Examination of the behavior of oxygen and hydrogen during weathering reactions shows that river dissolved load, although widely used, is an imperfect tracer of chemical denudation. At the current state of knowledge, none of the metrics for river total dissolved loads (such as the silicate-derived total dissolved solids, TDSsil = Ca2+ + Mg2+ + Na+ + K+ + SiO2, converted or not to equivalent oxides) account, in a mechanistic manner, for the transfer of oxygen and hydrogen between the solid and fluid phase during weathering reactions. We assess that chemical denudation derived from TDSsil will significantly overestimate the true chemical denudation for weathering of Ca-feldspar to kaolinite, whereas weathering of water-rich sedimentary rocks will be characterized by an underestimation of chemical denudation by TDSsil. For a handful of field sites, we estimate that the bias is lower than ±10%. The sign and extent of the bias depends on the nature of bedrock and on weathering conditions. Altogether, our analysis questions the broadly accepted concept of chemical denudation rate.