Med Oncol (2012) 29:1308–1313 DOI 10.1007/s12032-011-9840-0

ORIGINAL PAPER

Familial diabetes is associated with reduced risk of cancer in diabetic patients: a possible role for metformin Lev M. Berstein • Marina P. Boyarkina Svetlana Yu. Teslenko

•

Received: 17 December 2010 / Accepted: 19 January 2011 / Published online: 6 February 2011 Ó Springer Science+Business Media, LLC 2011

Abstract Type 2 diabetes mellitus (DM2) is a risk factor of a number of malignancies. Therefore, it is important to identify factors linking DM2 and cancer within family units and how current treatment regimens influence the development of cancer in DM2 patients. The present casecontrolled study was designed to assess DM2 prevalence among parents or siblings of (a) cancer patients who did not have diabetes (n = 77; age 59.3 ± 1.3 years) or (b) had overt (n = 197; 63.7 ± 0.6 years) or latent (n = 25; 61.5 ± 1.5 years) DM2 and (c) of female DM2 patients without cancer (n = 172; 61.7 ± 0.6 years). In the families of cancer-free DM2 women, DM2 was found to be significantly more frequent (30.8 ± 3.5%) than in families of cancer patients without diabetes (in all patients: 6.5 ± 2.8%; in female patients: 5.0 ± 3.4%). More importantly, DM2 in families of cancer-free DM2 women was more frequent than in the families of DM2 patients having mammary (9.5 ± 4.5%), endometrial (6.3 ± 4.1%) or any other cancer (in all: 15.2 ± 2.6%; in women: 12.9 ± 2.8%). Additionally, DM2 patients without cancer, who had parents or siblings with DM2, received biguanide metformin versus sulfonylurea derivatives more often than those with breast or endometrial cancer, either with or without family history of DM2. Our data indicate that familial DM2 may have a protective effect for some cancer

L. M. Berstein (&)
M. P. Boyarkina Laboratory of Oncoendocrinology, N.N. Petrov Research Institute of Oncology, Pesochny-2, St., 197758 Petersburg, Russia e-mail: [email protected] S. Yu. Teslenko L.G. Sokolov Clinical Hospital, St. Petersburg, Russia

123

types and that the type of anti-diabetes therapy may be a factor of influence in the associations observed. Keywords Cancer
Diabetes
Family history
Gender
Metformin

Introduction Type 2 diabetes mellitus (DM2) occurs at rather high frequency in many countries and increases the risk of several malignancies, including pancreas, endometrium, liver, breast, and colon. In addition, DM2 portends poor prognosis for cancer patients [1, 2]. The causes of these associations remain unclear. Both cancer and DM2 are regarded as multifactorial chronic conditions or main human non-communicable diseases [3–5]. Excess body weight, among other factors, can promote both diabetes and cancer [6] while factors like e.g. asbestos apparently contribute only to cancer [7]. On the other hand, there are poorly described factors that presumably exert unidirectional effects toward both cancer and DM2; however, the available evidence is inconclusive. The last group of factors may be exemplified by diabetes in the kin of DM2 patients. For example, a family history of DM2 is associated with an increased risk of developing DM2 [8, 9]. As mentioned earlier, DM2 increases risk of cancer, and many cancers also show familial clustering, which is not always genetic [10, 11]. Therefore, one could suspect that DM2 increases the relative risk of getting cancer in families having DM2 susceptibility. This notwithstanding, the data available on the role of familial DM2 as a modifier of cancer risk are scarce and ambiguous [12–14]. The objective of the present paper is to supplement the available evidence on this subject with the results of a case–control study of cancer and/or DM2 patients.

Med Oncol (2012) 29:1308–1313

1309

Patients and methods Data on familial diabetes history were obtained using a questionnaire in 2009–2010 among patients admitted to N.N. Petrov Research Institute of Oncology or followed up at the Outpatient Division of L.G. Sokolov Clinical Hospital, St. Petersburg, Russia. A total of 471 patients aged 62.3 ± 0.4 included 374 women and 97 men who could be broken down further as follows: (group a) 77 cancer patients without type 2 diabetes mellitus (DM2), among them 40 women (10 endometrial and 9 breast cancers, 4 melanoma, and 17 other malignancies) and 37 men (7 melanoma, 7 lymphoma, 5 prostate, 5 kidney, 4 lung, 4 gastrointestinal, and 5 other cancers); (group b) cancer patients with overt DM2 (139 women and 58 men) or with impaired glucose tolerance (23 women and 2 men); among these 222 patients were 162 women (48 breast, 37 endometrial, 12 colorectal, 10 ovarian, 8 gastric, 6 lung, 4 cervical cancers, 7 melanomas, 5 lymphomas, and 25 other—more rare—cancer cases) and 60 men (9 prostate, 8 gastric, 8 lung, 7 colorectal, 4 kidney cancers, 7 melanomas, 5 lymphomas, and 12 other cancer cases); (group c) 172 cancer-free postmenopausal women with DM2. The status of carbohydrate metabolism in the patients was assessed using WHO criteria [15]. All probands included in the study answered questions about DM2 in their mothers, fathers, siblings, etc. At the time, diabetic cancer patients and DM2 patients without cancer showed no signs of uncontrolled DM2, presumably due to diet, peroral antidiabetics and/or, rarely, insulin Table 1 Family history of diabetes (DM2) in study subjects

therapy. In some of the subjects, anthropometric parameters such as body weight, height and body mass index (BMI) were determined. In female DM2 patients without cancer, Gail coefficients were calculated to assess mammary cancer risk [16]; the results were compared against the presence or absence of positive family history of DM2. The results were presented as group mean ± standard errors (M ± m), and the differences were assessed with Student’s t-test using the software package SigmaPlot for Windows (SPSS Inc., Chicago, IL, USA). The study was approved by the local Ethics Committee.

Results In cancer-free DM2 women, the frequency of DM2 in family members was 30.8 ± 3.5%. By contrast, in cancer patients without DM2, the frequency of familial DM2 was only 6.5 ± 2.8% (5.0% in women, and 8.1% in men) (Table 1). Taken as point of departure, presented data show that if negligible differences in the age of study subjects are disregarded, the prevalence of family history of DM2 in cancer-free DM2 patients is also significantly higher than in the whole group of cancer patients with DM2 (15.2 ± 2.6%) and in diabetic women with different malignancies, 12.9 ± 2.8% (see Table 1 and below). Indeed, particularly in the postmenopausal DM2 patients with hormone-dependent tumors, such as endometrial and mammary cancer, the prevalence of DM2 among their kin was relatively low (6.3 ± 4.1% and 7.5 ± 4.2%,

Group Cancer patients without DM2

Prevalence of family diabetes, % (M ± m)

77

59.3 ± 1.3

6.5 ± 2.8

40

60.4 ± 1.7

5.0 ± 3.4

M

37

58.1 ± 2.0

8.1 ± 4.5

25

61.5 ± 1.5

20.0 ± 8.0 21.7 ± 8.6

P

\0.01

F

23

61.6 ± 1.6

M

2

60.5 ± 1.5

197

63.7 ± 0.6

15.2 ± 2.6

\0.01

F

139

63.7 ± 0.8

12.9 ± 2.8

\0.01

M

58

64.0 ± 0.8

18.9 ± 5.3

Cancer patients with DM2

Cancer patients with DM2 or IGT

P value is given versus group marked with asterisk (*)

Age (M ± m)

F Cancer patients with IGT

BC breast cancer, EC endometrial cancer, DM2 diabetes mellitus type 2, IGT impaired glucose tolerance, F females, M males, M mean value, m std. error of mean

Number

0

222

63.8 ± 0.6

15.8 ± 2.4

162 60

63.5 ± 0.7 64.6 ± 1.2

14.8 ± 2.8 18.3 ± 5.0

BC patients with DM2, all

42

59.5 ± 1,2

9.5 ± 4.5

\0.01

BC patients with DM2, postmenopausal

40

60.1 ± 1.2

7.5 ± 4.2

\0.01

EC patients with DM2 or IGT

37

62.3 ± 1.5

10.8 ± 5.1

\0.01

32

62.8 ± 1.7

6.3 ± 4.1

\0.01

172

61.7 ± 0.6

30.8 ± 3.5*

F M

EC patients with DM2 DM2 patients without cancer, postmenopausal women only

\0.01

123

1310

Med Oncol (2012) 29:1308–1313

Fig. 1 The prevalences of type 2 diabetes mellitus (DM2) in family histories of cancer and cancer-free patients. 1 Female cancer patients without DM2, 2 endometrial cancer patients with DM2, 3 cancer patients without DM2 (whole group), 4 postmenopausal breast cancer patients with DM2, 5 male cancer patients without DM2, 6 breast cancer patients with DM2 (whole group), 7 endometrial cancer patients with DM2 or impaired glucose tolerance (IGT), 8 female cancer patients with DM2, 9 female cancer patients with DM2 or IGT, 10 cancer patients with DM2 (whole group), 11 cancer patients with DM2 or IGT (whole group), 12 male cancer patients with DM2 or IGT, 13 male cancer patients with DM2, 14 cancer patients with IGT (whole group), 15 female cancer patients with IGT, 16 diabetic females without cancer

respectively). In other words, it was almost the same as in female cancer patients without DM2 (5.0 ± 3.4%); see additionally Fig. 1 showing main results in the ascending order. In this connection, it is worth mentioning that according to Gail methodology [16, 17] used to calculate mammary cancer risk in cancer-free postmenopausal DM2 female patients, the risk was higher than in healthy women without DM2; however, it was not significantly different from that in cancer-free postmenopausal DM2 patients with the family history of DM2 (Table 2). In 25 cancer patients with impaired glucose tolerance (23 women and 2 men), the prevalence of familial DM2 was somewhat higher (20.0 ± 8.0%) than in female cancer patients with overt DM2 (12.9 ± 2.8%); however, the

Table 2 The values of Gail coefficient (5-year prediction) in postmenopausal women Group

Gail coefficient, 5 years

Healthy women

1.37 ± 0.04 (n = 136)1

Diabetic women without cancer

1.58 ± 0.06 (n = 97)2

Diabetic women without cancer with family history of diabetes

1.66 ± 0.13 (n = 35)2

1

Data of M. Boiarkina et al. (17)

2

Present study

123

difference was not significant (P [ 0.3). Similarly, for male cancer patients with overt DM2 there was a higher prevalence of familial DM2 (18.9 ± 5.3%) when compared to male cancer patients without DM2 (8.1 ± 4.5%); however, the distinction was not statistically significant too (P [ 0.1). Our findings also show that parents of DM2-positive women with cancer had DM2 almost as often (76.2 ± 9.3% of diabetic kin) as those of DM2-positive women without cancer (78.4 ± 6.8%). However, while not significant (P [ 0.2) there is a trend toward mothers with DM2 being less prevalent in the group of women that are cancer-free and DM2 positive (56.8% of the kin vs. 71.4% in diabetic females with cancer). The mean body mass index (BMI) in postmenopausal DM2-positive women without cancer (28.9 ± 0.48) did not differ from that in the subgroup having DM2-positive kin (28.3 ± 0.77). In DM2-positive women with breast cancer, BMI was higher (31.9 ± 1.3, P \ 0.05); however, because of the small number of patients having DM2-positive kin among them, no distinct association with family DM2 history is apparent. Finally, to get a better insight into the causes of the relatively low prevalence of familial DM2 among diabetic cancer patients (Table 1; Fig. 1), we carried out a preliminary analysis of potential of an anti-diabetes therapy used for at least 12 months prior to study. Data were obtained from 96 cancer-free DM2 patients and 49 DM2 patients with mammary or endometrial cancer. Forty-three of them had family history of DM2. As a result, two main facts were established: (a) a ratio between the number of patients on biguanide metformin versus those using sulfonylurea derivatives or insulin was notably higher (2.22) in cancer-free DM2 patients than in DM2 patients with cancer (0.48); (b) cancer-free DM2 patients with DM2-positive kin used metformin 2.72 times more often than those with cancer; the latter value being only 1.57 when family DM2 association was absent (Fig. 2).

Discussion Our main conclusion is that although type 2 diabetes mellitus is recognized as a risk factor of cancer [1–3], familial predisposition to DM2 can, all other conditions being equal, contribute to relative reduction of cancer risk. The data on the issue are, to the best of our knowledge, scarce [12–14]. In a study carried out in Japan [13] (1997), stomach cancer demonstrated a familial covariation with diabetes and a borderline inverse covariation with hypertension. Those results prompted a suggestion that stomach cancer may share a common familial etiological factor with diabetes type 2 and hypotension [13]. A case–control study

Med Oncol (2012) 29:1308–1313

Fig. 2 Biguanide metformin use in cancer and cancer-free patients with or without family history of diabetes. Left panel The biguanide/ sulphonylurea preparations ratio: data on the treatment of DM2 patients with mammary or endometrial cancer (a) and cancer-free diabetics (b). Right panel The ratio of biguanide use in cancer-free DM2 patients versus DM2 patients with breast or endometrial cancer depending on the absence (c) or presence (d) of DM2-positive kin

[12] (2000) revealed a direct correlation between familial diabetes and colorectal adenomatous polyps and carcinoma, which was the highest among elderly (C60 years) probands, and still higher in cases of having diabetic siblings (odds ratio 5.7) rather than parents (OR 2.4). More recently, Hemminki et al. [14] examined about one half of all DM2 patients in Sweden and traced cancer incidence over time with medians ranging 9–15 years after the last hospitalization for type 2 diabetes. It was reported that, in DM2 patients, increased risk of cancers was associated with more than 20 localizations—mostly tumors of pancreas (OR 6.08), liver (OR 6.33), gall bladder (OR 2.62), extrahepatic bile ducts (OR 3, 63), kidney (OR 2.50), endometrium (OR 2.22), and thyroid. There were no detectable differences in total cancer incidence between diabetic patients featuring and not featuring family DM2 association (OR, respectively, 1.04 vs. 1.37). However, patients with familial DM2 revealed relatively lower risk of certain malignancies such as prostate and pancreas cancer, melanoma, small intestine carcinoma. Similar trends were identified in the incidence of mammary and endometrial cancer. Therefore, the results of our present case–control study coincide with the data obtained in the large-scale prospective study of Hemminki et al. [14] to a certain degree. Reasons for the discrepancies of the results of our study and the Swedish one [14], both demonstrating that familial diabetes may reduce cancer risk in DM2 patients, with the earlier studies, which suggest that familial DM2 somewhat increases cancer risk or causes a trend to such increase, are hard to figure out at present. One of the possible explanations is the focus on specific cancers, i.e. predominantly

1311

colon in [12] and stomach in [13]. According to Hemminki et al. [14], no association between DM2 and colorectal neoplasia was found, and the excess risk of stomach cancer was insignificantly lower in familial DM2 patients than in the whole sample (OR 1.19 vs. 1.51). The numbers of patients with these cancers in our study were too small to be conclusive. Apart from the problem of tumor types, there are other issues to be considered. First, the prevalence of familial DM2 in cancer-free DM2 patients (30.8%) is not uncommonly high as it may seem. It may vary depending on ethnic and other factors. In some cases, it was higher than 27–30% [9, 18], and at least in one report—over 40% [19]. Second, although our results do not point to any role of body mass index as a potential contributor to lower cancer risk in familial DM2 patients, it should be noted that, among obese humans without overt endocrine and metabolic disorders (the so-called metabolically healthy obese (MHO)), DM2-positive kin is relatively rare [20]. Considering the role of obesity as a risk factor for cancer [6, 21] as well as observations that in different forms of obesity, including MHO, their association with cancer risk [22, 23] may differ, it cannot be ruled out that these factors contribute to interrelationships between familial diabetes, body mass, and cancer risk. The size of kin (number of family members) with DM2 is known to significantly influence DM2 risk [9]. There are no data about its influence on cancer risk. As to the ‘‘composition’’ of this parameter, our data suggest that, in DM2 females with cancer, the prevalence of parents with DM2 was the same (about 75% of diabetic kin) as in cancer-free DM2 patients. Hence, there is no reason to suspect any changes in the parent/sibling ratio. We assessed the prevalence of familial DM2 in cancer patients of both sexes, while the group of cancer-free DM2 patients included only postmenopausal women. Surely, further studies should deal with data on men and women separately. However, the Swedish study showed that DM2 risk for men and women having DM2 kin is similar except for cases of two or more DM2 siblings associated with a higher DM2 risk for women [9]. On the other hand, variations revealed in the present study, such as somewhat higher prevalence of family DM2 in male cancer patients (Table 1 and Fig. 1), were small and insignificant suggesting that gender may be ignored in interpreting relatively lower cancer risk for DM2 patients having DM2-positive kin. Nevertheless, our team is now making an effort to collect data in male cancer-free diabetics required to carry out additional separate comparisons within male and female groups. Since families share many lifestyle features such as diet and exercise habits, it was not easy to establish whether genetic, behavioral, or environmental factors played a role

123

1312

in the present study. Calculations of Gail coefficient, an integral risk factor of mammary cancer, which is highly dependent on hereditary factors [16, 24], did not reveal any differences between cancer-free patients with or without family history of DM2 (Table 2). At the same time, our earlier data suggest that, in DM2-free women, the risk of mammary and endometrial cancer is increased by allelic polymorphisms associated with glucose intolerance and/or insulin resistance [25]. This evidence should be taken into account in analysis of the role of familial DM2 in cancer risk. When starting our study, we were not aware of any publications suggesting that differences in response to cancer treatment or antidiabetic therapy may be associated with the family history of diabetes. We also did not know whether the type of antidiabetic therapy modifies cancer risk in patients with family history of diabetes. The most popular antidiabetic therapies include biguanide metformin, different sulfonylurea derivatives, and insulin(s). A growing body of evidence suggests that cancer risk of DM2 is associated with the type of prior antidiabetic therapy. Several population studies demonstrated that, unlike other modalities of therapy, metformin can decrease cancer risk in DM2 patients [26, 27] and possibly modify the clinical course of cancer [28]. This is consistent with the present findings that showed that cancer-free DM2 patients used metformin rather than sulfonylureas to treat their diabetes, when compared with DM2 patients with cancer. Additionally, our preliminary observations suggest that due to the higher reliance on metformin in women with the family history of DM2 (Fig. 2), it is perhaps desirable that its use be promoted to reduce cancer risk in general as well as in DM2 patients with the familial aggregation of DM2, in particular. One of the limitations of our study is a relatively small number of patients with several different types of tumors, which makes it impossible to investigate thoroughly the results relating to any particular cancer type. However, the strengths of this study include (a) comparison of the two groups of cancer patients with or without DM2 as well as cancer-free DM2 patients on a single project; (b) confirmation of the Hemminki et al. [14] data with regard to a protective role of familial diabetes in relation to the risk of some cancers, and (c) a first suggestion of the importance of prior antidiabetic therapy (metformin vs. alternative treatment) in assessing the role of familial DM2 in cancer risk. The significance of main conclusions may also hold for other non-communicable diseases [5, 29] as evidenced by a report that maternal family history of diabetes is associated with reduced risk of cardiovascular diseases [18]. Very recently we discovered that according to O. Stephansson et al. [30] daughters of mothers with a

123

Med Oncol (2012) 29:1308–1313

lifetime history of diabetes were at a decreased risk of breast cancer which is in line with our main conclusion. Acknowledgments We thank Dr. Dmitry A. Vasilyev (N.N. Petrov Research Institute of Oncology, St. Petersburg) for help in data collection while Alexey Golubev (Research Institute of Experimental Medicine, St. Petersburg) and Robert A. Sikes (Center for Translational Cancer Research, University of Delaware, Newark, DE, USA) for valuable and friendly discussions. Conflict of interest

We declare that we have no conflict of interest.

References 1. Vigneri P, et al. Diabetes and cancer. Endocr Relat Cancer. 2009;16(4):1103–23. 2. Richardson LC, Pollack LA. Therapy insight: influence of type 2 diabetes on the development, treatment and outcomes of cancer. Natl Clin Pract Oncol. 2005;2(1):48–53. 3. Dilman VM. Development, aging and disease: a new rationale for an intervention strategy. Chur (Switzerland): Harwood Acad. Publ; 1994. p. 387. 4. Mathers CD, Loncar D. Projections of global mortality and burden of disease from 2002 to 2030. PLoS Med. 2006;3(11):e442. 5. Berstein LM. Cancer of hormone-dependent tissues in the frame of main non-communicable diseases. St. Petersburg: Aesculap; 2009. P. 180. 6. Hjarta˚ker A, Langseth H, Weiderpass E. Obesity and diabetes epidemics: cancer repercussions. Adv Exp Med Biol. 2008;630: 72–93. 7. Lotti M, Bergamo L, Murer B. Occupational toxicology of asbestos-related malignancies. Clin Toxicol (Philad). 2010;48(6): 485–96. 8. Hariri S, Yoon PW, Moonesinghe R, Valdez R, Khoury MJ. Evaluation of family history as a risk factor and screening tool for detecting undiagnosed diabetes in a nationally representative survey population. Genet Med. 2006;8(12):752–9. 9. Hemminki K, Li X, Sundquist K, Sundquist J. Familial risks for type 2 diabetes in Sweden. Diabetes Care. 2010;33(2): 293–7. 10. Wilson B, et al. Clinical utility of cancer family history collection in primary care. Evid Rep Technol Assess (Full Rep). 2009;179: 1–94. 11. Stein QP, Flanagan JD. Genetic and familial factors influencing breast, colon, prostate and lung cancers. S D Med. 2010:16–22. 12. Brauer PM, McKeown-Eyssen GE, Jazmaji V, et al. Familial aggregation of diabetes and hypertension in a case-control study of colorectal neoplasia/Amer. J Epidemiol. 2002;156(8):702–13. 13. Toyoshima H, Hayashi S, Hashimoto S, et al. Familial aggregation and covariation of diseases in a Japanese rural community: comparison of stomach cancer with other diseases. Ann Epidemiol. 1997;7(7):446–51. 14. Hemminki K, Li X, Sundquist J, Sundquist K. Risk of cancer following hospitalization for type 2 diabetes. Oncologist. 2010; 15(6):548–55. 15. Definition and Diagnosis of Diabetes Mellitus and Intermediate Hyperglycaemia. Report of a WHO/IDF consultation. Geneva, Switzerland; 2008. P. 50. 16. Gail MH, Brinton LA, Byar DP, et al. Projecting individualized probabilities of developing breast cancer for white females who are being examined annually. J Natl Cancer Inst. 1989;81(24): 1879–86.

Med Oncol (2012) 29:1308–1313 17. Boiarkina MP, Tsyrlina EV, Poroshina TE, Semiglazov VF, Bershteıˇn LM. Gail coefficient as a risk factor and prognosticator for the course of breast cancer: correlation with hormonal and metabolic parameters and clinical and morphologic features. Vopr Onkol. 2008;54(1):25–30. 18. Bruce DG, Van Minnen K, Davis WA et al. Maternal family history of diabetes is associated with a reduced risk of cardiovascular disease in women with type 2 diabetes: the Fremantle diabetes study. Diabetes Care. 2010 Apr 5. [Epub ahead of print]. 19. Papazafiropoulou A, Sotiropoulos A, Skliros E, et al. Familial history of diabetes and clinical characteristics in Greek subjects with type 2 diabetes. BMC Endocr Disord. 2009;9:12. 20. Velho S, Paccaud F, Waeber G, Vollenweider P, Marques-Vidal P. Metabolically healthy obesity: different prevalences using different criteria. Europ J Clin Nutr. 2010;64(10):1043–51. 21. Calle EE, Thun MJ. Obesity and cancer. Oncogene. 2004;23(38): 6365–78. 22. Berstein LM. Macrosomy, obesity and cancer. New York: Nova Science Publishers; 1997. p. 199. 23. Calori G, Lattuada G, Piemonti L et al. Prevalence, metabolic features and prognosis of metabolically healthy obese Italian individuals: the Cremona Study. Diabetes Care. 2010 Oct 11. [Epub ahead of print].

1313 24. Gail MH, Mai PL. Comparing breast cancer risk assessment models. J Natl Cancer Inst. 2010;102(10):665–8. 25. Ulybina YuM, Imyanitov EN, Vasilyev DA, Berstein LM. Polymorphic markers associated with genes responsible for lipid and carbohydrate metabolism disorders and insulin resistance in cancer patients. Mol Biol. 2008;42(6):947–56. 26. Bowker SL, Yasui Y, Veugelers P, Johnson JA. Glucose-lowering agents and cancer mortality rates in type 2 diabetes: assessing effects of time-varying exposure. Diabetologia. 2010;53(8): 1631–7. 27. Libby G, Donnely LA, Donnan PT, et al. New users of metformin are at low risk of incident cancer: a cohort study among people with type 2 diabetes. Diabetes Care. 2009;32(9):1620–5. 28. Berstein LM, Boyarkina MP, Tsyrlina EV, Turkevich EA, Semiglazov VF. More favorable progesterone receptor phenotype of breast cancer in diabetics treated with metformin. Med Oncology. 2010, May 20 [ePub ahead of print]. 29. O’Neill SM, Rubinstein WS, Wang C, et al. Familial risk for common diseases in primary care: the family healthware impact trial. Am J Prev Med. 2009;36(6):506–14. 30. Stephansson O, Granath F, Ekbom A, Michels KB. Risk of breast cancer among daughters of mothers with diabetes: a populationbased cohort study. Breast Cancer Res. 2010;12(1):R14.

123