High-dose oral vitamin D3 supplementation in the elderly
- 756 Downloads
Daily dosing with vitamin D often fails to achieve optimal outcomes, and it is uncertain what the target level of 25-hydroxyvitamin D should be. This study found that large loading doses of vitamin D3 rapidly and safely normalize 25OHD levels, and that monthly dosing is similarly effective after 3–5 months. With baseline 25OHD > 50 nmol/L, vitamin D supplementation does not reduce PTH levels.
There is concern that vitamin D supplementation doses are frequently inadequate, and that compliance with daily medication is likely to be suboptimal.
This randomized double-blind trial compares responses to three high-dose vitamin D3 regimens and estimates optimal 25-hydroxyvitamin D (25OHD) levels, from changes in parathyroid hormone (PTH), and procollagen type I amino-terminal propeptide (P1NP) in relation to baseline 25OHD. Sixty-three elderly participants were randomized to three regimens of vitamin D supplementation: a 500,000-IU loading dose; the loading dose plus 50,000 IU/month; or 50,000 IU/month.
The Loading and Loading + Monthly groups showed increases in 25OHD of 58 ± 28 nmol/L from baseline to 1 month. Thereafter, levels gradually declined to plateaus of 69 ± 5 nmol/L and 91 ± 4 nmol/l, respectively. In the Monthly group, 25OHD reached a plateau of ~80 ± 20 nmol/L at 3–5 months. There were no changes in serum calcium concentrations. PTH and P1NP were only suppressed by vitamin D treatment in those with baseline 25OHD levels <50 and <30 nmol/L, respectively.
Large loading doses of vitamin D3 rapidly and safely normalize 25OHD levels in the frail elderly. Monthly dosing is similarly effective and safe, but takes 3–5 months for plateau 25OHD levels to be reached.
KeywordsCholecalciferol Osteoporosis Parathyroid hormone Vitamin D
Adequate circulating concentrations of 25OHD are important for optimal bone health. Vitamin D status is related to bone density [1, 2, 3, 4, 5], bone turnover markers [6, 7], falls risk functional stability and strength [4, 8], and to the occurrence of hip fractures [9, 10, 11]. Intervention studies with vitamin D suggest beneficial effects on fractures, particularly in institutionalized individuals and when calcium is co-administered [12, 13, 14]. The elderly are particularly susceptible to vitamin D insufficiency [3, 15, 16] as a result of reductions in mobility, time spent outdoors, sun exposure (contributed to by increased skin coverage), intrinsic skin response to UV radiation, and dietary vitamin D intake [17, 18]. The consequences of vitamin D deficiency are potentially severe in the elderly since it is likely to contribute to their high fracture risk and to the mortality that is associated with fractures in this age group [19, 20].
Vitamin D deficiency is thought to contribute to bone loss by stimulating PTH secretion, resulting in increased bone resorption [21, 22]. Consequently, determinations of the optimal level of 25OHD have been based on estimates of the level of 25OHD required to minimize PTH concentrations, most frequently from cross-sectional studies. These analyses suggest that the optimal 25OHD concentration lies between 25 and 122 nmol/L . This variability may result from other factors that impact on PTH levels (such as dietary calcium intake, renal function) and from assumptions made in fitting curves through widely scattered data points. An approach that circumvents some of these problems is to study the impact of vitamin D supplementation on PTH concentrations longitudinally, in order to identify a level of 25OHD above which supplementation has no effect on PTH. There are only three such longitudinal investigations, and these have produced more consistent estimates of optimal 25OHD, ranging from 40 to 50 nmol/L [23, 24, 25]. Further corroboration of these results is needed.
Most people requiring vitamin D supplementation take this in the form of a daily tablet, but there is concern that the doses used are frequently inadequate, and that compliance with daily medication is likely to be suboptimal. The long half-life of 25OHD following oral calciferol supplementation  means that larger, less frequent doses are a practical alternative to daily supplementation. Despite this, systematic evaluation of regimens using high doses of vitamin D is lacking. The present study seeks to address this need by comparing 25OHD responses to three different high-dose cholecalciferol regimens. In addition, we have also used this dataset to determine the optimal level of 25OHD in these subjects, by assessing the relationships between baseline 25OHD and the responses of PTH and the bone formation marker P1NP (procollagen type I amino-terminal propeptide) to vitamin D supplementation.
Materials and methods
Patients aged ≥65 years in the general medical wards of a metropolitan hospital were recruited between October 2005 and May 2006. Exclusion criteria were a creatinine clearance <20 mL/min, current glucocorticoids use for >6 months, recent calciferol treatment at doses >600 IU/day, disorders or drugs that might influence vitamin D or PTH metabolism, and a life expectancy of <6 months. Sixty-three of 183 eligible participants agreed to participate. The main reason given for potential subjects declining was their and their spouses’ deteriorating health. All participants provided written informed consent and the study procedures were approved by the local ethics committee.
Participants were randomized into one of three treatment groups using a minimization algorithm to ensure balanced numbers of men and women between groups. One group (Loading) took 10 × 50,000 IU vitamin D3 tablets (12.5 mg) at the beginning of the study followed by a monthly placebo tablet for 8 months. A second group (Loading + Monthly) took 10 × 50,000 IU vitamin D3 tablets at study entry, followed by a 50,000 IU (1.25 mg) vitamin D3 tablet monthly. The third group (Monthly) took one vitamin D plus 9 placebo tablets at baseline followed by a vitamin D3 tablet monthly. The baseline doses were administered by study personnel, resulting in 100% compliance. Compliance with subsequent doses was assessed by tablet counts. Participants and investigators were blinded to the treatment assignments throughout. Trial medication was supplied by API Consumer Brands, Auckland
Plasma and serum were stored at −80°C until the completion of the study, when assays for the measurement of plasma intact PTH, serum 25OHD and serum P1NP were undertaken. PTH and P1NP were measured using automated Roche electrochemiluminescence immunoassays, and 25OHD was measured using the DiaSorin radioimmunoassay. Each participant’s samples were measured in the same batch. Inter-assay coefficients of variation were 6.1% for PTH, 7.1% for P1NP and 5.7% for 25OHD. Creatinine clearance was estimated using the Cockroft–Gault formula.
Details of medical and pharmacological history were taken from hospital medical records and during an interview conducted at baseline. Time spent outdoors, from an average of the times in mid-summer and mid-winter, was recorded. Calcium and vitamin D intakes were assessed using a food frequency questionnaire . Additional vitamin-D-containing foods were added to this questionnaire, and intake was quantified using food composition data from the Foodfiles 2004© (Revision 18) database (New Zealand Institute for Crop and Food Research Ltd., Christchurch, New Zealand) accessed through FoodWorks Professional Edition 4 (Xyris Software (Australia) Pty. Ltd., Brisbane, Australia). Calcium and vitamin D intakes include that derived from supplements. Grip strength was measured in triplicate in the dominant hand at baseline and at 9 months.
Data were analyzed on an intention to treat basis using SPSS version 14.0 (SPSS Inc., Chicago, IL, USA) and SAS version 9.1 (SAS Institute Inc., Cary, NC, USA). Two-way (month × dose) mixed model analysis of variance (ANOVA) was applied to examine changes in the dependent variables 25OHD, PTH, and P1NP. Differences in the areas under the curves for these variables were analyzed using one-way ANOVA. Post-hoc multiple comparisons are Bonferroni-adjusted. Linear relationships between 25OHD, PTH, P1NP and their changes are reported as Pearson’s correlation coefficients. Data are reported as mean ± SEM unless otherwise stated. All tests were two-tailed and P < 0.05 was considered significant.
Baseline characteristics of study participants
Loading (n = 19)
Loading + Monthly (n = 22)
Monthly (n = 22)
Total (n = 63)
83 ± 6
81 ± 8
81 ± 6
82 ± 7
Body mass index (kg/m2)
26 ± 4
26 ± 5
26 ± 4
26 ± 5
Physical activity (h/week)
7.0 ± 8.0
6.0 ± 6.5
6.4 ± 6.5
6.4 ± 6.8
Time outside (h/week)
6.5 ± 5.9
4.9 ± 7.2
4.9 ± 4.7
5.4 ± 6.0
Calcium intakea (mg/day)
1,150 ± 480
1,400 ± 590
1,350 ± 530
1,320 ± 540
Vitamin D intake (IU/day)
310 ± 260
290 ± 240
250 ± 220
280 ± 230
58 ± 25
66 ± 41
50 ± 25
58 ± 32
5.1 ± 2.4
5.7 ± 3.6
5.2 ± 2.7
5.3 ± 2.9
83 ± 79
58 ± 45
87 ± 61
76 ± 62
Serum adjusted calcium (mmol/L)
2.3 ± 0.1
2.4 ± 0.1
2.4 ± 0.1
2.4 ± 0.1
Serum phosphate (mmol/L)
1.1 ± 0.2
1.1 ± 0.2
1.0 ± 0.2
1.1 ± 0.2
Alkaline phosphatase (IU/L)
80 ± 31
107 ± 110
95 ± 36
95 ± 70
Creatinine clearance (mL/min)
48 ± 17
48 ± 17
49 ± 15
49 ± 16
Grip strength (kg)
24 ± 8
20 ± 10
19 ± 5
21 ± 8
Six participants withdrew within the first month because of ill health. Four withdrew subsequently for health reasons. One subject moved away from the area, and five died during the study. Of the 47 participants who completed the study, one completed a 4-week course of oral steroids for a bronchial infection and another was started on an oral bisphosphonate during the study. Their data are included in the analyses, except where noted otherwise. Compliance, defined as the number of monthly tablets taken as a percentage of the number of monthly tablets that should have been taken, was <80% in only two subjects, and the median was 100% in all groups.
25OHD, PTH and P1NP relationships at baseline
25OHD responses to supplementation
With the more widespread use of vitamin D supplementation, assessment of the response in non-deficient subjects is also of interest to determine the likelihood of hypercalcemia. Following the loading dose, the increase to 1 month was 50 (95% CI, 38–63) nmol/L, somewhat less than the increase of 71 (95% CI 58–84) nmol/L seen in the deficient group (P = 0.026). Mean 25OHD peaked at ~130 nmol/L, and the highest individual 25OHD level at 1 month was 220 nmol/L, in a subject with a baseline value of 136 nmol/L. At 3 months, the maximum level was 160 nmol/L. Serum calcium (corrected for albumin) was measured at baseline, 1, 5, and 9 months. There was no significant increase in serum calcium concentrations from baseline during the study, and no individual values were above the upper end of the reference range. The mean (±SD) values at these time-points were: 2.32 ± 0.13, 2.35 ± 0.12, 2.28 ± 0.11 and 2.24 ± 0.08 mmol/L in the Loading group; 2.36 ± 0.09, 2.31 ± 0.08, 2.31 ± 0.08, and 2.28 ± 0.09 in the Loading + Monthly group; 2.39 ± 0.09, 2.37 ± 0.09, 2.31 ± 0.10, and 2.28 ± 0.08 mmol/L in the Monthly group.
PTH and P1NP responses to supplementation as indicators of optimal 25OHD
Grip strength in the entire cohort increased 2.6 ± 0.6 kg from baseline to the end of the study (P < 0.01), but there were no between-groups differences. Increases in individual subjects were not related to the change in 25OHD (r = −0.1). In subjects in whom 25OHD increased by ≥20 nmol/L, there was no effect of baseline 25OHD levels (grouped as above or below 50 nmol/L) on the change in grip strength.
The present study demonstrates the efficacy and safety of three regimens utilizing a 50,000-IU tablet of vitamin D3. Individuals with initial 25OHD levels < 50 nmol/L who received a 500,000 IU loading dose of vitamin D3 showed increases in levels of 25OHD at 1 month to around 100 nmol/L. In contrast, with monthly 50,000 IU dosing, it took ~3 months for those with low initial vitamin D status to attain satisfactory levels. Both regimens are potentially useful. A loading dose is likely to be desirable for individuals with a high fracture risk and probability of poor vitamin D status, because it rectifies these low levels promptly, though there is still a 3–5 month delay in the bone marker response to the increased 25OHD levels. Rapid repletion of vitamin D is mandatory in deficient patients wishing to receive intravenous bisphosphonates post-fracture , since these drugs can cause sustained hypocalcemia in such individuals . Individuals with osteomalacia or myopathy secondary to vitamin D deficiency will also require rapid repletion. In contrast, many individuals whose fracture risk is lower or in whom vitamin D status is uncertain can be managed satisfactorily with 50,000 IU monthly. If compliance with a monthly regimen is in question, the present results suggest that single doses of 500,000 IU every 6–12 months might provide a similar outcome.
Several other studies have assessed the effects of large intermittent doses of vitamin D. We have previously shown that a single dose of 500,000 IU vitamin D3 is safe and effective in older women with serum 25OHD levels <25 nmol/L, and that it produces a 15% reduction in PTH concentrations . Dosing with 300,000 IU vitamin D3 leads to a rapid rise in 25OHD, peaking at 17 days and declining with a half-life of 90 days . Adams  administered vitamin D2 500,000 IU over 5 weeks to deficient patients resulting in a mean increase in 25OHD of 60 nmol/L, and Przybelski  gave 600,000 IU of D2 over 4 weeks with an increase of 47 nmol/L, both results similar to that found in the present study. There were no safety issues in either study. Diamond administered 600,000 IU of D3 as a single intramuscular injection to 50 subjects. Primary hyperparathyroidism was subsequently diagnosed in one person and mild hypercalcemia was found in two others at 12 months, raising uncertainty regarding its etiology . Others have shown that vitamin D2 7,000–10,000 IU/week will prevent falls  and 4–12-monthly doses of vitamin D may prevent fractures [35, 36].
Heaney’s group has investigated the time-course and dose–response to daily oral vitamin D3 in healthy young men with initial 25OHD levels of ~70 nmol/L [37, 38]. After 8–20 weeks dosing, 25OHD increased by 0.015–0.03 nmol/L for each IU/day of vitamin D3, similar to the increases seen with monthly dosing in the present study. Comparing these increments with those after the loading dose used here, suggests that this loading dose produces an increment at 1 month equivalent to a long-term daily dose of about 4,000 IU .
The number of subjects in the present study with normal 25OHD levels was unexpected, and is likely to indicate that they had been taking vitamin D supplements which they did not report, prior to their hospital admission. This has provided the opportunity to assess the effect of high-dose supplementation in individuals whose levels are ≥50 nmol/L. In this group, treatment with 50,000 IU monthly had a minimal effect, in contrast to those with baseline 25OHD < 50 nmol/L, in whom 25OHD increased by around 40 nmol/L. Providing replete individuals with a single 500,000-IU dose resulted in increases in 25OHD of around 50 nmol/L. Though substantial, these increases were again not as great as when baseline 25OHD was <50 nmol/L. A similar observation was made by Viljakainen et al. who provided vitamin D3 (200–800 IU/day) to elderly women and noted that the response was almost doubled when initial 25OHD levels were below the median level of 47 nmol/L . Vieth has suggested that homeostatic control mechanisms, possibly hepatic, prevent large increases in 25OHD . Our observations support this possibility, though uptake into adipose tissue could also contribute to this effect.
Levels of 25OHD which constitute a risk of hypercalcemia are not clear, and safety issues associated with high levels cannot be ignored. Vieth has pointed out that studies which document both vitamin D dose and 25OHD levels have demonstrated hypercalcemia at intakes >10,000–40,000 IU/day, and these intakes are usually associated with 25OHD levels >500 nmol/L [40, 41, 42]. Despite this, upper reference limits for 25OHD are often set at much lower levels of 150–250 nmol/L. The highest level observed in any individual in the present study was 220 nmol/L at 1 month, which appears to be well within the safe range, particularly since this peak was transient. This contention is supported by the normal serum calcium levels throughout this period. However, we would not recommend loading doses in subjects who are vitamin D sufficient. In contrast, in vitamin-D-deficient subjects, the loading dose was efficient in normalizing 25OHD and again did not result in values above the reference range.
The present data provide the opportunity to assess prospectively the optimal concentration of 25OHD, by identifying levels above which supplementation does not further suppress PTH and indices of bone turnover. The PTH data from the present study suggest that optimal levels of 25OHD are around 50 nmol/L, and the P1NP analyses suggest the optimal value might be even lower. Both these parameters might be affected by the subjects’ mobility, though only 11–19% had significant mobility impairment during the study period and their exclusion did not affect this analysis. This matter is subject to considerable controversy [43, 44, 45] and much of the evidence supporting higher levels is from cross-sectional studies. Such estimates are problematic because of the considerable variability in PTH levels within each study, meaning that the levels of 25OHD associated with a PTH nadir are not evident from simple inspection of scatter-plots of PTH versus 25OHD. As a result, statistical assumptions are necessary to develop models which define the relationship between PTH and 25OHD. The nature of these assumptions is often arbitrary, and will impact substantially on the results found.
Only three studies have used a longitudinal design to assess optimal 25OHD, as in the present study. All have arrived at optimal levels for 25OHD of around 50 nmol/L, even though they have varied in age group, race, country and duration [23, 24, 25]. Malabanan  assessed responses to 50,000 IU/week vitamin D2 over 8 weeks in 35 subjects with a mean age of 76 years. They found that PTH was not further suppressed in those whose baseline 25OHD concentrations were >50 nmol/L. Lips utilized data from the placebo group of a multi-national clinical trial in osteoporotic women, and found that in response to calcium (500 mg/day) and vitamin D3 (400–600 IU/day) over 6 months, subjects only showed a significant decrease in PTH if their baseline 25OHD was <50 nmol/L . Aloia studied healthy, postmenopausal African–American women given vitamin D3 800–2,000 IU/day over 3 years, and found that the threshold for PTH suppression was a 25OHD level between 40 and 50 nmol/L . Thus, there is substantial consistency between the present study and the others which have used a similar robust design in identifying a threshold of 25OHD for bone health. This does not mean that this is the optimal 25OHD level for the other tissues on which vitamin D is active , nor does it take into account the seasonal variation in 25OHD levels. We have recently shown that to ensure a nadir level of 25OHD in winter >50 nmol/L, the summer peak must be 60–70 nmol/L in women, and 70–90 nmol/L in men .
A limitation of this study is that there is no placebo-only group. When planning this study, we felt that it would be unethical to fail to provide any vitamin D treatment to frail elderly people who were deficient. It is therefore possible that some of the changes in 25OHD were due to factors other than the interventions provided, such as altered behavior relating to sun exposure or dietary modification. If this were the case, it is likely that the effect would be consistent across all treatment groups and small, given the clear differences in between-group responses. Furthermore, it is unlikely that increased ultraviolet exposure would have affected our results relating to changes in PTH and P1NP since these changes would have occurred irrespective of the cause of the 25OHD increases. In addition, we were not able to control calcium intake nor fluctuations in renal function, but these are likely to be lesser confounders in a longitudinal study than they would be using a cross-sectional design. The generalizability of these findings to less frail or younger individuals remains to be determined, though the studies of Lips and Aloia were in much younger individuals and yet found similar relationships.
In conclusion, our results show that a large single dose of vitamin D3 rapidly and safely normalizes 25OHD levels in the frail elderly. Monthly 50,000 IU dosing is similarly effective and safe, but takes 3–5 months for plateau 25OHD levels to be reached. The single high-dose regimen provides a convenient option for the rapid correction of vitamin D deficiency, and monthly dosing is likely to increase compliance in this population group.
The authors are grateful to API Consumer Brands for providing the study medication.
This study was supported by the Health Research Council of New Zealand
Conflicts of interest
- 3.Melin AL, Wilske J, Ringertz H, Saaf M (1999) Vitamin D status, parathyroid function and femoral bone density in an elderly Swedish population living at home. Aging Clin Exp Res 11:200–207Google Scholar
- 13.Avenell A, Gillespie WJ, Gillespie LD, O'Connell DL (2005) Vitamin D and vitamin D analogues for preventing fractures associated with involutional and post-menopausal osteoporosis. Cochrane Database Syst Rev CD000227Google Scholar
- 14.Boonen S, Lips P, Bouillon R, Bischoff-Ferrari HA, Vanderschueren D, Haentjens P (2007) Need for additional calcium to reduce the risk of hip fracture with vitamin D supplementation: evidence from a comparative metaanalysis of randomized controlled trials. J Clin Endocrinol Metab 92:1415–1423PubMedCrossRefGoogle Scholar