The cognitive approach, which is a key factor for the reuse of available spectrum in wireless networks, enables secondary users (SUs) to share the same band of primary users (PUs) by limiting the interference caused to the latter. Focusing on interference channels, this work develops a distributed power allocation (PA) strategy aimed at minimizing the coded packet error rate (PER) of noncooperative SUs’ links, with the constraint of satisfying the interference limit at PU receivers and considering SUs based on bit-interleaved-coded orthogonal frequency division multiplexing technology. The proposed PA is formulated as a noncooperative game, relaxed by applying a pricing mechanism and solved by an iterative algorithm. Then, given the PA strategy, an adaptive modulation and coding scheme to further improve performance is devised. Finally, numerical simulations show the effectiveness of the proposed algorithms.