Effects of resistance training on insulin sensitivity in the elderly: A meta-analysis of randomized controlled trials

We performed a systematic review and meta-analysis of randomized controlled trials (RCTs) assessing the effect of resistance training in older adults on insulin sensitivity.

Cochrane, Embase, PubMed, Web of Science and EBSCO were searched from inception to April 2021. We integrated randomized controlled trials published in English, and participants were non-athletic and aged ≥60 years. The outcome of interest was the change in insulin sensitivity, derived from the homeostatic model of insulin resistance (HOMA-IR) and hemoglobin A1c (HbA1c).

12 RCTs were included in this meta-analysis comparing resistance training (n = 232) with control (n = 209). Resistance exercise significantly reduced HOMA-IR level (d = -0.25, 95% CI, -0.43 to -0.06; P < 0.05) and HbA1c levels of (d = -0.51, 95% CI, -0.84 to -0.18; P < 0.05). Subgroup analysis of HOMA-IR revealed that the variables "population", "training intensity" and "period" had significant effects on HOMA-IR, with the largest effect sizes for high-intensity (d = -0.43, 95%CI, -0.85 to -0.22, P < 0.05) and long-term (more than 12 weeks) (d = -0.43, 95%CI, -0.85 to -0.22, P < 0.05) training programs in older adults without type 2 diabetes (T2D) (d = -0.23, 95%CI, -0.42 to -0.04, P < 0.05). Subgroup analysis of HbA1c suggested that resistance training programs with moderate intensity (d = -0.51, 95%CI, -0.90 to -0.12, P < 0.05) and short term (less than or equal to 12 weeks) (d = -0.49, 95%CI, -0.84 to -0.14, P < 0.05) have greater effects on HbA1c.

The findings of this meta-analysis suggest that resistance training is effective for inducing improvement in insulin sensitivity for elderly. Subgroup analysis showed that high intensity and long period of resistance exercise improve HOMA-IR in healthy old adults, and that resistance training with moderate intensity and short period improve HbA1c in T2D old people. More studies with high methodological qualities and large sample sizes need to be done to further confirm our conclusion.