Moderators of exercise effects on self-reported cognitive functioning in cancer survivors: an individual participant data meta-analysis

Anouk E Hiensch, Julia Beckhaus, Lenja Witlox, Evelyn M Monninkhof, Sanne B Schagen, Jonna K van Vulpen, Maike G Sweegers, Robert U Newton, Neil K Aaronson, Daniel A Galvão, Karen Steindorf, Martijn M Stuiver, Ilse Mesters, Hans Knoop, Martine M Goedendorp, Martin Bohus, Lene Thorsen, Karl-Heinz Schulz, Martina E Schmidt, Cornelia M UlrichGabe S Sonke, Wim H van Harten, Kerri M Winters-Stone, Miranda J Velthuis, Dennis R Taaffe, Willem van Mechelen, Marie José Kersten, Frans Nollet, Joachim Wiskemann, Laurien M Buffart, Anne M May

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

PURPOSE: This individual participant data meta-analysis (IPD-MA) assesses exercise effects on self-reported cognitive functioning (CF) and investigates whether effects differ by patient-, intervention-, and exercise-related characteristics.

METHODS: IPD from 16 exercise RCTs, including 1987 patients across multiple types of non-metastatic cancer, was pooled. A one-stage IPD-MA using linear mixed-effect models was performed to assess exercise effects on self-reported CF (z-score) and to identify whether the effect was moderated by sociodemographic, clinical, intervention- and exercise-related characteristics, or fatigue, depression, anxiety, and self-reported CF levels at start of the intervention (i.e., baseline). Models were adjusted for baseline CF and included a random intercept at study level to account for clustering of patients within studies. A sensitivity analysis was performed in patients who reported cognitive problems at baseline.

RESULTS: Minimal significant beneficial exercise effects on self-reported CF (β=-0.09 [-0.16; -0.02]) were observed, with slightly larger effects when the intervention was delivered post-treatment (n=745, β=-0.13 [-0.24; -0.02]), and no significant effect during cancer treatment (n=1,162, β=-0.08 [-0.18; 0.02]). Larger effects were observed in interventions of 12 weeks or shorter (β=-0.14 [-0.25; -0.04]) or 24 weeks or longer (β=-0.18 [-0.32; -0.02]), whereas no effects were observed in interventions of 12-24 weeks (β=0.01 [-0.13; 0.15]). Exercise interventions were most beneficial when provided to patients without anxiety symptoms (β=-0.10 [-0.19; -0.02]) or after completion of treatment in patients with cognitive problems (β=-0.19 [-0.31; -0.06]). No other significant moderators were identified.

CONCLUSIONS: This cross-cancer IPD meta-analysis observed small beneficial exercise effects on self-reported CF when the intervention was delivered post-treatment, especially in patients who reported cognitive problems at baseline.

IMPLICATIONS FOR CANCER SURVIVORS: This study provides some evidence to support the prescription of exercise to improve cognitive functioning. Sufficiently powered trials are warranted to make more definitive recommendations and include these in the exercise guidelines for cancer survivors.

Original languageEnglish
Pages (from-to)1492-1503
Number of pages12
JournalJournal of cancer survivorship : research and practice
Volume18
Issue number5
Early online date9 May 2023
DOIs
Publication statusPublished - Oct 2024

Keywords

  • Cancer
  • Cognitive functioning
  • Exercise
  • Individual participant data meta-analysis
  • Survivorship

Fingerprint

Dive into the research topics of 'Moderators of exercise effects on self-reported cognitive functioning in cancer survivors: an individual participant data meta-analysis'. Together they form a unique fingerprint.

Cite this