Exercise interventions for older adults: a systematic review of meta-analyses

Key points

1. 01Combined digital exercise-plus-nutrition programs improved handgrip strength (MD=2.21 kg) and quality of life (SMD=0.65) significantly more than exercise-only digital programs
2. 02Standalone digital exercise did not significantly improve appendicular skeletal muscle mass index, timed up-and-go test, or gait speed overall
3. 03Digital exercise improved total skeletal muscle mass (MD=0.90 kg, 95% CI 0.31-1.49) compared to maintaining original lifestyle or health education
4. 04Most included studies had high or some-concerns risk of bias, and GRADE certainty was low or very low for the majority of outcomes
5. 0568.75% of studies were conducted in China, limiting generalizability to other healthcare contexts

How it was conducted

Design

Systematic review and meta-analysis (PRISMA guidelines, PROSPERO CRD42024516930)

Databases searched

11 databases including Medline, Embase, Cochrane, CINAHL, PsycINFO, WOS, Scopus, and four Chinese databases, searched up to 1 October 2025

Participants

Older adults aged 60 and older with possible, confirmed, or severe sarcopenia

Included studies

16 studies (n=1,607 participants), published 2017-2025, from China (68.75%), Netherlands (12.5%), Korea (12.5%), and Italy (6.25%)

Interventions compared

Digital health exercise alone, digital exercise plus nutrition, digital exercise plus education, or digital nutrition alone, versus offline or no active control

Quality assessment

Cochrane RoB 2 and ROBINS-I for risk of bias; GRADE framework for certainty of evidence

What they found

• Handgrip strength (all digital vs control): MD=1.07 kg, 95% CI [0.30, 1.84], p=0.007, I2=95%; subgroup digital exercise plus nutrition vs control: MD=2.21 kg, 95% CI [1.33, 3.09]
• Appendicular skeletal muscle mass index (overall): MD=0.16 kg/m2, 95% CI [-0.03, 0.36], p=0.10, I2=82% (non-significant)
• Total skeletal muscle mass: MD=0.90 kg, 95% CI [0.31, 1.49], p=0.003, I2=54%
• Timed up-and-go test: SMD=-0.02, 95% CI [-0.40, 0.37], p=0.94, I2=80% (non-significant)
• Sit-to-stand test: MD=-0.52 s, 95% CI [-1.04, 0.00], p=0.05, I2=43%; after removing one influential study: MD=-0.71, p=0.003, I2=23%
• Gait speed: SMD=0.06, 95% CI [-0.44, 0.56], p=0.82, I2=84% (non-significant); after removing two outlier studies heterogeneity resolved (I2=0%), null effect confirmed (SMD=0.04, p=0.72)
• Quality of life: SMD=0.28, 95% CI [0.04, 0.51], p=0.02, I2=18%; subgroup digital exercise plus nutrition vs health education: SMD=0.65, 95% CI [0.29, 1.01]
• BMI: MD=-0.06 kg/m2, 95% CI [-1.28, 1.17], p=0.93, I2=90%; subgroup analysis found opposing directions: exercise-focused interventions reduced BMI while nutrition-focused interventions increased it (p<0.00001, I2=95.6%)
• Overall risk of bias for objective outcomes: 18.75% low risk, 37.5% some concerns, 43.75% high risk
• GRADE certainty: low or very low for most outcomes; high certainty found only for combined digital interventions improving quality of life vs health education alone

Limitations

• 43.75% of included studies were rated high risk of bias for objective outcomes and 56.25% for subjective outcomes, undermining confidence in pooled estimates
• Most interventions lasted only 12 weeks (56.25% of studies) with only 12.50% including follow-up, leaving long-term sustainability unknown
• Over two-thirds of studies were conducted in China with predominantly AWGS2 diagnostic criteria, limiting generalizability to Western populations and other healthcare systems
• Substantial and often extreme statistical heterogeneity (I2 up to 95%) for several key outcomes was not fully explained even by subgroup analysis

Why it matters

For patients

Older adults with sarcopenia may benefit most from programs that combine digital exercise with nutritional support rather than exercise alone, though evidence quality is low and access to such programs may vary.

For clinicians

Combined digital exercise-plus-nutrition interventions are the most evidence-supported digital option for improving handgrip strength and quality of life in sarcopenic older adults, but the low certainty of evidence means these should complement, not replace, face-to-face care where available.

For readers

This meta-analysis synthesizes a rapidly growing but methodologically fragile evidence base, identifying combined digital interventions as most promising while flagging that most pooled estimates come from low- or very-low-certainty studies with high heterogeneity.