25-Hydroxyvitamin D Response to Graded Vitamin D₃ Supplementation Among Obese Adults

Context:

Guidelines have suggested that obese adults need 2 to 3 times more vitamin D than lean adults to treat vitamin D deficiency, but few studies have evaluated the vitamin D dose response in obese subjects.

Objective:

The purpose of this study was to characterize the pharmacokinetics of 25-hydroxyvitamin D [25(OH)D] response to 3 different doses of vitamin D₃ (cholecalciferol) in a group of obese subjects and to quantify the 25(OH)D dose-response relationship.

Design, Setting, Intervention, Patients:

This was a randomized, single-blind study of 3 doses of oral vitamin D₃ (1000, 5000, or 10,000 IU) given daily to 67 obese subjects for 21 weeks during the winter months.

Main Outcome Measures:

Serum 25(OH)D levels were measured at baseline and after vitamin D replacement, and 25(OH)D pharmacokinetic parameters were determined, fitting the 25(OH)D concentrations to an exponential model.

Results:

Mean measured increments in 25(OH)D at week 21 were 12.4 ± 9.7 ng/mL in the 1000 IU/d group, 27.8 ± 10.2 ng/mL in the 5000 IU/d group, and 48.1 ± 19.6 ng/mL in the 10,000 IU/d group. Steady-state increments computed from the model were 20.6 ± 17.1, 35.2 ± 14.6, and 51.3 ± 22.0 ng/mL, respectively. There were no hypercalcuria or hypercalcemia events during the study.

Conclusion:

Our data show that in obese people, the 25(OH)D response to vitamin D₃ is directly related to dose and body size with ∼2.5 IU/kg required for every unit increment in 25(OH)D (nanograms per milliliter).

Obesity is associated with low 25-hydroxyvitamin D [25(OH)D] levels. Accordingly, it has been recommended that vitamin D replacement therapy be adjusted according to body size to achieve the desired serum 25(OH)D concentrations (1). Many explanations have been offered for the low 25(OH)D levels in obesity (2–7). Despite their plausibility, our recent work has shown that the most parsimonious explanation is simple dilution of 25(OH)D in the larger fat and tissue mass of obese patients (8).

Studies have been done in normal-weight people to predict the 25(OH)D dose response to a particular dose of vitamin D. One study calculated the anticipated rate of rise in serum 25(OH)D as ∼0.28 ng/mL/μg of vitamin D/d (9). This translates into a rise in serum 25(OH)D of ∼7 ng/mL for every 1000 IU/d of vitamin D ingested. Other articles have suggested an even more robust 25(OH)D response ranging from 0.4 to 0.78 ng/mL/μg of vitamin D/d and more (10–12).

Most of these vitamin D dose-response studies were conducted in populations that were normal to overweight (body mass index [BMI] up to 29.9 kg/m²), with very few including obese subjects (10, 11, 13). No previous studies have focused on the dose-response relationship between vitamin D dose and 25(OH)D response in subjects with obesity, including those with class III obesity (BMI ≥40 kg/m²) (14). The purpose of this study was to characterize the pharmacokinetics of the 25(OH)D response to 3 different doses of vitamin D₃ (cholecalciferol) in a group of obese subjects and to quantify the relationship.

Materials and Methods

Design

This was a randomized, single-blind study of 3 doses of oral vitamin D₃ (nominally 1000, 5000, or 10,000 IU) given daily to obese subjects for 21 weeks during the winter months of either 2009 or 2010. The institutional review board at Creighton University (Omaha, Nebraska) approved the study protocol, and all the participants signed an informed consent form.

Subjects and setting

The subjects were 67 women and men in good general health, aged 19 to 68 years with BMI ≥30.0 kg/m². (See Table 1 for subject demographics.) Volunteers were recruited among the Creighton University Medical Center and Medical School personnel and in the larger Omaha community using advertisements in local media and church bulletins. All subjects resided in Omaha, Nebraska, at a latitude of 41.2°N (approximately the same as that of Boston or southern Italy). They were recruited during the winter months of 2 consecutive years, assuring minimal cutaneous synthesis of vitamin D₃.

Table 1.

Subject Demographics

	Vitamin D3 Group
	1000 IU/d	5000 IU/d	10,000 IU/d
n	22	20	20
Sex (Male/Female)	9/13	9/11	7/13
Age, y	47.1 ± 12.5	45.7 ± 12.6	44.5 ± 12.9
Height, cm	169 ± 8	174 ± 12	167 ± 9
Weight, kg	105.8 ± 20.0	109.4 ± 22.2	106.5 ± 24.8
BMI, kg/m2	36.7 ± 4.6	36.1 ± 5.1	37.9 ± 7.2
Dietary vitamin D intake, IU/d	207 ± 277	203 ± 230	279 ± 313
Baseline 25(OH)D, ng/mL	20.3 ± 6.4	26.5 ± 6.7	23.2 ± 15.2

All values are means ± SD.

Subjects were questioned and excluded if they had a past or current hepatic or kidney disease or were taking medications that affect vitamin D metabolism (such as steroids and antiseizure medications). They were also excluded if they had any malabsorptive conditions from medical or surgical causes. Other exclusion criteria were a history of hypercalcemia, sarcoidosis, active kidney stones, a history of fractures, and current use of bisphosphonates. Participants were questioned specifically about their dietary vitamin D and vitamin D supplement use to determine vitamin D intake for eligibility. They could have had no more than 800 IU daily of vitamin D from food and supplements. They also could not have had an outdoor job during the previous summer or had plans to visit a sunny region during the study.

Randomization

Upon study entry, subjects were randomly assigned to 1 of 3 different doses of oral vitamin D_3, using random number generation (Microsoft Excel 2007; Microsoft).

Intervention

Subjects were assigned to take either 1000, 5000, or 10,000 IU of vitamin D₃ daily for a period of 21 weeks. Vitamin D₃ was supplied in a tablet form by Douglas Laboratories (nominally 1000-IU dose) or in capsule form by Tishcon Corporation (nominally 5000-IU capsules), so subjects randomly assigned to a 10,000-IU dose were given 2 of the 5000-IU capsules. The vitamin D content of the tablets or capsules was extracted and isolated for analysis by an HPLC method developed at Heartland Assays for food and modified for vitamin supplements (15). The vitamin D₃ content was confirmed to be as follows: 911 IU per labeled 1000-IU tablet; 5747 IU per labeled 5000-IU capsule (and, therefore, 11,495 IU in 2 capsules for the 10,000-IU group). Compliance was assessed by pill counts at each visit. The actual dosage for each subject was computed from the number of pills ingested over the entire period of treatment and the analyzed content of the dosage unit. This actual dose was used in the analysis.

Protocol

At baseline, medical history was assessed, and height and weight were measured. Height was measured 3 times using a Harpenden stadiometer (Seritex, Inc), and the average was used. Weight was measured 2 times using a Health-O-Meter balance beam scale (Continental Scale Corp), and the average was used. Blood was drawn for serum 25(OH)D, intact PTH, calcium, and creatinine. Two-hour fasting urine collections were obtained for calcium and creatinine at baseline and week 21. At follow-up visits at 1, 3, 6, and 10 weeks, subjects were weighed, and blood was drawn for serum calcium and 25(OH)D. At week 21, the final weight was obtained, and blood was drawn for serum calcium, creatinine, 25(OH)D, and intact PTH.

Analytical methods

Serum 25(OH)D levels were measured by RIA using the LIAISON 25 OH Vitamin D TOTAL Assay (DiaSorin). It was run in duplicate with a coefficient of variation of 2.6%, and the average was used. Serum calcium levels were measured using indirect potentiometry (Beckman Coulter). Serum and urine creatinine levels were measured using a modified Jaffe colorimetric method on an Express Plus Analyzer (Chiron). Intact PTH was measured by an immunoradiometric assay (DiaSorin). The urine calcium level was measured using atomic absorption spectrometry (AA100; PerkinElmer).

Statistical analysis

Descriptive statistics were generated using the statistics package PASW Statistics 18.0 (SPSS Inc) and Microsoft Excel 2007. Laboratory values within different vitamin D supplementation groups were compared using a one-way ANOVA. Post hoc comparisons were made using the Tukey honestly significant difference test. Within each vitamin D dose group, baseline laboratory values were compared with the final values using a paired t test. The subjects' serum 25(OH)D increment values at various times after the start of dosing were fitted individually (using Sigma Plot 10.0; Systat Software) to the following equation, modified from an approach derived and justified by Heaney et al (9):

$$y=a\left(1-e^{-bx}\right)$$

where y is the measured increment in serum 25(OH)D above baseline and x is the time in days. The a parameter is the increment at equilibrium, ie, the best estimate of the steady-state change from baseline that would have been reached had dosing continued without change; b is the rate constant, characterizing the rapidity with which the steady state was reached.

There was a wide variation in dose response at the lowest dose used (911 IU), making individual curve fitting difficult. Therefore, in this group only the mean values at each time point for the entire dosage group were fitted to the equation rather than the individual 25(OH)D increment values. The goodness of fit to this model at all 3 doses was excellent, with R² averaging 0.96 and ranging from 0.94 to 0.98 for the individual curves. The values for the steady-state increment (a) for each individual subject in the 5747- and 11,495-IU treatment groups and the increment in 25(OH)D for each individual subject in the 911-IU treatment group were plotted against the actual daily dose (measured dose and compliance) per kilogram of weight. We used linear regression to examine the effect of actual dose (international units per day) and weight (kilograms) on the 25(OH)D steady state (a parameter from equation 1 in nanomoles per liter) for each individual subject in the 5747- and 11,495-IU treatment groups and for the 25(OH)D increment for each individual subject in the 911-IU treatment group.

Using multiple linear backward stepwise regression, we examined the relationship of the 25(OH)D increment values with age, sex, weight, BMI, baseline 25-vitamin D, fat mass, and actual average daily dose based on compliance data (see Supplemental Table 1, and for histograms representing distribution of subjects across BMI ranges, see Supplemental Figures, all published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org). Only BMI and actual dose were significantly related (P < .001, R² = 0.70) and included in the final model.

Results

Subjects

We recruited and enrolled a total of 67 subjects. Of these, 5 withdrew from the study before or after the first visit for the following reasons: 1 was lost to follow-up; 2 were excluded because of comorbid illness; and 1 subject each was excluded because of the inability to obtain a blood sample and personal choice. Two withdrew from the study after visit 3 because of logistical problems with study visits, but their data to the latest available time point were included in the pharmacokinetic analysis. Two subjects were noncompliant before the last visit, and their last visit was excluded when we fitted their data to equation 1. Thus, we had a total of 62 subjects included in the analysis. All were Caucasian; 22 were in the 1000-IU vitamin D₃ group, and 20 were in each of the groups that took 5000 and 10,000 IU of vitamin D₃. The demographic data for each group are presented in Table 1. BMI ranged from 30 to 58 kg/m². (For histograms representing distribution of subjects across BMI ranges, see Supplemental Figures.) There were no statistically significant differences in age or BMI among the 3 groups. Overall compliance with vitamin D supplementation was 95%, ranging from 94% to 97%.

25(OH)D

Average baseline 25(OH)D levels were 23.3 ± 10.3 ng/mL for the entire cohort. Baseline 25(OH)D levels for each dose group are presented in Table 1. There was no significant difference in baseline 25(OH)D levels among the 3 groups.

The 1000-IU group had a mean measured increase in 25(OH)D of 12.4 ± 9.7 ng/mL at 21 weeks (range, 1.85–38.70 ng/mL) (P < .001, paired t test). The 5000-IU group had a mean measured increase of 27.8 ± 10.2 ng/mL at 21 weeks (range, 12.7–46.4 ng/mL) (P < .001, paired t test), and the 10,000-IU group had a mean measured increase of 50.7 ± 16.4 ng/mL at 21 weeks (range, 16.4–82.6 ng/mL) (P < .001, paired t test). The increments of 25(OH)D achieved were significantly different among the treatment groups (P < .001, one-way ANOVA and Tukey post hoc analysis). Using a multiple linear backward stepwise regression, BMI and actual dose were significant predictors (P < .001, R² = 0.70), and they accounted for 70% of the increase in 25(OH)D (Table 2), although with our small sample actual dose accounted for most of the effect. The data for the 25(OH)D values graphed as a function of actual dose (measured dose and compliance) per kilogram of weight is shown in Figure 1. The equation for this relationship is as follows:

$$25\left(\text{OH}\right)\text{D}\hspace{0.17em}\text{increment}\hspace{0.17em}\left(\text{nmol}/\text{L}\right)=1.0\times \left(\text{daily}\hspace{0.17em}\text{dose}\hspace{0.17em}\text{vitamin}\hspace{0.17em}{\text{D}}_3,\hspace{0.17em}\text{IU}/\text{kg}\right)+26.6\pm 35.2$$

Table 2.

Multiple Linear Regression of Variables Affecting 25(OH)D Increment (Nanomoles per Liter)

Variable	Coefficients (95% Confidence Interval)	P Value
BMI (kg/m2)	−1.887 (−3.896 to 0.121)	.065
Actual dose (IU/d)	0.009 (0.008 to 0.011)	<.0001

Figure 1.

Regression of the steady state in serum 25(OH)D concentration (a, from equation 1) for individuals in the 5747- and 11,495-IU treatment groups and for the increments in serum 25(OH)D concentration from individuals in the 911-IU treatment group on actual doses taken per kilogram of body weight. (Copyright by Laura A. G. Armas 2013, used with permission.)

25(OH)D pharmacokinetic parameters

The rate constant estimate (b) for the 10,000-IU treatment group was 3.43 ± 2.79 × 10⁻², for the 5000-IU treatment group, it was 1.85 ± 1.10 × 10⁻², and for the 1000-IU treatment group, it was 1.00 ± 5.93 × 10⁻². The time courses of the mean increments in 25(OH)D levels for the 3 treatment groups are presented in Figure 2. The steady-state increment (a) values [the increment of 25(OH)D that would be achieved if dosing were continued] were 20.6 ± 17.1 ng/mL in the 1000-IU group, 35.2 ± 14.6 ng/mL in the 5000-IU group, and 51.3 ± 22.0 ng/mL in the 10,000-IU group.

Figure 2.

The time course of the 25(OH)D increment in the 1000-IU group (○), the 5000-IU group (▵), and the 10,000-IU group (□). The error bars are 1 SEM. The regression lines are the least squares fit to the data (R² = 0.96). (Copyright by Laura A. G. Armas 2012, used with permission.)

Other results

Other laboratory results at baseline and changes therein at 21 weeks are presented in Table 3. There was no increase in serum calcium levels during the course of this study in any treatment group. Intact PTH levels trended down with vitamin D supplementation, and this decrease was statistically significant in the 1000- and 10,000-IU groups (P = .031 and .035, respectively, paired t tests), but not in the 5000-IU group. There was no change in the 2-hour urinary calcium to creatinine ratio during the study, and no values increased to >0.15. There was also no change in the subjects' weight at the end of study.

Table 3.

Laboratory Data

	Vitamin D3 Group
	1000 IU/d	5000 IU/d	10,000 IU/d
n	22	20	20
Baseline 25(OH)D, ng/mL	20.3 ± 6.4	26.5 ± 6.7	23.2 ± 15.2
Change 25(OH)D, ng/mLa	20.6 ± 17.1b	35.2 ± 14.6c	51.3 ± 22.0c
Baseline calcium, mg/dL	9.3 ± 0.3	9.3 ± 0.4	9.1 ± 0.3
Change in calcium, mg/dL	0.0 ± 0.2	0.1 ± 0.4	0.0 ± 0.2
Baseline PTH, ng/L	22.3 ± 8.6c	22.1 ± 13.0	28.7 ± 15.6d
Change in PTH, ng/L	−2.4 ± 4.8	−0.8 ± 8.0	−4.9 ± 9.4
Baseline urine calcium to creatinine ratio	0.07 ± 0.04	0.08 ± 0.06	0.10 ± 0.07
Change in urine calcium to creatinine ratio	0.02 ± 0.04	0.02 ± 0.06	−0.01 ± 0.05

All values are means ± SD.

^a

This is the a value from equation 1.

^b

The increment in 25(OH)D was significantly different among treatment groups (P < .001, one-way ANOVA).

^c

The decrease in PTH was significantly different from baseline (P = .031, paired t test).

^d

The decrease in PTH was significantly different from baseline (P = .031, paired t test).

Discussion

Increment in 25(OH)D

Our main focus in this study was to quantify the rise in 25(OH)D in obese subjects. This was best done using equation 2. The most important outcome of our study is that we were able to quantify the dose needed to raise a person's 25(OH)D level based on his or her weight. Solving equation 2 and converting to more clinically useful units (nanograms per milliliter) we get,

$$\text{Additional}\hspace{0.17em}\text{daily}\hspace{0.17em}\text{vitamin}\hspace{0.17em}{\text{D}}_3\hspace{0.17em}\text{dose}\hspace{0.17em}\left(\text{IU}\right)=\left[\text{weight}\hspace{0.17em}\left(\text{kg}\right)\times \text{desired}\hspace{0.17em}\text{change}\hspace{0.17em}\text{in}\hspace{0.17em}25\left(\text{OH}\right)\text{D}\times 2.5\right]-10$$

In other words, we require ∼2.5 IU/kg to raise 25(OH)D by 1 ng/mL. Thus, raising the 25(OH)D level of a 120-kg person by 10 ng/mL would require an average additional daily dose of 2990 IU of vitamin D₃. This equation provides a weight-based means to estimate the amount of vitamin D₃ (in addition to all prevailing inputs) needed to raise serum 25(OH)D by any desired amount.

The rate constant (b) derived from the curve fitting represents how quickly the steady state is reached and the fraction of the dose that is hydroxylated per unit time. It has been quantified in healthy Caucasian men (9), with estimates ranging from 2.46 to 2.99 × 10⁻². The rate constants from this study (1.00–3.43 × 10⁻²) were in the same general range as has been reported elsewhere with vitamin D₃ doses from 1000 to 10,000 IU/d in healthy, normal-weight subjects (9). This finding suggests that there are no appreciable differences in the rate of hydroxylation and utilization of 25(OH)D in obese vs lean subjects.

Figure 2 illustrates the incremental vitamin D dose-response curves in our study and Figure 3 compares these with data for healthy nonobese subjects previously published from our center (9). The nonobese subjects were men who were normal weight to overweight (mean BMI of 26 ± 2.4 kg/m²) and used daily doses of vitamin D₃ (836, 5500, and 11,000 IU, respectively) similar to those in this study. Although the nonobese subjects had slightly higher baseline 25(OH)D values (mean ∼28 ng/dL) than the obese subjects in our study, the vitamin D dose response was more robust in the nonobese subjects than in the obese subjects. The 25(OH)D increase expressed as a function of dose was 0.28 ng/mL/μg in the nonobese subjects (9) compared with an average 25(OH)D increase of 0.195 ng/mL/μg in the obese subjects in our current study. On the basis of comparison with historical data, it appears that the vitamin D dose response in obese subjects is about 30% lower than the response in nonobese subjects. In addition, the dose-response curves in obese and nonobese subjects are effectively parallel, further suggesting that dilution of vitamin D₃ in body tissue mass (in fat cell mass as well as extracellular fluid) rather than sequestration in the fat tissue is accounting for the differences in response, as we have shown previously (8).

Figure 3.

The time course of the 25(OH)D increment in the 1000-IU group (○), the 5000-IU group (▵), and the 10,000-IU group (□) in the current study, contrasted with a historical comparison from normal-weight males who were taking 5000 IU (▴) daily and 10,000 IU (■) daily (9). The error bars are 1 SEM. (Copyright by Laura A. G. Armas 2012, used with permission.)

To our knowledge, this is the first vitamin D dose-response study done exclusively in obese subjects with BMIs ranging from 30 to 58 kg/m². Similar to previous studies in normal-weight subjects, the time course of the 25(OH)D dose response to vitamin D is curvilinear. The reason for the curved time course is that induced increments in 25(OH)D lead to increased metabolic consumption, which rises until it equals input (9). However, compared with published data for normal-weight individuals from our unit, obese people had about a 30% lower response to the same dose of vitamin D₃. When we compare the dose-response data generated by other centers using different methods (10–12), the 25(OH)D lower response in our study was 50% to 75% less than that in normal-weight to overweight subjects.

Variability of 25(OH)D response

Although the obese subjects as a whole required more vitamin D supplementation than lean subjects, there was a great variability in vitamin D response among subjects in the same replacement group, as we and others have repeatedly noted (9, 11, 12, 16, 17). For instance, in the group of subjects who received 1000 IU of vitamin D₃, the 25(OH)D increment ranged from 1.85 to 38.7 ng/mL. In the group that received 5000 IU of vitamin D₃ daily, the 25(OH)D increment level ranged between 12.7 and 46.4 ng/mL, and in a group that received 10,000 IU of vitamin D₃ daily, the 25(OH)D increment ranged from 16.4 to 82.6 ng/mL.

Although our study was not designed to elucidate the mechanisms for the variability in response, it is known that 25(OH)D responses to given vitamin D inputs are in part due to genetic factors, including the activity of 25-hydroxylase, 24-hydroxylase, and synthesis of vitamin D–binding proteins (18–21). Thus, given the variability in response, 13 subjects receiving 1000 IU of vitamin D₃ daily achieved a 25(OH)D level >30 ng/mL, whereas 1 subject who received 10,000 IU daily did not reach a 25(OH)D level of 30 ng/mL. Therefore, although our data support the recommendation that obese patients need more vitamin D₃ than normal-weight patients (8), our data also show that there is value in obtaining a 25(OH)D level in a clinical setting to monitor the 25(OH)D response to vitamin D supplementation.

It is interesting to note that baseline 25(OH)D levels in our patients were somewhat higher than expected given the degree of their obesity, with the mean 25(OH)D level >20 ng/mL in all 3 groups. This finding also probably explains the absence of frankly elevated PTH concentrations in our subjects. Nevertheless, despite a normal baseline PTH level, PTH did decrease with vitamin D supplementation in most of the participants.

One strength of our study is the recruitment of subjects with a wide range of BMI values in the obese range including those with extreme obesity with a BMI as high as 58 kg/m². Another strength is that we used a broad range of vitamin D₃ doses. In addition, the input of exogenous vitamin D sources was minimized by studying subjects during the winter months. Another strength of the study is the comprehensive assessment of safety, through assessment not only of the serum calcium level but also of the risk of hypercalcuria through measurement of urinary calcium. The 2-hour urinary calcium to creatinine ratio >0.15 is an accepted indicator of hypercalcuria (22), and none of our subjects reached that value.

Our study has some limitations as well. We have not used a placebo but have included a smaller dose of vitamin D₃ of 1000 IU daily, allowing for adequate comparison of the dose response with the much larger dose of 10,000 IU daily. Finally, our comparisons with normal-weight individuals in this communication are based on published data not on concurrent measurements performed in parallel with those for the obese subjects. In addition, we have not included other ethnic groups, and we know that the 25(OH)D dose-response to oral vitamin D supplementation varies considerably in other races and ethnicities (23).

In conclusion, our data show that weight and dose are directly related to the steady-state 25(OH)D response and imply that obese people need more vitamin D₃ than normal-weight people to attain the same increment of 25(OH)D. This finding supports the recent recommendations of The Endocrine Society (1). A reasonable estimate of the vitamin D dose can be estimated by the following equation: additional daily vitamin D₃ dose (IU) = [weight (kg) × desired change in 25(OH)D × 2.5] − 10.

As noted in previous studies, we found that daily vitamin D₃ doses of up to 10,000 IU given for 21 weeks are not associated with hypercalcemia or hypercalcuria. Although there can be large individual variations in the 25(OH)D dose response and clinicians must rely on 25(OH)D measurements to assess the adequacy of dosing, this study provides a reasonable place for starting vitamin D replacement based on weight.

Acknowledgments

This work was supported by a Health Future Foundation Faculty Development Award, Creighton University (Omaha, NE) awarded to Andjela Drincic, MD.

Disclosure Summary: The authors have nothing to disclose.

Abbreviations

 
• BMI

  body mass index

 
• 25(OH)D

  25-hydroxyvitamin D.