Recent years have witnessed a growing interest in the application of wireless sensor networks in unattended environments. Nodes in such applications are equipped with limited energy supply and need careful management in order to extend their lifetime. In order to conserve energy, many of the routing protocols proposed for wireless sensor networks reduce the number of transmitted packets by pursuing in-network data aggregation. almost all of the aggregation schemes presented in the literature strive to save sensor's energy while considering unconstrained data traffic. However, aggregation extends the queuing delay at the relay nodes and can thus complicate the handling of latency-constrained data. In this paper, we analyze the conditions for effective aggregation of data traffic that is subject to end-to-end delay constraints. We present an algorithm for achieving maximal possible energy savings through data aggregation while meeting the desired level of timeliness. A Weighted Fair Queuing based mechanism for packet scheduling is employed at each node in order to perform service differentiation and ensure bounded delay for constrained traffic. The performance of the proposed approach is qualified via simulation.