Empirical studies in human-centered Explainable AI (XAI) showed that, even with explanations, users as decision-makers often over- or under-rely on AI advice. However, how explanation design is linked to user reliance remains undercharacterized. To address this gap, we analyzed 61 explanation design strategies with quantitative evaluation to examine their association with reliance behaviors. Our findings reveal local explanations, combined with textual/graphical outputs in recommendation tasks, as the most observed strategy associated with an increased appropriate reliance. In contrast, we identified no recurring strategies regarding a decreased over- and under-reliance. To support designers in taking into account reliance issues, we synthesized a concept matrix, which characterizes how different strategies of explanation formats, modalities, and XAI approaches are associated with reliance behaviors across contexts of use. Grounded on our findings, we identify several open challenges in the domain.