In this paper, achievable rates and energy efficiency in cognitive radio channels are analyzed in the presence of sensing errors. The cognitive secondary users are initially assumed to perform channel sensing and then enter into data transmission phase. Under sensing uncertainty and Gaussian primary user interference assumptions, additive disturbance experienced at the secondary receiver is seen to have a Gaussian mixture density. Achievable rates under such assumptions are examined and a closed-form rate expression is determined. An outage capacity formulation is also derived. Finally, energy efficiency is analyzed by investigating the minimum energy per bit and wideband slope expressions in the low-SNR regime. Through this analysis, tradeoffs and interactions between sensing parameters, sensing performance, throughput, transmission energy, and energy efficiency are identified.