Factors associated with sleep wake disturbances in adult survivors of childhood brain tumors.

Abstract

Sleep disturbances impact physical and mental health in brain tumor survivors, leading to sub-optimal participation in life activities. Technological advances in cancer treatments have improved five-year survival rates for childhood brain tumors to nearly 75%. With growing numbers of brain tumor survivors, mitigation of serious late sequelae from cancer treatments becomes increasingly important. This study evaluated factors associated with sleep quality in adult survivors of childhood brain tumors and a population-based comparison group. The Cancer-Related Factors Affecting Sleep model developed by Vena et al. (2004) provided a conceptual framework for this study. Participants were recruited from the University of Minnesota and St. Jude Children's Research Hospital clinical treatment databases. The comparison group was recruited from a national mailing database service. The first aim evaluated differences in global and component sleep quality between survivors and the comparison group. Using multiple variable linear regression and logistic regression, no differences in global sleep quality were detected between survivors and the comparison group. However, survivors were two and a half times more likely than the comparison group to have longer sleep latency, taking on average 33 minutes to fall asleep after going to bed. Females in both groups reported worse sleep quality, including more daytime dysfunction, and risk of poor sleep (PSQI global scores > 10). The second aim evaluated cancer treatment effects on sleep quality among brain tumor survivors. Tumor location and surgery were not significant predictors of sleep however, radiation to the hypothalamus approached significance among survivors for longer sleep latency. Exploratory analyses identified radiation to the hypothalamus and age at diagnosis < 5 as predictors of lower vigor subscale scores on the POMS. No treatment variables predicted abnormal BSI scores among the survivors. The third aim evaluated a path model proposing post-treatment variables including obesity, depression, and fatigue as mediators of global sleep quality. In addition, gender, age, radiation to the hypothalamus and age at diagnosis < 5 were included in the model as exogenous variables. Obesity and depression were not strong predictors of global sleep quality however, gender and age at diagnosis < 5 predicted higher fatigue subscale scores. Fatigue itself significantly predicted higher (worse) global sleep quality scores within the model. This study revealed new and surprising information about sleep latency in brain tumor survivors. It also confirmed known associations with sleep and gender and a strong relationship between fatigue and sleep quality. Given these findings, future studies can begin to focus on identifying additional factors impacting sleep in these survivors.