Cost-Effectiveness of Filgrastim and Pegfilgrastim as Primary Prophylaxis against Febrile Neutropenia in Lymphoma Patients Receiving R-CHOP Chemotherapy.

Non-Hodgkin Lymphoma (NHL) is the fifth most common type of malignancy in Canada. The most common subtype of NHL is diffuse large B-cell lymphoma (DLBCL). Initial standard treatment for DLBCL includes combination immuno-chemotherapy with rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP). This regimen is typically administered every 21 days for a total of 6 cycles. Febrile neutropenia (FN) is a serious toxicity of lymphoma chemotherapy and patients with this condition must be treated aggressively as it could lead to complications such as prolonged hospitalization and death. Granulocyte-colony stimulating factors such as filgrastim and pegfilgrastim are efficacious in preventing FN, yet their cost-effectiveness has not been evaluated in the publicly funded Canadian healthcare system.

A Markov model was constructed to evaluate the cost-effectiveness (cost-utility) of filgrastim and pegfilgrastim as primary prophylaxis versus no primary prophylaxis against FN in DLBCL patients receiving induction chemotherapy. Health states included in the model were hospitalization for FN, and receiving chemotherapy with no FN. It was assumed that patients in the no primary prophylaxis arm of the model who experienced a FN episode would receive secondary prophylaxis with filgrastim for all subsequent chemotherapy cycles. The time horizon of the model was 18 weeks, the time period over which the six cycles of chemotherapy are administered. The analysis was conducted from the hospital perspective. Costs are reported in 2009 Canadian dollars and outcomes in quality-adjusted life years (QALY). One-way sensitivity analyses were done on model parameters. A probabilistic sensitivity analysis was done to evaluate overall uncertainty in the model.

In the base case analysis costs associated with the no primary prophylaxis, filgrastim and pegfilgrastim interventions were $6044, $9450 and $15899, respectively. Quality-adjusted life years associated with the three interventions were 0.198, 0.200, and 0.202 respectively (over the 18-week model time horizon). The incremental cost-effectiveness ratio (ICER) of filgrastim compared to no primary prophylaxis was estimated to be $1.7 million (M)/QALY [95% confidence interval: -14M/QALY (dominated) to $15M/QALY]; for pegfilgrastim compared to filgrastim the ICER was estimated to be $4.4M/QALY [95% confidence interval: -25M/QALY (dominated) to $22.7M/QALY]. All one-way sensitivity analyses yielded ICERs of greater than $1M/QALY.

The ICERs for filgrastim and pegfilgrastim when used as primary prophylaxis against FN in DLBCL patients receiving induction chemotherapy are well above the usually accepted cost-effectiveness threshold of $50,000/QALY.

Mittmann:Amgen Canada: Unrestriced educational grant.