Estimating Proximity to Failure Using Repetitions in Reserve in Adults Treated for Cancer

BACKGROUND: There is a consensus from governing bodies that more evidence is needed in understanding how to properly prescribe, monitor and progress resistance training (RT) in oncology populations. Repetitions in reserve (RIR) has been demonstrated as a quick/easy tool to monitor proximity to failure and interventions using RIR have shown promise in improving muscle strength and mass in older adults and apparently healthy populations. However, the validity of RIR in RT in oncology has yet to be examined. Therefore, the purpose of this study was to assess the accuracy of RIR predications in individuals treated for cancer in a pragmatic setting. METHODS: Individuals treated for cancer (n=20) enrolled in an 8-week RT intervention, completed RIR testing sessions at weeks 2 and 8 of the study. During each testing session, participants first estimated a 2 RIR prediction and then continued performing repetitions until reaching momentary failure. Generalized linear mixed models (GLMM) were performed on both raw and absolute RIR difference values and a binary logistic regression model was performed on number of accurate predictions (±2). All statistical analyses were performed in R (Version 2024.12.0+467). RESULTS: Sample was 20 individuals treated for cancer (n=20, 65% breast, 20% colorectal cancers, age 61.7 ± 8.63 y,) with 55% of the sample identifying as black and 35% as white non-Hispanic origin, 5% Native Hawaiian/Pacific Islander, and 5% identifying as “Other”. The raw prediction (the difference between predicted RIR and actual RIR) was -2.27 at week-2, and was -1.94 at week-8, with no significant differences between timepoints. Preliminary GLMM data revealed time had negligible effects on raw RIR values (estimate = 0.325, p=0.265), however, time was a significant predictor of accurate absolute RIR predictions (estimate= -0.432, p=0.012). The binary logistic regression model demonstrated significant effects of time on the probability of accurate predictions (estimate=0.131, p=0.038). CONCLUSIONS: Raw RIR prediction at weeks 2 and 8 were outside the range of acceptable accuracy (±2), indicating a need for caution when solely relying on RIR to prescribe RT load in oncology. There was a small improvement in RIR accuracy across the intervention period, suggesting that training status and familiarity with the scale could improve RIR predictions. More research is warranted to understand RIR predictions with different training loads, anchoring procedures/scale familiarization, and intervention lengths.