Transient global amnesia (TGA) is a clinical syndrome characterized by the sudden onset of a massive episodic memory deficit that spares other cognitive functions. As such, it provides a unique human amnesia model for testing the enactment effect (i.e., better memory for performed actions than for verbally encoded sentences). Our main aim was to test whether the enactment effect is preserved in TGA patients, both to have a better understanding and to test the robustness of this effect in a massive amnesia. Object-action pairs were encoded under four conditions: verbal, experimenter-performed, and two enacted conditions (self-performed and self-performed with choice). We tested object-action pair retrieval using cued recall (CR) and recognition tasks, and source memory using a free recall task. We also assessed binding, executive functions, short-term memory, episodic memory, anxiety and mood. We run correlations to control for their putative effects on memory for action. Data were collected from 24 patients, 16 of whom were examined during the acute phase and eight the day-after, as well as from 18 healthy controls. The memory performances of the patients in the acute phase improved for both (i) the CR score, between the verbal, experimenter-performed and self-performed with choice conditions, and (ii) the total recognition score, between the verbal condition and the two enacted conditions. Correlations were found between self-performed task (SPT) enhancement and both the binding and anxiety. In spite of their severely impaired episodic memory, patients with TGA benefit from the enactment effect. These results are discussed in relation to the role of motor components and episodic integration in memory for actions. We suggest that enactment effect can be used in clinical practice and rehabilitation, possible even for patients with a massive memory impairment.