Background: The Food and Nutrition Board of the National
Academy of Sciences states that 95 ?g vitamin D/d is the lowest
observed adverse effect level (LOAEL).
Objective: Our objective was to assess the efficacy and safety
of prolonged vitamin D3intakes of 25 and 100 ?g (1000 and
4000 IU)/d. Efficacy was based on the lowest serum 25-hydroxy-
vitamin D [25(OH)D] concentration achieved by subjects taking
vitamin D3; potential toxicity was monitored by measuring
serum calcium concentrations and by calculating urinary cal-
Design: Healthy men and women (n = 61) aged 41 ± 9 y (x–± SD)
were randomly assigned to receive either 25 or 100 ?g vitamin
D3/d for 2–5 mo, starting between January and February. Serum
25(OH)D was measured by radioimmunoassay.
Results: Baseline serum 25(OH)D was 40.7 ± 15.4 nmol/L
(x–± SD). From 3 mo on, serum 25(OH)D plateaued at
68.7 ± 16.9 nmol/L in the 25-?g/d group and at 96.4 ± 14.6 nmol/L
in the 100-?g/d group. Summertime serum 25(OH)D concentra-
tions in 25 comparable subjects not taking vitamin D3were
46.7 ± 17.8 nmol/L. The minimum and maximum plateau serum
25(OH)D concentrations in subjects taking 25 and 100 ?g vita-
min D3/d were 40 and 100 nmol/L and 69 and 125 nmol/L,
respectively. Serum calcium and urinary calcium excretion did
not change significantly at either dosage during the study.
Conclusions: The 100-?g/d dosage of vitamin D3effectively
increased 25(OH)D to high-normal concentrations in practi-
cally all adults and serum 25(OH)D remained within the phys-
iologic range; therefore, we consider 100 ?g vitamin D3/d to be
a safe intake.
Am J Clin Nutr 2001;73:288–94.
hypercalciuria, toxicity, lowest observed adverse effect level,
Cholecalciferol, calcidiol, vitamin D3,
Food and Nutrition Board guidelines specify 50 ?g/d as the
highest vitamin D intake that healthy adults can consume with-
out risking hypercalcemia [it is the upper limit, or the no adverse
effect level (NOAEL)]. A prolonged intake of 95 ?g vitamin D/d
is said to be the lowest observed adverse effect level (LOAEL),
a dosage that causes hypercalcemia in healthy adults (1). These
intake limits have changed little from previous guidelines (2).
However, the current guidelines (1) are based on the data of
Narang et al (3), who reported that mean serum calcium concen-
trations were abnormally high in 6 healthy subjects who con-
sumed 95 ?g vitamin D/d for 3 mo. More recently, Adams and
Lee (4) reported a high urinary calcium-creatinine ratio in
4 patients taking nutritional supplements containing vitamin D2
(ergocalciferol). Substantial concern has been expressed about
the safety of consuming vitamin D at dosages greater than the
highest dosage available without prescription (25 ?g/d) (1, 5).
Directly related to this issue is the question of how much
vitamin D is needed to ensure target serum 25-hydroxyvitamin
D [25(OH)D] concentrations. According to the recommended
dietary allowances, persons should achieve “levels of intake of
essential nutrients considered. . . to be adequate to meet the
known nutritional needs of practically all healthy persons” (2,
6). Serum 25(OH)D is the appropriate index of vitamin D nutri-
tional adequacy (1). Therefore, the nutritional need for vitamin
D would be the amount of it needed to ensure for “practically
all healthy persons” that serum 25(OH)D concentrations are
maintained above a concentration considered adequate. Moder-
ate vitamin D malnutrition is based on what is now well docu-
mented—an inverse relation between serum 25(OH)D and
parathyroid hormone (PTH) concentrations (7–9). 25(OH)D
concentrations <40–50 nmol/L are considered to be insuffi-
cient (10–12). Because the suppression of PTH is seen as ben-
eficial for bone, many now regard serum 25(OH)D concentra-
tions ≥75–100 nmol/L as desirable. This is the concentration at
which PTH approaches a minimum in its relation with
25(OH)D (7, 9, 11, 13–15).
An intake of ≥100 ?g vitamin D3(cholecalciferol)/d may be
required to ensure desirable 25(OH)D concentrations (16). How-
ever, because such intakes exceed the LOAEL, it is not feasible
Am J Clin Nutr 2001;73:288–94 Printed in USA. © 2001 American Society for Clinical Nutrition
Efficacy and safety of vitamin D3intake exceeding the lowest
observed adverse effect level1–3
Reinhold Vieth, Pak-Cheung R Chan, and Gordon D MacFarlane
1From the Department of Pathology and Laboratory Medicine, the Uni-
versity of Toronto and Mount Sinai Hospital, Toronto, Canada, and DiaSorin
Inc, Stillwater, MN.
2Supported in part by a grant from the Dairy Farmers of Canada. P-CR
Chan was a recipient of a postdoctoral fellowship in Clinical Chemistry from
the Ontario Ministry of Health.
3Address reprint requests to R Vieth, Department of Pathology and Labo-
ratory Medicine, Mount Sinai Hospital, 600 University Avenue, Toronto,
Ontario, Canada M5G 1X5. E-mail: email@example.com.
Received January 10, 2000.
Accepted for publication July 21, 2000.
to use them in studies of healthy adults, especially in long-term
studies designed to evaluate health effects. Ethical review panels
may be hesitant to approve the use of a dosage that exceeds the
LOAEL, funding agencies may be hesitant to provide the funds
needed for such study, and study subjects themselves may have
or develop reservations that could lead to poor study compliance.
There are few data from which to establish vitamin D safety
and toxicity limits. Some studies that might be considered rele-
vant because they include data on serum calcium, urinary cal-
cium, or both have major shortcomings, eg, ≤6 subjects (3, 4,
17), follow-up times ≤3 mo (3, 18, 19), nonspecification of the
form of vitamin D used (vitamin D2or vitamin D3) (3, 4), or
nonverification of the accuracy of the stated dose (3, 4). The
objectives of the present study were to assess the efficacy of high
vitamin D intakes, in terms of the serum 25(OH)D concentra-
tions ensured for practically all adults, and to monitor the long-
term safety of vitamin D3intakes that exceed the LOAEL.
SUBJECTS AND METHODS
The study protocol was approved by an ethical review com-
mittee at the University of Toronto and subjects signed a form
indicating their informed consent. We recruited 73 generally
healthy volunteers, most of whom worked in the clinical labora-
tory departments of 2 Toronto hospitals (latitude 43??). The
characteristics of the subjects are summarized in Table 1. Sub-
jects were randomly assigned to receive a vitamin D3intake of
either 25 or 100 ?g/d; only data for subjects who consumed vita-
min D3for ≥1 mo are provided. The doses were assigned ran-
domly by distributing the 2 dose formats in equal numbers
through a partitioned rack from which each subject’s first coded
vial was taken, in sequence, without the giver’s or the subject’s
knowledge of the dose it contained. Of the 73 subjects enrolled
initially, 61 completed ≥1 mo of the protocol. The study began
between January and February (baseline, time 0). At 0, 0.5, 1, 2,
3, 4, and 5 mo of vitamin D3supplementation, a morning urine
sample (second void of the day) was collected and one tube of
blood was collected to prepare serum for biochemical testing. At
the time of each blood sampling, the vials that had contained the
vitamin D3solutions were collected to monitor compliance and
subjects were given fresh vials. Subjects were free to withdraw
from the study at any time and their withdrawal is reflected in the
number of points shown in the figures. When insufficient serum
samples were available from the clinical service laboratory,
25(OH)D concentrations were not measured; the measurement of
calcium was given priority. Blood was collected from 25 labora-
tory coworkers who did not take part in the vitamin D intake
study; this group served as an end-of-study reference group.
Vitamin D3was purchased in crystalline form from Sigma (St
Louis) and dissolved in US Pharmacopoeia–grade ethanol. The
molar concentration of vitamin D3was adjusted to 433 ?mol/L
[100 ?g (4000 IU) per 0.6-mL dose], which was based on an
absorbance at 265 nm [7.90 absorbance units (AU) with use of
an extinction coefficient of 18300 AU·mol?1·L?1] on an 8452A
diode array spectrophotometer (Hewlett-Packard, Palo Alto,
CA). The lower dose was adjusted to 108 ?mol/L [25 ?g (1000
IU) per 0.6-mL dose]. Appropriately blanked ultraviolet absorp-
tion spectra of the doses taken before, during, and after the study
remained identical. In addition, chromatographic analysis of the
dose preparations consistently indicated purity by showing only
the one peak appropriate for vitamin D3. Vitamin D3was con-
sumed daily by each subject by mixing 0.6 mL of the ethanolic
solution into juice or water just before drinking it (20).
Serum 25(OH)D was measured by radioimmunoassay (Dia-
Sorin, Stillwater, MN). Serum calcium and phosphate and uri-
nary calcium, phosphate, and creatinine were measured with an
Integra automatic chemistry analyzer (Roche, Basel, Switzer-
land). Urinary calcium excretion was assessed as a ratio of uri-
nary calcium to creatinine concentrations (4, 21).
Safety and efficacy were assessed by using repeated-measures
analysis of variance (ANOVA) followed by Dunnett’s multiple
(pairwise) comparison t test to determine the CI for the mean dif-
ferences from baseline at each time point (release 6.12; SAS Sta-
tistical Software, Cary, NC). A two-tailed P value < 0.05 was
considered significant. For regression lines plotted nonparamet-
rically, we used the locally weighted regression and smoothing
scatterplot (LOWESS) approach (22) using SPSS statistical soft-
ware (versions 8 or 10; SPSS Inc, Chicago). This software was
also used for paired t test comparisons with baseline values.
Criteria for safety and efficacy
The efficacy of a given dose of vitamin D3was based on the
final serum 25(OH)D concentration (not its change). A dose was
considered effective if it ensured a serum 25(OH)D concentra-
tion ≥75 nmol/L (7, 9, 11, 13–15). A dose was considered safe if
all of the following criteria were met:
1) a mean serum calcium concentration ≤2.75 mmol/L (11 mg/dL)
during vitamin D supplementation, the criterion used by oth-
ers to support the current LOAEL (1);
2) no increase in the relative number of subjects with hypercal-
cemia relative to baseline, when the subjects were tested
without vitamin D;
SAFETY OF VITAMIN D3EXCEEDING THE LOAEL289
Characteristics of the 2 groups of subjects1
25 ?g/d100 ?g/d
Number of subjects
Body weight (kg)2
Basal 25(OH)D (nmol/L)2
Subjects with basal 25(OH)D
<40 nmol/L (%)
Subjects with basal 25(OH)D
<25 nmol/L (%)
41.6 (18–53) 39.9 (23–56)
67.8 ± 11.9
66.4 ± 12.9
43.3 ± 16.8
37.9 ± 13.4
12.1 21.4 
1Range in parentheses; n in brackets 25(OH)D, 25-hydroxyvitamin D.
3) a mean urinary calcium-creatinine ratio ≤1.0 (when calcium
and creatinine are measured in mmol; ≤0.37 when measured
in mg) during vitamin D supplementation; and
4) no increase in the relative number of subjects with hypercal-
ciuria relative to baseline, when the subjects were tested
without vitamin D.
The subjects’mean (±SD) winter serum 25(OH)D concentration
before supplementation was 40.7 ± 15.4 nmol/L. The summer
serum 25(OH)D concentration in 25 comparable subjects not taking
vitamin D3was 46.7 ± 17.8 nmol/L. Of the 61 subjects, 28 (45.9%)
had low 25(OH)D concentrations (<40 nmol/L) and 10 (16.4%) had
concentrations in the osteomalacic range (<25 nmol/L).
Serum 25(OH)D concentrations during the supplementation
protocol are shown in Figure 1. With the 100-?g/d dosage,
repeated-measures ANOVA and post hoc comparisons with Dun-
nett’s test indicated that serum 25(OH)D increased significantly
beginning at 2 wk until the end of the study. With the 25-?g/d
dosage, repeated-measures ANOVA and post hoc comparisons
with Dunnett’s test indicated a significant increase from only
1 mo on. A conventional paired t test showed the expected signi-
ficant increase in 25(OH)D at 2 wk (P < 0.01). At 3 mo, serum
25(OH)D concentrations peaked at 68.7 ± 16.9 nmol/L in the
25-?g/d group and at 96.4 ± 14.6 nmol/L in the 100-?g/d group
and remained relatively stable at these concentrations for the
remainder of the study. For each group, the final 25(OH)D con-
centration attained was not significantly affected by either the
initial 25(OH)D concentration or by body weight (Figure 2).
From 3 mo on, the minimum and maximum of the plateau
serum 25(OH)D concentrations were 40 and 100 nmol/L in the
25-?g/d group and 69 and 125 nmol/L in the 100-?g/d groups.
The 25-?g/d dosage was effective at ensuring 25(OH)D concen-
trations of ≥75 nmol/L in 8 of the 23 (35%) subjects. The 100-?g/d
dosage was effective at ensuring 25(OH)D concentrations of
≥75 nmol/L in 22 of the 25 (88%) subjects.
The serum calcium concentrations and urinary calcium-crea-
tinine ratios measured during the study are shown in Figure 3. In
all subjects in the 25- and 100-?g/d groups, serum calcium con-
centrations remained within the reference range (2.2–2.6 mmol/L).
There was no significant change from baseline in serum calcium
at any time by repeated-measures ANOVA. CIs for the differ-
ences between serum calcium concentrations during the study
and baseline values were consistently ≤0.10 mmol/L. The upper
97.5% confidence limits for the mean difference in serum cal-
cium results are illustrated in Figure 3 and they show with sta-
tistical confidence (P < 0.05) that the values were consistently
<2.45 mmol/L. Similarly, there was no significant change from
baseline in urinary calcium-creatinine excretion ratios at any
time point by repeated-measures ANOVA. The upper confidence
limits for the differences between baseline urinary calcium-
creatinine excretion ratios and those during supplementation
were consistently <0.70 (Figure 3). There were more urinary
calcium-creatinine excretion ratios >1.0 in the 100-?g/d group
(in one subject, 2 of 6 values were >1.0 during treatment) than
in the 25-?g/d group. The relative number of occurrences of
hypercalciuria across the entire follow-up period was not signi-
ficantly different between the 2 dosage groups on the basis of
chi-square tests. Similarly, the relative number of treated sub-
jects with hypercalciuria was not significantly different from the
number of untreated subjects (baseline) with hypercalciuria.
More subjects than were used here will be required to show
whether the higher dose does in fact increase the occurrence of
hypercalciuria. The paired t test, which was used to compare
baseline values with those during treatment, is less appropriate
than is ANOVA with this study design, but is more sensitive for
detecting changes from baseline. Like the ANOVA, paired t tests
showed no significant changes in serum calcium concentrations
or urinary calcium-creatinine ratios at any time point during the
5-mo vitamin D intake period.
Prolonged consumption of vitamin D at a dosage of 100 ?g/d
resulted in plateau 25(OH)D concentrations averaging 96 nmol/L.
290VIETH ET AL
FIGURE 1. Serum 25-hydroxyvitamin [25(OH)D] concentrations
in healthy adults before and during supplementation with 25 (A) and
100 (B) ?g vitamin D3/d. The heavy lines represent the locally
weighted regression and smoothing scatter plot. Of 33 subjects starting
the protocol at 25 ?g/d, 26 completed ≥3 mo and 15 completed the
entire 5 mo of supplementation. Of 28 subjects starting the protocol at
100 ?g/d, 25 completed ≥3 mo and 15 completed the entire 5 mo of
On the basis of current nutritional guidelines, one would expect
25(OH)D concentrations much higher than this because of the
presumption that the 95-?g/d LOAEL will raise the average serum
calcium concentration to the hypercalcemic range (>2.75 mmol/L,
or 11 mg/dL) (1). In the present study, mean serum calcium con-
sistently remained <2.45 mmol/L. The vitamin D produced in the
skin as a result of sun exposure is not enough to cause hypercal-
cemia, even though sunshine can raise circulating concentrations
of 25(OH)D to >200 nmol/L (16, 24). The present results expand
on earlier work in which serum calcium was not affected signifi-
cantly by 2-mo intakes of 100 (18) or 1250 (19) ?g vitamin D3/d
or 12-mo intakes of 350 ?g vitamin D2/d (17). Together, the evi-
dence is overwhelming that the hypercalcemia evoked by Narang
et al (3) after 3 mo of treatment with 95 ?g vitamin D/d must
have been in error. We contend that without data on serum
25(OH)D concentrations, the hypercalcemia observed by Narang
et al was effectively a biological response indicating that they had
grossly underestimated the amount of vitamin D in the doses they
used. It is unfortunate that the study by Narang et al remains the
only study cited by the Food and Nutrition Board to support the
current LOAEL of 95 ?g/d (1).
The results of the present study differ slightly from those of
earlier studies in that we did not detect an effect of treatment on
urinary calcium excretion. Our urine samples were collected
from nonfasting subjects at the second void of the morning (col-
lected before 1000). Collection times in other studies were not
specified and the subject populations may have differed from
ours. Tjellesen et al (18) used the same dosage we did and showed
small but significant increases in urinary calcium-creatinine
excretion ratios. Arthur et al (17) reported a 50% increase in uri-
nary calcium-creatinine ratios in 6 women who took 350 ?g vita-
min D2/d for 16 mo (17). Adams and Lee (4) reported 4 cases of
SAFETY OF VITAMIN D3EXCEEDING THE LOAEL291
FIGURE 2. Effects of baseline (0 mo) body weight and serum 25-hydroxyvitamin D [25(OH)D] concentrations on final serum 25(OH)D concen-
trations in healthy adults after supplementation for 3–5 mo with 25 (A and C; n = 25) or 100 (B and D; n = 25) ?g vitamin D3/d. Because the 95%
confidence limits encompass the horizontal, they indicate that plateau 25(OH)D concentrations were not significantly affected by baseline 25(OH)D
concentrations or body weight (23), ie, there were no significant correlations (r2) for any of the correlations. A: y = (40.86 + 0.42) ? wt; r2= 0.08.
B: y = (89.89 + 0.10) ? wt; r2= 0.01. C: y = (55.36 + 0.33) ? baseline 25(OH)D; r2= 0.08. D: y = (90.91 + 0.15) ? baseline 25(OH)D; r2= 0.02.
vitamin D intoxication on the basis of only increased fasting uri-
nary calcium-creatinine ratios.
For people with serum calcium concentrations in the reference
range, the implications of high-normal urinary calcium concentra-
tions need to be understood better before this can be regarded as
an adverse response. Until there is more evidence to verify the
safety of intakes of 100 ?g vitamin D/d, it is simple and prudent
to continue to monitor calcium in randomly collected urine sam-
ples. Although randomly collected urine samples can be affected
by large calcium loads in the hours before collection (25, 26),
24-h collections are notoriously affected by improperly timed col-
lections, missed urine voids, and daily variations in calcium
intake. Calcium concentrations in randomly collected urine sam-
ples correlate well with calcium concentrations in 24-h urine sam-
ples (21, 27–30). The mean calcium-creatinine ratio in randomly
collected urine samples from nonfasting groups of healthy sub-
jects is ?0.40 (0.14 when measured in mg) (29, 31, 32). A 24-h
urinary calcium excretion >10 mmol/d (400 mg/d) is considered
to indicate hypercalciuria (33). Graphs of regression data for the
relation between random urinary calcium-creatinine ratios versus
24-h calcium excretion (21, 28) indicate that at a 24-h calcium
excretion of 10 mmol/d, the nonfasting random calcium-creatinine
ratio is ?1.0 (0.37 when measured in mg). In our study there were
more urinary calcium-creatinine ratios >1.0 in the 100-?g/d group
than in the 25-?g/d group, but the difference was not significant.
In the absence of hypercalcemia, urinary calcium excretion is
a minor contributor to renal stone disease. Stronger risk factors
for calcium oxalate stones include low urine volume, hyperox-
aluria, and hypocitraturia. Urinary calcium is useful for classify-
ing normocalcemic patients who already have stone disease (34),
but it cannot discriminate adults who form calcium oxalate
stones from those who do not form kidney stones (33, 35). Fur-
thermore, it is highly controversial whether isolated high-normal
calcium excretion contributes adversely to either stone disease or
bone health (33, 35–37). Because it is impossible to prove by sta-
tistical analysis that an agent has no harmful effect, we can only
conclude that an intake of 100 ?g vitamin D3/d is safe because
there was no convincing evidence of harm.
Throughout the history of vitamin D supplementation in North
America, high-dose preparations of the vitamin D2form have
generally been used. Vitamin D3is common in lower-dose regi-
mens, but in some parts of the world vitamin D2is the only form
licensed for use. In terms of rickets prevention, research from the
1930s was inconclusive at detecting a difference in efficacy
between the 2 forms of vitamin D; therefore, pharmacopoeias
continue to regard vitamin D2as being equivalent to vitamin D3,
292VIETH ET AL
FIGURE 3. Total serum calcium concentrations and urinary calcium-creatinine excretion ratios at baseline (0 mo) and during supplementation with
25 (A and C) and 100 (B and D) ?g vitamin D3/d. The heavy line in each panel is the nonparametric, locally weighted regression and smoothing scat-
ter plot. The dotted lines reflect the upper limit of the central 95% CI for the mean change. The value of each dotted line was calculated by adding the
upper limit value (97.5%) for the mean change from baseline at each time point to the mean baseline value (month 0) for each group (repeated-measures
ANOVA followed by Dunnett’s test).
even though the latter form is more effective at raising serum
25(OH)D concentrations (20). What seems to have been forgotten
is that the literature of half a century ago established that, at high
doses, there was a greater risk of toxicity with what was called
“the purely artificial” compound, vitamin D2(38–40). One expla-
nation for the difference in toxicity was the poorer stability and
greater impurity of vitamin D2than of vitamin D3preparations
(38). There are newer reasons why vitamin D2has a greater
potential for harm. First, vitamin D binding protein has a weaker
affinity for the vitamin D2 metabolites than for 25-hydroxyvita-
min D3and 1,25-dihydroxyvitamin D3(41–43). This means that
the proportions of free 25-hydroxyvitamin D2and 1,25-dihydrox-
yvitamin D2 are higher and more biologically available. Second,
unique biologically active metabolites are produced from vitamin
D2in humans and there are no analogous metabolites derived
from vitamin D3(44). There is no doubt that vitamin D2is a
synthetic analogue of vitamin D, with different characteristics.
It is an anachronism to regard vitamin D2as a vitamin. Future
research into the toxicology of this vitamin needs to focus on
vitamin D3 as being something distinct from vitamin D2, for
which almost all our current toxicity data relate to.
The working definition of the recommended dietary allowance
has been to ensure “levels of intake of essential nutrients consid-
ered, in the judgment of the Food and Nutrition Board on the
basis of available scientific knowledge, to be adequate to meet the
known nutritional needs of practically all healthy persons” (2, 6).
For vitamin D, the relevant question could be, what 25(OH)D
concentration is desirable and how much vitamin D is needed to
ensure that most adults attain this intake (45)? We did not meas-
ure PTH in this study, so we cannot address the question of what
the desirable vitamin D intake is. However, the present results do
provide insights into the lowest 25(OH)D concentrations that can
be reasonably ensured in adults consuming 25 and 100 ?g vita-
min D3/d. The 25-?g/d intake offered reasonable assurance that
serum 25(OH)D concentrations in adults would be >40 nmol/L,
but did not ensure that most subjects would attain serum
25(OH)D concentrations considered desirable (>75 nmol/L). The
100-?g/d intake offered reasonable assurance that 25(OH)D con-
centrations in adults would be >69 nmol/L (Figures 1 and 2),
close to the lower end of the desirable concentration.
If the serum 25(OH)D concentration is the appropriate
measure of vitamin D nutritional adequacy (1), then more of the
present type of specific data are needed to define the amounts of
vitamin D required to ensure that for “practically all healthy per-
sons” serum 25(OH)D concentrations are maintained above an
amount considered adequate. There are subgroups who require
≥25 ?g vitamin D/d to maintain acceptable 25(OH)D concentra-
tions. Gloth et al (46) reported that in older patients with
25(OH)D concentrations <25 nmol/L, vitamin D intakes ranged
as high as 29 ?g/d. Patients with cystic fibrosis require >20 ?g
vitamin D/d to maintain 25(OH)D concentrations >40 nmol/L
(47). Despite the greater number of subjects and the longer fol-
low-up in the present study than in previous comparable studies
(3, 18, 19), consumption of vitamin D3at intakes ≥100 ?g/d
causes no harm and effectively raises 25(OH)D to high-normal
concentrations in practically all adults.
1. Standing Committee on the Scientific Evaluation of Dietary Refer-
ence Intakes. Dietary reference intakes: calcium, phosphorus, mag-
nesium, vitamin D, and fluoride. Washington, DC: National Acad-
emy Press, 1997.
2. National Research Council. Recommended dietary allowances. 10th
ed. Washington, DC: National Academy Press, 1989:92–7.
3. Narang NK, Gupta RC, Jain MK, Aaronson K. Role of vitamin D in
pulmonary tuberculosis. J Assoc Physicians India 1984;32:185–6.
4. Adams JS, Lee G. Gains in bone mineral density with resolution of
vitamin D intoxication. Ann Intern Med 1997;127:203–6.
5. Marriott BM. Vitamin D supplementation: a word of caution. Ann
Intern Med 1997;127:231–3.
6. Yates AA. Process and development of dietary reference intakes:
basis, need, and application of recommended dietary allowances.
Nutr Rev 1998;56:S5–9.
7. Gallagher JC, Kinyamu HK, Fowler SE, Dawson-Hughes B,
Dalsky GP, Sherman SS. Calciotropic hormones and bone markers
in the elderly. J Bone Miner Res 1998;13:475–82.
8. Harris SS, Dawson-Hughes B. Seasonal changes in plasma 25-
hydroxyvitamin D concentrations of young American black and
white women. Am J Clin Nutr 1998;67:1232–6.
9. Chapuy MC, Preziosi P, Maamer M, et al. Prevalence of vitamin D
insufficiency in an adult normal population. Osteoporos Int
10. Liu BA, Gordon M, Labranche JM, Murray TM, Vieth R, Shear NH.
Seasonal prevalence of vitamin D deficiency in institutionalized
older adults. J Am Geriatr Soc 1997;45:598–603.
11. Thomas MK, Lloyd-Jones DM, Thadhani RI, et al. Hypovita-
minosis D in medical inpatients. N Engl J Med 1998;338:777–83.
12. Perry HM, Bernard M, Horowitz M, et al. The effect of aging on
bone mineral metabolism and bone mass in Native American
women. J Am Geriatr Soc 1998;46:1418–22.
13. Peacock M. Effects of calcium and vitamin D insufficiency on the
skeleton. Osteoporos Int 1998;8(suppl):S45–51.
14. McKenna MJ, Freaney R. Secondary hyperparathyroidism in the
elderly: means to defining hypovitaminosis D. Osteoporos Int 1998;
15. Heaney RP. Vitamin D: how much do we need, and how much is too
much? Osteoporos Int 2000;11:553–5.
16. Vieth R. Vitamin D supplementation, 25-hydroxyvitamin D concen-
trations, and safety. Am J Clin Nutr 1999;69:842–56.
17. Arthur RS, Piraino B, Candib D, Cooperstein L, Chen T, West CPJ.
Effect of low-dose calcitriol and calcium therapy on bone histomor-
phometry and urinary calcium excretion in osteopenic women.
Miner Electrolyte Metab 1990;16:385–90.
18. Tjellesen L, Hummer L, Christiansen C, Rodbro P. Serum concentra-
tion of vitamin D metabolites during treatment with vitamin D2and
D3in normal premenopausal women. Bone Miner 1986;1:407–13.
19. Barger-Lux MJ, Heaney RP, Dowell S, Chen TC, Holick MF. Vita-
min D and its major metabolites: serum levels after graded oral dos-
ing in healthy men. Osteoporos Int 1998;8:222–30.
20. Trang H, Cole DE, Rubin LA, Pierratos A, Siu S, Vieth R. Evidence
that vitamin D3increases serum 25-hydroxyvitamin D more effi-
ciently than does vitamin D2. Am J Clin Nutr 1998;68:854–48.
21. Gokce C, Gokce O, Baydinc C, et al. Use of random urine samples
to estimate total urinary calcium and phosphate excretion. Arch
Intern Med 1991;151:1587–8.
22. Cleveland WS. Robust locally weighted regression and smoothing
scatterplots. J Am Stat Assoc 1979;74:829–36.
23. Colton T. Regression and correlation. In: Colton T, ed. Statistics in
medicine. Boston: Little Brown and Company, 1974:189–216.
24. Holick MF. Environmental factors that influence the cutaneous pro-
duction of vitamin D. Am J Clin Nutr 1995;61(suppl):638S–45S.
25. McHenry CR, Rosen IB, Walfish PG, Pollard A. Oral calcium load
test: diagnostic and physiologic implications in hyperparathyroidism.
26. Guillemant J, Guillemant S. Effects on calcium and phosphate
metabolism and on parathyroid function of acute administration of
tricalcium phosphate. Bone 1991;12:383–6.
SAFETY OF VITAMIN D3EXCEEDING THE LOAEL 293
27. Siwach SB, Kalra OP, Sharma R, Singh V, Chopra JS. Estimation of Download full-text
24 hour protein excretion from single random urine specimen. Indian
J Med Res 1990;92:105–8.
28. Strohmaier WL, Hoelz KJ, Bichler KH. Spot urine samples for the
metabolic evaluation of urolithiasis patients. Eur Urol 1997;32:
29. Lavocat MP, Freycon MT, Muchrif M. Comparative study of 24-hour
calciuria and urinary calcium/creatinine ratio in children over 4 years
of age. Pediatrie (Bucur) 1992;47:565–8.
30. Matsushita K, Tanikawa K. Significance of the calcium to creatinine
concentration ratio of a single-voided urine specimen in patients
with hypercalciuric urolithiasis. Tokai J Exp Clin Med 1987;12:
31. Lakatos P, Takacs I, Buki B, Nemeth J, Horvath C. Urinary calcium
excretion. Normal values for urinary calcium/creatinine ratio in
Hungary. Multicenter study. Orv Hetil 1997;138:1405–9.
32. Ozcan T, Kaleli B, Ozeren M, Turan C, Zorlu G. Urinary calcium to
creatinine ratio for predicting preeclampsia. Am J Perinatol 1995;
33. Riess C, Hess B, Binswanger U. Questionable significance of the
chemical analysis of a single 24-hour urine sample in recurrent cal-
cium oxalate nephrolithiasis. Klin Wochenschr 1986;64:411–6.
34. Pak CY, Kaplan R, Bone H, Townsend J, Waters O. A simple test for
the diagnosis of absorptive, resorptive and renal hypercalciurias.
N Engl J Med 1975;292:497–500.
35. Ulmann A. Predictive value of lithogenic risk in hypercalciuria: should
24-hour urine calcium be measured? Nephrologie 1984;5:232–4.
36. Sterkel BB. Bone density and vitamin D intoxication. Ann Intern
Med 1998;128:507 (letter).
37. Adams JS. Bone density and vitamin D intoxication. Ann Intern
Med 1998;128:508 (letter).
38. Bicknell F, Prescott F. Vitamin D. The antirachitic or calcifying
vitamin. In: Bicknell F, Prescott F, eds. Vitamins in medicine. Lon-
don: Whitefriars Press, 1946:630–707.
39. Morgan AF. A comparison of the hypervitaminosis induced by irra-
diated ergosterol and fish liver oil concentrates. J Biol Chem
40. Becks H. Dangerous effects of vitamin D overdosage on dental and
paradental structures. J Am Dent Assoc 1942;29:1947–52.
41. Jones G, Byrnes B, Palma F, Segev D, Mazur Y. Displacement
potency of vitamin D2analogs in competitive protein-binding assays
for 25-hydroxyvitamin D3, 24,25-dihydroxyvitamin D3, and
1,25-dihydroxyvitamin D3. J Clin Endocrinol Metab 1980;50:773–5.
42. Hollis BW, Frank NE. Quantitation of vitamin D2, vitamin D3,
25-hydroxyvitamin D2, and 25-hydroxyvitamin D3in human milk.
Methods Enzymol 1986;123:167–76.
43. Nilsson SF, Ostberg L, Peterson PA. Binding of vitamin D to its
human carrier plasma protein. Biochem Biophys Res Commun 1972;
44. Mawer EB, Jones G, Davies M, et al. Unique 24-hydroxylated
metabolites represent a significant pathway of metabolism of vita-
min D2in humans: 24-hydroxyvitamin D2and 1,24-dihydroxyvita-
min D2detectable in human serum. J Clin Endocrinol Metab 1998;
45. Heaney RP. Lessons for nutritional science from vitamin D. Am J
Clin Nutr 1999;69:825–6.
46. Gloth FM, Tobin JD, Sherman SS, Hollis BW. Is the recommended
daily allowance for vitamin D too low for the homebound elderly?
J Am Geriatr Soc 1991;39:137–41.
47. Hanly JG, McKenna MJ, Quigley C, Freaney R, Muldowney FP,
FitzGerald MX. Hypovitaminosis D and response to supplementa-
tion in older patients with cystic fibrosis. Q J Med 1985;56:377–85.
294 VIETH ET AL