BACKGROUND: Little is known about the causal relationship between sarcopenia and social isolation. This study investigated the association between social isolation and sarcopenia and explored the mediating role of physical activity among American and Chinese older adults. METHODS: Data were from the Health and Retirement Study (HRS) and China Health and Retirement Longitudinal Study (CHARLS), encompassing 3,383 American and 6,996 Chinese. Sarcopenia was defined by the European Working Group on Sarcopenia in Older People 2 and Asian Working Group for Sarcopenia 2019 criteria. Social isolation was measured by the Steptoe Social Isolation Index for American and a 5-item scale for Chinese. Logistic regression models with inverse probability weighting were conducted to explore the effect of social isolation on sarcopenia risk. Causal mediation analysis was conducted to examine direct and indirect associations. RESULTS: Over 8-year and 4-year follow-ups, 363 Americans and 392 Chinese had sarcopenia. Social isolation was associated with higher risk of sarcopenia in HRS (OR = 1.14, 95% CI = 1.02-1.29) and CHARLS (OR = 1.24, 95% CI = 1.10-1.39). The indirect effects of social isolation on sarcopenia through physical activity were significant only in HRS, explaining 7.63% of the total effect. Handgrip strength and gait speed were significantly associated with social isolation for HRS and CHARLS, while muscle mass and chair-stand test performance were significantly associated only in CHARLS. CONCLUSIONS: Social isolation significantly increased the risk of sarcopenia. Physical activity partially mediated this association only in Americans. More attention should be paid on socially isolated older adults to prevent sarcopenia.