Breast cancer among women given X-ray therapy for acute postpartum mastitis

Acute postpartum mastitis (APM) is an inflammatory-infectious condition of the breast, occurring commonly at childbirth or during lactation. A series of 601 women who received x-ray therapy for APM during the 1940's or 1950's have been followed up by mail questionnaire, with medical verification of pertinent conditions, to ascertain their incidences of breast cancer. Control subjects consisted of a series with APM who did not receive irradiation, plus the female siblings of both the APM groups, for a total of 1,239 controls. The groups have been followed up to 45 years; the average was 29 years. The relative risk (RR) for breast cancer, adjusted for age and interval since irradiation (or an equivalent entry definition for controls), was 3.2 for the irradiated breasts; the 90% confidence interval (CI) was 2.3-4.3. For a linear multiplicative model, the risk increased by 0.4% per rad (90% Cl of 0.2-0.7). The dose-response curve appeared to be essentially linear, except for a diminution of risk at high doses (greater than or equal to 700 rad). The fact that there were no treated breasts with doses between 0 and 60 rad, however, means that it was not possible to evaluate the curvature with the maximum contrast between low and high doses. The dose fractionation analyses showed that neither the number of dose fractions, the number of days between fractions, nor the dose per fraction had any apparent effect on breast cancer risk when the variables were analyzed separately. Similarly, when the fractionation variables were considered jointly in a Cox regression analysis, none was significant once total breast dose was controlled for. Analyses of age at irradiation did not show appreciable differences between age groups, although the numbers were too small to be clear-cut (only 64 women greater than 34 yr old at irradiation). Other studies have shown diminished risk associated with an older age at irradiation. The lack of diminished risk in this study may occur because during pregnancy and lactation the breasts are under increased proliferative stimulation by hormones, by comparison with the normal condition of breasts at older ages. An analysis of the temporal relationship of radiation to breast cancer showed that the RR did not vary systematically with interval since irradiation, but the absolute risk increased over time. This finding agrees with other studies that have also suggested a better fit for the multiplicative model.