Energy efficiency is a crucial performance metric in sensor networks, imminently determining the network lifetime. Consequently, a key objective in WSN is to improve overall energy efficiency to extend the network lifetime. Its conservation influences the topology design of many WSN-based systems, especially the clustering of the network. Unlike other WSN clustering algorithms, that do not re-cluster the network after deployment, our hypothesis is that it is advisable, in terms of prolonging the network lifetime, to adaptively re-cluster specific regions that are triggered significantly more than other regions in the network. By doing so, it is possible to minimize or even prevent the premature death of CHs, which are heavily burdened with sensing and transmitting actions – much more than other parts of the WSN. In order to do so we introduce the Adaptive Clustering Refinement (ACR) algorithm, which is based on the Adaptive Mesh Refinement algorithm by Berger and Oliger [14] and the Hierarchical Control Clustering algorithm by Banerjee and Khuller [13]. We prove that the ACR algorithm complexity is linear in the total size of the graph, and that we manage to optimize the WSN cluster connectivity and prolong its lifetime. We also devise a local version of the algorithm with improved complexity.