Serum 25-hydroxyvitamin D and the risk of ovarian cancer

Methods We conducted a population-based longitudinal case–control study nested within the Finnish Maternity Cohort (FMC) which contains serum samples of virtually all pregnant women in Finland since 1983. Among them, 201 ovarian cancers diagnosed within 10 years of serum sampling were randomly selected as cases for this study. For each case, we selected two controls matched for age, parity and sampling season (±4 weeks) and one control matched for age and parity but for the opposite sampling season (6 months ± 4 weeks). Results The relative risks (estimated as odds ratio, OR) for ovarian cancer comparing the lowest quintile to the highest quintile of serum 25-OHD concentration were 1.8 (95% CI 0.9–3.5) among controls matched for the same season, and 1.1 (95% CI 0.6–2.2) among controls matched for the opposite season. The OR among women with insufficient (<75 nmol/L) serum 25-OHD concentration was 2.7 (95% CI 1.0–7.9, lower limit, 0.95) compared to that among those with sufficient (⩾75 nmol/L) serum 25-OHD concentration. No differences in the point estimates were observed between serous or mucinous histological subtypes of ovarian cancer. Conclusion Overall, we did not observe a significant association between serum 25-OHD concentrations and the risk of ovarian cancer. However, we found evidence suggestive of an increased risk among women with low to insufficient serum 25-OHD concentrations. Keywords Ovarian cancer Vitamin D 25-Hydroxyvitamin D Nested case–control Population-based 1 Introduction Ovarian cancer is the most lethal gynaecological malignancy and the fourth leading cause of cancer deaths among women worldwide. 1 About 191,000 new cases are diagnosed annually, 2 with the highest incidence rates found in Caucasians in northern and western Europe and in North America. 2 Despite its considerable burden to cancer mortality in women, very little is known about its aetiology. 1 Ecological studies have suggested that there may be an association between solar ultraviolet B radiation and risk of ovarian cancer. 3–6 It has been postulated that the protective effects of sunlight on ovarian cancer are mediated via vitamin D, which is a steroid-like hormone mainly produced through the action of sunlight on the skin. Of the six ecological studies published regarding ovarian cancer mortality rates and solar radiation, four 3–6 reported higher mortality rates with lower regional sunlight whereas two observed no relationship. 7,8 Solar ultraviolet B (UVB) and high solar irradiance were also shown to be inversely associated with incidence rates of ovarian cancer in a worldwide study 9 while there was no association in another multinational study. 10 Likewise, epidemiological studies examining the association of dietary vitamin D and ovarian cancer have been equivocal. 11–14 However, one recent study that evaluated the association of plasma vitamin D and ovarian cancer risk reported a beneficial effect among certain subgroups. 15 Vitamin D, by inducing cellular differentiation and apoptosis, inhibiting cellular proliferation, invasiveness and angiogenesis possesses anti-carcinogenic properties. 16 Vitamin D exerts its action through the vitamin D receptor (VDR) and VDRs are present in both normal ovarian tissues and ovarian cancer cells. 17–19 In this nested case–control study, we examined the relationship between serum vitamin D concentration and risk of ovarian cancer exploiting the resources of the nationwide Finnish Maternity Cohort (FMC), established in 1983. 20 2 Materials and methods 2.1 Finnish Maternity Cohort We carried out a case–control study nested within the Finnish Maternity Cohort (FMC). The FMC was established by the National Institute for Health and Welfare, Finland in 1983. 20 Following an informed consent, first trimester blood samples are withdrawn from pregnant women at the municipal maternity care units to screen for intrauterine infections. After the screening has been done, the remaining sample (1–3 mL of serum) is stored at −25 °C in polypropylene cryo vials in a well-protected biorepository at the National Institute for Health and Welfare in Oulu. More than 98% of pregnant women in Finland have donated blood samples to the cohort since 1983 and currently over 1.3 million samples are kept in storage. Each year about 60,000 new serum samples are added to the repository. 2.2 Identification of cases and controls Incident ovarian cancer cases were identified by the population-based Finnish Cancer Registry (FCR). All cancer cases diagnosed in Finland since 1953 are reported to the FCR (reporting mandatory since 1961). The coverage of the FCR is virtually complete with no losses to follow-up. 21 Every resident of Finland has a unique personal identity code that is also used in official health registries such as the FMC and FCR. Our study cohort was record linked with the cancer registry data by using the personal identity code. Two-hundred and one random ovarian cancer cases diagnosed within 10 years of serum sampling were selected for this study. If a case had donated more than one sample within this 10-year period, the sample donated closest to cancer diagnosis was selected. Time between serum donation and cancer diagnosis ranged between 2 years and 10 years with a median of 5 years. Among the 201 cancer cases, 195 had histological confirmation. Of the 195 with histological confirmation, 76 (39%) were serous, 70 (35.9%) mucinous, 15 (7.7%) sex cord stromal tumours, 11 (5.6%) endometrioid, 9 (4.6%) germ cell tumours, 5 (2.6%) clear cell tumours and 9 (4.6%) others. The tumours were left-sided in 74 (45.4%) cases, right-sided in 66 (40.5%) cases and bilateral in 23 (14.1%) cases. The cases were matched with two sets of controls. The first set of controls consisted of women from the FMC who were alive and free of cancer at the time of diagnosis of the case and were matched for (i) age at sample withdrawal ±1 year, (ii) parity and (iii) date of index blood sampling ±4 weeks (same season as case). For each case, 2 controls were selected and 400 serum samples were available for the laboratory analysis. The second set of controls consisted of women from the cohort, who were alive and free of cancer at the time of diagnosis of the cases and whose serum samples were taken at the opposite seasons to the cases (approximately six months ± 4 weeks apart). These controls were matched for (i) age at sample withdrawal ±1 year and (ii) parity. One control per case was selected. In all, 201 cases, 398 same season controls and 199 opposite season controls were available. All study samples were previously unthawed. The study was approved by the ethical committee of the National Institute for Health and Welfare, Finland. 2.3 Laboratory analysis Quantification of 25-hydroxyvitamin D (25-OHD) was done at the Department of Medical Biosciences, University of Umeå, Umeå, Sweden using a 25-OHD radioimmunoassay (RIA) from IDS Ltd., Boldon, UK. The manufacturer stated a specificity (% cross-reactivity) of 100% for 25-OHD3, 75% for 25-OHD2, 100% for 24, 25-OH2D3, and less than 0.01% and 0.3% for cholecalciferol (D3) and ergocalciferol (D2), respectively. The within, between and total coefficients of variations (CVs) of the assay were 7.8%, 9.6% and 12.4% at level 28.4 nmol/L 25-OHD, and 4.1%, 7.4% and 8.5% at 107 nmol/L 25-OHD, respectively. Case and control samples belonging to each other study set were assayed together, ordered randomly and labelled to mask case–control status. 2.4 Statistical analysis Two of the samples had 25-OHD concentrations of less than 2 nmol/L, which were considered to be outliers and set to missing. Descriptive data of cases and controls with regard to age at sample withdrawal and cancer diagnosis, age at last full term pregnancy, number of pregnancies, gestational days and bench lag-time were recorded. Quintile cut-off points were determined using 25-OHD concentrations of the controls. Quintiles of 25-OHD concentrations were used to estimate the relative risk of ovarian cancer and their 95% confidence intervals (CIs) by conditional logistic regression and are expressed as odds ratio (OR). The multivariate model was adjusted for age at last full term pregnancy and bench lag-time (days between sample withdrawal and freezing it for storage). Separate analyses were conducted for the main different histological subgroups of ovarian cancer available vis-a-vis, serous and mucinous tumours using cases and controls that were matched for the same season. Risk of ovarian cancer was also compared among women with serum 25-OHD levels above 75 nmol/L and those with serum 25-OHD levels below 75 nmol/L. 22,23 All statistical analyses were performed using SPSS 15 for windows (SPSS Inc., Chicago, IL). Two-sided p < 0.05 was considered statistically significant. 3 Results The age range for cases and controls was from 17.5 to 44 years. The median ages at the time of serum sampling were 31.5 and 31.4 years for cases and controls, respectively. The median duration between serum sampling and cancer diagnosis was 5 years. The median number of pregnancies was 2 with a range of 1–8 for cases and 1–10 for controls. Median serum 25-OHD concentrations were 35.0, 35.8 and 34.3 nmol/L among cases, same season controls and opposite season controls, respectively ( Table 1 ). We first compared the relative risk of ovarian cancer by quintile groups of serum 25-OHD among cases and controls whose serum samples were taken within the same season and had been matched for age and parity. Low levels of 25-OHD appeared to be associated with increased risk of ovarian cancer. The strongest association did, however, not reach statistical significance (1st quintile versus 5th quintile: OR, 1.8, CI 0.9–3.6; p trend = 0.26) age at last full term pregnancy and bench lag-time adjusted ( Table 2 ). Women with insufficient serum vitamin D concentrations had a threefold increased risk of ovarian cancer compared to women with sufficient serum concentrations (OR 2.8, 95% CI 1.0–7.9, lower limit 0.95, p -value = 0.06) ( Table 3 ). Different point estimates were obtained when control samples from opposite seasons were used. The point estimates were closer to unit risk (1st quintile versus 5th quintile: OR, 1.1, CI O.6–2.2; p trend = 0.49) ( Table 2 ). No significant associations were observed between quintiles of 25-OHD concentrations and ovarian cancer risk when the analyses were carried out for the two main histological groups. Comparing the 1st to the 5th quintiles, the ORs were 1.3 (95% CI 0.7–3.2) for serous tumours, and 1.2 (95% CI 0.4–3.1) for mucinous tumours. Stratification for seasonality did not show any further differences (data not shown). 4 Discussion In this longitudinal population-based nested case–control study, we observed no significant relationship between serum 25-hydroxyvitamin D concentrations and the risk of ovarian cancer. However, there appeared to be a tendency to increased risk among those who had low or insufficient serum 25-hydroxyvitamin D concentrations compared to those who had sufficient concentrations. Perhaps, one of the reasons why an association between 25-OHD and ovarian cancer was not apparent using the quintile distribution of 25-OHD was the generally relatively low serum 25-OHD concentrations in the cohort. Recently, due to a greater understanding of the interplay between vitamin D and its regulatory hormones, there have been discussions on what constitutes vitamin D deficiency. 22–24 A 25-OHD concentration of less than 50 nmol/L is considered to represent vitamin D deficiency, and that between 50 and 72 nmol/L is considered to represent relative insufficiency while concentrations above 75 nmol/L are thought to indicate vitamin D sufficiency. 22 Using these cut-off values, we found evidence suggestive of an increased risk of ovarian cancer among women with insufficient serum vitamin D concentrations compared with women with sufficient vitamin D concentrations. This supports results from the experimental studies which suggest that high vitamin D concentrations may be needed to prevent against ovarian cancer. 25,26 The results from our study are similar to those of another recent prospective study by Tworoger and colleagues. 15 The authors did not observe any association between plasma 25-hydroxyvitamin D and ovarian cancer risks in their overall analysis but women with adequate versus inadequate 25-hydroxyvitamin D levels had a decreased risk of borderline significance (RR, 0.7; 95% CI, 0.4–1.1). In addition, Tworoger and colleagues reported a stronger effect among overweight women. Unfortunately, data on body weight were not available for the women in our sub-sample of the FMC to explore this hypothesis. The ovary together with other tissues such as the kidney, breast, colon and prostate express the enzyme, 1 alpha-hydroxylase, (1aOHase). 25 This enzyme converts 25-OHD to 1, 25-(OH) 2 D, thus, the biologically active form of vitamin D is available locally in the ovaries and other tissues that express the enzyme. Since the ovaries also express VDR 18,19 and the effects of vitamin D are exerted through the VDR, it is biologically plausible that vitamin D may have effects on the ovaries. Ovarian cancer is a heterogeneous disease. The histological subtypes have different risk factors (both genetic and environmental) and dissimilar epidemiological and clinical characteristics. 27 In our study, we did not observe any differences in risk related to vitamin D concentration by histology. Apart from serous and mucinous cancers, the other histological subgroups, however, had very few cases. The following limitations of our study need to be taken into account: (i) oral contraceptive use increases serum 25-OHD concentrations 28,29 ; and use of oral contraceptives also protects against ovarian cancer. 30 Hence, use of oral contraceptives is a confounder in the association between 25-OHD and ovarian cancer. While this was controlled for by sample withdrawal during 12th to 14th week of pregnancy (when use of oral contraceptive must have been stopped for sometime), we did not have information on the use of oral contraceptives among our study subjects. (ii) Our study subjects were limited to fertile women in their reproductive ages who were matched for parity. Therefore we do not know if similar results will be obtained in infertile and/or post-menopausal women since they were not represented in our study. Serum 25-OHD concentration is a reflection of endogenous synthesis taking place in the skin under the influence of UVB irradiation and exogenous production from dietary intake. 22,31 Vitamin D first undergoes hydroxylation in the liver to 25-hydroxyvitamin D (25-OHD) before undergoing further hydroxylation to 1, 25-dihydroxyvitamin D (1, 25-(OH) 2 D) in the kidneys. 22 While 1, 25-(OH) 2 D is the most biologically active form of vitamin D, 25-OHD is considered as the best gauge of individual vitamin D status, 22,31,32 making measurement of serum/plasma 25-OHD concentrations the ideal way to determine vitamin D status. Strengths of our study include (i) its population-based nature; the FMC covers 98% of pregnant Finnish women. 20 (ii) Its prospective nature; the blood samples were collected before cancer diagnosis. Furthermore, the study samples were also collected before the start of vitamin D supplementation among pregnant women in Finland. In addition, we were able to explore the possible effects of matching for season of blood collection on the results. The reason why we selected the opposite season control group was to determine how much variation in risk estimates can occur when such controls are used as compared to when same season controls are used. We believe that this may have important repercussions on studies examining vitamin D/disease associations especially in serial sample studies. In this study, deliberate opposite season controls showed lower, close to unit risk estimates when compared to season-matched controls. This could mean that if an effect exists, not taking into account the input of seasonality probably confounds the estimation. In epidemiological studies ascertaining vitamin D disease associations, adequate matching of cases and controls for time of sample collection and/or statistical control is essential. In conclusion, although we did not observe statistically significant association between serum 25-OHD concentrations and risk of ovarian cancer, low serum 25-OHD concentrations may be associated with an increased risk of developing the disease. As few modifiable factors are known to reduce ovarian cancer risk, further prospective data on vitamin D and ovarian cancer are of interest. Conflict of interest statement None declared. Acknowledgements This study was supported by a grant from the Cancer Control using Population-based Registries and Biobanks (CCPRB) European Union Network . We thank Tapio Luostarinen of the Finnish Cancer Registry and Aini Bloigu of the National Institute for Health and Welfare for the advice during the preparation of this manuscript. References 1 M.N. Barnes W.E. Grizzle J.G. Clinton E.E. Patridge Paradigms for primary prevention of ovarian carcinoma CA Cancer J Clin 52 2002 216 225 2 I.G. Runnebaum E. Stickeler Epidemiological and molecular aspects of ovarian cancer risk J Cancer Res Clin Oncol 127 2001 73 79 3 D.M. Freedman M. Dosemeci K. McGlynn Sunlight and mortality from breast, ovarian, colon, prostate, and non-melanoma skin cancer: a composite death certificate based case–control study Occup Environ Med 59 2002 257 262 4 W.B. Grant An estimate of premature cancer mortality in the US due to inadequate doses of solar ultraviolet-B radiation Cancer 59 2002 257 262 5 E.S. Lefkowitz C.F. Garland Sunlight, Vitamin D and ovarian cancer mortality rates in US women Int J Epidemiol 23 1994 1133 1136 6 W.B. Grant Ecological studies of solar UV-B radiation and cancer mortality rates Recent Results Cancer Res 164 2003 371 377 7 T. Mizoue Ecological studies of solar radiation and cancer mortality in Japan Health Phys 87 2004 532 538 8 T.E. Robsahm S. Trelti A. Dahlback J. Moan Vitamin D3 from sunlight may improve the prognosis of breast, colon, and prostate cancer (Norway) Cancer Causes Control 15 2004 149 158 9 C.F. Garland S.B. Morh E.D. Gorham W.B. Grant F.C. Garland Role of ultraviolet B irradiance and vitamin D in prevention of ovarian cancer Am J Prev Med 31 6 2006 512 514 10 P. Tuohimaa E. Pukkala G. Scelo Dose solar exposure, as indicated by the non-melanoma skin cancers, protect from solid cancers: vitamin D as a possible explanation Eur J Cancer 43 2007 1701 1712 11 E. Bidoli C. La Vecchia R. Talamani Micronutrients and ovarian cancer: a case–control study in Italy Ann Oncol 12 2001 1589 1593 12 E. Salazar-Martinez E.C. Lacano-Ponce G. Gonzalez Lira-Lira P. Escudero-De los Rios M. Hernandez-Avila Nutritional determinants of epithelial ovarian cancer risk: a case–control study in Mexico Oncology 63 2002 151 157 13 J.M. Gerkinger D.J. Hunter D. Spiegelman Diary products and ovarian cancer: a pooled analysis of 12 cohort studies Cancer Epidemiol Biomarkers Prev 15 2006 364 372 14 M.T. Goodman A.H. Wu K.H. Tung Association of diary products, lactose, and calcium with the risk of ovarian cancer Am J Epidemiol 156 2002 148 157 15 S.S. Tworoger I.-M. Lee J.E. Buring B. Rosner B.W. Hollis S.E. Hankinson Plasma 25-hydroxyvitamin D and 1, 25 dihydroxyvitamin D and the risk of incident ovarian cancer Cancer Epidemiol Biomarkers Prev 16 2007 783 788 16 E. Giovanucci The epidemiology of vitamin D and cancer incidence and mortality: a review (United States) Cancer Causes Control 16 2005 83 95 17 M.H. Ahonen Y.H. Zhuang R. Aine T. Ylikomi P. Tuohimaa Androgen receptor and vitamin D receptor in human ovarian cancer: growth stimulation and inhibition by ligands Int J Cancer 86 2000 40 46 18 A. Agic H. Xu C. Altgassen Relative expression of 1, 25-dihydroxyvitamin D3 receptor, vitamin D1 alpha-hydroxylase, vitamin D 24-hydroxylase, and vitamin 25-hydroxylase in endometriosis and gynaecological cancers Reprod Sci 14 2007 486 497 19 A.W. Norman From vitamin D to hormone D: fundamentals of the vitamin D endocrine system essential for good health Am J Clin Nutr 88 Suppl. 2008 491s 499s 20 P. Koskela T. Antilla T. Bjørge Chlamydia trachomatis infection as a risk factor for invasive cervical cancer Int J Cancer 85 2000 35 39 21 L. Teppo E. Pukkala M. Lehtonen Data quality and quality control of a population-based cancer registry. Experience in Finland Acta Oncol 33 1994 365 369 22 M.F. Hollick Vitamin D deficiency N Engl J Med 357 2007 266 281 23 R. Vieth H. Bischoff-Ferrari B.J. Boucher The urgent need to recommend an intake of vitamin D that is effective Am J Clin Nutr 85 2007 649 650 24 B. Dawson-Hughes R.P. Heaney M.F. Holick P. Lips P.J. Meunier R. Vieth Estimates of optimal vitamin D status Osteoporos Int 16 2005 713 716 25 S. Miettinen M.H. Ahonen Y.-.L. Lou Role of 24-hydroxylase in vitamin D3 growth response of OVCAR-3 ovarian cancer cells Int J Cancer 108 2004 367 373 26 S.F. Rashid J.C. Mountford A.F. Gombart M.J. Campbell 1 alpha, 25-dihydroxyvitamin D(3) displays divergent growth effects in both normal and malignant cells Steroids 66 2001 433 440 27 V.W. Chen B. Ruiz J.L. Killeen Pathology and classification of ovarian tumours Cancer 97 2003 2631 2642 28 S.S. Harris B. Dawson-Hughes The association of oral contraceptive use with plasma 25-hydroxyvitamin D levels J Am Coll Nutr 17 3 1998 282 284 29 C.W. Thane C.J. Bates A. Prentice Oral contraceptives and nutritional status in adolescent British girls Nutr Res 22 2002 449 462 30 Collaborative Group on Epidemiological Studies of Ovarian Cancer V. Beral R. Doll C. Hermon R. Peto G. Reeves Ovarian cancer and oral contraceptives: collaborative reanalysis of data from 45 epidemiological studies including 23 257 women with ovarian cancer and 87 303 controls Lancet 371 9609 2008 303 314 31 G.E. Mullin A. Dobs Vitamin D and its role in cancer and immunity: a prescription for sunlight Nutr Clin Pract 22 2007 305 322 32 International Agency for Research on Cancer, World Health Organization. Vitamin D and cancer , vol. 5. IARC working group report, 2008.