Abstract
PURPOSE: Understanding important features for performance in non-disabled bench press and Paralympic powerlifting may inform talent identification and transfer models. The aim of this scoping review was to describe features associated with bench press performance. METHODS: We conducted a systematic search of three electronic databases (PubMed, SportDiscus and EMBASE) to identify studies involving non-disabled and Para athlete populations that investigated features related to bench press one-repetition maximum (1RM), across six domains (anthropometric, body composition, demographic, technical, disability and neuromuscular). Search terms included "resistance training", "para powerlifting' and "one repetition max". No date restrictions were include in searches. Studies using adult participants, with at least six-months of bench press experience, who were able to bench press their body mass were included. RESULTS: Thirty-two studies met our inclusion criteria. Twenty-four studies involved non-disabled athletes (total n = 2,686; 21.9% female) and eight involved Para athletes (total n = 2,364; 39.4% female). Anthropometric (17 studies) and body composition (12 studies) features were most studied; half of the 32 studies investigated features from a single domain. Of anthropometric variables, upper arm circumference shared the strongest association with bench press 1RM in non-disabled (r = 0.87) and para-athletes (r = 0.65). Upper limb fat free mass (r = 0.91) and body mass index (r = 0.46) were the body composition variables sharing the strongest association with bench press 1RM in non-disabled and para-athletes. Few studies considering the uncertainty of their results. Practices of open and transparent science, such as pre-registration and data sharing, were absent. CONCLUSION: The development of bench press talent identification and sport transfer models will require future studies to investigate both non-training and training modifiable features, across multiple domains. Large longitudinal studies using information from athlete monitoring databases and multivariable model approaches are needed to understand the interacting features associated with bench press performance, and for the development of talent identification models.