640 IU/d was required for ≥97.5% of subjects to reach 20 ng/mL. Hypercalcemia was not seen at these doses in a subset of participants with available data, which was consistent with the 2011 IOM report (11). To our knowledge, this is the largest randomized, placebo-controlled trial to evaluate vitamin D dosing in African Americans.
Our study filled an important knowledge gap because previous randomized trials of vitamin D have been largely confined to whites and have included few, if any, African Americans. A recent trial randomly assigned 163 postmenopausal white women with 25(OH)D concentrations ≤20 ng/mL to 7 doses of vitamin D3 compared with a placebo for 12 mo and determined the RDA to be 800 IU/d for a target of 20 ng/mL and 1600 IU/d for a target of 30 ng/mL (32). The authors subsequently published results of 79 postmenopausal African American women within their study and showed that the RDA was similar to that for whites at 800 IU/d for the IOM target of 20 ng/mL and 1600 IU/d for 30 ng/mL (34). In contrast, in a separate trial of 79 younger African American women conducted by the same group, an estimated RDA of 1200 IU/d was reported (35), which was more consistent with our findings. Potential reasons for the disparate results, despite similar statistical methods, included a much smaller sample size, lower baseline 25(OH)D concentrations, different age groups, and a longer period of supplementation in the Gallagher et al (32, 34) populations compared with in the current study.
Our findings are relevant in light of the 2011 IOM report that recommended an RDA of 600 IU/d for adults ≤70 y old to achieve a target 25(OH)D concentration of 20 ng/mL on the basis of studies of hip fractures in whites (11). The IOM acknowledged that additional research was needed to clarify the vitamin D dose-response relation and impact on skeletal health in African Americans. High rates of vitamin D deficiency have long been well documented in African Americans (36). Beyond differences in the melanin content, disparities in BMI and lifestyle behaviors also contribute with higher rates of obesity (37), lower intake of vitamin D-rich foods (38), and less sun exposure (39) in African Americans. An underlying germline variation in genes associated with vitamin D metabolism may also play a role (EM Awumey, BW Hollis, NH Bell, AW Norman, R Bouillon, and M Thomasset, unpublished observations, 1997) (40). Indeed, data have suggested that serum concentrations of vitamin D–binding protein are significantly lower in African Americans than whites (41, 42), which lead to lower total 25(OH)D concentrations.
However, the clinical significance of low 25(OH)D in African Americans is unknown because the frequency of fractures is lower in this population (43, 45). Possible explanations for this relation include anatomical differences in hip geometry, with increased femoral cortical thickness and a shorter hip-axis length compared with in whites (44) and a blunted skeletal response to parathyroid hormone that leads to lower rates of bone resorption (45). Other potential factors area higher bone mineral density (46) and prevalence of obesity (43). Indeed, a randomized, placebo-controlled trial of vitamin D in 208 African American women reported no beneficial effect on bone mineral density after 3 y of treatment (47).
Independent of bone health, observational studies have implicated vitamin D as a mechanism that potentially underlies racial disparities in CVD and cancer (17, 18), although this hypothesis is still controversial. In large, prospective cohort studies of predominantly white populations, a 25(OH)D concentration ∼33 ng/mL has been associated with decreased CVD and cancer risk and improved survival (7–10, 28). In the NHANES III cohort, participants with 25(OH)D concentrations ≥37 ng/mL had a lower prevalence of several cardiovascular risk factors (28). In contrast, a randomized, placebo-controlled, clinical trial of elderly women reported higher rates of falls and fractures with massive yearly doses of vitamin D3 (500,000 IU) compared with a placebo, and an observational study reported increased risk of pancreatic cancer with 25(OH)D concentrations ≥40 ng/mL. Therefore, the IOM concluded that the effect of vitamin D on cancer, immune disorders, and CVD has not been sufficiently established. An ongoing study, the Vitamin D and Omega-3 Trial (VITAL), is comparing 2000 IU vitamin D3/d to a placebo for disease prevention with a planned oversampling of African American participants.
Our study had several strengths. The randomized, placebo-controlled design allowed for an objective assessment of effects of the vitamin D3 dose on plasma 25(OH)D while controlling for confounding factors. Serum calcium concentrations were monitored, and no adverse effects were seen. Pill compliance and follow-up rates for blood draws and questionnaires were also high. There was minimal confounding of the effect of vitamin D supplementation by UVB exposure because the study was conducted in the Northeastern United States during winter, travel to sunnier regions was infrequent, and regular supplement use at baseline was low. The limited use of vitamin D supplements after the treatment period also allowed us to examine the natural trend in plasma 25(OH)D once supplementation ended. Finally, our study was able to quantify skin pigmentation in subjects and showed that it was not a significant predictor of plasma 25(OH)D.
Several limitations also deserve comment. Our cohort was supplemented for only 3 mo, which could have affected the estimation of the dose needed to reach certain thresholds of plasma 25(OH)D, and the endpoint of the study was not the long-term benefit of supplementation, which is currently being addressed by the Vitamin D and Omega-3 Trial (VITAL). However, data have suggested that a steady state of 25(OH)D is reached at 3 mo (48). Indeed, in a study of 50 colorectal cancer patients treated with 2000 IU vitamin D/d for 6 mo, the mean 25(OH)D concentration was 31.6 ng/mL at 3 mo and did not increase substantially at 6 mo (mean: 33.8 ng/mL) (49). Although we selected a target plasma 25(OH)D concentration ≥33 ng/mL on the basis of prospective observational studies, clinical trials that have assessed the effect of achieving this concentration are not available to our knowledge. In addition, we studied only 3 doses of vitamin D3, with the highest being 4000 IU/d; therefore, we were not able to evaluate the influence of more-intermediate or higher doses on plasma 25(OH)D concentrations.
In conclusion, in African Americans residing in a public-housing community, 1640 IU vitamin D3/d was necessary to achieve a target plasma concentration 25(OH)D of 20 ng/mL, whereas 4000 IU/d was needed to raise plasma 25(OH)D to a concentration potentially associated with reduced cancer and CVD risk in selected prospective observational studies. These findings may be helpful when designing future studies of disease prevention in African Americans.
Supplementary Material
Acknowledgments
We thank the following members of the data and safety monitoring board for their contributions to the project and manuscript: Debra Prothrow-Stith, John A Baron, and Caren G Solomon. We also thank Walter C Willett for his input on the design of the study, data interpretation, and manuscript. We gratefully acknowledge the efforts of the Vitamin D Research Team (Jennifer Bido, Christian Brown, Tamara Buchanan, Lindsay Cloutier, Molly Coeling, Ruth Lederman, Kamala Loscocco, Yudy Muneton, Bianka Recinos, Sanghamitra Savadatti, and Qi Wang) and Dana-Farber Cancer Institute phlebotomists (Jean Epps and Merline McKenzie). Finally, we thank the Resident Service Coordinators at collaborating housing sites, Maxi Hudson from Jubilee Christian Church, and all participants in the study.
The authors’ responsibilities were as follows—KN, BFD, ATC, PDC, GGB, ELG, JAM, KME, and CSF: designed the research; KN, JBS, BFD, ATC, BWH, PDC, EG-S, KME, and CSF: conducted the research; JBS, EG-S, KME, and CSF: provided essential materials and subjects; KN, JBS, ELG, KME, and CSF: wrote the manuscript; KN: had primary responsibility for the final content of the manuscript; and all authors: analyzed data and read and approved the final manuscript. BWS has received support from DiaSorin SpA for serving as an academic consultant. KN, JBS, BFD, ATC, PDC, GGB, ELG, EG-S, JAM, KME, and CSF reported no conflicts of interest.
Footnotes
Abbreviations used: CVD, cardiovascular disease; IOM, Institute of Medicine; PCP, primary care physician; RDA, Recommended Dietary Allowance; 25(OH)D, 25-hydroxyvitamin D.
REFERENCES
- 1.Wang TJ, Pencina MJ, Booth SL, Jacques PF, Ingelsson E, Lanier K, Benjamin EJ, D'Agostino RB, Wolf M, Vasan RS. Vitamin D deficiency and risk of cardiovascular disease. Circulation 2008;117:503–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Giovannucci E, Liu Y, Hollis BW, Rimm EB. 25-hydroxyvitamin D and risk of myocardial infarction in men: a prospective study. Arch Intern Med 2008;168:1174–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Pittas AG, Sun Q, Manson JE, Dawson-Hughes B, Hu FB. Plasma 25-hydroxyvitamin D concentration and risk of incident type 2 diabetes in women. Diabetes Care 2010;33:2021–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Liu E, Meigs JB, Pittas AG, Economos CD, McKeown NM, Booth SL, Jacques PF. Predicted 25-hydroxyvitamin D score and incident type 2 diabetes in the Framingham Offspring Study. Am J Clin Nutr 2010;91:1627–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Lappe JM, Travers-Gustafson D, Davies KM, Recker RR, Heaney RP. Vitamin D and calcium supplementation reduces cancer risk: results of a randomized trial. Am J Clin Nutr 2007;85:1586–91. [DOI] [PubMed] [Google Scholar]
- 6.Mohr SB, Gorham ED, Alcaraz JE, Kane CJ, Macera CA, Parsons JK, Wingard DL, Garland CF. Serum 25-hydroxyvitamin D and prevention of breast cancer: pooled analysis. Anticancer Res 2011;31:2939–48. [PubMed] [Google Scholar]
- 7.Gorham ED, Garland CF, Garland FC, Grant WB, Mohr SB, Lipkin M, Newmark HL, Giovannucci E, Wei M, Holick MF. Optimal vitamin D status for colorectal cancer prevention: a quantitative meta analysis. Am J Prev Med 2007;32:210–6. [DOI] [PubMed] [Google Scholar]
- 8.Ng K, Meyerhardt JA, Wu K, Feskanich D, Hollis BW, Giovannucci EL, Fuchs CS. Circulating 25-hydroxyvitamin d levels and survival in patients with colorectal cancer. J Clin Oncol 2008;26:2984–91. [DOI] [PubMed] [Google Scholar]
- 9.Tretli S, Schwartz GG, Torjesen PA, Robsahm TE. Serum levels of 25-hydroxyvitamin D and survival in Norwegian patients with cancer of breast, colon, lung, and lymphoma: a population-based study. Cancer Causes Control 2012;23:363–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Tretli S, Hernes E, Berg JP, Hestvik UE, Robsahm TE. Association between serum 25(OH)D and death from prostate cancer. Br J Cancer 2009;100:450–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ross AC, Manson JE, Abrams SA, Aloia JF, Brannon PM, Clinton SK, Durazo-Arvizu RA, Gallagher JC, Gallo RL, Jones G, et al. 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 2011;96:53–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Harris SS, Soteriades E, Coolidge JA, Mudgal S, Dawson-Hughes B. Vitamin D insufficiency and hyperparathyroidism in a low income, multiracial, elderly population. J Clin Endocrinol Metab 2000;85:4125–30. [DOI] [PubMed] [Google Scholar]
- 13.Nesby-O'Dell S, Scanlon KS, Cogswell ME, Gillespie C, Hollis BW, Looker AC, Allen C, Doughertly C, Gunter EW, Bowman BA. Hypovitaminosis D prevalence and determinants among African American and white women of reproductive age: third National Health and Nutrition Examination Survey, 1988-1994. Am J Clin Nutr 2002;76:187–92. [DOI] [PubMed] [Google Scholar]
- 14.Clemens TL, Adams JS, Henderson SL, Holick MF. Increased skin pigment reduces the capacity of skin to synthesise vitamin D3. Lancet 1982;1:74–6. [DOI] [PubMed] [Google Scholar]
- 15.Siegel R, Ward E, Brawley O, Jemal A. Cancer statistics, 2011: the impact of eliminating socioeconomic and racial disparities on premature cancer deaths. CA Cancer J Clin 2011;61:212–36. [DOI] [PubMed] [Google Scholar]
- 16.Harris SS. Does vitamin D deficiency contribute to increased rates of cardiovascular disease and type 2 diabetes in African Americans? Am J Clin Nutr 2011;93:1175S–8S. [DOI] [PubMed] [Google Scholar]
- 17.Fiscella K, Winters P, Tancredi D, Hendren S, Franks P. Racial disparity in death from colorectal cancer: does vitamin D deficiency contribute? Cancer 2011;117:1061–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Fiscella K, Franks P. Vitamin D, race, and cardiovascular mortality: findings from a national US sample. Ann Fam Med 2010;8:11–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.McNeill LH, Coeling M, Puleo E, Suarez EG, Bennett GG, Emmons KM. Colorectal cancer prevention for low-income, sociodemographically-diverse adults in public housing: baseline findings of a randomized controlled trial. BMC Public Health 2009;9:353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.US Bureau of the Census. Overview of race and Hispanic origin. Census 2000 brief. Washington, DC: US Bureau of the Census, 2001. [Google Scholar]
- 21.Williams DR. Race/ethnicity and socioeconomic status: measurement and methodological issues. Int J Health Serv 1996;26:483–505. [DOI] [PubMed] [Google Scholar]
- 22.Holick MF. Resurrection of vitamin D deficiency and rickets. J Clin Invest 2006;116:2062–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Hollis BW. Quantitation of 25-hydroxyvitamin D and 1,25-dihydroxyvitamin D by radioimmunoassay using radioiodinated tracers. Methods Enzymol 1997;282:174–86. [DOI] [PubMed] [Google Scholar]
- 24.Willett WC, Reynolds RD, Cottrell-Hoehner S, Sampson L, Browne ML. Validation of a semi-quantitative food frequency questionnaire: comparison with a 1-year diet record. J Am Diet Assoc 1987;87:43–7. [PubMed] [Google Scholar]
- 25.Clarys P, Alewaeters K, Lambrecht R, Barel AO. Skin color measurements: comparison between three instruments: the Chromameter(R), the DermaSpectrometer(R) and the Mexameter(R). Skin Res Technol 2000;6:230–8. [DOI] [PubMed] [Google Scholar]
- 26.Parra EJ, Kittles RA, Shriver MD. Implications of correlations between skin color and genetic ancestry for biomedical research. Nat Genet 2004;36:S54–60. [DOI] [PubMed] [Google Scholar]
- 27.Shriver MD, Parra EJ. Comparison of narrow-band reflectance spectroscopy and tristimulus colorimetry for measurements of skin and hair color in persons of different biological ancestry. Am J Phys Anthropol 2000;112:17–27. [DOI] [PubMed] [Google Scholar]
- 28.Martins D, Wolf M, Pan D, Zadshir A, Tareen N, Thadhani R, Felsenfeld A, Levine B, Mehrotra R, Norris K. Prevalence of cardiovascular risk factors and the serum levels of 25-hydroxyvitamin D in the United States: data from the Third National Health and Nutrition Examination Survey. Arch Intern Med 2007;167:1159–65. [DOI] [PubMed] [Google Scholar]
- 29.Kipnis V, Midthune D, Buckman DW, Dodd KW, Guenther PM, Krebs-Smith SM, Subar AF, Tooze JA, Carroll RJ, Freedman LS. Modeling data with excess zeros and measurement error: application to evaluating relationships between episodically consumed foods and health outcomes. Biometrics 2009;65:1003–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Feskanich D, Ma J, Fuchs CS, Kirkner GJ, Hankinson SE, Hollis BW, Giovannucci EL. Plasma vitamin D metabolites and risk of colorectal cancer in women. Cancer Epidemiol Biomarkers Prev 2004;13:1502–8. [PubMed] [Google Scholar]
- 31.Holick MF, Binkley NC, Bischoff-Ferrari HA, Gordon CM, Hanley DA, Heaney RP, Murad MH, Weaver CM. Evaluation, treatment, and prevention of vitamin D deficiency: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab 2011;96:1911–30. [DOI] [PubMed] [Google Scholar]
- 32.Gallagher JC, Sai A, Templin T, 2nd, Smith L. Dose response to vitamin d supplementation in postmenopausal women: a randomized trial. Ann Intern Med 2012;156:425–37. [DOI] [PubMed] [Google Scholar]
- 33.Garland CF, French CB, Baggerly LL, Heaney RP. Vitamin D supplement doses and serum 25-hydroxyvitamin D in the range associated with cancer prevention. Anticancer Res 2011;31:607–11. [PubMed] [Google Scholar]
- 34.Gallagher JC, Peacock M, Yalamanchili V, Smith LM. Effects of vitamin D supplementation in older African American women. J Clin Endocrinol Metab 2013;98:1137–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Gallagher JC, Jindal PS, Smith LM. Vitamin D supplementation in young Caucasian and African American women. J Bone Miner Res (Epub ahead of print 12 June 2013). [DOI] [PubMed] [Google Scholar]
- 36.Ginde AA, Liu MC, Camargo CA., Jr Demographic differences and trends of vitamin D insufficiency in the US population, 1988-2004. Arch Intern Med 2009;169:626–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Taylor HA, Jr, Coady SA, Levy D, Walker ER, Vasan RS, Liu J, Akylbekova EL, Garrison RJ, Fox C. Relationships of BMI to cardiovascular risk factors differ by ethnicity. Obesity (Silver Spring) 2010;18:1638–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Moore CE, Murphy MM, Holick MF. Vitamin D intakes by children and adults in the United States differ among ethnic groups. J Nutr 2005;135:2478–85. [DOI] [PubMed] [Google Scholar]
- 39.Santmyire BR, Feldman SR, Fleischer AB., Jr Lifestyle high-risk behaviors and demographics may predict the level of participation in sun-protection behaviors and skin cancer primary prevention in the United States: results of the 1998 National Health Interview Survey. Cancer 2001;92:1315–24. [DOI] [PubMed] [Google Scholar]
- 40.Signorello LB, Shi J, Cai Q, Zheng W, Williams SM, Long J, Cohen SS, Li G, Hollis BW, Smith JR, Blot WJ: Common variation in vitamin D pathway genes predicts circulating 25-hydroxyvitamin D Levels among African Americans. PLoS 2011, 6:e28623. Epub 22011 Dec 28621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Powe CE, Seely EW, Rana S, Bhan I, Ecker J, Karumanchi SA, Thadhani R: First trimester vitamin D, vitamin D binding protein, and subsequent preeclampsia. Hypertension 2010;56:758-763. Epub 2010 Aug 23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Powe CE, Evans MK, Wenger J, Zonderman AB, Berg AH, Nalls M, Tamez H, Zhang D, Bhan I, Karumanchi SA, et al. Vitamin D-Binding Protein and Vitamin D Status of Black Americans and White Americans. N Engl J Med 2013;369:1991–2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Aloia JF. African Americans, 25-hydroxyvitamin D, and osteoporosis: a paradox. Am J Clin Nutr 2008;88:545S–50S. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Theobald TM, Cauley JA, Gluer CC, Bunker CH, Ukoli FA, Genant HK. Black-white differences in hip geometry. Study of Osteoporotic Fractures Research Group. Osteoporos Int 1998;8:61–7. [DOI] [PubMed] [Google Scholar]
- 45.Cosman F, Morgan DC, Nieves JW, Shen V, Luckey MM, Dempster DW, Lindsay R, Parisien M. Resistance to bone resorbing effects of PTH in black women. J Bone Miner Res 1997;12:958–66. [DOI] [PubMed] [Google Scholar]
- 46.Looker AC, Wahner HW, Dunn WL, Calvo MS, Harris TB, Heyse SP, Johnston CC, Jr, Lindsay R. Updated data on proximal femur bone mineral levels of US adults. Osteoporos Int 1998;8:468–89. [DOI] [PubMed] [Google Scholar]
- 47.Aloia JF, Talwar SA, Pollack S, Yeh J. A randomized controlled trial of vitamin D3 supplementation in African American women. Arch Intern Med 2005;165:1618–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Vieth R, Chan PC, MacFarlane GD. Efficacy and safety of vitamin D3 intake exceeding the lowest observed adverse effect level. Am J Clin Nutr 2001;73:288–94. [DOI] [PubMed] [Google Scholar]
- 49.Fakih MG, Andrews C, McMahon J, Muindi JR: A prospective clinical trial of cholecalciferol 2000 IU/day in colorectal cancer patients: evidence of a chemotherapy-response interaction. Anticancer Res 2012;32:1333-1338. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.