Prevalence of Cognitive Impairment and Association With Survival Among Older Patients With Hematologic Cancers.

Importance: As the population ages, cognitive impairment has promised to become increasingly common among patients with cancer. Little is known about how specific domains of cognitive impairment may be associated with survival among older patients with hematologic cancers. Objective: To determine the prevalence of domain-specific cognitive impairment and its association with overall survival among older patients with blood cancer. Design, Setting, and Participants: This prospective observational cohort study included all patients 75 years and older who presented for initial consultation in the leukemia, myeloma, or lymphoma clinics of a large tertiary hospital in Boston, Massachusetts, from February 1, 2015, to March 31, 2017. Patients underwent screening for frailty and cognitive dysfunction and were followed up for survival. Exposures: The Clock-in-the-Box (CIB) test was used to screen for executive dysfunction. A 5-word delayed recall test was used to screen for impairment in working memory. The Fried frailty phenotype and Rockwood cumulative deficit model of frailty were also assessed to characterize participants as robust, prefrail, or frail. Results: Among 420 consecutive patients approached, 360 (85.7%) agreed to undergo frailty assessment (232 men [64.4%] and 128 women [35.6%]; mean [SD] age, 79.8 [3.9] years), and 341 of those (94.7%) completed both cognitive screening tests. One hundred twenty-seven patients (35.3%) had probable executive dysfunction on the CIB, and 62 (17.2%) had probable impairment in working memory on the 5-word delayed recall. Impairment in either domain was modestly correlated with the Fried frailty phenotype (CIB, ρ = 0.177; delayed recall, ρ = 0.170; P = .01 for both), and many phenotypically robust patients also had probable cognitive impairment (24 of 104 [23.1%] on CIB and 9 of 104 [8.7%] on delayed recall). Patients with impaired working memory had worse median survival (10.9 [SD, 12.9] vs 12.2 [SD, 14.7] months; log-rank P < .001), including when stratified by indolent cancer (log-rank P = .01) and aggressive cancer (P < .001) and in multivariate analysis when adjusting for age, comorbidities, and disease aggressiveness (odds ratio, 0.26; 95% CI, 0.13-0.50). Impaired working memory was also associated with worse survival for those undergoing intensive treatment (log-rank P < .001). Executive dysfunction was associated with worse survival only among patients who underwent intensive treatment (log-rank P = .03). Conclusions and Relevance: These data suggest that domains of cognitive dysfunction may be prevalent in older patients with blood cancer and may have differential predictive value for survival. Targeted interventions are needed for this vulnerable patient population.