Motivated by scenario requirements for 5G cellular networks, we study one of the candidate protocols for massive random access: the family of random access methods known as Coded Slotted ALOHA (CSA). A recent trend in research has explored aspects of such methods in various contexts, but one aspect has not been fully taken into account: the impact of path loss, which is a major design constraint in long-range wireless networks. In this article, we explore the behavior of CSA, by focusing on the path loss component correlated to the distance to the base station. Path loss provides opportunities for capture, improving the performance of CSA. We revise methods for estimating CSA behavior, provide bounds of performance, and then, focusing on the achievable throughput, we extensively explore the key parameters, and their associated gain (experimentally). Our results shed light on the behavior of the optimal distribution of repetitions in actual wireless networks.