BACKGROUND: Age-related macular degeneration (AMD) is the primary cause of vision loss in the elderly and results in significant economic and humanistic burden. The selective vascular endothelial growth factor inhibitor, pegaptanib (Macugen) is indicated for patients with neovascular AMD. Guidance is needed regarding the cost effectiveness of treatment, any variation between sub-populations of differing clinical characteristics and the optimum duration of treatment.
OBJECTIVE: To estimate the cost effectiveness of pegaptanib versus best supportive care (BSC) for AMD from the perspective of the UK government, and to evaluate the impact of patient characteristics and differing treatment discontinuation scenarios.
METHODS: A cohort of 1000 patients aged >45 years with a best-corrected visual acuity (VA) in their better-seeing eye of < or =6/12 was modelled. Patients were either treated with pegaptanib (0.3mg every 6 weeks for a maximum of 2 years in their better-seeing eye only) or received BSC (no active treatment). Supportive services were provided for patients with a VA < or =6/60. A 10-year Markov model composed of 12 VA states (defined by individual Snellen lines) and a dead state was constructed. The model reported herein was used to support submissions to the National Institute for Health and Clinical Excellence (NICE) and the Scottish Medicines Consortium (SMC). NICE guidance is expected to be available in October 2007 and the SMC advice was issued on 7 July 2006. SMC accepted pegaptanib for use in patients with visual acuity between 6/12 and 6/60 (inclusive) and should be stopped if visual acuity falls below 6/60 during treatment or where severe visual loss is experienced. Time-dependent transition probabilities for the loss and gain of Snellen lines were derived from parametric survival curves fitted to patient-level data from the VISION trials. Survival curves were fitted with treatment and baseline Snellen scores as covariates; additional curves were fitted with the addition of age, gender, lesion type or lesion size as covariates. Mortality rates were adjusted for the age, gender and VA of the population. Cost effectiveness was expressed as the incremental cost (IC) per vision-year saved and IC/QALY. Uncertainty was explored by probabilistic and univariate sensitivity analysis. Costs (year 2005 values) and outcomes were discounted at 3.5% per anum.
RESULTS: In the base-case analysis, treatment was targeted to patients with a VA of 6/12 to 6/95 and discontinued after 2 years, or earlier if VA fell below 6/95 or by > or =6 lines. The IC/QALY was estimated as 8023 pounds(upper 95% CI 20,641 pounds). Cost effectiveness varied by age (age <75 years = 2033 pounds/QALY; age > or =75 years = 11,657 pounds/QALY) and by pre-treatment VA (6/12-6/95 = 8023 pounds/QALY; 6/12-6/60 = 6664 pounds/QALY; 6/12-6/24 = 1920 pounds/QALY). Gender and lesion type or size had little effect. Cost effectiveness was not sensitive to precise rules for treatment discontinuation, but was maximised if treatment was discontinued in patients no longer likely to benefit.
CONCLUSIONS: The results suggest that pegaptanib treatment is likely to be cost effective across all groups studied, and marginally more cost effective in younger patients and those with better pre-treatment VA. Cost effectiveness appears to be optimised if treatment is discontinued after 1 year if individual patients' VA has dropped by > or =6 lines from pre-treatment levels, or at any time if it drops below 6/95. However, strict application of discontinuation rules does not appear to be necessary for pegaptanib to be cost effective. Clinical judgement and patient preference should be an important determinant in decisions about stopping treatment
