Abstract
Purpose :
Myopia is a refractive error that develops during adolescence and is a known risk factor for rhegmatogenous retinal detachment (RRD). It is predicted that half of the world’s population will be myopic by 2050, with as much as 10% being highly myopic. Therefore, we aimed to determine the trend in incidence of RRD in high myopes in the United States over ten years.
Methods :
A retrospective cohort study. The IBM MarketScan database composed of privately insured medical claims from over 150 million beneficiaries from all 50 United States. International Diagnosis Codes (ICD) were used to identify high myopes and to identify those diagnosed with RRD. We specified Cox proportional hazards models to calculate the risk of RRD given age, sex, high myopia diagnosis, and interactions between each of those variables. We ran survival models and calculated risks after restricting the sample to 1 year of continuous insurance enrollment, 2 years, etc. up to 10 years. Data were analyzed using Stata 14.2 (Stata Corp. College Station, TX, US).
Results :
Of 85,476781 commercially insured patients who met inclusion criteria, 1,896,659 (2.22%) carried a diagnosis of myopia, 108,158 (0.13%) of high myopia and 106,823 (0.12%) of RRD. The rate of myopia increased from 32.42 high myopes (653.93 myopes) per 100,000 person-years in 2007 to 48.09 high myopes (1307.19 myopes) per 100,000 person-years in 2016. Importantly, the rate of RRD in high myopes increased from 44.09 in high myopes per 100,000 person-years in 2007 to 62.40 in 2016. High myopia increased the risk of RRD more than 51-fold (HR 51.525, 95% CI 48.709 to 54.503). Using a model with 5-years minimum follow-up, the 5-year risk of RRD was shown to increase with age among high myopes.
Conclusions :
High myopes in the United States are at far higher risk of RRD. Additionally, the risk of RRD increases with age in high myopes. Notably, the magnitude of this increased risk varies substantially according to the minimum follow-up period in our models and should be accounted for in big data analyses.
This is a 2020 ARVO Annual Meeting abstract.