3.3.2. Effects of HMB supplementation on testosterone concentration
As outlined in Table 4 , our preliminary analysis indicated that HMB consumption did not significantly improve serum testosterone concentration (WMD= 0.75 nmol/L; 95% CI: -0.09, 1.58; P= 0.079). Also, no significant heterogeneity was observed among studies (Cochran’s Q test = 3.33, P= 0.853, I2 = 0.00%). Subgroup analysis was conducted to check as to how the effect of HMB supplementation on serum testosterone is different according to type of exercise activities (Figure 3 ) or duration of the interventions (Figure 4 ). This analysis revealed that the pooled effect of HMB supplementation on testosterone concentration was influenced by the duration of the supplementation (<6 weeks versus ≥6 weeks), in which a significant increase in testosterone concentration was found in studies that lasted longer than 6 weeks (WMD= 1.60 nmol/L; 95% CI: 0.08, 3.12; P= 0.039). Also, subgroup analysis based on the type of exercise training indicated that the increasing effect of HMB supplementation on serum testosterone is significant in aerobic and anaerobic combined activities (WMD= 1.60 nmol/L; 95% CI: 0.08, 3.12; P= 0.039). Subgroup analysis didn’t show any significant effect of HMB supplementation on testosterone concentration in studies lasting lower than six weeks (WMD= 0.38; 95% CI: -0.62, 1.38; P= 0.455) or on resistance training exercise (WMD= 0.38 nmol/L; 95% CI: -0.63, 1.38; P= 0.464).