Today, with the increasing popularity of Internet of Things (IoT), there is a huge demand of the wireless devices and is creating the problem of spectrum scarcity. On the other hand, spectrum utilization statistics says that the spectrum usage is inefficient causing white spaces; the unutilized portions of the spectrum. Dynamic spectrum usage with Cognitive Radio (CR) devices is the solution of these problems. But the variations in availability of spectrum pose many challenges in designing the protocols at all the layers of the protocol suite. Also, at network layer routing is much difficult. In this paper, a new routing scheme called Hybrid Proactive Reactive Routing (HPRR) for CR Networks (CRNs) has been introduced. It is implemented using clustering. The proposed algorithm is hybrid, as it considers the combination of proactive and reactive mechanisms. The algorithm is proactive by considering that it constructs a routing table by exchanging routing information, but if destination node goes beyond some specific number of hops then it keep the routing information with some in between node until and unless the destination node receives it. Here, far destination node receives it through route request and reply messages. Simulation results of HPRR show enhanced performance in terms of network characteristics like average end to end delay, average throughput and routing overheads. The performance of HPRR is compared against other two routing schemes in CRNs namely Adhoc On-demand Distance Vector (AODV) and Cluster based Cognitive Radio Routing Protocol with Link stability (C2R2M-LS).