Of mobile phones and morbidity.

The effect of strenuous physical activity on the incidence of benign breast disease and benign tumours of the female reproductive system is unknown. Previous findings of a lower prevalence of breast cancer and cancers of the reproductive system among former college athletes compared to non-athletes (Frisch et al., 1985) suggested that former athletes and non-athletes might also differ in the prevalence of benign tumours of these tissues. We collected detailed medical, reproductive, diet and activity histories from 5,398 women: 2,622 former college athletes and 2,776 non-athletes ranging in age from 21 to 80 yr. We report data on the prevalence (lifetime occurrence) of benign breast diseases and benign tumours of the reproductive system among these women. Details of the procedures for obtaining rosters of athletes and non-athletes, the sports included, and the questions on length and intensity of college, pre-college and current physical activity have been previously reported (Frisch et al., 1985). In addition to the questions on athletics, the 14-page questionnaire requested detailed medical history, reproductive history from menarche through the menopause, including births and pregnancy outcome, smoking history, current health problems, height, weight, weight changes, and current diet. The questions pertaining to benign tumours were: 1. Did you ever have any of the following medical procedures?: Breast biopsy, endometrial biopsy or gynaecological biopsy? If 'yes', for each biopsy: Age occurred; inpatient or outpatient; diagnosis; and treatment, if any. 2. Did you ever have a benign tumour(s)? If 'yes': Age occurred; type or site; and treatment, if any. The literature on benign breast disease (BBD) universally mentions the ill-defined nature of BBD, the wide variety of diseases and non-diseases included under this rubric and the need for diagnostic criteria. In the analysis that follows we defined BBD as at least one benign breast tumour or at least one breast biopsy (Ernster, 1981; Love et al., 1982). The relative risks were adjusted for potential confounding factors by multiple logistic regression, using backward selection (Kleinbaum et al., 1982). In addition to athlete or non-athlete, these factors included: age, leanness, ever-pregnant, age of menarche, ever used oral contraceptives (yes/no), use of hormones for menopausal symptoms (yes/no), family history of cancer (yes/no), and smoking (ever/never). For benign breast disease, ever having had an endometrial biopsy was also included. Variables are retained in the logistic model if the significance level is 0.30 or less. Age of menarche in the logistic regression model was coded …