Risk factors for breast cancer in women undergoing mammography.

To determine risk factors for carcinoma of the breast, we compared women with cancer on screening and diagnostic mammography with those in whom cancer was not detected. For 39 months, medical histories were collected by mammography technologists on 3492 women having routine screenings or diagnostic mammograms at our institution. Potential risk factors of women with biopsy-proved breast cancer were compared with those in women who had normal findings on mammograms or negative biopsy results (control subjects). Of the 3492 women, 49 had biopsy-proved breast cancer. There were 3361 patients in the control group, including those women with normal findings on mammograms (3294) and those with negative biopsy results (67). Eighty-two women had incomplete questionnaires or were lost to follow up. Nearly all of the patients with breast cancer were postmenopausal compared with 68% of the control subjects. The mean length of lactation for breast cancer patients was significantly less than for control subjects: 5.6 vs 7.5 weeks (p = .015). This was true also for the postmenopausal patients: 8.1 vs 6.1 weeks (p = .041). Postmenopausal breast cancer patients had menstruated significantly more years (p = .016) than the postmenopausal control subjects: 34 vs 31 years, although the mean age at menarche was not different. When corrected for age, there was no significant difference in the total duration of menstruation in the postmenopausal cancer patients compared with the postmenopausal control subjects. Postmenopausal breast cancer patients had a significantly greater (p = .021) average body weight than postmenopausal control subjects: 71.7 vs 66.7 kg, although body weight was the same when all patients were considered. Similar results were found when Quetelet's index for obesity (weight in kg/height in cm2) (p = .004) was calculated for postmenopausal patients: 28 for cancer patients and 26 for control subjects. There was no significant difference in height between the cancer patients and control subjects when all patients or just the postmenopausal patients were considered. History of oral contraceptive use was significantly less common among postmenopausal breast cancer patients than among postmenopausal control subjects: 9% vs 20%. Patients with breast cancer had lower parity than the control subjects. In our series of patients, women in whom breast cancer was detected on mammography lactated less, showed no significant difference in years of menstruation when corrected for age, had a greater average body weight, used oral contraceptives less often, and had fewer children than women in whom no cancer was detected on mammography.