Spinal fusion (SF) comprises surgical procedures for several pathologies that affect different spinal levels, and different cages are employed in SF surgery. Few clinical studies highlight the role of cages in complications beyond the outcomes. The aim of this systematic review is to collect the last 10 years’ worth of clinical studies that include cages in SF surgery, focusing on complications. Three databases are employed, and 21 clinical studies are included. The most-performed SF procedure was anterior cervical discectomy and fusion (ACDF), followed by lumbar SF. The polyetheretherketone (PEEK) cage was the most-used, and it was usually associated with autograft or calcium phosphate ceramics (hydroxyapatite (HA) and tricalcium phosphate (βTCP)). For lumbar SF procedures, the highest percentages of subsidence and pseudoarthrosis were observed with PEEK filled with bone morphogenetic protein 2 (BMP2) and βTCP. For ACDF procedures, PEEK filled with autograft showed the highest percentages of subsidence and pseudoarthrosis. Most studies highlighted the role of surgical techniques in patient complications. There are many interacting events that contextually affect the rate of clinical success or failure. Therefore, in future clinical studies, attention should focus on cages to improve knowledge of chemical, biological and topographical characteristics to improve bone growth and to counteract complications such as cage loosening or breaking and infections.