The incidence of breast cancer (BC) increases with age, as does the occurrence of cognitive impairment (CI). However, little is known about the effects of CI on BC survival or chronic medication adherence. A retrospective cohort study was conducted using data from the National Cancer Institute’s SEER (Surveillance, Epidemiology, and End Results) Medicare Linked Database, which is a population-based cancer registry that includes 28% of the U.S. population, to compare cancer-specific, noncancer, and all-cause mortality among BC survivors with or without CI; to determine the effects that a BC diagnosis has on chronic medication adherence (i.e., adherence to cardiovascular and diabetes medications) in patients with or without CI; and to examine the effects that chronic medication adherence has on mediating or moderating the association between CI and noncancer mortality.
Women were included in the study if their initial cancer was a primary BC diagnosed between January 1, 2008 and December 31, 2013; if they were enrolled in Medicare Parts A and B for 24 months prior to and for at least 12 months following their BC diagnosis and in Medicare Part D for at least 12 months prior and 18 months after their BC diagnosis; and if they were aged 67 years or older at the time of their BC diagnosis. Exclusion criteria included enrollment in a Health Maintenance Organization, missing cancer data, BC diagnosis upon autopsy, or lack of documentation of BC treatment.
Medicare Part D data were used to assess the association between CI and chronic medication adherence. In order to avoid possible interference due to adverse effects of cancer treatment, a 6-month washout period was built into the study protocol to account for potential periods of surgery and/or chemotherapy initiation. A diagnosis of CI was made based on ICD-9 CM codes found in any Medicare file for dementia, mild CI, or Alzheimer’s disease in any Medicare file. Cohort selection and propensity score matching were performed to minimize bias.
Outcome measures included overall survival, cancer-specific mortality, noncancer mortality, and all-cause mortality during the study period until the censor date, which was December 31, 2014. Chronic medication adherence, focusing on cardiovascular and diabetes medications, was assessed pre- and postcancer diagnosis and was determined using the proportion of days covered (PDC) method. Chronic medication adherence was estimated using the average PDC for all drugs, and adherence was defined as a PDC >0.8. The impact of changing the PDC to >0.7 or >0.9 was also examined.
Investigators found that among the 67,565 female patients aged >67 years who were eligible for participation, there was a preexisting diagnosis of CI among 8.38%. After matching, the study population was comprised of two groups, each with 5,542 patients, who either had CI or did not have CI for a total sample size of 11,084. Those with preexisting CI were less likely to be diagnosed at an early stage of disease (31.83% vs. 38.94%), and their tumor size was more likely to be >2 cm at the time of diagnosis (48.05% vs. 39.91%) compared with women without CI. Women with CI were also less likely to receive conservative surgery, radiation therapy, chemotherapy, hormone therapy, or targeted therapy than their cognitively intact counterparts. These differences were all statistically significant.
CI was associated with lower survival and increased cancer-specific (hazard ratio = 1.3, CI, 1.04-1.23), noncancer specific (hazard ratio = 1.16, CI, 1.11-1.21), and all-cause mortality (hazard ratio = 1.30, CI, 1.23-1.38). Mortality was 22% in those with CI compared to 18% in cognitively intact women during the first year following a BC diagnosis.
Adherence among women who received at least two prescriptions for cardiovascular and/or diabetes medications was low, at about 55%, in both the cognitively impaired and cognitively intact groups prior to a BC diagnosis. Both groups demonstrated a significant increase in adherence following the BC diagnosis, but there was no significant difference between the groups either before or after the cancer diagnosis. Patients who were nonadherent had a higher risk of mortality from noncancer causes. CI was also associated with a higher risk of noncancer mortality. Both CI and medication adherence were strong predictors of an increased risk of noncancer mortality, but medication adherence did not mediate or moderate the relationship between CI and noncancer mortality.
Pharmacists should be cognizant of these findings when managing older patients with BC who have concomitant CI, as these patients may require closer follow-up and monitoring.
The content contained in this article is for informational purposes only. The content is not intended to be a substitute for professional advice. Reliance on any information provided in this article is solely at your own risk.
« Click here to return to Breast Cancer Update.