Gynecologic Oncology 121 (2011) 369–375

Contents lists available at ScienceDirect

Gynecologic Oncology
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y g y n o

Meta-analysis: Circulating vitamin D and ovarian cancer risk

Lu Yin, Norma Grandi, Elke Raum, Ulrike Haug, Volker Arndt, Hermann Brenner ⁎
Division of Clinical Epidemiology and Aging Research, German Cancer Research Center, Bergheimer Strasse 20, D-69115 Heidelberg, Germany

a r t i c l e i n f o a b s t r a c t

Article history: Objective. To review and summarize evidence from longitudinal studies on the association between
Received 13 December 2010 circulating 25 hydroxyvitamin D (25(OH)D) and the risk of ovarian cancer (OC).
Available online 15 February 2011 Methods. Relevant prospective cohort studies and nested case-control studies were identified by
systematically searching Ovid Medline, EMBASE, and ISI Web of Knowledge databases and by cross-
Keywords:
referencing. The following data were extracted in a standardized manner from eligible studies: first author,
Circulating vitamin D
publication year, country, study design, characteristics of the study population, duration of follow-up, OC
Ovarian cancer
Meta-analysis
incidence according to circulating vitamin D status and the respective relative risks, and covariates adjusted
for in the analysis. Due to the heterogeneity of studies in categorizing circulating vitamin D levels, all results
were recalculated for an increase of circulating 25(OH)D by 20 ng/ml. Summary relative risks (RRs) were
calculated using meta-analysis methods.
Results. Overall, ten individual-level studies were included that reported on the association between
circulating vitamin D levels and OC incidence. Meta-analysis of studies on OC incidence resulted in a summary
RR (95% confidence interval, CI) of 0.83 (0.63–1.08) for an increase of 25(OH)D by 20 ng/ml (P = 0.160). No
indication for heterogeneity and publication bias was found.
Conclusions. A tentative inverse association of circulating 25(OH)D with OC incidence was found, which
did not reach statistical significance but which requires clarification by additional studies due to potentially
high clinical and public health impact.
© 2011 Elsevier Inc. All rights reserved.

Introduction evaluating the association between circulating 25(OH)D levels and OC

Although vitamin D is obtained from diet and dietary supplements,
the main source for vitamin D is its production in the skin under the
influence of solar ultraviolet B (UV-B) radiation. In 1980, Garland and
Garland [1] hypothesized that lower levels of vitamin D resulting from
much weaker UV-B radiation at higher latitudes may account for the
striking geographical pattern of cancer mortality. Partly stimulated by
this article, further research in this area has been conducted in
observational studies over the past 20 years [2–4].
Several epidemiologic studies addressing the association between
vitamin D and ovarian cancer (OC) have assessed dietary vitamin D
intake, but results have been equivocal [5–9]. Most evidence from
ecological studies of solar UV-B and ovarian cancer mortality and
incidence support the role of vitamin D on ovarian carcinogenesis
[10–13], but no association was found in a multinational study [14]. In
recent years, several individual-level studies have addressed the
association of OC risk with circulating 25(OH)D levels representing an
integrated measure for vitamin D from diet, dietary supplements, and
skin production [15]. In this paper, we aimed to provide a systematic
review and meta-analysis of observational epidemiological studies
⁎ Corresponding author. Fax: +49 6221 548142.
E-mail address: h.brenner@dkfz.de (H. Brenner).
risk by using methods for comprehensive trend estimation from
summarized dose–response data [16].
Material and methods
Identification of studies and study selection
A literature search was conducted to identify longitudinal studies
assessing the association between circulating levels of 25(OH)D and
OC incidence or mortality. We searched Ovid (Ovid Technologies, Inc.,
New York, 1950–August 6, 2010), EMBASE (Elsevier, Amsterdam, the
Netherlands, 1980–August 16, 2010), and ISI Web of Knowledge
(Thomson Scientific Technical Support, New York, 1945–August 14,
2010) databases for relevant articles by various combinations of
relevant terms in the article including 25-OH-D, cholecalciferol,
calcidiol, calcitriol, 25-hydroxyvitamin D, Hydroxycholecalciferols,
25-Hydroxyvitamin D3 1-alpha-Hydroxylase, 1,25-dihydroxyvitamin
D, vitamin D, ovarian, ovary, ovariate, cancer, tumor, tumour, and
neoplasm. Duplicate publications were deleted. Each title and abstract
was checked for relevance. The full text was reviewed if the abstract
indicated that the article reported associations between circulating
vitamin D and OC risk. Only original longitudinal studies conducted
among humans were considered for the review. Cross-referencing
was employed to complement the study identification process.

0090-8258/$ – see front matter © 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.ygyno.2011.01.023
370 L. Yin et al. / Gynecologic Oncology 121 (2011) 369–375
Data extraction
From each of the eligible studies, the first author (Yin L) and the
second author (Grandi N) extracted the following data independently
in a standardized manner: author(s), publication year, country, study
design, characteristics of the study population, duration of follow-up,
OC incidence or mortality according to circulating vitamin D status
and the respective measures of relative risk (see below), as well as
covariates adjusted for in the analysis. Any initial disagreement was
resolved by consensus after additional review of the articles.
Meta-analyses
Main outcome variables were measures of relative risks for the
association between circulating 25(OH)D levels and OC risk. When
such data were not explicitly reported, they were derived from data
provided in the articles or requested from the authors through
personal contacts, wherever possible. For consistency, circulating
concentrations of 25(OH)D given in nmol/l were converted to ng/ml,
Fig. 1. Flow diagram of the literature search process.
using the pertinent conversion factor (1 ng/ml = 2.5 nmol/l). In most
studies, OC incidence was reported stratified by various categories of
25(OH)D. Depending on available information, median, midpoints or
means of the categories were used for meta-analysis. Due to the
different categorization of 25(OH)D levels across studies, all results
were recalculated for an increase of circulating 25(OH)D by 20 ng/ml,
both within studies (taking possible correlations resulting from a
common reference category into account), as well as across studies
using the methods suggested by Greenland and Longnecker [16].
Summary RRs from fixed and random effects models were calculated
using standard meta-analysis methods [17].
In a conservative approach, the random effects estimates, which
allow for variation of true effects across studies, were taken as “main
results” [18]. Random effects estimates were derived using the
DerSimonian–Laird method [19,20]. Heterogeneity was assessed by
the I2 statistic, and standardized deleted residual analysis was done to
identify outliers. The funnel plot, Begg and Mazumdar rank correla-
tion test and Egger's test of the intercept were employed to assess
indications of publication bias [21]. The R/S plus software, version
Table 1
Nested case-control studies reporting on the association of circulating 25(OH)D levels with incidence of ovarian cancer.

Ref. Author (s), year Study population RR (95% CI) of ovarian cancer Adjustment factors
according to 25(OH)D
Country (blood sampling, follow-up) No. of participants Age (mean) Setting
(range or median) (ng/ml) ‡,†
Case Control

[22] Tworoger et al., 2007 NHS/NHSII⁎ (1976–1990; 1990–2004) 159 476 34–69 Female nurses b20.6: 1.00 Adjusted for: ever use of postmenopausal hormones, BMI⁎
(56) 20.6–26.4: 0.93 (0.56, 1.55) at blood collection, parity, lactose intake, duration of oral
26.5–32.4: 0.93 (0.53, 1.62) contraceptive use, season of blood collection, the interaction
≥32.5: 0.84 (0.47, 1.52) between study with both duration of oral contraceptive use
and BMI⁎ at blood collection.

WHS⁎ (1992; 1993–2004) 61 122 45–73 Female professionals b17.4: 1.00

(56) 17.4–21.9: 1.45 (0.52, 4.05)
22.0–27.6: 3.04 (1.06, 8.78) Matched for: menopausal status at baseline and diagnosis,
≥27.7: 0.88 (0.28, 2.82) age, month of blood collection, time of day of blood
collection, fasting status, and
postmenopausal hormone use at blood collection.

NYUWHS⁎ USA (1985–1991; 1985–2005) b14.68:

L. Yin et al. / Gynecologic Oncology 121 (2011) 369–375

[23] Arslan et al., 2009 71 123 34–65 Women based 1.00 Adjusted for: oral contraceptive use and parity.
(55) 14.68-23.11: 1.13 (0.39, 3.27)
≥23.12: 1.50 (0.53, 4.23) Matched for: cohort, age at entry, and date of blood collection.

NSHDS⁎ Sweden (1985–2005) 97 191 30–65 b13.60: 1.00

(51) 13.60–17.87: 0.54 (0.25, 1.17)
≥17.88: 0.83 (0.38, 1.81)

[24] Toriola et al., 2010 FMC⁎ Finland (1983–now) 201 398 17–44 Pregnant women based b10.6: 1.8 (0.9, 3.5) Adjusted for: age at last full term pregnancy and
(31) 10.7–13.2: 1.6 (0.8, 3.1) bench lag-time (days between sample withdrawal
13.3–16.0: 1.5 (0.8, 2.8) and freezing it for storage).
16.1–21.2: 1.6 (0.9, 2.9)
≥21.3: 1.0 Matched for: season of blood collection, age and parity

[25] Zheng et al., 2010 CLUE⁎ USA (1974–1989; 1975–2007) 102 102 58 Population based b10: 1.18 (0.52, 2.68) Adjusted for: duration of oral contraceptive use and
(45–70) 10–29: 1.00 number of pregnancies.
≥30: 1.02 (0.47, 2.24) Matched for: age, race/ethnicity, date of blood collection.

PLCO⁎ USA (1992–2001; 1993–2006) 74 74 58 Population based b10: 1.90 (0.63, 5.71)
(55–74) 10–29: 1.00
≥30: 1.87 (0.62, 5.64)

MEC⁎ USA (1993–1996; 1994–2003) 18 18 58 Population based b10: 0.79 (0.10, 6.09)
(N/A) 10–29: 1.00
≥30: 0.49 (0.03, 8.36)

CPS-II⁎ USA (1982–1992:1992–2002) 27 27 58 Population based b10: 0.67 (0.12, 3.68)

(N/A) 10–29: 1.00
≥30: –

SWHS⁎ China (1996–2000; 1996–2005) 74 74 58 Population based b10: 1.24 (0.54, 2.87)
(40–70) 10–29: 1.00
≥30: –
‡
For consistency, circulating concentrations of 25(OH)D in nmol/l were converted to ng/ml using the conversion factor, 1 ng/ml = 2.5 nmol/l.
†
RR: relative risk; CI: confidence interval.
⁎BMI: Body Mass Index; NHS: the Nurses' Health Study; WHS: the Women's Health Study; NYUWHS: The New York University Women's Health Study; NSHDS: The Northern Sweden Health and Disease Study; FMC: Finish Maternity Cohort;
CLUE: Connective Tissue Disease Leg Ulcer Etiology study; PLCO: Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial; MEC: Multiethnic Cohort Study of Diet and Cancer; CPS-II: Cancer Prevention Study II Nutrition Cohort; SWHS:
Shanghai Women's Health Study. N/A: not available.

371
372 L. Yin et al. / Gynecologic Oncology 121 (2011) 369–375
2.8.1, and the statistics software SAS©, version 9.1 (SAS Institute Inc.,
Cary, N.C., USA), were used for the analysis.
Results
A flow diagram of the search process is given in Fig. 1. Total
searches yielded 770 entries. Following removal of 231 duplicates,
539 titles and abstracts were assessed and 22 articles appeared to be
potentially relevant for inclusion into the review. 18 articles were
excluded for the following reasons: no original articles but editorials,
comments, reviews (N = 10), ecological studies (N = 3), only vitamin
D intake reported (N = 2), only mortality among ovarian cancer
patients assessed (N = 1), case-control studies (N = 1), and repeated
studies from the same study population (N = 1). The references of
excluded studies are provided in Appendix 1.
Although only 4 articles were left for inclusion in this review and
meta-analysis, these articles reported on 10 different studies. Details
on the respective study design, the study populations, the study
results, and covariates adjusted for all included studies are shown in
Table 1 and summarized below and in Fig. 2.
The first nested case-control study was reported by Tworoger et al.
[22] based on three prospective cohorts: the Nurses' Health Study
(NHS), NHSII, and the Women's Health Study (WHS). This analysis
included 224 cases (161 from NHS/NHSII and 63 from WHS) and 603
controls (matching ratio: 1:3 for NHS/NHSII and 1:2 for WHS).
Aslan et al. [23] conducted a nested case-control study within two
prospective cohorts, the New York University Women's Health Study
(NYUWHS) and the Northern Sweden Health and Disease Study
(NSHDS), to evaluate the association between circulating levels of 25
(OH)D and risk of subsequent invasive epithelial OC. This study
included 168 OC cases (71 from NYUWHS and 97 from NSHDS) and
316 controls (125 from NYUWHS and 191 from NSHDS).
Toriola et al. [24] conducted a nested case-control study on the
association between circulating 25(OH)D levels and OC risk within
the Finnish Maternity Cohort (FMC), which was established by
National Institute for Health and Welfare, Finland in 1983. This
study included 201 OC cases and 398 controls, aged from 17 to
44 years. The cases were matched with two sets of controls. The first
set of controls were matched by age at blood collection (±1 years),
parity and date of blood collection (±4 weeks) with matching ratio of
1:2. The second set of controls were matched by age at blood
collection (±1 years) and parity with matching ratio of 1:1.
A recent nested case-control study, which was conducted by
Zheng et al. [25] in the Cohort Consortium Vitamin D Pooling Project
of Rarer Cancers (VDPP), combined seven prospective cohort studies
from different countries to evaluate the circulating 25(OH)D levels in
relation to epithelial ovarian cancer risk. Among these seven studies,
two (Nurses' Health Study: NHS; New York University Women's
Health study: NYU-WHS) were excluded because they had smaller
sample sizes and came from the same cohorts for which results had
been published in detail in two earlier articles [22,23]. Five studies
from the Cohort Consortium Vitamin D Pooling Project of Rarer
Fig. 2. Relative risk and 95% CIs of ovarian cancer risk according to circulating 25(OH)D levels † and ‡. †: Depending on available information, median, midpoints or means of the
categories were used for definition of study specific levels of circulating 25(OH)D categories. ‡: NHS: the Nurses' Health Study; WHS: the Women's Health Study; NYUWHS: The New
York University Women's Health Study; NSHDS: The Northern Sweden Health and Disease Study; FMC: Finish Maternity Cohort; CLUE: Connective Tissue Disease Leg Ulcer Etiology
study; PLCO: Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial; MEC: Multiethnic Cohort Study of Diet and Cancer; CPS-II: Cancer Prevention Study II Nutrition Cohort;
SWHS: Shanghai Women's Health Study.
Fig. 2-1. Tworoger et al. [22], 2007 (USA: NHS/NHSII).
Fig. 2-2. Tworoger et al. [22], 2007 (USA: WHS).
Fig. 2-3. Arslan et al. [23], 2009 (USA: NYUWHS).
Fig. 2-4 Arslan et al. [23], 2009 (Sweden: NSHDS).
Fig. 2-5. Toriola et al. [24], 2010 (Finland: FMC).
Fig. 2-6. Zheng et al. [25], 2010 (USA: CLUE).
Fig. 2-7. Zheng et al. [25], 2010 (USA: PLCO).
Fig. 2-8. Zheng et al. [25], 2010 (USA: MEC).
Fig. 2-9. Zheng et al. [25], 2010 (USA: CPS-II).
Fig. 2-10. Zheng et al. [25], 2010 (China: SWHS).
Cancers (VDPP) with a total number of 295 cases and 295 controls
were included in our meta-analysis (Connective Tissue Disease Leg
Ulcer Etiology study: CLUE; Prostate, Lung, Colorectal, and Ovarian
Cancer Screening Trial: PLCO; Multiethnic Cohort Study: MEC; Cancer
Prevention Study II Nutrition Cohort: CPS-II; Shanghai Women's
Health Study: SWHS) [25].
As illustrated in Fig. 2, analyses of dose–response trends across
studies did not reveal consistent patterns. The results of meta-
analyses on the association between circulating 25(OH)D levels and
OC incidence are shown in Fig. 3. All RRs refer to an increase of 25(OH)
D by 20 ng/ml. Seven of ten studies showed an inverse association for
an increase in 25(OH)D levels, but this association was not significant
in any of the studies. A non-significant inverse association was
observed in pooled analyses using both a fixed effects model and a
random effects model (RR, 0.83; 95% CI, 0.63–1.08; P = 0.166). No
statistical heterogeneity among these ten included studies was
observed (I2 = 0.0%; P = 0.87). The funnel plot did not show evidence
of publication bias (Kendall's tau = 0.20; P = 0.48; Egger's t
value = 0.77, P = 0.44). In standardized deleted residual analysis, no
study was identified as an outlier.
Discussion
In this review and meta-analysis, we summarize the results of ten
longitudinal studies published so far on the association between
circulating 25(OH)D and OC incidence, which comprised a total of
2488 subjects including 883 OC cases. Although seven out of ten
studies found a tentatively reduced risk of OC incidence with
increasing 25(OH)D levels, none of these associations was statistically
significant. The pooled estimate from our meta-analysis also suggests
a possible inverse association, but this association did likewise not
reach statistical significance.
A previous systematic review of vitamin D and ovarian cancer,
which included ecological studies and case-control studies with blood
sample collection after diagnosis, did likewise not find conclusive
evidence for an association between circulating 25(OH)D levels and
OC risk [26]. In contrast to this article, we restricted our analysis to
longitudinal studies which are less prone to bias from reverse causality
(such as lower vitamin D levels due to hospitalization after diagnosis).
Furthermore, six additional longitudinal studies which were published
only very recently were included and results were summarized in a
quantitative manner by comprehensive trend estimate and meta-
analyses. While our study focused on the association of OC risk with
circulating levels of 25(OH)D, three previous studies also reported a
tentative, but not significant association of OC risk with polymorphism
in the vitamin D receptor (VDR) [27–29].
Interestingly, a significant inverse association between circulating
25(OH)D levels and OC risk has been observed among women with a
body mass index (BMI) of ≥ 25 kg/m2 in the earliest study as well as in
the pooled results of the VDPP [22,25]. One explanation is that the
amount of adipose tissue in humans is inversely related to circulating
25(OH)D levels, likely due to uptake of vitamin D by fat cells [30–32],
 L. Yin et al. / Gynecologic Oncology 121 (2011) 369–375 373

which suggested that adipose tissue may have an important role on in the Finnish population [24]. Therefore, more comprehensive
long-term vitamin D status. However, no such association was investigation is needed to explore this potential interrelationship.
observed in the study by Arslan et al. [23] involving populations Ovarian cancer is a heterogeneous disease. Two recent articles
from Sweden and New York, and BMI-specific data were not reported evaluated associations by histological subtype, but they did not
374 L. Yin et al. / Gynecologic Oncology 121 (2011) 369–375

Fig. 3. Meta-analyses: Relative risk of ovarian cancer risk per 20 ng/ml increase in circulating 25(OH)D º. º: NHS: the Nurses' Health Study; WHS: the Women's Health Study;
NYUWHS: The New York University Women's Health Study; NSHDS: The Northern Sweden Health and Disease Study; FMC: Finish Maternity Cohort; CLUE: Connective Tissue
Disease Leg Ulcer Etiology study; PLCO: Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial; MEC: Multiethnic Cohort Study of Diet and Cancer; CPS-II: Cancer Prevention
Study II Nutrition Cohort; SWHS: Shanghai Women's Health Study.

observe any relevant differences [24,25]. However, case numbers for
individual subtypes were small and more and larger studies are
needed to draw firm conclusions.
In recent years, an increasing number of studies have addressed
associations of circulating 25(OH)D levels with a variety of cancers.
Whereas inverse associations were rather consistently found for
colorectal cancer [33] and colorectal adenoma [34], results were more
heterogeneous for breast cancer, where strong inverse association has
been found in case-control studies, but not in longitudinal studies
[35]. No associations have been found for other cancers, such as
prostate cancer [36], but risk increase has been suggested at both the
lower and the upper levels of the 25(OH)D distributions [37,38].
Analyses of the 25(OH)D-cancer association by cancer types and
subtypes may point to potential common or divergent underlying
mechanisms.
Our analysis has specific strengths and limitations. A major
strength of our study is the application of advanced techniques of
statistical analysis that allowed to summarize adjusted associations
across studies and over the entire range of circulating 25(OH)D
values, despite the very heterogeneous categorization of 25(OH)D
levels in the individual studies. Our study also has important
limitations. Although our review searched three databases, i.e., Ovid
Medline, EMBASE, and ISI Web of Knowledge, and extensive checks
for completeness by cross-referencing was employed, we cannot
exclude having missed a relevant study. Analyses are limited by the
data provided by the individual studies. Depending on the results
reported, median, midpoints and mean 25(OH)D levels of the group
had to be used for pooling. As a result, estimates of risk may be less
accurate than if individual-level data had been available. Also,
different studies used different methods of measuring circulating
vitamin D, which might affect comparability of studies and introduce
heterogeneity between studies. Furthermore, despite the lack of
indication of major publication bias in the formal evaluations
employed, potential publication bias is impossible to be excluded
completely, especially in the light of the low number of studies.
Finally, our meta-analysis focused on a potential linear trend of
decreased OC risk with increasing 25(OH)D levels. Non-monotone
associations, such as U-shaped associations, have recently been
 L. Yin et al. / Gynecologic Oncology 121 (2011) 369–375
suggested for other types of cancers, such as prostate cancer [37–40],
but not for ovarian cancer.
Conclusions
The results of this meta-analysis are consistent with the hypoth-
esis of a potential weak inverse association between circulating
vitamin D and OC risk. However, available data are still sparse and the
tentative inverse association was not statistically significant in any of
the 10 included studies, or in the meta-analysis of all studies.
Nevertheless, even a modest inverse association between circulating
vitamin D and OC risk in the order of magnitude estimate in this
analysis could be of substantial clinical and public health impact, in
particular in the light of increasing, albeit equivocal evidence of
association of vitamin D with other cancers and other chronic diseases
[33,35,36,41]. In-depth analyses of the assessed associations in the
context of additional longitudinal studies are highly desirable to
enable more precise estimates and a better understanding of the role
of vitamin D in OC carcinogenesis, especially regarding subgroups
such as overweight and obese women.
Supplementary materials related to this article can be found online
at doi:10.1016/j.ygyno.2011.01.023.
Conflict of interest statement
No potential conflicts of interest were disclosed.
Acknowledgment
The work of Lu Yin was supported by a scholarship from the
German Research Foundation (Deutsche Forschungsgemeinschaft)
within the framework of a PhD program (Graduiertenkolleg 793).
References
[1] Garland CF, Garland FC. Do sunlight and vitamin D reduce the likelihood of colon
cancer? Int J Epidemiol 1980;9:227–31.
[2] Holick MF. Vitamin D: importance in the prevention of cancers, type 1 diabetes,
heart disease, and osteoporosis. Am J Clin Nutr 2004;79:362–71.
[3] Giovannucci E. The epidemiology of vitamin D and cancer incidence and mortality:
a review (United States). Cancer Causes Control 2005;16:83–95.
[4] Garland CF, Grant WB, Mohr SB, Gorham ED, Garland FC. What is the dose–
response relationship between vitamin D and cancer risk? Nutr Rev 2007;65:
S91–5.
[5] Bidoli E, La Vecchia C, Talamini R, et al. Micronutrients and ovarian cancer: a case-
control study in Italy. Ann Oncol 2001;12:1589–93.
[6] Genkinger JM, Hunter DJ, Spiegelman D, et al. Dairy products and ovarian cancer: a
pooled analysis of 12 cohort studies. Cancer Epidemiol Biomark Prev 2006;15:
364–72.
[7] Goodman MT, Wu AH, Tung KH, et al. Association of dairy products, lactose, and
calcium with the risk of ovarian cancer. Am J Epidemiol 2002;156:148–57.
[8] Koralek DO, Bertone-Johnson ER, Leitzmann MF, et al. Relationship between
calcium, lactose, vitamin D, and dairy products and ovarian cancer. Nutr Cancer
2006;56:22–30.
[9] Salazar-Martinez E, Lazcano-Ponce EC, Gonzalez Lira-Lira G, Escudero-De los Rios
P, Hernandez-Avila M. Nutritional determinants of epithelial ovarian cancer risk: a
case-control study in Mexico. Oncology 2002;63:151–7.
[10] Lefkowitz ES, Garland CF. Sunlight, vitamin D, and ovarian cancer mortality rates
in US women. Int J Epidemiol 1994;23:1133–6.
[11] Freedman DM, Dosemeci M, McGlynn K. Sunlight and mortality from breast,
ovarian, colon, prostate, and non-melanoma skin cancer: a composite death
certificate based case-control study. Occup Environ Med 2002;59:257–62.
[12] Grant WB, Garland CF. The association of solar ultraviolet B (UVB) with reducing
risk of cancer: multifactorial ecologic analysis of geographic variation in age-
adjusted cancer mortality rates. Anticancer Res 2006;26:2687–99.
375
[13] Garland CF, Mohr SB, Gorham ED, Grant WB, Garland FC. Role of ultraviolet B
irradiance and vitamin D in prevention of ovarian cancer. Am J Prev Med 2006;31:
512–4.
[14] Waltz P, Chodick G. Assessment of ecological regression in the study of colon,
breast, ovary, non-Hodgkin's lymphoma, or prostate cancer and residential UV.
Eur J Cancer Prev 2008;17:279–86.
[15] Hollis BW, Kamerud JQ, Selvaag SR, Lorenz JD, Napoli JL. Determination of vitamin
D status by radioimmunoassay with an 125I-labeled tracer. Clin Chem 1993;39:
529–33.
[16] Greenland S, Longnecker MP. Methods for trend estimation from summarized
dose–response data, with applications to meta-analysis. Am J Epidemiol
1992;135:1301–9.
[17] Higgins J, Green S, editors. updated September 2006. The Cochrane Library, Issue
4, 2006Chichester, UK: John Wiley & Sons, Ltd.; September 2006.
[18] Normand SL. Meta-analysis: formulating, evaluating, combining, and reporting.
Stat Med 1999;18:321–59.
[19] DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:
177–88.
[20] Lipsey M, Wilson D. Practical meta-analysis: Thousand Oaks. CA: Sage; 2001.
[21] Rothstein HR, Sutton AJ, Borenstein M. Publication bias in meta-analysis:
prevention, assessment and adjustments: Chichester. England: Wiley; 2005.
[22] Tworoger SS, Lee IM, Buring JE, Rosner B, Hollis BW, Hankinson SE. Plasma 25-
hydroxyvitamin D and 1, 25-dihydroxyvitamin D and risk of incident ovarian
cancer. Cancer Epidemiol Biomark Prev 2007;16:783–8.
[23] Arslan AA, Clendenen TV, Koenig KL, et al. Circulating vitamin d and risk of
epithelial ovarian cancer. J Oncol 2009;2009:672492.
[24] Toriola AT, Surcel HM, Agborsangaya C, et al. Serum 25-hydroxyvitamin D and the
risk of ovarian cancer. Eur J Cancer 2010;46:364–9.
[25] Zheng W, Danforth KN, Tworoger SS, et al. Circulating 25-hydroxyvitamin D and
risk of epithelial ovarian cancer. Am J Epidemiol 2010;172:70–80.
[26] Cook LS, Neilson HK, Lorenzetti DL, Lee RC. A systematic literature review of
vitamin D and ovarian cancer. Am J Obstet Gynecol 2010;203:70.e1–8.
[27] Lurie G, Wilkens LR, Thompson PJ, et al. Vitamin D receptor gene polymorphisms
and epithelial ovarian cancer risk. Cancer Epidemiol Biomark Prev 2007;16:
2566–71.
[28] Clendenen TV, Arslan AA, Koenig KL, et al. Vitamin D receptor polymorphisms and
risk of epithelial ovarian cancer. Cancer Lett 2008;260:209–15.
[29] Tworoger SS, Gates MA, Lee IM, et al. Polymorphisms in the vitamin D receptor
and risk of ovarian cancer in four studies. Cancer Res 2009;69:1885–91.
[30] Arunabh S, Pollack S, Yeh J, Aloia JF. Body fat content and 25-hydroxyvitamin D
levels in healthy women. J Clin Endocrinol Metab 2003;88:157–61.
[31] Gomez JM, Maravall FJ, Gomez N, Navarro MA, Casamitjana R, Soler J. Relationship
between 25-(OH) D3, the IGF-I system, leptin, anthropometric and body
composition variables in a healthy, randomly selected population. Horm Metab
Res 2004;36:48–53.
[32] Lucas JA, Bolland MJ, Grey AB, et al. Determinants of vitamin D status in older
women living in a subtropical climate. Osteoporos Int 2005;16:1641–8.
[33] Yin L, Grandi N, Raum E, Haug U, Arndt V, Brenner H. Meta-analysis: longitudinal
studies of serum vitamin D and colorectal cancer risk. Aliment Pharmacol Ther
2009;30:113–25.
[34] Wei MY, Garland CF, Gorham ED, Mohr SB, Giovannucci E. Vitamin D and
prevention of colorectal adenoma: a meta-analysis. Cancer Epidemiol Biomark
Prev 2008;17:2958–69.
[35] Yin L, Grandi N, Raum E, Haug U, Arndt V, Brenner H. Meta-analysis: serum
vitamin D and breast cancer risk. Eur J Cancer 2010;46:2196–205.
[36] Yin L, Raum E, Haug U, Arndt V, Brenner H. Meta-analysis of longitudinal
studies: serum vitamin D and prostate cancer risk. Cancer Epidemiol 2009;33:
435–45.
[37] Tuohimaa P, Tenkanen L, Ahonen M, et al. Both high and low levels of blood
vitamin D are associated with a higher prostate cancer risk: a longitudinal, nested
case-control study in the Nordic countries. Int J Cancer 2004;108:104–8.
[38] Tuohimaa P, Tenkanen L, Syvala H, et al. Interaction of factors related to the
metabolic syndrome and vitamin D on risk of prostate cancer. Cancer Epidemiol
Biomark Prev 2007;16:302–7.
[39] Faupel-Badger JM, Diaw L, Albanes D, Virtamo J, Woodson K, Tangrea JA. Lack of
association between serum levels of 25-hydroxyvitamin D and the subsequent
risk of prostate cancer in Finnish men. Cancer Epidemiol Biomark Prev 2007;16:
2784–6.
[40] Platz EA, Leitzmann MF, Hollis BW, Willett WC, Giovannucci E. Plasma 1, 25-
dihydroxy- and 25-hydroxyvitamin D and subsequent risk of prostate cancer.
Cancer Causes Control 2004;15:255–65.
[41] Grandi NC, Breitling LP, Brenner H. Vitamin D and cardiovascular disease:
systematic review and meta-analysis of prospective studies. Prev Med 2010;51:
228–33.