© 2019 ACM.Severe characteristics and convergecast nature of underwater acoustic channels make underwater sensor networks (USNs) vulnerable to malicious attacks. One of the most malicious attacks in USNs is the sinkhole attack, where an adversary first captures a sensor node and then lures the surrounding network traffic by using false routing information. Later, the captured node can forward the captured network traffic to the intruder or drop the packets. Sinkhole attacks negatively affect the lifetime, end-to-end latency, and energy-efficiency of USNs since lured nodes spend energy in an unbalanced manner and the forwarding process introduces additional latency. In this work, we investigate the lifetime, end-to-end latency, and energy consumption performances of USNs under sinkhole attacks within an integer programming (IP) model which maximizes USNs lifetime. Our results show that if half of the nodes in the network are lured by a sinkhole node, the network lifetime decreases at a minimum of 71%; the end-to-end latency and energy consumption are increased at least by 89% and 77% as compared to the performance metrics which are obtained in the case of no sinkhole attacks.