Definition-specific prevalence estimates for sarcopenia in an Australian population: the Geelong Osteoporosis Study

Background

We aimed to compare prevalence estimates for sarcopenia in an Australian sample of older men and women by using different criteria.

Methods

Women (n = 323) and men (n = 342) aged 60–96 years from the Geelong Osteoporosis Study were included. Handgrip strength (HGS) was measured by dynamometer (Jamar or Vernier) and appendicular lean mass (ALM) by whole-body densitometry (Lunar). Sarcopenia definitions included European Working Group on Sarcopenia in Older People (EWGSOP) 1, EWGSOP2, and US Foundation for the National Institutes of Health (FNIH). Sarcopenia was identified as low HGS and low ALM/height² or low muscle performance, and low HGS and low ALM/body mass index (BMI). Prevalence estimates were standardized to the Australian population and agreement between definitions assessed using the Cohen kappa statistic (κ).

Results

Low HGS was identified in 13.7–29.9% of women and 2.1–14.1% of men. Low ALM/height² was identified in 7.1–11.8% of women and 6.0–8.4% of men, while 21.7% of women and 21.1% of men had low ALM/BMI. Mean age-standardized prevalence estimates for sarcopenia were 5.9% (95% confidence interval 3.4–8.4) for women and 2.9% (1.9–4.0) for men (EWGSOP1), 2.3% (1.1–3.4) for women and 0.5% (0.2–0.9) for men (EWGSOP2), and 4.0% (2.1–5.8) for women and 1.1% (0.6–1.5) for men (FNIH). There was moderate agreement between EWGSOP1 and EWGSOP2 (κ = 0.58 women, 0.30 men) and poor agreement between FNIH and EWGSOP1 (κ = 0.16 women, 0.12 men) and EWGSOP2 (κ = 0.19 women, 0 men).

Conclusions

Sarcopenia prevalence differed according to definition applied. Point estimates for sarcopenia prevalence according to EWGSOP2 identified fewer individuals than EWGSOP1, with FNIH estimates between the two; however, there were overlapping 95% confidence intervals across definitions.