Sex differences in absolute and relative changes in muscle size following resistance training in healthy adults: a systematic review with Bayesian meta-analysis

Abstract

Background: Muscle hypertrophy may be influenced by biological differences between males and females. This meta-analysis investigated absolute and relative changes in muscle size following resistance training (RT) between males and females and whether measures of muscle size, body region assessed, muscle fibre type, and RT experience moderate the results.

Methods: Studies were included if male and female participants were healthy (18-45 years old) adults that completed the same RT intervention, and a measure of pre- to post-intervention changes in muscle size was included. Out of 2,720 screened studies, 29 studies were included in the statistical analysis. Bayesian methods were used to estimate a standardised mean difference (SMD), log response ratio (lnRR) with exponentiated percentage change (Exp. % Change of lnRR), and probability of direction (pd) for each outcome.

Results: Absolute increases in muscle size slightly favoured males compared to females (SMD = 0.19 (95% HDI: 0.11 to 0.28); pd = 100%), however, relative increases in muscle size were similar between sexes (Exp. % Change of lnRR = 0.69% (95% HDI: -1.50% to 2.88%)). Outcomes were minimally influenced by the measure of muscle size and not influenced by RT experience of participants. Absolute hypertrophy of upper-body but not lower-body regions was favoured in males. Type I muscle fibre hypertrophy slightly favoured males, but Type II muscle fibre hypertrophy was similar between sexes.

Conclusion: Our findings strengthen the understanding that females have a similar potential to induce muscle hypertrophy as males (particularly when considering relative increases in muscle size from baseline) and findings of our secondary analyses should inform future research that investigates sex differences in highly trained participants and muscle fibre type-specific hypertrophy.

Keywords: Gender difference; Muscle hypertrophy; Muscle size; Resistance training; Sex difference.

Figure 1. PRISMA flow chart.

Summary of systematic literature search and article selection process.

Figure 2. Meta-analysis of standardised mean differences to assess absolute changes in muscle size from pre- to post-intervention between males and females.

Positive values favour greater increases in muscle size for male participants. Point (mean) estimates and 95% high density credible intervals are shown by the point and interval line below each posterior distribution. Red vertical lines represent the point estimate (solid) and width of the highest density credible interval (dotted) for the pooled effect size. Standardised mean differences shown are adjusted towards the overall mean, known as shrinkage. Note: Walsh et al. (2009), Sterczala et al. (2024), Schwanbeck et al. (2020), Rissanen et al. (2022), Ribeiro et al. (2014), Reece et al. (2023), Psilander et al. (2019), Peterson et al. (2011), O’Hagan et al. (1995), Nunes et al. (2020), Moesgaard et al. (2022), McMahon et al. (2018), Lundberg et al. (2019), Kosek et al. (2006), Kojic, Mandic & Ilic (2021), Ivey et al. (2000), Hurlbut et al. (2002), Hubal et al. (2005), Hammarstrom et al. (2020), Hakkinen et al. (2001), Hakkinen et al. (1998), Grandperrin et al. (2024), Fernandez-Gonzalo et al. (2014), Cureton et al. (1988), Coratella et al. (2018), Alway et al. (1992), Abou Sawan et al. (2021), Abe et al. (2000).

Figure 3. Meta-analysis of log response ratios (converted to exponentiated percentage changes) to assess relative changes in muscle size from pre- to post-intervention between males and females.

Positive values favour greater increases in muscle size for male participants. Point (mean) estimates and 95% high density credible intervals are shown by the point and interval line below each posterior distribution. Red vertical lines represent the point estimate (solid) and width of the highest density credible interval (dotted) for the pooled effect size. Exponentiated log response ratios are adjusted towards the overall mean, known as shrinkage. Note: Walsh et al. (2009), Sterczala et al. (2024), Schwanbeck et al. (2020), Rissanen et al. (2022), Ribeiro et al. (2014), Reece et al. (2023), Psilander et al. (2019), Peterson et al. (2011), O’Hagan et al. (1995), Nunes et al. (2020), Moesgaard et al. (2022), McMahon et al. (2018), Lundberg et al. (2019), Kosek et al. (2006), Kojic, Mandic & Ilic (2021), Ivey et al. (2000), Hurlbut et al. (2002), Hubal et al. (2005), Hammarstrom et al. (2020), Hakkinen et al. (2001), Hakkinen et al. (1998), Grandperrin et al. (2024), Fernandez-Gonzalo et al. (2014), Cureton et al. (1988), Coratella et al. (2018), Alway et al. (1992), Abou Sawan et al. (2021), Abe et al. (2000).