Risk of memory impairment in gynecologic cancer survivors in the Women's Health Initiative: Untangling effects of age, comorbidities, and cancer.

OBJECTIVE

To assess whether gynecologic cancer history predicts worsened self-reported memory outcomes, accounting for the interaction of age and comorbidities.

METHODS

Nested case-control study of cancer survivors enrolled in the ancillary Life and Longevity After Cancer (LILAC) study of the Women's Health Initiative (WHI) and matched cancer-free WHI participants, longitudinally followed after initial enrollment from 1993 to 1998. Postmenopausal women with incident endometrial or ovarian cancer during the WHI were compared to cancer-free women matched up to 5:1 to survivors. Participants with pre-existing cognitive dysfunction, neurodegenerative disease, or multiple cancers were excluded. The primary outcome was self-reported "moderate/severe memory problems" ≥12 months following cancer diagnosis of the index survivor. Associations of pre-diagnosis conditions with later memory problems were examined through cause-specific hazards models, accounting for competing mortality risk.

RESULTS

Primary analyses included 1395 survivors and 5364 cancer-free controls. In multivariable analysis, cancer history was associated with decreased self-reported memory problems (cause-specific hazard ratio [csHR] 0.71 [0.57-0.87], p = 0.001). The only factors independently predictive of future memory concerns were increasing age (5-year increase; csHR 1.96 [1.84-2.09], p < 0.001), cardiovascular disease (hypertension csHR 1.24 [1.03-1.50]; major cardiovascular events csHR 1.85 [1.28-2.68], p = 0.002), diabetes (csHR 1.47 [1.10-1.96], p = 0.01), and depression (csHR 1.47 [1.11-1.95], p = 0.008).

CONCLUSIONS

After accounting for mortality differences, gynecologic cancer survivors experienced reduced hazard for self-reported moderate/severe memory problems. In this population, significant memory symptoms may represent aging and comorbidity effects more than cancer-related outcomes.