Risk factors for breakthrough vitreous hemorrhage after intravitreal tissue plasminogen activator and gas injection for submacular hemorrhage associated with age related macular degeneration

Purpose We investigated risk factors for breakthrough vitreous hemorrhage (VH) after an intravitreal tissue plasminogen activator (tPA) and gas injection in patients with submacular hemorrhage (SMH) associated with age-related macular degeneration (AMD). Methods The medical records of patients diagnosed with SMH associated with AMD who received an intravitreal tPA (50 μg/0.05 mL) and perfluoropropane gas (0.3 mL) injection were reviewed retrospectively. We analyzed the associations of breakthrough VH with age, sex, best-corrected visual acuity, intraocular pressure, AMD subtype, accompanying sub-retinal pigment epithelium (RPE) hemorrhage, history of cataract surgery, history of hypertension and diabetes mellitus, history of drinking and smoking, and history of antiplatelet or anticoagulant medication. We also examined the relationships between various parameters, including the area ratio of the SMH to the optic disc (AHD) and the height of the SMH obtained from optical coherence tomography. Results In total, 52 eyes from 52 patients were enrolled in this study; 16 eyes (30%) showed breakthrough VH. The proportions of patients with a current smoking history were 75.0% in the VH group and 22.2% in the non-VH group (p = 0.010). Other factors did not differ significantly between the two groups. The proportion of cases with accompanying sub-RPE hemorrhage was 50.0% and 58.3% in the VH and non-VH groups, respectively (p = 0.763). The AHD (p = 0.001) and SMH height (p < 0.001) were significantly greater in the VH group. In a receiver operating characteristic curve analysis, the cut-off values of AHD and SMH height were 20.1 and 1208 μm, respectively. According to logistic regression analysis, when the AHD and SMH height were greater than the individual cut-off values, the odds ratio of VH increased by 10.286 fold (95% confidence interval [CI], 2.452–43.148; p = 0.001) and 75.400 fold (95% CI, 7.991–711.441; p < 0.001), respectively, with respect to their respective reference groups (less than the cut-off value). Among the significant factors associated with VH occurrence, including current smoking, AHD, and SMH height, only current smoking and SMH height were found to be significant in multiple regression analysis (p = 0.040, 0.016). Conclusions The incidence of breakthrough VH was significantly higher in those with current smoking status and for SMH with a larger AHD and greater height. The height of the SMH was more predictable of the possibility of VH than AHD.