Image-based deblurring methods encounter difficulties in handling complex motion blur under real-world conditions, primarily due to significant information loss. Event-based methods alleviate this issue by using events to provide detailed motion information. However, discrepancies in spatial resolution between images and events pose a major challenge for event-based deblurring. To tackle this problem, we propose a Cross-Modal Scale-Aware Fusion Network (CSFNet), inspired by the fact that events predominantly capture high-contrast edge information. The network includes a Spatial Scale-Aware Super-Resolution Module (SASR), which adaptively aligns event resolution with image resolution by incorporating refined image features. Furthermore, a Global Interaction Fusion Module (GIFM) is introduced to improve cross-modal feature integration through global spatial aggregation and channel-wise correlation. Experimental evaluations on both synthetic and real-world datasets validate the effectiveness of our approach, which achieves a PSNR gain of 0.12 dB on the Ev-REDS dataset.