Insulin-Like Growth Factor 1 and Prostate Cancer Risk: A Population-Based, Case-Control Study

Journal of the National Cancer Institute Alicja Wolk 0 2 3 Christos S. Mantzoros 0 2 3 Swen-Olof Andersson 0 2 3 Reinhold Bergstr om 0 2 3 Lisa B. Signorello 0 2 3 Pagona Lagiou 0 2 3 Hans-Olov Adami 0 2 3 Dimitrios Trichopoulos 0 1 2 3 0 Oxford University Press 1 Affiliations of authors: A. Wolk, R. Bergstro m, Department of Medical Epidemiology, Karolinska Institutet , Stockholm , Sweden; C. S. Mantzoros , De- partment of Epidemiology, Harvard School of Pub- lic Health, and Division of Endocrinology, Beth Is- rael Deaconess Medical Center, Harvard Medical School , Boston, MA; P. Lagiou , Department of Epi- demiology, Harvard School of Public Health; S.-O. Andersson, Department of Medical Epidemiology, Karolinska Institutet, and Department of Urology , O 2 Journal of the National Cancer Institute , Vol. 90, No. 12, June 17, 1998 3 Supported by grants from the Swedish Cancer So- ciety and O - Background: Recent epidemiologic investigations have suggested an association between increased blood levels of insulin-like growth factor 1 (IGF-1) and increased risk of prostate cancer. Our goal was to determine whether an association exists between serum levels of IGF-1 and one of its binding proteins, insulin-like growth factorbinding protein 3 (IGFBP-3), and prostate cancer risk. Methods: An immunoradiometric assay was used to quantify IGF-1 levels and IGFBP-3 levels in serum samples as part of a populationbased, casecontrol study in Sweden. The study population comprised 210 patients with newly diagnosed, untreated prostate cancer and 224 frequency-matched control subjects. Data were analyzed by use of unconditional logistic regression to calculate odds ratios (ORs) and 95% confidence intervals (CIs). Reported P values are twosided. Results: The mean serum IGF-1 level for case patients (158.4 ng/mL) was significantly higher than that for c o n t r o l s u b j e c t s ( 1 4 7 . 4 n g / m L ) (P = .02); corresponding mean serum IGFBP-3 levels were not significantly different between case patients (2668 ng/mL) and control subjects (2518 ng/ mL) (P = .09). We found a moderately strong and statistically significant (P = .04) positive association between serum levels of IGF-1 levels and risk of prostate cancer (OR = 1.51; 95% CI = 1.02.26 per 100 ng/mL increment); the association was particularly strong for men younger than 70 years of age (OR = 2.93; 95% CI = 1.435.97). No association was found between serum IGF-1 levels and disease stage. Serum IGFBP-3 levels were not significantly associated with increased risk of disease, and adjustment for IGFBP-3 had little effect on the association between IGF-1 levels and risk of prostate cancer. Conclusion: Elevated serum IGF-1 levels may be an important predictor of risk for prostate cancer. However, our results do not support an important role for serum IGFBP-3 as a predictor of risk for this disease. [J Natl Cancer Inst 1998;90:9115] Until recently, the search for endocrine factors that may be involved in prostate carcinogenesis has focused on sex steroid hormones and on sex hormone-binding globulin. Although supported by animal studies (1), epidemiologic investigations have not unequivocally supported the hypothesis that sex hormones and their associated receptors and/or binding proteins are the most important endocrine factors involved in prostate cancer (24). Therefore, it is likely that other factors are involved in development of the disease. Insulin-like growth factor 1 (IGF-1) has mitogenic and antiapoptotic effects on prostate epithelial cells in vitro (5,6). These results prompted epidemiologic studies to determine the role of IGF-1 in human prostate carcinogenesis. In a study of 52 prostate cancer patients and an equal number of control subjects in Greece, Mantzoros et al. (7) reported that increased serum IGF-1 level (by an increment of about one standard deviation; 60 ng/mL in their study) was associated with a doubling of the disease risk. Chan et al. (8) also determined that an increased risk of prostate cancer was associated with increased blood levels of IGF-1 in a nested casecontrol study within the Physicians Health Study, using prospectively collected blood samples from 152 case subjects and 152 control subjects. Chan et al. also found that one major circulating IGF1-binding protein, insulin-like growth factor-binding protein 3 (IGFBP-3), after adjustment for IGF-1, was inversely related to the risk of prostate cancer. They attributed this effect to the reduction of bioavailable IGF-1 with increasing levels of IGFBP-3. If the association between increased serum levels of IGF-1 and increased prostate cancer risk were confirmed by independent investigations, this relationship could represent an important finding with considerable diagnostic and therapeutic value. To study the association of IGF-1 and IGFBP-3 serum levels with prostate cancer risk, we made use of data from a large, population-based, casecontrol study in Sweden consisting of newly diagnosed prostate cancer cases that were cytologically and histologically confirmed. Subjects and Methods All men under the age of 80 years, born in Sweden and living in O rebro County, Sweden, at any time from January 1989 through September 1991, formed the study base. Patients in this population with newly diagnosed prostate cancers, cytologically and histologically confirmed, were eligible to participate in the study. Clinical records from the three participating hospitals ( Orebro Medical Center and hospitals in Karlskoga and Lindesberg) and the Department of Pathology at Orebro Medical Center allowed complete case ascertainment, confirmed through cross-checking the clinical records of case subjects with the regional cancer registry (9,10). All tumors were staged clinically in accordance with the tumornodemetastasis classification system (11); among them, 26.6% were surgically staged. Control subjects were identified contemporaneously with case subjects. Selected every 3rd month from the county population register, control subjects were frequency-matched to case subjects in 10-year age groups. All potential control subjects underwent a digital rectal examination by one of us (S.-O. Andersson). Men with a palpable nodule and/or serum levels of prostate-specific antigen higher than 10 ng/mL underwent further diagnostic testing through ultrasound-guided biopsy (four to six random samples). If interpretation of the initial biopsy specimens failed to confirm a diagnosis of cancer, the procedure was repeated 6 months later. Only those individuals whose biopsy specimens showed no evidence of cancer were deemed eligible as control subjects (10). Less than 3% of potential control subjects who were diagnosed with cancer by analysis of tissue obtained through biopsy were included in the current study as case subjects. Subjects eligible for the study were mailed a food-frequency questionnaire for self-administration at home. A majority of prostate cancer patients receive (...truncated)