ORIGINAL
ARTICLE
E n d o c r i n e
C a r e
Long-Term Dietary Vitamin D Intake and Risk of
Fracture and Osteoporosis: A Longitudinal Cohort
Study of Swedish Middle-aged and Elderly Women
Greta Snellman, Liisa Byberg, Eva Warensjö Lemming, Håkan Melhus,
Rolf Gedeborg, Hans Mallmin, Alicja Wolk, and Karl Michaëlsson
Department of Surgical Sciences (G.S., L.B., E.W.L., K.M., H.Ma.), Section of Orthopedics; Department of
Medical Sciences (H.Me.), Section of Clinical Pharmacology; and Department of Surgical Sciences (R.G.),
Section of Anesthesiology and Intensive Care, Uppsala University, SE-751 85 Uppsala, Sweden; and
Department of Nutritional Epidemiology (A.W.), Institute of Environmental Medicine, Karolinska
Institutet, SE-171 77 Stockholm, Sweden
Context: The importance of dietary vitamin D for osteoporotic fracture prevention is uncertain.
Objective: Our objective was to investigate associations between dietary vitamin D intake with risk
of fracture and osteoporosis.
Design and Participants: In the population-based Swedish Mammography Cohort (including
61 433 women followed for 19 years), diet was assessed by repeated food frequency
questionnaires.
Setting: The study was conducted in 2 municipalities in central Sweden.
Main Outcome Measure: Incident fractures were identified from registry data. In a subcohort (n ⫽
5022), bone mineral density was determined by dual-energy x-ray absorptiometry and serum
25-hydroxyvitamin D was measured using HPLC-tandem mass spectrometry.
Results: A total of 14 738 women experienced any type of first fracture during follow-up, and 3871
had a hip fracture. Multivariable-adjusted hazard ratio (HR) for any first fracture was 0.96 (95%
confidence interval, 0.92–1.01) for the lowest (mean, 3.1 ␮g/d) and 1.02 (0.96 –1.07) for the highest
(mean, 6.9 ␮g/d) quintile compared with the third quintile of vitamin D intake. The corresponding
HR for a first hip fracture was 1.02 (0.96 –1.08) for the lowest and 1.14 (1.03–1.26) for the highest
quintile. Intakes ⬎10 ␮g/d, compared with ⬍5 ␮g/d, conferred an HR of 1.02 (0.92–1.13) for any
fracture and an HR of 1.27 (1.03–1.57) for hip fracture. The intake of vitamin D did not affect the
odds for osteoporosis, although higher levels were associated with higher bone mineral density
(0.3%–2%, P ⬍ .0001). A positive association was observed between vitamin D intake and serum
25-hydroxyvitamin D.
Conclusions: Dietary intakes of vitamin D seem of minor importance for the occurrence of fractures
and osteoporosis in community-dwelling Swedish women. (J Clin Endocrinol Metab 99: 781–790,
2014)
A
large proportion of women and men will have 1 or
more osteoporotic fractures during their lifetime
(1). Moreover, the worldwide incidence of hip fractures is
expected to increase almost 4-fold by 2050 (2). Therefore,
fracture prevention is essential. Vitamin D is imperative
for bone mineralization in that it enables absorption of
dietary calcium and has a negative feedback on PTH secretion. Vitamin D insufficiency may therefore contribute
ISSN Print 0021-972X ISSN Online 1945-7197
Printed in U.S.A.
Copyright © 2014 by the Endocrine Society
Received March 23, 2013. Accepted December 3, 2013.
Abbreviations: BMD, bone mineral density; BMI, body mass index; CI, confidence interval;
FFQ, food frequency and lifestyle questionnaire; HR, hazard ratio; NHS, Nurses Health
Study; OR, odds ratio; S-25(OH)D, 25-hydroxyvitamin D; SMC, Swedish Mammography
Cohort; SMCC, SMC Clinical.
doi: 10.1210/jc.2013-1738
J Clin Endocrinol Metab, March 2014, 99(3):781–790
jcem.endojournals.org
781
782
Snellman et al
Dietary Vitamin D Intake and Bone Health
to calcium loss from bone with osteoporosis and fractures
as a final consequence (3).
Vitamin D status is primarily determined by genetic
constitution and UV-B radiation, although the dietary intake of vitamin D also contributes to circulating vitamin D
levels, but the influence seems to be of modest importance
(4 – 6). Therefore, weak associations between dietary vitamin D intakes and future fracture risk are to be expected.
The significance and optimal level of dietary vitamin D
intake for the prevention of osteoporosis and fractures
have, nevertheless, been much debated and remain unclear
(7, 8). This uncertainty is reflected by the wide range of
official and nonofficial intake recommendations for middle-aged and older individuals: from 5 ␮g vitamin D/d to
50 ␮g/d (9 –11).
Recommendations regarding vitamin D intake based
on the results from clinical trials and previous cohort studies are challenging. Meta-analyses of randomized trials
have found that supplemental vitamin D alone in doses of
10 to 20 ␮g/d provides no reduction in fracture rate (12,
13). Randomized controlled trials on fracture prevention
with vitamin D in combination with calcium supplementation in community-dwelling women found no certain
beneficial effects, although there is evidence that the combination can prevent fractures in old, frail women (8, 14).
To what extent long-term dietary vitamin D intake promotes bone health is not well understood. Prospective observational studies are few, and those that have been done
have yielded conflicting results (15–17). Against this background, we aimed to investigate associations between
long-term dietary intake of vitamin D with risk of fracture
of any type, with hip fractures, with osteoporosis, and
finally with serum 25-hydroxyvitamin D (S-25(OH)D)
levels in a large population-based prospective study of
Swedish middle-aged and elderly women.
J Clin Endocrinol Metab, March 2014, 99(3):781–790
Baseline invitees (n=90 303)
FFQ1, 1987-1990
Non-responders (n=23 652)
Responders (n=66 651)
Exclusions* (n=5218)
Baseline sample (n=66 433)
1997 invitees (n=56 030)
FFQ2, 1997
Non-responders (n=16 803)
Responders (n=39 227)
xcluded if energy intake was implausible (n=243)
1997 sample (n=38 984)
hort FFQ3, 2003-2009 (n=5 022)
Figure 1. Flow chart of the participants in the SMC. *, Excluded were
persons with an erroneous personal identification number, date of
moving out of the study area, or death, questionnaires not properly
dated, implausible energy intake (⫾3 SD from the mean value of the
log-transformed energy intake), and a cancer diagnosis (except
nonmelanoma skin cancer) before baseline.
Fracture identification
Fracture events were collated through linkage with the Swedish National Patient Registry. Data on outpatient-treated fractures were identified from outpatient registers. A complete deterministic record linkage was achieved by use of the unique
personal identification number assigned to all Swedish permanent residents. Any first fracture event after cohort entry was
defined as a hospital admission or an outpatient visit with an
International Classification of Diseases (ICD-10) diagnosis code
of S12, S22, S32, S42, S52, S62, S72, S82, or S92. Hip fracture
cases were defined by the codes S720, S721, and S722. Incident
fracture admissions were separated from readmissions from a
previous fracture event by the use of a validated method (19).
Repeated fractures were also assessed.
The SMC Clinical (SMCC)
Subjects and Methods
The Swedish Mammography Cohort
The Swedish Mammography Cohort (SMC) is a populationbased cohort in central Sweden (latitude, 60° N). All women
born between 1914 and 1948 living in Uppsala County (n ⫽
48 517) and all women born between 1917 and 1948 living in
Västmanland County (n ⫽ 41 786) were asked to respond to a
comprehensive 6-page food frequency and lifestyle questionnaire (FFQ) when invited to a mammography screening (March
1987 to December 1990). Completed questionnaires were obtained from 66 651 (74%) women in the source population.
After exclusions, 61 433 women remained in the cohort (18). In
1997, a second expanded questionnaire was sent to all participants who still lived in the study area (n ⫽ 56 030). Of these,
38 984 (70%) returned the second questionnaire (Figure 1).
In the period 2003 to 2009, a randomly selected subgroup
from the SMC was invited to participate in a physical examination, and 5022 women (65%) accepted to participate. A third
FFQ was answered 1 to 3 months before the clinical examination. Fasting blood samples were collected at the research visit.
Bone mineral density (BMD) at the hip, the lumbar spine (L1–
L4), and of the total body was also measured using dual-energy
x-ray absorptiometry (Lunar Prodigy; Lunar Corp). Osteoporosis was defined as a BMD of 2.5 SD or more below the mean
of U.S. white female reference populations, aged 20 to 40 years
(20). The short-term precision measurement error, based on duplicate measurements with repositioning according to recommendations from the International Society for Clinical Densitometry, varied between 0.8% and 1.5% depending on the site.
The long-term coefficient of variation was ⬍1% for a spine
phantom (21).
doi: 10.1210/jc.2013-1738
Determination of S-25(OH)D with
HPLC-atmospheric pressure chemical ionization
(APCI)-tandem mass spectrometry
In 2012, determination of 25(OH)D2, 3-epi-25(OH)D3 and
25(OH)D3 in plasma with HPLC-APCI-MS/MS was done at
Vitas, Oslo, Norway (www.vitas.no). Samples from 5019 of
5022 women in the SMC Clinical (SMCC) were analyzed. The
samples had been stored at ⫺80°C protected from light and were
not thawed before analysis. The coefficient of variation for
25(OH)D was 9.7% (36.94nM) and 10.2% (48.08nM) using
plasma quality control samples at 2 levels analyzed in series with
study samples.
Description of the FFQ
Nutrient intake was estimated by multiplying the frequency
of consumption of each food item by the nutrient content of
age-specific portion sizes and by use of data from the Swedish
National Food Administration database. All nutrients were adjusted for total energy intake (mean 1700 kcal in the study population) using the residual method (22). In the second and third
FFQ (the 1997 expanded questionnaire and the questionnaire
completed by the SMCC subcohort), the lifetime use of dietary
supplements was reported. Total vitamin D intake included supplemental vitamin D. Reported frequency of supplement use
within the cohort during the first years of follow-up was ⬍5%
(23). Low-fat dairy products and margarine are fortified with
vitamin D and, together with fatty fish, they constitute the major
food sources of vitamin D in Sweden. Validation of vitamin D
intake from the FFQ was carried out with four 7-day food records every third month in 104 of the women (r ⫽ 0.72). There
are only small systematic errors related to intake level between
the methods with an average difference of only 0.11 ␮g higher
vitamin D values with the food records (95% confidence interval
(CI), ⫺0.15 to 0.38), an indication of good accuracy for the FFQ.
Additional information
The questionnaires provided lifestyle information, including
use of postmenopausal estrogen therapy, menopausal status,
parity information, medication use, weight and height, smoking
habits, and leisure-time physical activity during the past year
(with 5 predefined levels ranging from 1 h/wk to ⬎5 h/wk).
Physical activity collected in the 1997 questionnaire is valid
when compared with activity records and accelerometer data
(24). The educational level was categorized as ⱕ9 years, 10 to 12
years, ⬎12 years, and other education (such as vocational). ICD
diagnosis codes were collated from the Swedish National Patient
Registry to calculate Charlson’s comorbidity index (25).
Statistical analyses
The present study was made up of 2 study samples: the SMC
with the primary outcomes any first fracture and first hip fracture
and the SMCC with the secondary outcomes osteoporosis,
BMD, and S-25(OH)D levels.
For each woman in the SMC, person-years of follow-up were
calculated from the date they entered the cohort until the date of
fracture, date of death, the date of emigration, or the end of the
follow-up period (December 31, 2008), whichever occurred
first. We estimated age-adjusted and multivariable-adjusted hazard ratios (HRs) of fracture by Cox proportional hazard regression and odds ratios (ORs) of osteoporosis by logistic regression.
jcem.endojournals.org
783
To better account for changes in diet during follow-up and to
better represent long-term dietary intake, vitamin D intake and
the covariates were treated as time-updated variables in the Cox
models and as cumulative average intakes in the logistic regression models (26). The proportional hazard assumptions in the
Cox models were confirmed graphically by comparing NelsonAalen plots and by Schoenfeld’s test. We estimated associations
for dietary and for total vitamin D intake, including both diet and
supplement use. In addition, we examined the relation between
quintiles of vitamin D intake and risk of fracture and osteoporosis. To facilitate comparisons of the estimates, the same quintile cutoffs were used in the SMCC as in the entire cohort. We
further analyzed the rate of fracture after stratification by calcium intake (⬍800 mg/d vs ⱖ800 mg/d). To obtain a clearer view
of the shape of the association between vitamin D intake and
fracture risk, we used restricted cubic spline models, with 4 knots
located at percentiles 5, 35, 65, and 95 of vitamin D intake. This
approach allowed us to obtain a smoothed dose-risk curve with
95% CI.
In addition, we contrasted women who consistently reported
lowest quintile of vitamin D intake to women who reported a
constant intake within the highest quintile for all outcomes. Finally, we broadened our categorical exposure range into 4 vitamin D intake categories (⬍5.0, 5.0 –7.5, 7.5–10, and ⬎10 ␮g/d),
which reflects different current dietary recommendations, and
also into 5 categories (⬍3.5, 3.5– 6, 6 –9, 9 –12.5, and ⬎12.5
␮g/d) as previously described (15).
In the multivariable model, we included age; total energy intake; alcohol, retinol, protein, potassium, and calcium intake;
body mass index (BMI); height (all continuous); education level
(ⱕ9, 12, ⬎12 years, and other); nulliparity (yes or no); vitamin
D supplement use (yes or no); calcium supplements (yes or no);
physical activity (5 categories); smoking status (never, former, or
current); previous fracture before study start (yes or no); and
Charlson’s comorbidity index (continuous, 1–16). In the analysis of SMCC, we also included estrogen, bisphosphonate, and
cortisone use as separate marker variables in the multivariable
model. Covariates not assessed in the baseline FFQ (such as
smoking habits and physical activity) were imputed by the
Markov chain Monte Carlo multiple imputation procedure. Sensitivity analysis with restriction to nonmissing data did not alter
our interpretation of the results.
Given that both low and high vitamin D intake might be
related to prevalent disease and mortality, we compared cumulative incidence curves with Kaplan-Meier failure curves to address the potential competing risk problem from mortality (27).
The curves indicated that our results would not be compromised
by competing risk from mortality (Supplemental Figure 1, published on The Endocrine Society’s Journals Online website at
http://jcem.endojournals.org).
The association between total S-25(OH)D (the sum of
S-25(OH)D2 and S-25(OH)D3 although not including 3-epi25(OH)D3) and vitamin D intake from diet and supplements was
analyzed by linear regression analyses, adjusting for age, weight,
height, and season). The statistical analyses were performed with
STATA version 11 (StataCorp).
Ethical approval
The study was approved by the regional ethical review board
in Stockholm, Sweden, and all participants gave their informed
consent.
784
Snellman et al
Dietary Vitamin D Intake and Bone Health
Results
The basic characteristics of the study participants by quintiles
of dietary vitamin D intake are presented in Table 1. With
increasing intake of vitamin D, the average age and BMI of
the women as well as intake of calcium and retinol were
higher, whereas the intake of alcohol was lower. However,
Table 1.
Intakea
J Clin Endocrinol Metab, March 2014, 99(3):781–790
level of education, physical activity, and smoking status were
similar across categories of vitamin D intake.
The women were followed for a median time of 19.2
years, which contributed to 996 800 person-years at risk.
During this time, 14 738 women suffered any type of first
fracture, of which 5043 had had 1 or more additional
fractures during follow-up. In addition, 3871 of the
Characteristics of the Total Cohort (SMC) and the Subcohort (SMCC) by Quintiles of Dietary Vitamin D
Quintile
Number of participants
Vitamin D intake, ␮g/d
SMC
Age at entry, y
BMI at entry, kg/m2)
Dietary vitamin D intake, ␮g/d
Total vitamin D intake,b ␮g/d
Vitamin D supplement use n (%)c,d
Dietary calcium intake, mg/d
Total calcium intake,b mg/d
Energy intake, kcal/d
Dietary retinol intake, ␮g/d
Alcohol intake, g/d
Nulliparity, n (%)
Leisure time PA level, n (%)d
1 (lowest)
2
3
4
5 (highest)
Smoking status, n (%)d
Current
Former
Never
Education level, n (%)e
ⱕ9 y
9 –12 y
⬎12 y
Other
SMCC
Age at DXA investigation, y
BMI, kg/m2
Energy intake, kcal/d
Vitamin D intake, ␮g/d
Vitamin D supplement use, n (%)c,d
Total vitamin D intake,b ␮g/d
Calcium intake, mg/d
Supplemental calcium intake, mg/d (n ⫽ 610)
Total calcium intake,b mg/d
Retinol intake, ␮g/d
Alcohol intake, g/d
Q1
Q2
Q3
Q4
Q5
12 286
⬍3.3
12 287
3.3– 4.0
12 287
4.0 – 4.7
12 282
4.7–5.4
12 291
⬎5.4
56.1 (10.2)
24.4 (4.3)
2.6 (0.5)
3.1 (1.7)
1622 (13)
919 (275)
969 (413)
1642 (531)
267 (467)
3.4 (5.2)
2245 (11)
55.9 (10.2)
24.6 (4.2)
3.7 (0.2)
4.1 (1.8)
1828 (15)
907 (255)
956 (394)
1661 (485)
319 (531)
3.4 (4.5)
2061 (10)
56.5 (10.3)
24.7 (3.8)
4.3 (0.2)
4.8 (1.6)
1730 (14)
927 (260)
974 (390)
1646 (475)
362 (601)
3.3 (4.3)
2006 (10)
57.2 (10.3)
25.1 (4.1)
5.0 (0.2)
5.4 (1.5)
1816 (15)
983 (259)
1025 (364)
1640 (470)
359 (648)
2.9 (3.8)
1914 (10)
59.0 (10.7)
25.7 (4.6)
6.5 (1.5)
6.9 (2.2)
1604 (13)
1067 (320)
1109 (417)
1609 (614)
421 (752)
2.6 (3.8)
2239 (11)
392 (20)
4806 (24)
6336 (32)
2773 (14)
1971 (10)
3844 (19)
5128 (26)
6483 (32)
2698 (13)
1893 (10)
3837 (19)
5209 (25)
667 (33)
2728 (13)
2012 (10)
362 (19)
4952 (26)
6326 (33)
2637 (14)
1831 (8)
4072 (20)
5115 (25)
6646 (32)
2845 (14)
2065 (9)
4245 (21)
5425 (27)
10 136 (51)
4113 (21)
5593 (28)
1034 (52)
4178 (20)
5583 (27)
10 695 (52)
3902 (20)
5131 (26)
10 333 (53)
4590 (22)
5674 (27)
10 479 (51)
14 976 (76)
1579 (8)
2267 (11)
984 (5)
15 232 (76)
1589 (8)
2195 (11)
103 (5)
15 808 (77)
155 (8)
2095 (10)
1003 (5)
15 215 (79)
1400 (7)
1722 (9)
1029 (5)
16 657 (80)
1328 (6)
1737 (8)
1021 (5)
66.5 (6.4)
23.6 (3.2)
1698 (442)
2.8 (0.5)
112 (13.7)
4.9 (2.2)
976 (209)
348 (187)
1033 (233)
409 (265)
6.0 (5.7)
66.5 (6.5)
24.0 (3.5)
1735 (398)
3.7 (0.2)
68 (13.9)
4.0 (2.6)
991 (188)
360 (185)
1028 (248)
449 (262)
5.4 (4.7)
67.0 (6.5)
24.3 (3.2)
1714 (385)
4.3 (0.2)
164 (14.2)
5.6 (2.3)
982 (184)
362 (185)
1028 (228)
482 (239)
5.5 (4.8)
67.6 (6.7)
24.9 (3.4)
1722 (388)
5.0 (0.2)
228 (16.9)
6.3 (3.0)
1028 (204)
382 (180)
1068 (244)
539 (252)
5.1 (4.4)
69.5 (7.1)
25.4 (3.8)
1692 (377)
6.2 (0.8)
201 (16.6)
7.4 (2.4)
1064 (223)
381 (179)
1108 (262)
590 (342)
4.8 (4.7)
Abbreviations: DXA, dual-energy x-ray absorptiometry; PA, physical activity.
a
Data are shown as mean ⫾ SD or as otherwise indicated. Intake per day refers to the energy adjusted intake in the subcohort (SMCC).
b
Total intake is the sum of dietary and supplemental intakes.
c
Supplemental vitamin D alone in combination with calcium or as multivitamins.
d
Information only available in the 1997 questionnaire and in the DXA questionnaire.
e
Educational level: ’Other’ refers to vocational or other education.
doi: 10.1210/jc.2013-1738
jcem.endojournals.org
785
Table 2. HRs With 95% CIs for the First of Any Type of Fracture and Hip Fracture by Quintiles of Vitamin D Intake
and ORs With 95% CIs for Osteoporosis in the Subcohort SMCC by Quintiles of Vitamin Da
Vitamin D intake, ␮g/d
First event any fracture
Dietary intake of vitamin D
Number of fractures
Person-years at risk
Rate per 1000 person years
Age-adjusted HR (95% CI)
Multivariable-adjusted HR (95% CI)b
Total intake of vitamin D
Number of fractures
Person-years at risk
Rate per 1000 person years
Age-adjusted HR (95% CI)
Multivariable-adjusted HR (95% CI)b
First event of hip fracture
Dietary intake of vitamin D
Number of fractures
Person-years at risk
Rate per 1000 person years
Age-adjusted HR (95% CI)
Multivariable-adjusted HR (95% CI)b
Total intake of vitamin D
Number of fractures
Person-years at risk
Rate per 1000 person years
Age-adjusted HR (95% CI)
Multivariable-adjusted HR (95% CI)b
Osteoporosis
Dietary intake of vitamin D
Number of women (%)
Women without osteoporosis n (%)
Women with osteoporosis, n (%)
Age-adjusted OR (95% CI)
Multivariable-adjusted OR (95% CI)c
Total intake of vitamin D
Number of women (%)
Women without osteoporosis n (%)
Women with osteoporosis n (%)
Age-adjusted OR (95% CI)
Multivariable-adjusted OR (95% CI)c
a
Quintile 1
Quintile 2
Quintile 3
Quintile 4
Quintile 5
⬍3.3
3.3– 4.0
4.0 – 4.7
4.7–5.4
⬎5.4
2884
212 880
13.5 (13.1–14.1)
1.01 (0.96 –1.07)
0.96 (0.92–1.01)
2813
212 240
13.3 (12.8 –13.8)
0.99 (0.94 –1.05)
0.98 (0.93–1.03)
2952
213 360
13.8 (13.3–14.3)
1 (reference)
1 (reference)
2855
202 290
14.1 (13.6 –14.6)
1.01 (0.96 –1.06)
1.01 (0.96 –1.07)
3234
210 510
15.4 (14.8 –15.9)
1.00 (0.95–1.05)
1.02 (0.96 –1.07)
2630
197 560
13.3 (12.8 –13.8)
1.02 (0.97–1.08)
0.97 (0.92–1.03)
2528
197 120
12.8 (12.3–13.3)
0.99 (0.93–1.04)
0.97 (0.92–1.02)
2652
197 700
13.4 (12.9 –13.9)
1 (reference)
1 (reference)
2598
189 800
13.7 (13.2–14.2)
1.00 (0.95–1.06)
1.01 (0.95–1.06)
4330
269 100
16.1 (15.6 –16.6)
0.94 (0.90 – 0.99)
1.01 (0.96 –1.07)
734
228 580
3.2 (3.0 –3.5)
1.10 (0.99 –1.21)
1.02 (0.96 –1.08)
686
227 490
3.0 (2.8 –3.2)
1.03 (0.93–1.14)
0.94 (0.93– 0.96)
723
230 000
3.1 (2.9 –3.4)
1 (reference)
1 (reference)
774
217 700
3.6 (3.3–3.8)
1.09 (0.98 –1.20)
1.10 (0.99 –1.21)
954
227 830
4.2 (3.9 – 4.5)
1.07 (0.97–1.18)
1.14 (1.03–1.26)
682
211 440
3.2 (3.0 –3.5)
1.06 (0.96 –1.18)
0.94 (0.84 –1.04)
631
210 240
3.0 (2.8 –3.2)
0.99 (0.89 –1.10)
0.93 (0.84 –1.04)
685
212 240
3.2 (3.0 –3.5)
1 (reference)
1 (reference)
718
203 590
3.5 (3.3–3.8)
1.04 (0.94 –1.15)
1.06 (0.95–1.18)
1155
294 090
3.9 (3.7– 4.2)
0.85 (0.77– 0.93)
1.09 (0.99 –1.21)
395 (7.9)
308 (78.0)
87 (22.0)
1.27 (0.95–1.69)
1.15 (0.85–1.55)
819 (16.3)
641 (78.3)
178 (21.7)
1.24 (0.99 –1.55)
1.13 (0.89 –1.44)
1277 (25.4)
1033 (80.9)
244 (19.1)
1 (reference)
1 (reference)
1421 (28.3)
1146 (80.7)
275 (19.4)
0.96 (0.79 –1.17)
1.10 (0.89 –1.35)
1110 (22.1)
882 (79.5)
228 (20.5)
0.87 (0.70 –1.07)
1.08 (0.85–1.37)
280 (5.6)
213 (76.1)
67 (23.9)
1.37 (0.98 –1.90)
1.20 (0.85–1.71)
568 (11.3)
437 (76.9)
131 (23.1)
1.36 (1.04 –1.76)
1.21 (0.92–1.60)
929 (18.5)
754 (81.2)
175 (18.8)
1 (reference)
1 (reference)
1105 (22.0)
902 (81.6)
203 (18.4)
0.94 (0.75–1.19)
1.03 (0.81–1.32)
2140 (42.6)
1704 (79.6)
436 (20.4)
0.95 (0.78 –1.16)
0.99 (0.78 –1.25)
Energy-adjusted mean nutrient data were estimated with data from the baseline and the 1997 questionnaire.
b
The multivariable-adjusted HRs were adjusted for age; BMI; height; total energy intake; retinol, potassium, protein, alcohol, and calcium intake;
calcium and vitamin D supplementation; nulliparity; educational and physical activity level; smoking status; previous fracture of any type before
baseline; and Charlson’s comorbidity index.
c
Osteoporosis was defined in subjects when the T-score, determined at the total hip, femoral neck, or lumbar spine, was ⱕ⫺2.5 SD below the
mean of a young adult reference range. The multivariable-adjusted HRs were adjusted for age; BMI; height; total energy intake; retinol, potassium,
protein, alcohol, and calcium intake; calcium and vitamin D supplementation; nulliparity; educational and physical activity level; smoking status;
previous fracture of any type before baseline; Charlson’s comorbidity index; estrogen replacement therapy; and cortisone and bisphosphonate use.
women had a first hip fracture during follow-up and 1368
experienced multiple hip fractures. In the SMCC subcohort, 20% (n ⫽ 1012) of the women were classified as
osteoporotic.
Vitamin D intake and fractures
The HRs for a first fracture by quintiles of dietary and
total vitamin D intake are presented in Table 2. Despite a
more than 2-fold difference in median vitamin D intake
between lowest and highest quintiles, the rate of fracture
was not higher among women with the lowest intake levels. On the contrary, a somewhat higher rate for a first hip
fracture was discerned within the highest dietary quintile
of vitamin D intake (HR, 1.14; 95% CI, 1.03–1.26), an
observation confirmed when analyzing multiple hip fracture events (Supplemental Table 1). The risks with more
extreme intakes of vitamin D are displayed by the spline
curves illustrated in Figure 2, A and B, and by the categories shown in Supplemental Tables 2 and 3. In comparison with a vitamin D intake below 5 ␮g/d, women with an
intake higher than 10 ␮g/d did not have a lower rate of
fracture of any type (HR 1.02; 0.92–1.13), and they even
had a higher rate of hip fracture (HR, 1.27; 1.03–1.57).
Even more extreme intakes (⬍3.5 vs ⬎12.5 ␮g) did not
alter our results (Figure 3 and Supplemental Table 3).
Calcium intake (higher or less than 800 mg daily, Supplemental Table 4) did not modify the association between
Hazard ratio of any first fracture (adjusted)
786
Snellman et al
Dietary Vitamin D Intake and Bone Health
J Clin Endocrinol Metab, March 2014, 99(3):781–790
Dietary vitamin D intake (>10 µg vs <5 µg/day)
2.0
A
Hip fracture
1.5
Any fracture
1.0
Total vitamin D intake (>10 µg vs <5 µg/day)
Hip fracture
0.5
Any fracture
0.0
0
5
10
15
20
Total vitamin D intake (>12.5 µg vs <3.5 µg/day)
Dietary vitamin D intake (µg/day)
Hip fracture
Any fracture
Hazard ratio of first hip fracture (adjusted)
2.0
B
0
1
1.5
2
HR with 95% CI
1.5
Figure 3. HR (square) and 95% CI (error bar) for fracture comparing
extremes of dietary and total vitamin D intake. Adjustments were
made for age; BMI; height; total energy intake; retinol, potassium,
protein, alcohol, and calcium intake; calcium and vitamin D
supplementation; nulliparity; educational and physical activity level;
smoking status; previous fracture of any type before baseline; and
Charlson’s comorbidity index.
1.0
0.5
0.0
0
5
10
15
20
Dietary vitamin D intake (µg/day)
6.0
Odds ratio of osteoporosis (adjusted)
0.5
C
5.0
4.0
3.0
2.0
1.0
0.0
0
5
10
15
20
Dietary vitamin D intake (µg/day)
Figure 2. Multivariable-adjusted spline curves for the relation
between dietary intake of vitamin D and time to the first of any type of
fracture (A), a first hip fracture (B), or osteoporosis (C). The HR is
indicated by the solid line and the 95% CI by dashed lines.
Adjustments were made for age; BMI; height; total energy intake;
retinol, potassium, protein, alcohol, and calcium intake; calcium and
vitamin D supplementation; nulliparity; educational and physical
activity level; smoking status; previous fracture of any type before
baseline; and Charlson’s comorbidity index. In C, we also included use
of bisphosphonates, cortisone, and estrogen replacement therapy in
the model. The reference was set at 4.3 ␮g, which is the mean dietary
vitamin D intake in the cohort.
low vitamin D intake and fracture risk, nor did phosphorus intake (P values for interaction ⬎ .3). Moreover, the
women who were in the lowest quintile of vitamin D intake at both dietary surveys did not have a higher rate of
any fracture or of hip fracture (Supplemental Table 5).
Vitamin D intake, osteoporosis, and BMD
Vitamin D intake in the SMCC was similar to the intake
in the entire cohort (Table 1). The risk for osteoporosis
was not correlated to vitamin D intake in quintiles (Table
2), although a trend toward a reduced risk with higher
intake could be detected in the spline analysis (Figure 2C).
Multivariable-adjusted BMD values, however, were 2%
higher at the lumbar spine and 0.3% higher at the total hip
in women with highest compared with lowest quintile intake of vitamin D (P ⬍ .0001, Figure 4), with mean differences in multivariable-adjusted BMD of 0.022 g/cm2
(95% CI, 0.021– 0.023) for the lumbar spine, 0.004 g/cm2
(95% CI, 0.004 – 0.005) for the total body, and 0.003
g/cm2 (95% CI, 0.003– 0.004) for the total hip. The differences in BMD when comparing women who were in the
lowest quintile of vitamin D intake at all 3 dietary surveys
with the women with a constant high intake were similar
to the estimates provided above. Thus, we found a multivariable-adjusted difference in BMD of 0.029 g/cm2
(95% CI, 0.025– 0.034) at the lumbar spine and 0.010
g/cm2 (95% CI, 0.008 – 0.012) of the total body, whereas
the average BMD at the total hip was slightly higher in
doi: 10.1210/jc.2013-1738
jcem.endojournals.org
787
50 nmol/L did not have a more prominent relation between vitamin D intake and BMD than women with serum
concentrations higher than 50 nmol/L, not even during the
dark season (Supplemental Figure 2).
Discussion
Figure 4. Adjusted mean BMD at total body, lumbar spine, and total
hip by quintiles of vitamin D intake. The error bars indicate 95%
confidence intervals. Adjustments were made for age; BMI; height;
total energy intake; retinol, potassium, protein, alcohol, and calcium
intake; calcium and vitamin D supplementation; nulliparity; educational
and physical activity level; smoking status; previous fracture of any type
before baseline; and Charlson’s comorbidity index.
women with low vitamin D intake (difference: 0.005
g/cm2; 95% CI, 0.002– 0.008).
Vitamin D intake and S-25(OH)D
Dietary and supplemental vitamin D intake was positively associated with S-25(OH)D. The adjusted mean
S-25(OH)D level was 52.9 nmol/L (95% CI, 50.4 –55.4)
among the women with a dietary vitamin D intake less
than 2.5 ␮g (Supplemental Table 6). The corresponding
level among women with a dietary intake more than 10
␮g/d was 60.2 nmol/L (95% CI, 58.1– 62.4). On average,
S-25(OH)D increased 1.2 nmol/L (95% CI, 0.7–1.8) for
every 3 ␮g of dietary intake. Interestingly, the increase
was, however, nonlinear with the strongest association
between the dietary intake of vitamin D and S-25(OH)D
among women with a low dietary vitamin D intake (⬍10
␮g/d), with an increase of 2.5 nmol/L (95% CI, 1.6 –3.3)
per 3 ␮g intake. Increasing intakes among women with a
high dietary vitamin D intake (⬎10 ␮g/d) did not confer an
additional increase in serum levels (parameter estimate
⫺1.3 nmol/L (95% CI, ⫺2.7 to 0.1) per 3 ␮g intake. The
use of 10 ␮g vitamin D supplement daily was, however,
associated with 13.3 nmol/L (95% CI 10.7–13.8) higher
S-25(OH)D.
Finally, we considered whether vitamin D status
modified the association between vitamin D intake and
BMD. Women with S-25(OH)D concentrations below
The present findings suggest that long-term consumption
of vitamin D in the range normally consumed, either from
the diet or from diet and supplements, do not substantially
affect the risk of fracture or of osteoporosis in Swedish
middle-aged and elderly women. Although not reflected in
the fracture rate, women with high vitamin D intake
tended to have a slightly higher BMD and higher
S-25(OH)D levels.
Our results are not surprising in light of the weak associations found in randomized trials between vitamin D
and bone health (8, 12, 13). In addition, vitamin D supplementation without calcium has not been proven effective for the prevention of osteoporotic fractures, and supplementation with vitamin D in combination with calcium
has mainly a clear fracture preventive effect only in frail
elderly women living in nursing homes (13, 14, 28, 29).
Nonetheless, higher intake doses of vitamin D, not normally achieved even by food fortification of the diet or by
customary supplement doses used in randomized trials,
might be needed (30, 31).
On the other hand, the Nurses Health Study (NHS) (15)
reported a 43% risk reduction in hip fracture in women
with high dietary vitamin D intake (ⱖ6.25 vs ⬍2.50 ␮g/d).
In addition, a high total vitamin D intake (ⱖ12.50 vs
⬍3.50 ␮g/d) conferred a 37% risk reduction. We cannot
readily explain the differences in results between our study
and the NHS study. Both studies used repeated FFQs, although biannual measurements were used in the NHS
study. The number of fractures included in our population-based study was larger and was ascertained from hospital records, whereas fractures were self-reported in the
NHS study. Few women in our study had an intake level
above 12.5 ␮g/d, with relatively wide CIs as a consequence. Although Sweden is a country with limited sunlight in the winter season, analyses of serum vitamin D
concentrations in populations in Europe have shown that
the highest levels are found in Scandinavia, even during the
dark season of winter (32). Due to vitamin D food fortification of the diet, mean vitamin D intake in Scandinavia
has been estimated to be twice that in other European
countries (33), even though these international comparisons are hampered by methodological differences in the
collection of data. Theoretically, too few women in our
study, despite its large size, might have had a dietary
788
Snellman et al
Dietary Vitamin D Intake and Bone Health
vitamin D intake low enough to influence their risk of
fracture. Genetic predisposition might also explain international differences in S-25(OH)D (6). Indeed, the
S-25(OH)D values found in Sweden seem to be higher
than those found in North American populations (32),
even though the average dietary intake of vitamin D in our
study and NHS was similar, slightly higher than 4 ␮g/d. In
some of our analysis, we used exactly the same cut-offs as
in the NHS study without detecting an association with
fracture risk. Our results may reflect a situation when a
moderate intake of vitamin D combined with satisfactory
intake of other micronutrients together with a genetic predisposition to provide a sufficient vitamin D status is adequate to meet the structural and functional demands of
the skeleton. The results might therefore not apply to other
settings. Moreover, we noted with interest that among the
women in our study who had an adequate dietary vitamin D intake (⬎10 ␮g/d), higher intake levels did not
result in further increases in S-25(OH)D. On the contrary, use of vitamin D supplements (10 ␮g/d) was associated with higher S-25(OH)D. A high intake of dairy
products, in Sweden fortified with vitamin D, raises
serum levels of IGF-1 (34), which in turn can lead to a
downregulation of the endogenous production of vitamin D (35). Dairy intake is also associated with higher
phosphorous intake, and the main regulator of serum
phosphate, fibroblast growth factor 23, acts to suppress
renal phosphate reabsorption and also 1,25-dihydroxyvitamin D synthesis (36).
In accordance with our previous finding of calcium intake (18), we found a slightly higher rate of hip fracture
among women with high vitamin D intake. This might
theoretically be explained by a reverse causation, ie,
women with a higher predisposition for osteoporosis may
have deliberately increased their intake of vitamin D-rich
foods. We tried to avoid this bias by restricting our analysis to women with first fracture events. If present, this
bias would probably have also been reflected in a higher
rate of other types of fracture (not only hip fractures).
Furthermore, few participants had knowledge of their
BMD (which could have influenced dietary habits) because general screening of osteoporosis with bone scans
does not exist in Sweden. Our finding therefore remains
unexplained. Intriguingly, and also unexpectedly, however, an annual injection of vitamin D increased the risk of
fractures in 2 recent large, placebo-controlled randomized
studies (37, 38).
In the Institute of Medicine report (8), new dietary reference intakes for vitamin D and calcium are presented.
These recommendations are based not only on bone health
but also on the total scientific field for vitamin D and
health outcomes. The review panel’s conclusion is that the
J Clin Endocrinol Metab, March 2014, 99(3):781–790
estimated average requirement for older women and men
without any sun exposure should be 10 ␮g of vitamin D
intake per day and that the recommended dietary allowance should be 20 ␮g/d. Despite mandatory vitamin D
food fortification of low-fat milk products and margarine
in Sweden, less than 1% of our female population reached
an intake level higher than 10 ␮g/d. The vitamin D content
in fortified low-fat milk products in Sweden is 0.45 ␮g/100
g, and the content in margarine is 10 ␮g/100 g. In a recently
published reanalysis of randomized vitamin D supplementation trials, it was concluded that an intake of ⱖ800 IU
(20 ␮g/d) vitamin D daily is needed to prevent nonvertebral fractures in women and men aged 65 years or older.
The analysis was based on adherent therapies and the total
intake from supplements (39). The relatively low intake in
our study, despite food fortification, could therefore explain the lack of fracture association in our study. Only 3
women had a dietary intake ⱖ20 ␮g/d, and 123 women
had a total intake of ⱖ20 ␮g/d. To achieve intake levels of
20 ␮g/d or higher, a general recommendation of vitamin
D supplementation is needed or a substantial increase in
the amount of vitamin D fortified in Swedish foods.
Strengths of our study are the size of the cohort, the
population-based design, the unique number of fractures,
and the repeated and validated FFQs. Because of the
unique personal identification number of all Swedish residents, in combination with national healthcare registers,
we are capable of a complete fracture ascertainment. Our
results can be applied to Swedish middle-aged and elderly
women, a population with a high incidence of osteoporotic fractures. Our study has limitations that should be
considered when interpreting the results. First, one limitation common to observational studies is that they preclude conclusions regarding causality. Our results might
not apply to other people of different ethnic origins or to
men. Our estimates were adjusted for several important
covariates, but residual confounding still remains a possible limitation. Even though we considered only first fracture events, we cannot rule out the possibility of confounding by indication. Women with a family history of
fracture, which renders them a higher risk of fracture, even
without a previous fracture themselves, might have been
prone to take preventive measures by increasing their intake of foods containing vitamin D. Unfortunately, we
could not take family history of fracture into account in
the full cohort analysis. However, reanalysis of a previously published validation study (40) of the SMC revealed
that women who reported a family history of osteoporotic
fractures did not (P ⫽ .45) have a higher dietary vitamin
D intake (mean, 4.1 ␮g/d; 95% CI, 3.6 – 4.7) than women
without such information (mean, 4.4 ␮g/d; 95% CI, 4.1–
4.6). Moreover, based on reanalysis of another previously
doi: 10.1210/jc.2013-1738
published population-based cohort study (41), men with
a family history of osteoporotic fractures did not (P ⫽ .79)
have higher dietary vitamin D intakes (mean, 5.8 ␮g/d;
95% CI, 5.2– 6.4) compared with men without a family
history of osteoporotic fractures (mean, 5.8; 95% CI, 5.6 –
5.9). Therefore, we found it unlikely that a family history
of fracture confounded our results.
Dietary assessment methods are subject to a number of
limitations that affect both the precision and accuracy of
the measurement. Nevertheless, an FFQ is used to assess
the habitual intake of diet in larger studies, and a recent
review concluded that it is a valid method for assessing
dietary vitamin D intake (42), as also indicated by our
validation. The CIs of the estimates at the extremes of
vitamin D intake are wide, and these estimates should be
cautiously interpreted.
We conclude that vitamin D intake at levels customarily
used in the Swedish diet is not an important determinant
for fracture or osteoporosis in Swedish middle-aged and
elderly women. A high intake of vitamin D was associated
with a somewhat higher BMD in our study but not to the
extent that it affects the clinically important endpoint,
fractures.
Acknowledgments
Address all correspondence and requests for reprints to: Greta
Snellman, Department of Surgical Sciences, Section of Orthopedics, Uppsala University SE-751 85 Uppsala, Sweden. E-mail:
[email protected].
The Swedish Research Council supported this work through
Grants 2008-2202 and 2009-6281. The founding source was not
involved in the design, conduct, or interpretation of the study or
in the writing of the submitted work.
R.G. is employed at the Medical Products Agency in Sweden.
The views expressed in this paper do not necessarily represent the
views of the agency.
G.S., L.B., and K.M. designed the study, analyzed and interpreted the data, and drafted the manuscript. H.M. contributed to
the study design, analysis and interpretation of the data, and
writing of the manuscript. E.W., R.G., and A.W. interpreted the
data and made significant contributions to drafts of the manuscript. All authors had full access to all data (including statistical
reports and tables) in the study and take responsibility for the
integrity of the data and the accuracy of the data analysis. G.S.
and K.M. are guarantors. All authors read and approved the final
manuscript.
Disclosure Summary: None of the authors have any conflict
of interest or any financial interests to report.
References
1. Chrischilles EA, Butler CD, Davis CS, Wallace RB. A model of
lifetime osteoporosis impact. Arch Intern Med. 1991;151:2026 –
2032.
jcem.endojournals.org
789
2. Cummings SR, Melton LJ. Epidemiology and outcomes of osteoporotic fractures. Lancet. 2002;359:1761–1767.
3. Lips P. Vitamin D physiology. Prog Biophys Mol Biol. 2006;92:
4 – 8.
4. Burgaz A, Akesson A, Oster A, Michaëlsson K, Wolk A. Associations of diet, supplement use, and ultraviolet B radiation exposure
with vitamin D status in Swedish women during winter. Am J Clin
Nutr. 2007;86:1399 –1404.
5. Hunter D, De Lange M, Snieder H, et al. Genetic contribution to
bone metabolism, calcium excretion, and vitamin D and parathyroid
hormone regulation. J Bone Miner Res. 2001;16:371–378.
6. Snellman G, Melhus H, Gedeborg R, et al. Seasonal genetic influence
on serum 25-hydroxyvitamin D levels: a twin study. PLoS One.
2009;4:e7747.
7. Grey A, Bolland M. Vitamin D: a place in the sun? Arch Intern Med.
2010;170:1099 –1100.
8. Ross C, Abrams S, Aloia J, et al. Review of potential indicators of
adequacy and selection of indicators: calcium and vitamin D. In:
Ross C, Taylor C, Yaktine A, Valle HD, eds. Dietary Reference
Intake for Calcium and Vitamin D. Washington, DC: National Academics Press; 2011:1–187.
9. Garland CF, Gorham ED, Mohr SB, Garland FC. Vitamin D for
cancer prevention: global perspective. Ann Epidemiol. 2009;19:
468 – 483.
10. Ovesen L, Andersen R, Jakobsen J. Geographical differences in vitamin D status, with particular reference to European countries.
Proc Nutr Soc. 2003;62:813– 821.
11. Pedersen JI. Vitamin D requirement and setting recommendation
levels - current Nordic view. Nutr Rev. 2008;66:S165–S169.
12. DIPART (Vitamin D Individual Patient Analysis of Randomized
Trials) Group. Patient level pooled analysis of 68 500 patients from
seven major vitamin D fracture trials in US and Europe. BMJ. 2010;
340:b5463.
13. Avenell A, Gillespie WJ, Gillespie LD, O’Connell D. Vitamin D and
vitamin D analogues for preventing fractures associated with involutional and post-menopausal osteoporosis. Cochrane Database
Syst Rev. 2009;2:CD000227.
14. Chapuy MC, Arlot ME, Delmas PD, Meunier PJ. Effect of calcium
and cholecalciferol treatment for three years on hip fractures in
elderly women. BMJ. 1994;308:1081–1082.
15. Feskanich D, Willett WC, Colditz GA. Calcium, vitamin D, milk
consumption, and hip fractures: a prospective study among postmenopausal women. Am J Clin Nutr. 2003;77:504 –511.
16. Michaëlsson K, Melhus H, Bellocco R, Wolk A. Dietary calcium and
vitamin D intake in relation to osteoporotic fracture risk. Bone.
2003;32:694 –703.
17. Nieves JW, Barrett-Connor E, Siris ES, Zion M, Barlas S, Chen YT.
Calcium and vitamin D intake influence bone mass, but not shortterm fracture risk, in Caucasian postmenopausal women from the
National Osteoporosis Risk Assessment (NORA) study. Osteoporos Int. 2008;19:673– 679.
18. Warensjo E, Byberg L, Melhus H, et al. Dietary calcium intake and
risk of fracture and osteoporosis: prospective longitudinal cohort
study. BMJ. 2011;342:d1473.
19. Gedeborg R, Engquist H, Berglund L, Michaelsson K. Identification
of incident injuries in hospital discharge registers. Epidemiology.
2008;19:860 – 867.
20. Kanis JA. Assessment of fracture risk and its application to screening
for postmenopausal osteoporosis: synopsis of a WHO report. WHO
Study Group. Osteoporos Int. 1994;4:368 –381.
21. Melhus H, Snellman G, Gedeborg R, et al. Plasma 25-hydroxyvitamin D levels and fracture risk in a community-based cohort of
elderly men in Sweden. J Clin Endocrinol Metab. 2010;95:2637–
2645.
22. Willett WC, Howe GR, Kushi LH. Adjustment for total energy intake in epidemiologic studies. Am J Clin Nutr. 1997;65:1220S–
1228S; discussion 1229S–1231S
23. Michaëlsson K, Holmberg L, Mallmin H, et al. Diet and hip fracture
790
24.
25.
26.
27.
28.
29.
30.
31.
32.
Snellman et al
Dietary Vitamin D Intake and Bone Health
risk: a case-control study. Study Group of the Multiple Risk Survey
on Swedish Women for Eating Assessment. Int J Epidemiol. 1995;
24:771–782.
Orsini N, Bellocco R, Bottai M, et al. Validity of self-reported total
physical activity questionnaire among older women. Eur J Epidemiol. 2008;23:661– 667.
Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of
classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–383.
Hu FB, Stampfer MJ, Rimm E, et al. Dietary fat and coronary heart
disease: a comparison of approaches for adjusting for total energy
intake and modeling repeated dietary measurements. Am J Epidemiol. 1999;149:531–540.
Lin DY. Non-parametric inference for cumulative incidence functions in competing risks studies. Stat Med. 1997;16:901–910.
Freyschuss B, Ljunggren O, Saaf M, Mellstrom D, Avenell A Calcium and vitamin D for prevention of osteoporotic fractures. Lancet.
2007;370:2098 –2099; author reply 2099.
Tang BM, Eslick GD, Nowson C, Smith C, Bensoussan A. Use of
calcium or calcium in combination with vitamin D supplementation
to prevent fractures and bone loss in people aged 50 years and older:
a meta-analysis. Lancet. 2007;370:657– 666.
Manson JE, Bassuk SS, Lee IM, et al. The VITamin D and OmegA-3
TriaL (VITAL): rationale and design of a large randomized controlled trial of vitamin D and marine omega-3 fatty acid supplements
for the primary prevention of cancer and cardiovascular disease.
Contemporary clinical trials. 2012;33:159 –171.
Barrett EL, Richardson DS. Sex differences in telomeres and lifespan. Aging cell. 2011;10:913–921.
Lips P, Duong T, Oleksik A, et al. A global study of vitamin D status
and parathyroid function in postmenopausal women with osteopo-
J Clin Endocrinol Metab, March 2014, 99(3):781–790
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
rosis: baseline data from the multiple outcomes of raloxifene evaluation clinical trial. J Clin Endocrinol Metab. 2001;86:1212–1221.
Lips P. Vitamin D status and nutrition in Europe and Asia. J Steroid
Biochem Mol Biol. 2007;103:620 – 625.
Ma J, Giovannucci E, Pollak M, et al. Milk intake, circulating levels
of insulin-like growth factor-I, and risk of colorectal cancer in men.
J Natl Cancer Inst. 2001;93:1330 –1336.
Tuohimaa P. Vitamin D and aging. J Steroid Biochem Mol Biol.
2009;114:78 – 84.
Kuro-o M. Endocrine FGFs and Klothos: emerging concepts. Trends
Endocrinol Metab. 2008;19:239 –245.
Sanders KM, Stuart AL, Williamson EJ, et al. Annual high-dose oral
vitamin D and falls and fractures in older women: a randomized
controlled trial. JAMA. 2010;303:1815–1822.
Smith H, Anderson F, Raphael H, Maslin P, Crozier S, Cooper C.
Effect of annual intramuscular vitamin D on fracture risk in elderly
men and women–a population-based, randomized, double-blind,
placebo-controlled trial. Rheumatology (Oxford). 2007;46:1852–
1857.
Bischoff-Ferrari HA, Willett WC, Orav EJ, et al. A pooled analysis
of vitamin D dose requirements for fracture prevention. N Engl
J Med. 2012;367:40 – 49.
Michaëlsson K, Holmberg L, Mallmin H, Wolk A, Bergström R,
Ljunghall S. Diet, bone mass, and osteocalcin: a cross-sectional
study. Calcif Tissue Int. 1995;57:86 –93.
Melhus H, Snellman G, Gedeborg R, et al. Plasma 25-hydroxyvitamin D levels and fracture risk in a community-based cohort of
elderly men in Sweden. J Clin Endocrinol Metab. 2010;95:2637–
2645.
Henriquez-Sanchez P, Sanchez-Villegas A, Doreste-Alonso J, OrtizAndrellucchi A, Pfrimer K, Serra-Majem L. Dietary assessment
methods for micronutrient intake: a systematic review on vitamins.
Br J Nutr. 2009;102(Suppl 1):S10 –S37.
Members have FREE online access to the journal
Hormones and Cancer.
www.endocrine.org/HC