Abstract
PURPOSE: To estimate incidence rates of adverse events associated with bone-conduction hearing implants from primary literature and to compare rates among different technological designs. METHODS: A systematic literature review and meta-regression was conducted to estimate incidence rates of minor and major complications and their consequences (i.e., revision surgery, explantation, re-implantation and becoming a non-user) while testing for effects of device design, age group, mean follow-up time and study type. These four designs of bone-conduction systems were included: 1) active transcutaneous with electromagnetic transducer (aBCIem), 2) active transcutaneous with piezoelectric transducer (aBCIpz), 3) passive transcutaneous (tBAHA), and 4) passive percutaneous (pBAHA). RESULTS: The final dataset included 170 articles reporting on 6451 implantations and 1847 minor and 668 major events. Mean follow-up time was a significant predictor of incidence rates (p < 0.001), with lower rates reported in studies with longer follow-up times. After adjusting to the median follow-up time, the pooled incidence rate of minor complications was significantly lower in aBCIem (p < 0.05) compared to other designs. For both major events and revision surgery, pooled incidence rates were significantly higher in pBAHA compared to aBCIem (p < 0.001) and tBAHA (p < 0.001), but not compared to aBCIpz (Major: p = 0.197; Revision: p = 0.248). Becoming a non-user occurred significantly more frequently in tBAHA compared to other designs (p < 0.005). No statistically significant differences were found in rates of explantation and explantation with re-implantation. CONCLUSION: When comparing across multiple studies, adverse event rates should be adjusted for different lengths of follow-up. Synthesizing published evidence without considering follow-up time may lead to false conclusions.