Meal conditions affect the absorption of supplemental vitamin D3 but not the plasma 25-hydroxyvitamin D response to supplementation


  • Bess Dawson-Hughes,

    Corresponding author
    • Jean Mayer United States Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, Boston, MA, USA
    Search for more papers by this author
  • Susan S Harris,

    1. Jean Mayer United States Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, Boston, MA, USA
    Search for more papers by this author
  • Nancy J Palermo,

    1. Jean Mayer United States Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, Boston, MA, USA
    Search for more papers by this author
  • Lisa Ceglia,

    1. Jean Mayer United States Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, Boston, MA, USA
    2. Division of Endocrinology, Diabetes, and Metabolism, Tufts Medical Center, Boston, MA, USA
    Search for more papers by this author
  • Helen Rasmussen

    1. Jean Mayer United States Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, Boston, MA, USA
    Search for more papers by this author

  • This material is based upon work supported by the U.S. Department of Agriculture, under agreement No. 58-1950-0-014. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture.

Address correspondence to: Bess Dawson-Hughes, MD, Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University, 711 Washington St., Boston, MA 02111, USA. E-mail:


It is sometimes assumed that dietary fat is required for vitamin D absorption, although the impact of different amounts of dietary fat on vitamin D absorption is not established. This study was conducted to determine whether the presence of a meal and the fat content of the meal influences vitamin D absorption or the 25-hydroxyvitamin D [25(OH)D] response to supplemental vitamin D3. Based on earlier studies in rats we postulated that absorption would be greatest in the low-fat meal group. Sixty-two healthy older men and women were randomly assigned to one of three meal groups: no meal, high-fat meal, or low-fat meal; each was given a monthly 50,000 IU vitamin D3 supplement with the test breakfast meal (or after a fast for the no-meal group) and followed for 90 days. Plasma vitamin D3 was measured by liquid chromatography–mass spectroscopy (LC/MS) before and 12 hours after the first dose; plasma 25(OH)D was measured by radioimmunoassay at baseline and after 30 and 90 days. The mean 12-hour increments in vitamin D3, after adjusting for age and sex, were 200.9 nmol/L in the no-meal group, 207.4 nmol/L in the high-fat meal group, and 241.1 nmol/L in the low-fat meal group (p = 0.038), with the increase in the low-fat group being significantly greater than the increases in the other two groups. However, increments in 25(OH)D levels at 30 and 90 days did not differ significantly in the three groups. We conclude that absorption was increased when a 50,000 IU dose of vitamin D was taken with a low-fat meal, compared with a high-fat meal and no meal, but that the greater absorption did not result in higher plasma 25(OH)D levels in the low-fat meal group.


The 25(OH)D response to supplementation with vitamin D varies widely among individuals and this complicates the process of selecting the dose of vitamin D needed to reach a specific targeted serum 25-hydroxyvitamin D [25(OH)D] level. Several factors are known to contribute to this variability, including genetic variants in the vitamin D binding proteins[1-3] and in the enzyme that hydroxylates vitamin D in the liver to form 25(OH)D,[4] and subject characteristics, including body mass index (BMI)[5] and the starting level of serum 25(OH)D.[6] The presence of a meal and the fat content of that meal may also be important.

Gut perfusion studies in the unanesthetized rat reveal that vitamin D is absorbed by passive diffusion in the proximal jejunum and the distal ileum.[7] The influence of fat intake on vitamin D absorption is complex. In rats, the absorption of D3 appeared to be enhanced by the presence of small amounts of fat in the gut, presumably because the fat stimulated bile acid secretion.[7] Similarly, absorption of other fat soluble vitamins, E and K, was enhanced by the presence of small amounts of dietary fat.[7] In contrast, larger amounts of fat in the gut appeared to impair vitamin D absorption and reduce increments in 25(OH)D.[7] Further work led to the observation that the polyunsaturated fatty acids (PUFAs), linoleic and linolenic acids, were particularly effective in decreasing vitamin D absorption.[8]

In the human, it is sometimes assumed that dietary fat is required for vitamin D absorption, although the impact of different amounts and types of dietary fat on vitamin D absorption is not established. Tangpricha and colleagues[9] compared the change in serum 25(OH)D level after ingestion of a fat-free beverage (orange juice that had been fortified with calcium and 1000 IU vitamin D3), with the change after taking calcium-fortified orange juice alone. Serum 25(OH)D increased by 150% in the vitamin D–supplemented juice group and by 45% in the control juice group over a 12-week period. The increase in the control group was attributed to season and the significantly larger increase in the vitamin D group indicated greater absorption, suggesting that fat is not essential for vitamin D absorption.[9] A recent clinical report indicated that serum 25(OH)D levels increased by an average of 57% over a 2-month to 3-month period in 17 clinic patients after they were instructed to take their usual doses of vitamin D with the largest meal of the day as opposed to receiving no advice on when to take their supplements.[10] There was no control group in this study and so the effects of season, sun exposure, and other factors on serum 25(OH)D levels are unknown.

Nonetheless, the findings are consistent with increased absorption of vitamin D when the vitamin is taken with a meal. Raimundo and colleagues[11] recently reported that the increment in serum 25(OH)D two weeks after taking a single 50,000 IU dose of vitamin D3 was greater when the dose was given with a high-fat than with a low-fat meal. Neither of these studies measured vitamin D absorption.

Drawing mainly on earlier work in the rat,[7, 8] we hypothesized that vitamin D would be better absorbed when taken with a meal than after fasting and that a low-fat meal would have a more favorable effect on absorption than a high-fat meal. We tested these hypotheses in a study in which changes in vitamin D absorption and in plasma 25(OH)D levels were compared in healthy older adults who took supplemental vitamin D with no meal (after a 12-hour fast), a high-fat meal, or a low-fat meal.

Subjects and Methods


Sixty-two healthy men and postmenopausal women, age 50 to 69 years were enrolled and 62 completed the study. Exclusion criteria were any of the following: a screening 25(OH)D level ≤8 or ≥25 ng/mL; BMI <18.5 or >27.9 kg/m2; an abnormal serum or urine calcium; or use of estrogen, antiepileptic drugs, drugs that alter gastric acid (including proton pump inhibitors, H2 blockers, and antacids), and drugs that alter fat and cholesterol absorption. The participants agreed not to take more than 400 IU of their own vitamin D or more than 1000 mg of supplemental calcium per day during the study. The protocol was approved by the Investigational Review Board at Tufts University, and written informed consent was obtained from each subject.

Subjects were prescreened by means of an opt-in telephone questionnaire. Interested and potentially eligible candidates were invited to the Center for a screening medical history, physical examination, and blood and routine urine tests. Eligible and interested subjects were enrolled.

Study design

Healthy men and postmenopausal women age 50 to 69 years were enrolled in this 3-month study. Subjects were randomly assigned to one of three groups: no-meal, low-fat meal, or high-fat meal. At baseline and each month, subjects came to the Center after a 12-hour fast. They had a blood draw (except on day 60), consumed their test meal (or continued to fast if in the no-meal group), took a 50,000 IU vitamin D3 tablet, and then were free to leave the Center, but agreed to fast (except for water) for the next 4 hours. A vitamin D3 absorption test was performed on day 1 and plasma 25(OH)D concentration was measured on days 1, 30, and 90. Subjects were enrolled between October 1 and March 31 to limit skin synthesis of vitamin D. Subjects agreed not to travel south of latitude 34 degrees North during the study. Compliance with the monthly test meal and vitamin D dose was assessed by direct observation. Plasma vitamin D3 absorption on day 1, and 1-month and 3-month changes in plasma 25(OH)D levels were compared in the three groups. Staff members who made the outcome measurements were blinded to meal assignment.


The nutrient content of the test meals is shown in Table 1. Both meals included an egg or egg-white frittata with varying amounts of olive oil and vegetables, turkey bacon, and Parmesan cheese topping, a slice of toast with either jelly (low-fat meal group) or butter, and a mango milk smoothie (mango, milk, ginger ale, with or without cream). Calories in the two meal groups were balanced by adjusting the proportions of diet and regular ginger ale. The high-fat (50% of calories) and low-fat (10%) contents of the breakfast meals simulated the extremes of fat intake in adult American diets (1st and 85th percentiles, respectively, according to the Institute of Medicine Dietary Reference Intakes).[12]

Table 1. Content of the Test Meals
 Fasting (no meal)High-fat mealLow-fat meal
  1. MUFA = monosaturated fatty acid; PUFA = polyunsaturated fatty acid; SFA = saturated fatty acid.
Total fat, g35.211.1
MUFA, g15.84.2
PUFA, g4.42.3
SFA, g11.63.5
Carbohydrates, g51.5118.9
Protein, g29.529.6
Total fiber, g3.74.3
Energy, kJ28412841
Cholesterol, mg41761


The 50,000 IU vitamin D3 tablets were purchased from BioTech Pharmacal, Inc. (Fayetteville, AR, USA). Independent analysis by Covance (Madison, WI, USA) revealed that the tablets contained 57,000 IU of vitamin D3.

Vitamin D3 absorption

Vitamin D absorption was estimated on study day 1 by the method of Lo and colleagues.[13] This method assesses change in serum parent vitamin D3 concentration before and 12 hours after (the time of the peak vitamin D level) an oral 50,000 IU dose of vitamin D3. In that study, the timing of the peak vitamin D level after ingestion of vitamin D mixed with food versus in capsule form was similar.[13] This method enabled us to evaluate whether the presence of a meal and/or the fat content of that meal affected the absorption of the oral vitamin D3.

Intake assessment

Intake of vitamin D and other nutrients over the last 3 months was assessed on day 1 with use of a Fred Hutchinson food frequency questionnaire.[14] Subjects were reminded not to change their usual intake of vitamin D or calcium during the study.


Total body dual-energy X-ray absorptiometry (DXA) scans were performed on a GE Lunar Prodigy scanner (Madison, WI, USA). Total body fat was measured with a precision of 0.94%.[15]

Biochemical assays

Blood was drawn in the morning after the subjects had fasted for 12 hours. All samples from individual subjects were batched for analysis. Urinary creatinine was measured on an automated clinical chemistry analyzer (Olympus AU400; Olympus America Inc., Melville, NY, USA) with a coefficient of variation (CV) of 3.0%. Urinary calcium was measured by direct-current plasma emission spectroscopy (Beckman SpectraSpan VI Direct Current Plasma Emission Spectrophotometer; Beckman Instruments, Fullerton, CA, USA) with a CV of 3% to 5%. Pepsinogen 1, an indirect indicator of gastric acid status, was measured in serum by an ELISA kit manufactured by ALPCO Diagnostics (Salem, NH, USA); this assay utilizes a two-site “sandwich” technique with two selected monoclonal antibodies that bind to different epitopes of human pepsinogen 1 without cross-reaction to human pepsinogen II. The intraassay and interassay CVs are 5.1% and 6.3%, respectively. For the parent vitamin D3 assay, standards were added to 1 mL of plasma and the vitamin D3 was extracted twice with hexane-ethyl acetate. The extracts were dried and reconstituted in methylene chloride and methanol, and injected onto liquid chromatography–mass spectrometry (LC/MS) columns. Parent vitamin D3 was identified at tandem mass spectrometry (MS/MS) transition m/z 385 to m/z 107, with a peak at 11.3 minutes, separate from the metabolites, 1,25(OH)2D and 25OHD3. Plasma 25(OH)D was measured by radioimmunoassay with commercial kits (DiaSorin, Stillwater, MN, USA) with intraassay and interassay CVs of 8.6% to 12.5% and 8.2% to 11.0%, respectively.

Statistical analyses

Data were examined graphically to rule out the presence of outliers and to evaluate the linearity of bivariate associations. Subject characteristics and other means were compared across groups by analysis of variance for unadjusted values and by analysis of covariance with least-squares means for adjusted values. Proportions were compared with the chi-square test. Pearson correlation coefficients were calculated to describe selected bivariate associations. Values of p <0.05 were considered to indicate statistical significance. Analyses were conducted with SPSS, version 19 (SPSS Inc., Chicago, IL, USA).


The clinical characteristics of the 62 subjects by meal group are shown in Table 2. The groups did not differ significantly in characteristics known to affect the response to supplementation, including age, BMI, or starting 25(OH)D levels. There were more men than women in the study and the proportion of men was greater in the no-meal and the high-fat meal groups.

Table 2. Baseline Characteristics of the 62 Subjects
MeasureNo mealHigh-fat mealLow-fat meal
  1. BMI = body mass index; DXA = dual-energy X-ray absorptiometry; 25(OH)D = 25-hydroxyvitamin D; Ca/Cr = calcium/creatine ratio.
Subjects, n212120
Age, years57.1 ± 5.058.2 ± 5.959.9 ± 4.9
Female, %28.633.345.0
Height, cm172.7 ± 8.7168.9 ± 7.6170.7 ± 5.2
Weight, kg75.4 ± 9.371.5 ± 8.468.0 ± 8.6
BMI, kg/m2, V125.2 ± 2.025.0 ± 2.123.3 ± 2.3
DXA, total body fat, g19468.5 ± 6224.721225.9 ± 6836.818013.1 ± 6832.1
Energy, kcal1514.6 ± 589.21761.1 ± 743.51571.8 ± 542.0
Total fat intake, g62.4 ± 27.772.7 ± 35.368.8 ± 22.6
Protein intake, g63.0 ± 26.774.3 ± 35.766.6 ± 27.1
Calcium intake food, mg757.7 ± 440.9896.2 ± 632.8817.1 ± 485.0
Calcium intake supplements, mg59.9 ± 152.064.0 ± 184.395.1 ± 255.9
Plasma vitamin D3, nmol/L23.6 ± 11.425.3 ± 11.521.2 ± 9.7
Plasma 25(OH)D, nmol/L49.9 ± 16.648.4 ± 14.847.4 ± 11.0
Urine, Ca/Cr, mmol/mol0.16 ± 0.100.15 ± 0.100.18 ± 0.08
Serum pepsinogen 1, µg/L138.2 ± 55.0132.3 ± 58.9121.8 ± 46.6

Vitamin D3 absorption, defined as the increment in plasma vitamin D3 over the first 12 hours after dosing, differed significantly in the three groups, after adjusting for age and sex (p = 0.038). The mean increments were 200.9 nmol/L in the no-meal group, 207.4 nmol/L in the high-fat meal group, and 241.1 nmol/L in the low-fat meal group (see Fig. 1). The increment in the low-fat meal group was significantly greater than the increments in both the high-fat meal (p = 0.040) and no-meal groups (p < 0.017). Absorption in the no-meal and high-fat meal groups did not differ significantly from one another. Neither BMI nor starting level of vitamin D3 or 25(OH)D significantly influenced absorption.

Figure 1.

Effect of meal condition on absorption of vitamin D3. The increase in plasma vitamin D3 differed significantly in the three meal groups after adjustment for age and sex (p = 0.038). The increment was significantly greater in the low-fat meal group than in either the high-fat (0.040) or the no meal groups (0.017). Like letters on the figure bars indicate these differences.

Changes in plasma 25(OH)D over the study period in the three groups, after adjustment for age, sex, and baseline 25(OH)D level are shown in Fig. 2. Of the increments observed on day 90, about one-half had occurred by day 30. There were no significant group differences in starting 25(OH)D levels or in the 25(OH)D increments at either 30 days or 90 days. There was a significant positive correlation of baseline pepsinogen 1 with baseline 25(OH)D level (r = 0.253, p = 0.047) but not with increment in 25(OH)D level. The 12-hour change in vitamin D3 on day 1 was not significantly correlated with 30-day change in serum 25(OH)D in any of the groups or in all subjects combined.

Figure 2.

Changes in plasma 25(OH)D by meal group, adjusted for age, sex, and baseline 25(OH)D level. Differences between groups were not statistically significant.


In these healthy older men and women with no evidence of malabsorption, vitamin D3 absorption occurred in the no-meal group, clearly indicating that a meal is not essential for vitamin D absorption. The presence of a meal did improve absorption, however, at least when the meal had a low-fat content. Absorption after the low-fat meal was 20% higher when compared with no meal; it was 16% higher when compared with the high-fat meal. This is consistent with the findings in the rat studies of Hollander and colleagues.[7] The basis for lower absorption in the presence of the high-fat compared with the low-fat meal is speculative. Hollander and colleagues.[7] offered several potential explanations. First, more fat in the intestinal lumen may have increased the solubility of vitamin D in the micelles and changed the partition coefficient such that the vitamin D stayed in the micelle. Alternatively, it may have increased the size of the micelle and thereby reduced its diffusion rate and increased its difficulty in crossing the unstirred water layer lining the intestinal mucosa.[8]

Several factors known or thought to influence vitamin D absorption should be considered in interpreting our findings. Drugs known to alter gastric acid production were exclusions because of the possibility that gastric acid production would influence the absorption of fats and fat-soluble vitamins by reducing bile secretion.[16, 17] Naturally occurring hypochlorhydria or achlorhydria may impair absorption by a similar mechanism. However, the subjects in our meal groups appeared to be balanced in gastric acid production, at least as indicated by their similar pepsinogen 1 levels.[18] We have previously observed that healthy older men and women who reported consuming diets with a low monounsaturated fatty acid (MUFA)/PUFA ratio had smaller increments in 25(OH)D in response to supplementation with 700 IU of D3 per day than subjects consuming diets with a high MUFA/PUFA ratio.[19] The reduced increment in 25(OH)D could reflect reduced absorption or altered metabolism of the vitamin. In the current study, the MUFA/PUFA ratio was actually lower in the low-fat meal than in the high-fat meal and thus may possibly have masked some of the favorable effect of the low-fat meal on absorption. A minor limitation of the study is the lack of parathyroid hormone (PTH) measurements, because PTH can alter vitamin D metabolism; however, the likelihood of parathyroid disorder is reduced by the fact that all subjects had normal serum and urinary calcium levels at entry. Finally, becasue we tested only one dose of vitamin D, 50,000 IU per month, we have no way to know whether similar results would have been observed had we used a lower dose. We chose the 50,000 IU dose because it was the dose in which the tolerance test was originally described.[13] It will be important to assess the impact of meal conditions on absorption of lower doses of vitamin D in the future.

It is hard to compare our findings directly with those of Mulligan and Licata[10] and difficult to reconcile ours with those of Raimundo and colleagues.[11] The 17 subjects in the Mulligan and Licata[10] study were a mixed group of patients with malabsorption and other conditions, not healthy volunteers, and no information was provided on factors that may have influenced the 25(OH)D increments they reported, including the composition of the meals with which the patients took their vitamin D, the season of measurement, or their body weight. Raimundo and colleagues[11] compared the 25(OH)D responses to a single dose of 50,000 IU of vitamin D3 given with a high-fat (25.6 g) or a low-fat (1.7 g) meal at 7 and 14 days after the dose. Surprisingly, there was no increase (and even a small decline) in serum 25(OH)D in the low-fat group over the 2 weeks after dosing. In the high-fat group, the pattern of change in serum 25(OH)D was unexpected. Armas and colleagues[20] administered a similar single oral dose of 50,000 IU of vitamin D3 and found that over 90% of the increment in 25(OH)D observed at 14 days was present at 7 days. In contrast, in the Raimundo and colleagues[11] study, only 30% of the increment observed at day 14 was present on day 7. The relatively large increase between days 7 and 14 raises the possibility of inadvertent sun exposure or supplemental vitamin D ingestion in the second half of that study.

The importance of the impact of a meal and its fat content on vitamin D homeostasis may be questioned by our second finding, which was that the increment in plasma 25(OH)D after 1 and 3 months of supplementation did not differ in the three groups. We cannot exclude the possibility that this null finding was related to the relatively high vitamin D dose we used, and that a favorable effect of higher absorption would have been evident in the 25(OH)D level had a lower dose of vitamin D been used. This null finding cannot be attributed to group differences in starting 25(OH)D levels, in BMI levels, or in presumed gastric acid production (as assessed by pepsinogen 1 levels), which were balanced in the three groups at entry. The observed positive correlation between baseline serum pepsinogen 1 and plasma 25(OH)D is consistent with a positive role for gastric acid in vitamin D absorption under ambient conditions. This is plausible because gastric acid increases the rate of food decomposition[21] and gastric emptying, stimulating cholecystokinin, which then increases the secretion of bile, which is essential for micelle formation.[17, 21, 22]

In conclusion, a meal is not necessary for vitamin D absorption; however, absorption of vitamin D after a 50,000 IU dose was greater when the supplement was taken with a low-fat meal than when taken with no meal (after a 12-hour fast) or with a high-fat meal. Improved absorption from the 50,000 IU dose, however, did not translate to greater increases in plasma 25(OH)D levels from monthly dosing at either 30 or 90 days. The impact of the meal condition on vitamin D absorption and increment in plasma 25(OH)D after lower doses of vitamin D remains to be determined.


All authors state that they have no conflicts of interest.


This study was funded by an Investigator Initiated Research grant from Pfizer Inc (IIR Grant No. WS877392). We are indebted to the study participants and to the staff of the Metabolic Research Unit at the USDA Human Nutrition Research Center on Aging at Tufts University for their dedication and commitment.

Authors' roles: Study concept and design: BD-H, SSH. Acquisition of data: NP, LC, and HR. Data analysis: SSH, NP, and HR. Data interpretation: BD-H, SSH, HR, and LC. Drafting manuscript: BD-H and SSH. Approving manuscript: BD-H, SSH, NP, LC, and HR. SSH takes responsibility for the integrity of the data analysis.