The relay-aided multicast network is considered, where an NT -antenna source multicasts confidential messages to N single-antenna legitimate users via a self-sustainable M-antenna regenerative relay. In particular, the relay is powered by the energy harvested from the radio signal of the source, and there are K unauthorized eavesdroppers wiretapping the channel. Assuming the knowledge of statistical channel state information of eavesdroppers, we aim to minimize the source transmission power via energy-efficient beamforming, subject to the signal-to-noise ratios of legitimate users/relay, the power constraint at the relay, and the outage constraints of the eavesdroppers. An efficient algorithm is developed by using the iterative first-order Taylor expansion and successive convex approximation, where the original nonconvex problem is transformed and solved.