Conserving energy has been known as the most significant problem in all facets of sensor network operation. Particularly in routing, researchers were concerned with the problems of using topology control to achieve power efficient routes, as well as finding alternate routes to extend the lifetime of the network. Current schemes though, fail in one or more of the following areas: 1) provide solutions that treat both the aforementioned problems, 2) provide distributed algorithms, 3) account for the overhead of the algorithms 4) compare the results with theoretically computed optimums, 5) account for the effect of the MAC layer. We propose and evaluate a unicast routing algorithm that exploits the ability of the nodes to transmit at multiple power levels. It can find the optimal power-efficient route between two nodes, with less energy and time overheads than the distributed Bellman-Ford algorithm, as well as use alternate routes to extend the total lifetime of the network, up to 87% of the theoretical optimum lifetime, taking into account all overheads. Furthermore, we address many practical considerations in the context of sensor networks. We built our case around three main pillars: i) a distributed algorithm, ii) extensive evaluation of the overhead of the algorithm, and iii) account for the effect of the MAC layer.