A decline in breast cancer incidence has been attributed to the reduction in hormone replacement therapy (HRT) prescriptions since the publication of the landmark WHIT paper in 2003. Concurrently, a relationship between HRT and cerebrovascular disease incidence has also been suggested. No generalized analysis of HRT prescription rates and breast cancer incidence rates that included more than seven years of data. We hypothesized that detailed analysis of SEER data would clarify the relationship between HRT use and breast cancer incidence. Given the large decline in HRT prescription rates uncovered, analyses of potential complications were also conducted, with the understanding that a small effect or one limited to a subpopulation, such as a single race, might not be detected.
Incidence rates (per 100,000 women) and standard errors for ductal and lobular breast carcinomas, and endometrioid /endometrial carcinomas in women over 50 years were obtained from the Surveillance, Epidemiology, and End Results (SEER) database 1992–2012. From the Medical Expenditure Panel Survey 1996–2012 weighted counts and standard errors of hormone replacement therapy (HRT) prescriptions for women over 50 years were obtained. Using the National Hospital Discharge Survey (NHDS), 1996–2010 weighted counts and standard errors of femoral neck fractures, total hip replacements, acute myocardial infarctions, and cerebral infarctions were obtained for 50+ year men and women. Weighted counts and standard errors were divided by US census figures and multiplied by 100,000. Joinpoint regression was used to analyze rates.
Beginning 2001, HRT prescription rates dropped dramatically, 2001–2012 AAPC -14.9 (95% CI -17.4, -12.4). Breast cancer rates, which began to decline in 1999, increased after 2003; 2012 rates were similar to those seen in 2001 for both ductal, AAPC 0.1 (-0.4, 0.6) and lobular, AAPC 0.5 (-0.4, 1.5), carcinoma. Endometrial carcinoma rates increased, 2001–2012 AAPC 3.5 (3.1, 3.8), arguing against a negative effect of HRT discontinuation of endometrial carcinoma. Tests for parallelism failed to detect APC differences among genders for femoral neck fractures (P = 0.24), for total hip replacements (P = 0.11), for myocardial infarctions (P = 0.10), or for cerebral infarctions (P = 0.19), precluding any assignment of general effect on these disorders by HRT.
Using SEER data, we demonstrated that changes in breast cancer rates cannot be explained by HRT prescription rate changes.