The Cost-effectiveness of Eltrombopag for the Treatment of Immune Thrombocytopenia in the United States.

PURPOSE

Eltrombopag was evaluated as a second-line treatment for adult chronic immune thrombocytopenia (ITP) in the 2006 Phase III RAISE (Eltrombopag for Management of Chronic Immune Thrombocytopenia) randomized, placebo-controlled trial. More than 80% of patients reached satisfactory platelet counts within 2 weeks. However, the economic value of eltrombopag as a second-line treatment for ITP remains to be formally assessed. This study aimed to estimate the cost-effectiveness of treating ITP with a comparable thrombopoietin receptor agonist (eltrombopag vs romiplostim).

METHODS

A Markov model was implemented over a lifetime time horizon to estimate the benefits and costs of each treatment. The model featured 3 health states based on current guidelines: (1) on treatment; (2) treatment failure/discontinuation; and (3) mortality. In line with therapeutic goals in ITP, model patients could experience 3 events: no bleeding, mild/moderate bleeding, or severe bleeding. Data on eltrombopag use were obtained from an open-label extension of previous Phase II/III trials, including RAISE. Romiplostim data were obtained from Phase III trials and an extension study. Lifetime overall survival was extrapolated by using treatment-specific mortality rates derived from severe bleeding and natural mortality rates. The costs of drugs, routine care, bleeding episodes, adverse events, and mortality were represented in the model.

FINDINGS

Eltrombopag-treated patients gained 17.58 life years and 14.68 quality-adjusted life years, whereas romiplostim-treated patients gained 17.52 life years and 14.67 quality-adjusted life years. The total lifetime cost of eltrombopag treatment was estimated at $1.58 million versus $2.13 million for romiplostim. Sensitivity analyses supported base case findings. Deterministic sensitivity analysis predicted the greatest sensitivity to the rates of severe bleeding, discontinuation, and natural mortality. Probabilistic sensitivity analysis showed that eltrombopag would be an efficient use of resources at a $50,000 threshold in 52.8% of cases. In all probabilistic iterations, the total cost of eltrombopag treatment was lower than with romiplostim, primarily because of lower drug costs.

IMPLICATIONS

Clinical data were applied in an economic analysis, and eltrombopag exhibited economic dominance compared with romiplostim, driven largely by the reduced costs of primary therapy. This model was limited by a lack of specific patient-level data and robust data on the duration of secondary therapy, as well as by the fact that utilization values are likely conservative estimates for routine care use.