Postoperative Trajectories of Self-perceived Cognitive Function and Associated Factors in Patients with Primary Brain Tumors

BACKGROUND
Cognitive dysfunction profoundly impacts daily lives of patients with brain tumors. Standardized tests are key to its objective assessment, but fail to fully capture patients’ experience of cognitive (dys-)function in day-to-day activities. The way in which this self-perceived cognitive function (SPCF) changes over the disease course is not clear, and limited insights into factors that contribute to SPCF trajectories hinder early risk identification and intervention. This study aimed to identify latent trajectories of SPCF and associated factors in patients with meningioma or glioma.

MATERIAL AND METHODS
Patients with meningioma or glioma reported SPCF through the Cognitive Failures Questionnaire (CFQ) as part of a neuropsychological screening within two weeks before surgery, three months post-surgery and one year post-surgery. A higher CFQ score indicated more severe complaints. An unsupervised latent class growth (LCG) model was used to identify classes of patients based on the longitudinal trajectory of their CFQ scores. For comparative purposes, a second LCG model was applied to longitudinal data with similar intervals from healthy controls (n=158). Diagnosis, sociodemographics, anxiety and depression symptomatology, general and activity-related fatigue, and the presence of cognitive impairment (Z-score <-1.5 on ≥ 1 cognitive test) before surgery were adopted as external predictors of patients’ trajectories, based on proportional classification.

RESULTS
In total, 159 patients (meningioma, n=103; glioma, n=56 [60% IDH-1 mutated]) were included based on availability of ≥2 consecutive CFQ assessments. A four-class LCG model, i.e., four distinct trajectories, showed the best fit among patients. Trajectories were interpreted as follows: a moderate complaint load showing a mild increase across each timepoint (62% of the sample), a low and stable complaint load (20%), a high load with a late increase between 3 and 12 month follow-up (11%) and a low load with a similarly late increase (8%). In controls, three trajectories with low (46%), low-moderate (31%) and moderate (23%) complaint loads existed, all of which showing a mild decrease over time. After false discovery rate correction, pre-surgical general and activity-related fatigue were related to patients’ CFQ trajectories, with stronger fatigue particularly distinguishing the two trajectories with late increases from the others.

CONCLUSION
Patients showed four distinct trajectories of SPCF up to one year after surgery. Trajectories were characterized by different, but mostly increasing, complaint loads. While similar across diagnoses, patients’ trajectories were seemingly different from those of controls. Among early factors, fatigue types seem particularly related to the course of SPCF. Potentially diverse trajectories of domain-specific complaints should be investigated.