Abstract In the last decade, the heat energy dissipated in a unit volume of material per cycle (the Q parameter) has been adopted by the authors as a fatigue damage indicator of metallic materials. The advantage of using such a parameter is that it can be readily and in-situ measured at a point or a component undergoing fatigue solicitations. Geometrical, mean stress and variable amplitude (limited to two stress-level tests) effects have been successfully analysed by using the Q parameter. Concerning geometrical effects, approximately 160 experimental results generated from constant amplitude, completely reversed, stress- or strain-controlled fatigue tests on plain or notched hot rolled as well as cold drawn stainless steel specimens have been rationalised. Afterwards, the heat-energy based approach was extended to include the mean stress effect, by using a thermodynamic fatigue damage variable that combines two parameters, i.e. Q and the thermoelastic temperature achieved by the material at the maximum stress of the load cycle. Finally, Q was used to rationalise two stress-level fatigue test results, by using the Q-based fatigue curve combined with Miner’s rule. In this paper, the theoretical background and the application of the energy-based approach are reviewed in order to analyse all previously mentioned effects, focusing mainly on the mean stress and the variable amplitude, two stress-level effects.