The Effect of Weekly Set Volume on Strength Gain: A Meta-Analysis

Abstract

Background: Strength training set organisation and its relationship to the development of muscular strength have yet to be clearly defined. Current meta-analytical research suggests that different population groups have distinctive muscular adaptations, primarily due to the prescription of the strength training set dose.

Objectives: We conducted a meta-analysis with restrictive inclusion criteria and examined the potential effects of low (LWS), medium (MWS) or high weekly set (HWS) strength training on muscular strength per exercise. Secondly, we examined strength gain variations when performing multi-joint or isolation exercises, and probed for a potential relationship between weekly set number and stage of subjects' training (trained versus untrained).

Methods: Computerised searches were performed on PubMed, MEDLINE, SWETSWISE, EMBASE and SPORTDiscus™ using the terms 'strength training', 'resistance training', 'single sets', 'multiple sets' and 'volume'. As of September 2016, 6962 potentially relevant studies were identified. After review, nine studies were deemed eligible per pre-set inclusion criteria. Primary data were pooled using a random-effect model. Outcomes for strength gain, strength gain with multi-joint and isolation exercise were analysed for main effects. Sensitivity analyses were calculated for several subgroups by separating the data set and by calculation of separate analyses for each subgroup. Heterogeneity between studies was assessed using the Cochran Q and I ² statistics.

Results: Pre- versus post-training strength analysis comprised 61 treatment groups from nine studies. For combined multi-joint and isolation exercises, pre- versus post- training strength gains were greater with HWS compared with LWS [mean effect size (ES) 0.18; 95% CI 0.06-0.30; p = 0.003]. The mean ES for LWS was 0.82 (95% CI 0.47-1.17). The mean ES for HWS was 1.01 (95% CI 0.70-1.32). Separate analysis of the effects of pre- versus post-training strength for LWS or MWS observed marginally greater strength gains with MWS compared with LWS (ES 0.15; 95% CI 0.01-0.30; p = 0.04). The mean ES for LWS was 0.83 (95% CI 0.53-1.13). The mean ES for MWS was 0.98 (95% CI 0.62-1.34). For multi-joint exercises, greater strength gains were observed with HWS compared with LWS (ES 0.18; 95% CI 0.01-0.34; p = 0.04). The mean ES for LWS was 0.81 (95% CI 0.65-0.97). The mean ES for HWS was 1.00 (95% CI 0.77-1.23). For isolation exercises, greater strength gains were observed with HWS compared with LWS (ES 0.23; 95% CI 0.06-0.40; p = 0.008). The mean ES for LWS was 0.95 (95% CI 0.30-1.60). The mean ES for HWS was 1.10 (95% CI 0.26-1.94). For multi-joint and isolation exercise-specific one repetition maximum (1 RM), marginally greater strength gains were observed with HWS compared with LWS (ES 0.14; 95% CI -0.01 to 0.29; p = 0.06). The mean ES for LWS was 0.80 (95% CI 0.47-1.13). The mean ES for HWS was 0.97 (95% CI 0.68-1.26).

Conclusion: This meta-analysis presents additional evidence regarding a graded dose-response relationship between weekly sets performed and strength gain. The use of MWS and HWS was more effective than LWS, with LWS producing the smallest pre- to post-training strength difference. For novice and intermediate male trainees, the findings suggest that LWSs do not lead to strength gains compared with MWS or HWS training. For those trainees in the middle ground, not a novice and not advanced, the existing data provide a relationship between weekly sets and strength gain as set configurations produced different pre- to post-training strength increases. For well trained individuals, the use of either MWS or HWS may be an appropriate dose to produce strength gains.

Fig. 1

Flow of journal articles through the systematic review process

Fig. 2

Galbraith plot used to examine study heterogeneity (pre- vs post-intervention strength change). Each open circle represents one pre- vs post-intervention study datum. Two pre- vs post-intervention study data of Reid [24] identified as outliers (solid filled black circles)

Fig. 3

Galbraith plot with the removal of two pre- vs post-intervention study outliers (Reid [24]). Each open circle represents one pre- vs post-intervention study datum

Fig. 4

Funnel plot of standard error (SE) by mean difference (MD) for assessment of publication bias. Each open circle denotes a study included in the meta-analysis. The dashed vertical line represents the overall effect calculated with the random-effects model

Fig. 5

Forest plot of LWS vs HWS (MWS and HWS combined) on multi-joint and isolation exercise by study. The vertical line indicates the overall estimate of combined multi-joint and isolation studies mean effect size. The horizontal line indicates 95% CI, squares indicate estimates, whereas square size is proportional to sample size, and rhombus indicates meta-analytically pooled estimates 95% CI. 95% CI 95% confidence interval, HWS high weekly sets per exercise (≥10), IV inverse variance, LWS low weekly sets per exercise (≤5), MWS medium weekly sets per exercise (5–9)

Fig. 6

Forest plot of LWS vs MWS on multi-joint and isolation exercise by study. The vertical line indicates the overall estimate of combined multi-joint and isolation studies mean effect size. The horizontal line indicates 95% CI, squares indicate estimates, whereas square size is proportional to sample size, and rhombus indicates meta-analytically pooled estimates 95% CI. 95% CI 95% confidence interval, IV inverse variance, LWS low weekly sets per exercise (≤5), MWS medium weekly sets per exercise (5–9)

Fig. 7

Forest plot of LWS vs HWS (MWS and HWS combined) on multi-joint exercises by study. The vertical dashed line indicates the overall estimate of multi-joint studies mean effect size. Horizontal lines indicate 95% CI, squares indicate estimates, whereas square size is proportional to sample size, and rhombus indicates meta-analytically pooled estimates 95% CI. 95% CI 95% confidence interval, HWS high weekly sets per exercise (≥10), IV inverse variance, LWS low weekly sets per exercise (≤5), MWS medium weekly sets per exercise (5–9)

Fig. 8

Forest plot of LWS vs MWS on isolation exercises by study. The vertical dashed line indicates the overall estimate of isolation studies mean effect size. Horizontal lines indicate 95% CI, squares indicate estimates, whereas square size is proportional to sample size, and rhombus indicates meta-analytically pooled estimates 95% CI. 95% CI 95% confidence interval, IV inverse variance, LWS low weekly sets per exercise (≤5), MWS medium weekly sets per exercise (5–9)

Fig. 9

Forest plot of LWS vs HWS (MWS and HWS combined) on multi-joint and isolation exercise-specific 1RM by study. The vertical dashed line indicates the overall estimate of combined multi-joint and isolation studies mean effect size. Horizontal lines indicate 95% CI, squares indicate estimates, whereas square size is proportional to sample size, and rhombus indicates meta-analytically pooled estimates 95% CI. 1RM one repetition maximum, 95% CI 95% confidence interval, HWS high weekly sets per exercise (≥10), IV inverse variance, LWS low weekly sets per exercise (≤5), MWS medium weekly sets per exercise (5–9)