Study Design. A retrospective cohort study was conducted in 5573 female patients with scoliosis who were referred for treatment at 14 orthopedic medical centers in the United States. Patients were less than 20 years of age at diagnosis which occurred between 1912 and 1965.
Objectives. To evaluate patterns in breast cancer mortality among women with scoliosis, with special emphasis on risk associated with diagnostic radiograph exposures.
Summary of Background Data. A pilot study of 1030 women with scoliosis revealed a nearly twofold statistically significant increased risk for incident breast cancer. Although based on only 11 cases, findings were consistent with radiation as a causative factor.
Methods. Medical records were reviewed for information on personal characteristics and scoliosis history. Diagnostic radiograph exposures were tabulated based on review of radiographs, radiology reports in the medical records, radiograph jackets, and radiology log books. Radiation doses were estimated for individual examinations. The mortality rate of the cohort through January 1, 1997, was determined by using state and national vital statistics records and was compared with that of women in the general U. S. population.
Results. Nearly 138,000 radiographic examinations were recorded. The average number of examinations per patient was 24.7 (range, 0–618); mean estimated cumulative radiation dose to the breast was 10.8 cGy (range, 0–170). After excluding patients with missing information, 5466 patients were included in breast cancer mortality analyses. Their mean age at diagnosis was 10.6 years and average length of follow-up was 40.1 years. There were 77 breast cancer deaths observed compared with the 45.6 deaths expected on the basis of U.S. mortality rates (standardized mortality ratio [SMR] = 1.69; 95% confidence interval [CI] = 1.3–2.1). Risk increased significantly with increasing number of radiograph exposures and with cumulative radiation dose. The unadjusted excess relative risk per Gy was 5.4 (95% CI = 1.2–14.1); when analyses were restricted to patients who had undergone at least one radiographic examination, the risk estimate was 2.7 (95% CI = −0.2–9.3).
Conclusions. These data suggest that exposure to multiple diagnostic radiographic examinations during childhood and adolescence may increase the risk of breast cancer among women with scoliosis; however, potential confounding between radiation dose and severity of disease and thus with reproductive history may explain some of the increased risk observed.
Studies of the atomic bomb survivors, 28 patients with tuberculosis who underwent multiple fluoroscopic examinations during treatment, 3,14,23 and patients who underwent therapeutic irradiation for a variety of conditions, including benign breast disease, 22 acute postpartum mastitis, 32 enlarged thymus, 12 Hodgkin’s disease, 2 and skin hemangiomas, 21 have established that exposure to ionizing radiation increases the risk of breast cancer. Age at exposure has been found to modify response to radiation, with childhood exposures conferring the highest risk and exposure after age 40 imparting low or minimal risk. 4,18 It has been hypothesized that the period between Tanner stage breast-2 33 and the onset of menarche may be a particularly sensitive biologic stage with respect to the action of radiation and other environmental carcinogens. 15 No such pattern has been apparent among atomic bomb survivors and other populations exposed during childhood and adolescence, 17,34,35 but published analyses have not focused on the narrow range of exposure ages relevant to the hypothesis.
Study of women with scoliosis may provide additional insight into the risk of breast cancer after low-dose radiation exposures received during childhood and adolescence. Depending on the magnitude of the spinal curvature, patients typically undergo routine diagnostic radiographic examinations of the spine throughout the growth spurt to monitor for curve progression. An initial pilot study of 1030 women, diagnosed in any of three hospitals or one clinic in Minneapolis–St. Paul, Minnesota, identified a nearly twofold risk of incident breast cancer that was consistent with a radiation effect. 13 The number of breast cancer cases was small (n = 11), however, and did not allow for adjustment for possible confounders, factors that might be related both to scoliosis and breast cancer risk, such as reproductive history.
The pilot study was expanded to improve statistical power and obtain information from living subjects on potential confounding factors. Approximately 4500 female patients were enrolled from 10 of the largest orthopedic medical centers in the United States. The current article presents findings from an evaluation of breast cancer mortality after diagnostic radiograph exposures.
From the *Radiation Epidemiology Branch, National Cancer Institute, National Institutes of Health, Bethesda, Maryland; the †Twin Cities Spine Center, Minneapolis, Minnesota; the ‡Department of Radiation Physics, The University of Texas, M. D. Anderson Cancer Center, Houston, Texas; and Information Management Services, Inc., Silver Spring, Maryland.
This research was supported in part by contracts N01-CP-85651, N02-CP-33013, N01-CP-40535, and N02-CP-81005 with the National Cancer Institute, U. S. Public Health Service, Bethesda, Maryland.
Acknowledgment date: July 31, 1998.
First revision date: February 12, 1999.
Acceptance date: December 6, 1999.
Address reprint requests to
Michele M. Doody
Radiation Epidemiology Branch
National Cancer Institute
Executive Plaza South, Room 7088
Bethesda, MD 20892
The authors assume complete responsibility for the content and conclusions of this article.
Device status category: 1.
Conflict of interest category: 14.