In the context of geographic routing in wireless sensor networks linked by fading communication channels, energy-efficient transmission is important to extend the network lifetime. To this end, we propose a novel method to minimize the energy consumed by one bit of information per meter and per second toward the destination in fading channels. Using the outage probability as a measure to maximize the amount of information delivered within a given time interval we decide energy-efficient geographic routing between admissible nodes in a wireless sensor network. We present three different approaches, the first is optimal and is obtained by varying both transmission rate and power, the other two are sub-optimal since only one of them is tuned. Simulation examples comparing the energy costs for the different strategies illustrate the theoretical analysis in the cases of log-normal and Nakagami shadow fading. With the method proposed it is possible to obtain significant energy savings (up to ten times) with respect to fixed transmission rate and power.