Joanne Wuu

Diet and cancer of the prostate: a case‐control study in Greece

The nutritional aetiology of prostate cancer was evaluated in Athens, Greece, through a case‐control study that included 320 patients with histologically confirmed incident prostate cancer and 246 controls without history or symptomatology of benign prostatic hyperplasia or prostate cancer, treated in the same hospital as the cases for minor diseases or conditions. Among major food groups, milk and dairy products as well as added lipids were marginally positively associated with risk for prostate cancer. Among added lipids, seed oils were significantly and butter and margarine non‐significantly positively associated with prostate cancer risk, whereas olive oil was unrelated to this risk. Cooked tomatoes and to a lesser extent raw tomatoes were inversely associated with the risk for prostate cancer. In analyses focusing on nutrients, rather than foods, polyunsaturated fats were positively and vitamin E inversely associated with prostate cancer. We conclude that several nutrition‐related processes jointly contribute to prostate carcinogenesis. Int. J. Cancer 80:704–708, 1999.

Transient increase in breast cancer risk after giving birth: postpartum period with the highest risk (Sweden)

Objective: Identify time-points when the elevated postpartum maternal breast cancer risk peaks. Methods: A case–control study nested within the Swedish Fertility Register included 34,018 breast cancer cases from the Swedish Cancer Register between 1961 and 1995. From the Fertility Register, 170,001 eligible subjects matched to the cases by age were selected as controls. Analysis contrasted risk between uniparous (7084 cases and 31,703 controls) and nulliparous (5411 cases and 22,580 controls) women and between biparous (13,239 cases and 65,858 controls) and uniparous women. Logistic regression analysis included indicator variables representing each year of age, ages at delivery, and time since delivery. Results: Comparing uniparous with nulliparous women the transient increase in maternal breast cancer risk peaked 5 years following delivery (odds ratio = 1.49, 95% confidence interval 1.01–2.20) and leveled off 15 years postpartum. Biparous women had a transient increase in risk that was lower at its peak than that of uniparous women, occurring about 3 years following second delivery. Conclusions: A time window of 5 years postpartum when maternal breast cancer risk is highest was observed. Establishing timing of peak transient increase in postpartum breast cancer risk may define the latent period required for pregnancy hormones in promoting progression of breast cells that have undergone early stages of malignant transformation.

The role of cell cycle regulatory proteins, cyclin D1, cyclin E, and p27 in thyroid carcinogenesis.

The cell cycle is controlled in part by cyclin-dependent kinases (CDKs), which are activated by forming complexes with cyclins. CDKs phosphorylate certain substrates to facilitate the proliferating cells through the cell cycle. CDK inhibitors (CDKIs) such as p27 inhibit cyclin-CDK complexes and function as a negative cell cycle regulator. The overexpression of the positive regulators (cyclins) or the underexpression of the negative regulators including p27 has been seen in a variety of neoplasms, but their role and interaction in thyroid carcinogenesis is yet to be established. We studied the expression of cyclins D1 and E, and the CDKI, p27 by immunohistochemistry in 116 cases, including 59 cases of follicular variant of papillary carcinoma (FVPC) and 57 cases of follicular adenoma (FA). The positive staining was divided into four grades: 1+ if less than 10%, 2+ if 11% to 25%, 3+ if 26% to 50%, and 4+ if greater than 50% of the nuclei of tumor cells stained positively. Cyclin D1 expression was seen in 37 (63%) FVPC and 34 (60%) FA. Cyclin E-positive cells were seen in 51 (86%) FVPC and 47 (82%) FA. No significant differences in the grade of cyclins D1 (P = .261) and E (P = .284) staining was seen between FVPC and FA. Of the 59 FVPC, 53 (89%) showed p27-positive cells; of these, 33 were 1+, nine were 2+, seven were 3+ and only four were 4+ positive. Conversely, all 57 FA were p27 positive, 53 were 4+, and four were 3+ positive. This difference in the grade of p27 staining between FVPC and FA was statistically significant (P < .001). This study shows a significant underexpression of p27 in FVPC compared with FA, suggesting that a decrease in p27 expression plays a more important role than overexpression of cyclins D1 and E alone in thyroid carcinogenesis and that p27 immunostaining may be helpful in the diagnosis of FVPC.

DIET AND BENIGN PROSTATIC HYPERPLASIA: A STUDY IN GREECE

OBJECTIVES To evaluate the nutritional etiology of benign prostatic hyperplasia (BPH) by conducting a case-control study in Athens, Greece. Despite the high morbidity and substantial human suffering produced by BPH, little research has been undertaken concerning the nutritional etiology of this disease. METHODS The study sample consisted of 184 patients with histologically confirmed BPH and 246 control patients without clinical evidence of prostate disease. All patients and controls were permanent residents of the greater Athens area. The data were modeled through unconditional logistic regression. RESULTS Among the food groups, fruits were inversely related to BPH risk, with a logistic regression-derived odds ratio of 0.79 per quintile increase and 95% confidence interval 0.67 to 0.93. Increased consumption of both butter and margarine was positively associated with BPH risk, and a marginally significant positive association was also evident for seed oils. No overall association was found with respect to consumption of olive oil. In analyses evaluating the role of nutrients rather than foods, zinc, an element selectively concentrated in the prostate gland, was significantly positively associated with BPH risk. CONCLUSIONS Our study provides evidence that, among added lipids, butter and margarine may increase the risk of BPH, and fruit intake may reduce this risk. Dietary zinc may play an important role in the etiology of BPH.

Childbearing at older age and endometrial cancer risk (Sweden)

Objectives: Several studies have found an inverse association between older age at last birth and endometrial cancer risk. A nested case-control study was undertaken to examine the influence of this and other aspects of reproductive patterns on the risk of developing endometrial cancer.Methods: Among women born in 1925 and later, 4,839 eligible patients were identified in the Swedish Cancer Register. For each case, five individually age-matched controls were randomly selected from a population-based Fertility Register. Relative risks were estimated from odds ratios obtained from conditional logistic regression analyses.Results: Compared to uniparous women, childless women were at a higher risk of endometrial cancer (odds ratio [OR]=1.38, 95% confidence interval [CI]=1.25–1.52). This association was stronger in younger (<50 years) than in older (50+ years) women. At all ages of first birth, a delivery was associated with a reduced risk of endometrial cancer that slowly diminished with time. Among parous women, the risk decreased by almost 20% for each additional live birth (OR=0.81, CI=0.78–0.84). In an analysis limited to women with two or more births that compared the independent effects of age at first and at last birth, only older age at last birth was associated with a lowered risk of endometrial cancer. The risk decreased at a rate of about 15% per five-year delay of last birth.Conclusions: Endometrial cancer is often referred to as the prototype hormonally-determined disease in women. However, our findings give further support to the hypothesis that a birth may not only affect risk through hormonal influences, but possibly also through mechanical shedding of cells that have undergone malignant transformation.

Pregnancy and risk of renal cell cancer: a population-based study in Sweden

Epidemiological findings indicate that hormonal influences may play a role in the etiology of renal cell cancer (RCC). The possible effect of childbearing remains enigmatic; while some investigators have reported a positive association between number of births and renal cell cancer risk, others have not. A case–control study, nested within a nation-wide Fertility Register covering Swedish women born 1925 and later, was undertaken to explore possible associations between parity and age at first birth and the risk of renal cell cancer. Among these women a total of 1465 cases of RCC were identified in the Swedish Cancer Register between 1958 and 1992 and information on the number of live childbirths and age at each birth was obtained by linkage to the Fertility Database. For each case, five age-matched controls were randomly selected from the same register. Compared to nulliparous women, ever-parous women were at a 40% increased risk of RCC (Odds Ratio [OR]=1.42; 95% CI 1.19-1.69). The corresponding OR for women of high parity (five or more live births) was 1.91 (95% CI 1.40–2.62). After controlling for age at first birth among parous women, each additional birth was associated with a 15% increase in risk (OR=1.15; 95% CI 1.08–1.22). The observed positive association between parity and renal cell cancer risk is unlikely to be fully explained by uncontrolled confounding, but warrants further evaluation in large studies, with allowance for body mass index.

Correlates of pregnancy oestrogen, progesterone and sex hormone-binding globulin in the USA and China

The objective of this study is to examine perinatal correlates of oestradiol (E2), oestriol (E3), progesterone and sex hormone-binding globulin (SHBG) among pregnant women in the USA and China. Three hundred and four Caucasian women in Boston and 335 Chinese women in Shanghai were studied. Levels of E2, E3, progesterone and SHBG were measured in maternal blood at weeks 16 and 27 of gestation, and correlated with maternal, gestational and perinatal characteristics. Height, weight and body mass index (BMI) before pregnancy is inversely associated with E2 and SHBG, whereas E3 is inversely associated with height and progesterone is inversely associated with weight and BMI. A previous live birth is associated with lower E2 and SHBG in the index pregnancy. Total gestation duration is inversely associated with E2, E3 and progesterone, whereas weight gain during pregnancy is inversely associated with progesterone and SHBG. In the US, pregnancies with female fetuses are characterized by significantly reduced progesterone. Pregnancy hormones are associated with several maternal, gestational and neonatal characteristics.

Coexpression of granulocyte colony stimulating factor and its receptor in primary ovarian carcinomas

Immunohistochemistry was used to determine the expression of granulocyte colony-stimulating factor (G-CSF) and its receptor (G-CSFR) in primary ovarian carcinomas. The expression of G-CSFR was observed in the malignant cells of each of the 46 primary carcinomas examined; G-CSF was coexpressed in both the malignant epithelial cells and the stroma of 56.5% of the specimens. Thus the majority of ovarian carcinomas harbor both potential autocrine and paracrine G-CSF axes. In 37% of the samples, G-CSF was expressed only within stromal cells, suggesting that only a potential paracrine system is in place. In a preliminary, retrospective, evaluation, the survival of patients whose tumors expressed only the apparent paracrine loop was significantly worse than patients whose tumors expressed both potential autocrine and paracrine G-CSF-based regulatory loops (14.5 vs. 42.5 months, respectively). Studies on the potential function of G-CSF were performed using the G-CSFR-expressing OVCAR-3 ovarian carcinoma line. As a single agent, rhG-CSF failed to stimulate [3H]-thymidine incorporation in these cells, but enhanced the mitogenic action of epidermal growth factor (EGF) in a dose-dependent manner. Thus, potential autocrine and/or paracrine loops involving G-CSF and its receptor occur in over 90% of primary ovarian carcinomas, and may act to modulate the action of growth factors.

Lifestyle factors and insulin-like growth factor 1 levels among elderly men.

Insulin-like growth factor 1 (IGF-1) is a potentially important determinant of disease; hence epidemiological identification of factors that influence circulating IGF-1 is merited. We therefore analysed data collected in Greece to determine the relationship between anthropometric, lifestyle and dietary variables and serum levels of IGF-1 among elderly men. We identified 51 men with prostate cancer, 50 men with benign prostatic hyperplasia, and 52 apparently healthy elderly men (controls), all matched for age (+/- 1 year). These 153 men provided blood specimens and were interviewed using a validated lifestyle and food frequency questionnaire. We performed multivariate linear regression to identify potential predictors of circulating IGF-1. After controlling for age, body mass index, smoking habits, alcohol drinking and coffee consumption, each 5 cm increase in height predicted a 13.0% increase in IGF-1 (95% CI 0.4-27.2%) among the controls and a 11.3% increase in IGF-1 (95% CI 4.5-18.6%) among the entire study group. None of the investigated dietary factors (total fat, carbohydrate, protein, dairy products, tomatoes, calcium) were strongly related to IGF-1 levels. The positive association between IGF-1 and height integrates the empirical evidence linking IGF-1 and height with prostate cancer risk.

Parity and the risk of pancreatic cancer: A nested case-control study

Smoking is the only generally accepted risk factor for pancreatic cancer. Reproductive history has in recent studies been associated with pancreatic cancer, but with contradictory results. In order to evaluate a possible association between age at first birth and the number of births and pancreatic cancer, we conducted a nested case‐control study by linking 2 Swedish nationwide registries: the Cancer Registry and The Fertility Registry. Among women born between 1925 and 1970, 1,015 patients with pancreatic cancer were compared with 5,073 age‐matched controls. No association between pancreatic cancer and number of births was found. Age at first birth was inversely related with the risk of pancreatic cancer (OR per 5 years = 0.90; 95% CI 0.83<0R>–<0B>0.97; p<0B> = 0.01), an association mainly confined to women with a diagnosis of pancreatic cancer before 50 years of age (OR per 5 years = 0.85; 95% CI 0.73<0R>–<0B>1.00; p<0B> = 0.04). This trend remained after adjustment for parity, but was less prominent. Young age at first birth and high parity in Sweden are, however, associated with an increased frequency of smoking, thus at least some of the increased risk for pancreatic cancer in women with young age at first birth is likely to be explained by smoking acting as a confounder. Int. J. Cancer 77:224–227, 1998.