Factors Influencing Weight Loss Practices in Italian Boxers: A Cluster Analysis

Abstract

It is common practice in combat sports that athletes rapidly lose body weight before a match, by applying different practices-some safer and others possibly dangerous. The factors behind the choice of practices utilised have not been fully studied. This study aimed to investigate the weight loss strategies used by Italian boxers and to look at the difference between higher and lower risk practice adaptors. A modified version of a validated questionnaire has been sent to 164 amateur (88%) and professional (12%) boxers by email. A heatmap with hierarchical clustering was used to explore the presence of subgroups. Weight loss strategies were used by 88% of the athletes. Two clusters were found, defined by the severity of weight loss behaviours. Professional fighters, high-level athletes and females were more represented in Cluster 2, the one with more severe weight-loss practices. These athletes were characterised by a higher weight loss magnitude and frequency throughout the season and reported being more influenced by physicians and nutritionists, compared with the boxers in Cluster 1. Not all the weight loss practices are used with the same frequency by all boxers. The level of the athlete and the boxing style have an influence on the weight-cutting practices.

Keywords: boxing; cluster analysis; combat sports; professional athletes; weight cutting.

Figure 1

Heatmap with weight-cutting strategies (in columns) and subjects (in rows). An empty line separates the two clusters (Cluster 1 at the top, Cluster 2 at the bottom). White colour is associated with the ‘never used’ answer, while purple represents the ‘always’ answer.

Figure 2

Frequency of use of the different weight-cutting strategies in Cluster 1 (upper bars) and Cluster 2 (lower bars), for each weight-cutting method. Association strength is reported as Cramer’s V.

Figure 3

Rapid Weight Loss Score (RWLS) in the two clusters (C1 and C2). The thick line represents the median, while the thin ones represent first and third quartiles; Asterisk indicate statistically significant differences.