Abstract
Purpose :
Post-vitrectomy ocular hemorrhage (PVOH) has been linked to use of oral antithrombotics (OAT). However, there are no large-scale commercial claims data to support this association. This retrospective analysis of patient data characterized the impact of OAT subtypes on the incidence of PVOH among adults in the United States.
Methods :
Data from MarketScan® Commercial Claims and Medicare Supplemental databases were analyzed to identify patients who underwent vitrectomy between January 1, 2009 and September 30, 2015, with/without prior OAT use. OAT subtypes included warfarin, novel oral anticoagulants (NOACs), and antiplatelets; their use was defined as having ≥1 prescription in the 90 days prior to and including the index (surgery) date. The primary outcome measure was the occurrence of ocular hemorrhage (vitreous, preretinal, retinal, subretinal, retrobulbar, suprachoroidal) within 30 days of the index date. Multivariate regression analysis assessed the association between OAT subtype use and PVOH, adjusting for covariates including surgical indication, diabetes, and hypertension.
Results :
This analysis included 73,020 patients aged ≥50 years who underwent vitrectomy in the study period. The incidences of PVOH per 100 cases among patients prescribed (n=7356; 10.1%) vs not prescribed OAT were 8.35 (95% CI 7.71–8.98) vs 4.77 (4.61–4.93), respectively (P < 0.0001). Specifically, the incidences of PVOH per 100 cases across those prescribed warfarin (n=2614; 3.6%), NOACs (n=527; 0.7%), antiplatelets (n=3999; 5.5%), and multiple OAT (n=216; 0.3%) were 8.15 (95% CI 7.10–9.20), 5.31 (3.39–7.23), 8.75 (7.88–9.63), and 10.65 (6.50–14.79), respectively (P = 0.03). Among OAT subtypes, the odds of PVOH were significantly greater in patients prescribed warfarin (adjusted odds ratio 1.27, 95% CI 1.09–1.48; P = 0.0026) or antiplatelets (1.21, 1.07–1.37; P = 0.0025), but not greater in those on NOACs (1.08, 0.72–1.61; P = 0.7149), compared with those not prescribed OAT. Preoperative vitreous hemorrhage, vitrectomies requiring laser, diabetes, and hypertension (all P < 0.001) also were associated with increased odds of PVOH.
Conclusions :
A prescription for warfarin and antiplatelets, but not NOACs, was significantly associated with an increased risk of PVOH. Given the potential visual consequences of ocular hemorrhage, these results may provide guidance for the informed consent process in the event of vitrectomy surgery.
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.