Lactase persistence and ovarian carcinoma risk in Finland, Poland and Sweden



Ovarian carcinoma is the fourth most common cause of cancer death in women. The cause and pathogenesis of this disease has remained obscure. Galactose, the hydrolyzing product of the milk sugar lactose, has been hypothesized to be toxic to ovarian epithelial cells and consumption of dairy products and lactase persistence has been suggested to be a risk factor for ovarian carcinoma. In adults, downregulation of lactase depends on a variant C/T−13910 at the 5′ end of the lactase gene. To explore whether lactase persistence is related to the risk of ovarian carcinoma we determined the C/T−13910 genotype in a cohort of 782 women with ovarian carcinoma. The C/T−13910 genotype was defined by solid phase minisequencing from 327 Finnish, 303 Polish, 152 Swedish patients and 938 Finnish, 296 Polish and 97 Swedish healthy individuals served as controls. Lactase persistence did not associate significantly with increased risk for ovarian carcinoma in the Finnish (odds ratio [OR] = 0.77, 95% confidence interval [CI] = 0.57–1.05, p = 0.097), in the Polish (OR = 0.95, 95% CI = 0.68–1.33, p = 0.75), or in the Swedish populations (OR = 1.63, 95% CI = 0.65–4.08, p = 0.29). Our results do not support the hypothesis that lactase persistence increases the ovarian carcinoma risk. On the contrary, lactase persistence may decrease the ovarian carcinoma risk at least in the Finnish population. © 2005 Wiley-Liss, Inc.

Ovarian cancer is the sixth most common cancer among women worldwide with the second highest mortality rate among gynecological malignancies.1 The highest ovarian cancer incidence rates are found in Northern and Western Europe.2 More than 90% of the cases of ovarian cancer develop sporadically.3 The molecular pathogenesis is poorly known but inactivation of p53 tumor suppressor gene has been reported frequently.4 Depending on the population, 3.3–13.5% of epithelial ovarian cancer cases are explained by germline mutations of the BRCA1 gene, the BRCA2 gene probably contributes to a lesser extent.5, 6, 7, 8, 9, 10, 11, 12, 13 Multiparity, lactation, use of oral contraceptives, tubal ligation and hysterectomy are associated with decreased risk of ovarian carcinoma.14, 15 The etiology of ovarian carcinoma is unknown but frequent ovulation and hormonal factors have been suggested to play a role in its pathogenesis.16 Also, dietary factors have been implicated as causal or contributing factors.17

Disaccharide lactose is found in milk and other dairy products. It is hydrolyzed to galactose and glucose by lactase-phlorizin hydrolase (LPH) in the intestinal mucosa. Women with galactosemia suffer from early ovarian failure.18 Galactosemia is an autosomal recessive disorder with almost total absence of galactose-1-phosphate uridyltransferase (GALT) activity leading to the metabolites of galactose accumulating in the ovary and other tissues. GALT belongs to the Leloir pathway together with several enzymes that are responsible for the conversion of galactose to glucose.19, 20 Premature ovarian failure has been suggested to be due to direct toxicity of galactose, its metabolites or elevated gonadotropin levels.21, 22, 23 Additionally, animal studies have shown that galactose is poisonous to oocytes.24, 25 Lactose, a main source of galactose, was then hypothesized to play a role in ovarian tumorigenesis.

Lactase activity is high in infants but may decrease remarkably during childhood. This state is known as adult-type hypolactasia or lactase non-persistence and represents a normal physiological condition after weaning. The majority of northern Europeans have the ability to maintain lactase activity and digest lactose throughout life (lactase persistence).26 Based on epidemiological observations there is higher incidence of ovarian cancer in populations where lactase persistence is common.27 This was supported by a case-control study where the consumption of dairy products and lactase persistence were observed to elevate ovarian carcinoma risk28 but 2 studies reported no association.29, 30 It is noteworthy that the effect of lactase persistence on ovarian carcinoma has been based on indirect lactose tolerance tests or self-reports whose accuracy has been challenged.31, 32

A variant C/T−13910, locating 13.9 kb upstream of the LPH gene, has been shown to associate with the adult-type hypolactasia trait. The CC−13910 genotype is associated with lactase non-persistence whereas CT−13910 and TT−13910 genotypes are associated with lactase persistence.33 The developmental downregulation of lactase activity operates at the transcriptional level. The T−13910 allele allows lactase gene expression and maintains lactase activity whereas the C−13910 allele lets the downregulation operate normally.34 In addition, in vitro studies have shown that the T−13910 allele increases the lactase promoter activity.35, 36 The frequency of the CC−13910 genotype is in agreement with the reported prevalence of adult-type hypolactasia in several populations including Finns, French, North Americans, African-Americans and Somalians.32, 33

To assess whether lactase persistence is a risk factor for ovarian carcinoma we have determined the C/T−13910 variant in Finnish, Polish and Swedish women with ovarian carcinoma and defined its relation to their corresponding control populations by solid phase minisequencing.37

Material and methods


The Finnish sample set consisted of 327 patients with epithelial ovarian cancer treated at the Department of Obstetrics and Gynaecology, Helsinki University Central Hospital between 1989–1998. The clinical diagnosis of ovarian carcinoma in the Finnish patients was made at the mean age of 56 years ± 14. About a quarter of the Finns reside in the Helsinki region and due to recent migration patterns in Finland the patients have ancestral origins throughout Finland from early and late settlements regions. Finnish controls consisted of 938 anonymous blood donors from Western and Eastern Finland reflecting Finnish population mixture in early and late settlements regions. The analysis of the Finnish control material was reported by Enattah et al.33

Polish samples were obtained from 303 consecutive patients with ovarian carcinoma treated at the Department of Chemotherapy, Regional Oncology Centre, Szczecin, between 2001–2004. The mean age of diagnosis was 54 (range = 21–80) years. The control material consisted of 296 unselected adults in the region of Szczecin representing the same population as the cases. The mean age of controls was 70.8 (range = 22–80) years.

The 152 Swedish patients lived in the southern Swedish health care region and were diagnosed with ovarian carcinoma between 1998–2000. The mean age of diagnosis was 59 (range = 22–82). This patient material has previously been analyzed for germline mutations in BRCA1 and BRCA2.7 The control samples were obtained from healthy individuals and from patients without known malignancy operated on at the Dept. of Orthopaedics at the University Hospital in Lund, representing the southern Swedish population.

Histological classification

Ovarian tumors were classified as serous, mucinous, endometrioid or other (includes undifferentiated, mixed epithelial tumors). The histopathological characteristics of these cases are illustrated in Table I. All patient materials were approved by the local ethical committees and written informed consent was obtained from the patients and controls.

Table I. Lactase Persistence, Defined the C/T−13910 Variant, and its Risk for Ovarian Carcinoma Pathogenesis in the Finnish, Polish and Swedish Populations and Histopathological Characteristics
Finnish patientsnLactase persistence (%)Genotype for C/T−13910OR95% CIp-trend
  • a

    Mantel-Haenszel common OR estimates, CI.

  • b

    Fisher's exact test, otherwise Pearson's χ2 test, CC−13910, lactase non-persistence, CT−13910 and TT−13910, lactase persistence.

 Serous179 (55%)77.7 (139/179)4067720.770.52–1.140.185
 Mucinous56 (17%)80.4 (45/56)1126190.910.46–1.790.78
 Endometrioid34 (10%)79.4 (27/34)715120.850.37–1.990.72
 Other58 (17%)74.1 (43/58)1528150.640.35–1.170.14
 Total327 (100%)77.7 (254/327)731361180.770.57–1.050.097
 Control population93881.9 (768/938)170446322   
Polish patients
 Serous138 (45.5%)63.0 (87/138)5167200.840.55–1.290.43
 Mucinous23 (8%)56.5 (13/23)10940.640.27–1.520.31
 Endometrioid13 (4%)61.5 (8/13)5530.790.25–2.480.77b
 Other129 (42.5%)70.5 (91/129)3867241.190.76–1.860.46
 Total303 (100%)65.7 (199/303)104148510.950.68–1.330.75
 Control population29666.9 (198/296)9813662   
Swedish patients
 Serous85 (56%)90.6 (77/85)829481.110.42–2.950.84
 Mucinous8 (5%)100 (8/8)0440.900.84–0.961.00b
 Endometrioid20 (13%)100 (20/20)04160.900.84–0.960.21b
 Other39 (26%)94.9 (37/39)213242.130.44–10.180.51b
 Total152 (100%)93.4 (142/152)1050921.630.65–4.080.29
 Control population9789.7 (87/97)103255   
All patientsa
 Serous402 (51%)75.4 (303/402)991631400.830.63–1.090.17
 Mucinous87 (11%)75.9 (66/87)2139270.850.50–1.430.53
 Endometrioid67 (9%)82.1 (55/67)1224311.030.53–1.990.94
 Other226 (29%)75.7 (171/226)55108631.020.71–1.450.93
 Total782 (100%)76.1 (595/782)1873342610.880.71–1.100.26
All controls1,33179.1 (1053/1331)278614439   

PCR and solid phase minisequencing

PCR reactions were carried out in a total volume of 50 μl containing PCR buffer (Dynazyme; Finnzymes, Espoo, Finland) containing 10 mmol/l Tris-HCl (pH = 8.8 at 25°C), 1.5 mmol/l MgCl2, 50 mmol/l KCl, and 0.1% Triton X-100, 10 nmol deoxynucleotide triphosphates (dNTP) (Amersham Pharmacia Biotech, Little Chalfont, Buckinghamshire, UK), 50 pmol forward (5′-CCTCGTTAATACCCACTGACCTA-3′) and 5 pmol biotinylated reverse (5′-GTCACTTTGATATGATGAGAGCA-3′) primers amplifying the C/T−13910 variant. All described PCR-reactions were initiated by a hot start with 0.8 U of Taq polymerase and carried out in the following conditions: after denaturation at +94°C for 4 min the PCR was initiated by a hot start, the samples were cycled 34 times with a 30-sec annealing step at +53°C, a 75-sec elongation step at +72°C and a 30-sec denaturing step at +94°C followed by the last 30-sec annealing step at +53°C and 10 min extension at +72°C. PCR reactions were electrophoresed with 1 kb DNA ladder size standards (GeneRuler, Vilnius, Lithuania) on a 1.5% agarose gel with ethidium bromide.

Ten microliters of the biotinylated PCR products were immobilized on a streptavidin-coated solid support by incubation with 40 μl of 0.15 mol/l NaCl, 20 mmol/l Na2SO4 (pH 7.5), and 0.1% Tween 20 in streptavidin coated microtiter wells for 90 min at 37°C (Titramax 1000, Heidolph, Schwabach, Germany). The wells were washed 6 times with 350 μl of 40 mmol/l Tris-HCl, pH = 8.8, 1 mmol/l EDTA, 50 mmol/l NaCl and 0.1% Tween 20 using an automated washer (BW50, Biohit Plc, Helsinki,Finland). To remove the unbiotinylated complementary strand of the immobilized PCR product, the wells were treated with 100 μl of 50 mmol/l NaOH for 5 min and washed 6 times as described earlier. Fifty microliters of minisequencing reaction mix containing 20 pmol of the detection primer (5′-GGCAATACAGATAAGATAATGTAG-3′), 0.5 U of Taq polymerase, 2 pmol tritium labeled dNTP: 3H-dTTP (111 Ci/mmol) or 3H-dCTP (64 Ci/mmol) (Amersham Pharmacia Biotech) in PCR buffer (Dynazyme, Finnzymes, Espoo, Finland). The minisequencing reaction was allowed to continue for 15 min at 56°C. The wells were washed 6 times as described above and treated with 60 μl of 50 mmol/l NaOH for 5 min to release tritium labeled detection primers. Fifty microliters of NaOH solution containing eluted detection primers were added to 100 μl of scintillation liquid (Perkin Elmer, Wellesley, MA). The incorporated radioactivity of 3H-labeled nucleotides was measured using a scintillation counter (1450 Microbeta, Perkin Elmer).

Statistical analysis

Lactase persistence and its relation to ovarian carcinoma were evaluated by unconditional logistic regression and expressed as odds ratio (OR) with 95% confidence interval (CI). Common OR were estimated by Mantel-Haenszel statistics. Furthermore, the risk was evaluated by 2-sided Pearson's χ2 or Fischer exact tests. Statistical analyses were carried out by SPSS for Windows, Release 11.5.1 (SPSS Inc., Chicago, IL).


The distribution of lactase genotypes in ovarian carcinoma patients and their corresponding controls according to histological subtype is shown in Table I. In the Finnish patients, lactase persistence had a modestly decreased effect on ovarian cancer risk (OR = 0.77, 95% CI = 0.57–1.05). The probability for the trend (0.097), however, was not statistically significant.

Lactase persistence had no significant effect on the risk of ovarian carcinoma in the Polish patients (OR = 0.95, 95% CI = 0.68–1.33, p = 0.75) or the Swedish patients (OR = 1.63, 95% CI = 0.65–4.08, p = 0.29). When OR was evaluated for all 3 populations lactase persistence showed a tendency of decreased, although not significant, risk for ovarian carcinoma (OR = 0.88, 95% CI = 0.71–1.10, p = 0.26) The results were not dependent on tumor histology (Table I).

The frequency of lactase non-persistence, defined by the CC−13910 genotype, in the Polish controls was observed to be 33.1%. This frequency figure is a somewhat lower than the previously published (37.5%) in Poland by breath hydrogen assay.38 The frequency of lactase non-persistence was 10.3% in the Swedish control population, which is in agreement with the frequency observed recently (9.6%) in Swedish Caucasian children (T.K. Nilsson, personal communication).


Our results indicate that lactase persistence, defined by the C/T−13910 variant associated with the lactase persistence/non-persistence trait,32, 33, 34, 39 may decrease the ovarian carcinoma risk in the Finnish population. Lactase persistence did not have effect on the ovarian carcinoma risk in the Polish or Swedish populations. Earlier, results on lactase persistence and ovarian carcinoma risk that have been obtained using lactose tolerance tests, self-reports and milk products consumption data have been conflicting. Lactase persistence and lactose consumption were shown to associate with increased ovarian carcinoma risk in Sardinia, Italy28 and lactose intolerance has been suggested to indicate protective effect for ovarian carcinoma at an early age in eastern United States40 but in studies with subjects from Ontario, Canada29 and with Caucasians from Washington (United States)30 no such association was found.

Our results are in agreement with previous studies in which lactose consumption had a protective effect on ovarian carcinoma.41, 42, 43, 44 There is evidence that milk consumption is more abundant among subjects with lactase persistence than in subjects with lactase non-persistence.32, 45 In Finland with the highest milk consumption rate (177.9 kg/year/person, which equals to 23.4 g of lactose/day/person) the association between lactase persistence and ovarian carcinoma was negative. The corresponding figures of milk consumption are 142.4 kg/year/person in Sweden and 62.7 kg/year/person in Poland.46 In 2 cohort studies lactose consumption was associated with increased ovarian carcinoma risk47, 48 especially for that of the serous subtype.48 Several studies, however, have failed to find a clear relation between lactose consumption and ovarian carcinoma.29, 30, 49, 50, 51, 52, 53

Milk is an important source of vitamin D (0.5 μg/100 g of milk). About a half of the recommended daily intake of vitamin D (5 μg),54 added in non- or low-fat milk, is satisfied by the use of fluid milk in Finland. It is of interest that vitamin D has been shown to inhibit the growth of a human ovarian cancer cell lines.55, 56 There are studies that dietary vitamin D42 with the highest quintile57 and sunlight-induced vitamin D formation58 are protective factors for ovarian carcinoma. Also, calcium intake is closely related to milk consumption (120 mg/100 g of milk). Recommended daily intake of calcium is 800–900 mg in women but increasing the calcium intake to 500–1,000 mg per day may prevent osteoporosis.54 Calcium consumption has been detected to inversely associate with ovarian carcinoma risk.41, 57 Not all studies have confirmed the possible protective effect of vitamin D or calcium.41, 42, 47, 50, 59 Calcium as well as dietary vitamin D and endogenous vitamin D formation downregulate the production of parathyroid hormone (PTH).60, 61 PTH participates in dietary calcium absorption stimulating 1,25 dihydroxyvitamin D formation. It has been speculated that PTH can promote cancer indirectly, increasing IGF-1 synthesis or directly, acting as a co-mitogen in preneoplastic lesions expressing the PTH receptor.62 Lactase persistence and high consumption of milk might partly explain our observed protective influence for ovarian carcinoma in the Finnish population. Unfortunately, due to lack of data for milk consumed by the study subjects we could not draw further conclusions.

GALT activity has been observed to be lower in patients with ovarian carcinoma than in controls. Furthermore, reduced activity of GALT together with lactose consumption has been shown to associate with ovarian carcinoma.63 Later studies have not been able to confirm these observations30, 40 except among women with abnormally low GALT activity.52 Polymorphism N314D of GALT, can reduce the activity to 74–43% from normal GALT activity.64, 65 N314D frequency was higher in women with serous subtype40 and endometrioid or clear cell ovarian cancers than controls.66 The N314D polymorphism showed little evidence for ovarian carcinoma risk44 together with galactose intake.67

Lactase persistence allows the efficient use of milk as nutrition after childhood. It has been thought to have been a benefit since dairying began around 9,000 years ago especially in Northern Europe where the amount of sunlight is limited. We found that women with lactase persistence may have a moderately decreased risk for the development of ovarian carcinoma in the Finnish population. We detected no association in the Polish or Swedish populations. The results do not support the hypothesis that lactase persistence increases ovarian carcinoma risk.


H. Komu is acknowledged for her technical assistance. The financial support of the Finnish Cultural Foundation, the Sigrid Jusélius Foundation, Helsinki, Helsinki University Research Funding, the Helsinki University Science Foundation, and the Swedish Cancer Fund are gratefully acknowledged.