This study aims the throughput maximization of carrier sense multiple access (CSMA) in random access cognitive radio networks (CRNs) by using cooperative spectrum sensing. In the cooperative spectrum sensing, multiple cognitive radio (CR) users individually monitor the spectrum for the primary users (PUs)' activities and send their local decisions to the CR access point. Upon receiving the individual decisions, the CR access point gives the final decision about the PU's presence at the spectrum and sends it to all CR users in the network. At decision stage, CR access point utilizes the k-out-of-N decision fusion rule with optimal k value in order to obtain maximum CRN throughput. In this paper, we model a CRN utilizing CSMA medium access control scheme and focus on analyzing the throughput performance of it. Throughput performances of our CRN model which separately utilizes AND, OR, MAJORITY fusion rules and k-out-of-N decision fusion rule with the optimum k value are obtained and compared under the same network level probability of detection constraint. The throughput results obtained show that our CRN model utilizing k-out-of-N decision fusion rule with the optimum k value achieves higher throughput performance than AND, OR and MAJORITY fusion rules when they all have the same CRN load.