Safety, effectiveness, and cost among Texas Medicaid patients with Diabetic Macular Edema (DME) or Age-Related Macular Degeneration (AMD)

Although bevacizumab is one of the most commonly used treatments for DME and AMD, there are concerns regarding safety and effectiveness due to its off-label use. The study objectives were to determine if: 1) the risk of cardiovascular/ hemorrhagic events (safety) and visual impairment (effectiveness) differed by bevacizumab use (i.e., use vs. non-use and number of treatments) among DME and AMD patients; and 2) direct medical costs differed between DME and DME control patients. A retrospective cohort analysis was conducted with Texas Medicaid medical and prescription data (9/1/07-12/31/12) for patients: 18- 63 years, continuously enrolled 1-year pre- and post-index, and diagnosed with DME or AMD. The index date was the first date of diagnosis. The dependent variables were: 1) cardiovascular/hemorrhagic risk; 2) visual impairment; 3) direct medical costs. The independent variables were bevacizumab use and number of bevacizumab treatments. Covariates were disease state, Charlson Comorbidity Index (CCI) score, total medication use, number of laser treatments, and demographics. Propensity scoring technique was used to match: 1) bevacizumab users and non-users; and 2) DME and DME control cohorts. Descriptive analyses, logistic regression, Cox-regression, and generalized linear models were employed. A final cohort of 3,647 DME, 297 AMD, and 57,897 DME control patients were included. The majority (DME and AMD) was between 45-63 years of age (86.6%), Hispanic (54.0%), and female (65.1%). The mean total number of unique medications and mean CCI were 2.7 ± 3.4 and 6.0 ± 3.3, respectively. Total direct medical costs/person (Mean (±SD)) incurred by DME, DME control, and AMD subjects in the post-index period were $6,704(±9,338), $5,495(±10,153), and $4,935(±12,702), respectively. No differences in cardiovascular/ hemorrhagic risk were found between bevacizumab users and non-users. The claims data lacks the detail to determine the effectiveness of bevacizumab. DME control patients had lower overall direct medical costs than DME patients (p<0.0001). In conclusion, although bevacizumab is a less expensive off-label alternative of ranibizumab, the choice between bevacizumab and ranibizumab should be made through careful consideration. However, as the use of anti-VEGF agent increases, further research should be conducted to determine if any changes in cardiovascular adverse events occur.