Late-onset Cognitive Impairment and Modifiable Risk Factors in Adult Childhood Cancer Survivors

Abstract

Importance: Long-term survivors of childhood cancer may be at elevated risk for new neurocognitive impairment and decline as they age into adulthood.

Objective: To determine whether aging adult childhood cancer survivors report more new-onset neurocognitive impairments compared with their siblings and to identify risk factors associated with such impairments.

Design, setting, and participants: Participants of this cohort study included adult survivors of childhood cancer from the Childhood Cancer Survivor Study and their siblings as a control group. The original cohort included survivors who received a diagnosis between January 1, 1970, and December 31, 1986, for whom longitudinal neurocognitive assessment was available. This study examined the prevalence of new-onset neurocognitive impairment between baseline (23.4 years after diagnosis) and follow-up (35.0 years after diagnosis). The analysis was performed from January 2021 to May 2022.

Exposures: Cancer treatment exposures were abstracted from medical records. Chronic health conditions were graded using Common Terminology Criteria for Adverse Events version 4.03.

Main outcomes and measures: The primary outcome was new-onset (present at follow-up, but not present at baseline) neurocognitive impairment (defined as a score in the worst 10% of the sibling cohort). Impairment was assessed using the Childhood Cancer Survivor Study Neurocognitive questionnaire. Relative risks (RRs) and 95% CIs were used to estimate associations of neurocognitive impairment with treatment and health behaviors and conditions using generalized linear models.

Results: The cohort comprised 2375 survivors (mean [SD] age at evaluation, 31.8 [7.5] years; 1298 women [54.6%]) of childhood cancer, including acute lymphoblastic leukemia (ALL; 1316 participants), central nervous system (CNS) tumors (488 participants), and Hodgkin lymphoma (HL; 571 participants). A total of 232 siblings (mean [SD] age at evaluation, 34.2 [8.4] years; 134 women [57.8%]) were included. Compared with siblings, a higher proportion of survivors with no impairment in memory at baseline had new-onset memory impairment at follow-up: siblings proportion, 7.8% (95% CI, 4.3%-11.4%); ALL survivors treated with chemotherapy only, 14.0% (95% CI, 10.7%-17.4%); ALL survivors treated with cranial radiation (CRT), 25.8% (95% CI, 22.6%-29.0%); CNS tumor survivors, 34.7% (95% CI, 30.0%-39.5%); and HL survivors, 16.6% (95% CI, 13.4%-19.8%). New-onset memory impairment was associated with CRT in CNS tumor survivors (RR, 1.97; 95% CI, 1.33-2.90) and alkylator chemotherapy greater than or equal to 8000 mg/m2 in ALL survivors treated without CRT (RR, 2.80; 95% CI, 1.28-6.12). Neurologic conditions mediated the impact of CRT on new-onset memory impairment in CNS survivors. Smoking, low educational attainment, and low physical activity were associated with elevated risk for new-onset memory impairment.

Conclusions and relevance: These findings suggest that adult survivors of childhood cancer are at elevated risk for late-onset memory impairment related to modifiable risk factors identified early in survivorship.

Figure 1.. Participant Enrollment Flowchart

ALL indicates acute lymphoblastic leukemia; CCSS, Childhood Cancer Survivor Study; CNS, central nervous system; CRT, cranial radiation therapy; HL, Hodgkin lymphoma; NF1, neurofibromatosis type 1.

Figure 2.. Prevalence and Relative Risk of Neurocognitive Impairment in Survivors Compared With Sibling Controls at Follow-up

Cancer survivors who did not have impairment at baseline on a given cognitive domain demonstrate significantly greater risk of decline to cognitive impairment for that domain compared with sibling controls after an 11-year interval. Panel A shows that a higher percentage of survivors reported impairment on all domains at baseline compared with siblings. This increased rate of new-onset impairment can be seen in all measured cognitive domains. Panel B shows the relative risk of impairment by diagnosis group compared with siblings for all measured cognitive domains at follow-up among those not impaired on that measure at baseline, with 95% CIs adjusted for sex and age at baseline. Baseline was 23.4 years and follow-up was 35.0 since diagnosis. ALL indicates acute lymphoblastic leukemia; CNS, central nervous system; CRT, cranial radiation therapy; HL, Hodgkin lymphoma.

Figure 3.. Relative Risk of New Onset Cognitive Impairment by Chronic Condition Onset

Graphs show data for acute lymphoblastic leukemia (ALL) without cranial radiation therapy (CRT) (A), ALL with CRT (B), central nervous system (CNS) tumors (C), and Hodgkin lymphoma (HL) (D). Relative risk is based on models for conditions with significant results, adjusted for age and sex. Each model is shown for survivors who were unimpaired in the neurocognitive domain at baseline. Established indicates survivors with grade 3 or 4 condition at baseline; new onset indicates grade 3 or 4 condition developed between baseline and follow-up, with referent as no grade 3 or 4 condition by follow-up. Blank spaces indicate model for given domain had no significant chronic conditions.