Figure 1.

The decomposition of effects for the relationship between groups (control vs. long-COVID-19 syndrome) and muscle strength parameters, mediated through appendicular lean mass index. Nodes represent the variables being analyzed in the mediation model. Group in the equations below represents a dichotomous variable (long-COVID-19 syndrome = 1, control = 0). The total effect, denoted by path c, is the lineal regression coefficient of groups (control vs. long-COVID-19 syndrome) on the outcomes (muscle strength parameters) without controlling for the mediator. The direct effect, denoted by path c′, is the effect coefficient of exposure groups in a lineal regression with groups (control vs. long-COVID-19 syndrome), and the mediator on the outcomes. Path a denotes the effect coefficient of exposure groups on appendicular lean mass index. Path b denotes the effect coefficient of appendicular lean mass index on the outcomes, in the model with both the exposure and mediator. The indirect effect is the component of the total effect that is mediated through appendicular lean mass index (c–c′) in the relationship between exposure group and the outcomes; it is denoted by β₁ × β₃.