Prophylactic antibiotics for anterior nasal packing in emergency department: A systematic review and meta-analysis of clinically-significant infections.

ABSTRACT

BACKGROUND: Patients presenting to emergency departments with spontaneous anterior epistaxis may undergo anterior nasal packing and sometimes receive systemic prophylactic antibiotics. There has not been sufficient evidence to support or refute this practice. The main objective of this study was to compare the likelihood of clinically significant infection (CSI) between patients with or without prophylactic antibiotics for anterior nasal packing due to spontaneous epistaxis. METHODS: We performed a meta-analysis of the literature to assess whether prophylactic antibiotics prevented CSI among patients with anterior nasal packing by searching PubMed, Embase, and Scopus databases for original articles. We also looked at the secondary outcome of non-infectious complications. We reported the outcomes using random effect models. Human studies in English, randomized control trials, quasi-randomized trials, clinical trials, retrospective studies, and case series were included. We excluded studies involving patients undergoing otolaryngologic surgeries. Statistical heterogeneity was examined using the DerSimonian and Laird Q test and I2 statistic. RESULTS: A total of 281 articles were identified. Of these, 5 articles met inclusion criteria, with 383 patients receiving anterior nasal packing. One hundred sixty (42%) patients did not receive prophylactic antibiotics while 223 (58%) received antibiotics. The proportion of CSI in the pooled cohort was 0.8% (95% CI 0.2-1.9), resulting in a number needed to treat (NNT) to prevent one infection of 571. The rate of non-infectious complications associated with epistaxis was 20% (95% CI 10-32). CONCLUSIONS: This meta-analysis suggests that prescribing prophylactic antibiotics for anterior nasal packing may not be necessary due to the low proportion of CSIs across heterogenous patient populations. Further high-quality randomized trials are needed to support this finding.