2.1 Model description
The present study consists of a pharmacoecomomic model that aims to analyze the cost-effectiveness of adding rituximab in combination with the fludarabine plus cyclophosphamide protocol (R-FC) in comparison with the fludarabin plus cyclophosphamide (FC) regimen, as a therapeutic option for CLL naïve treatment patients. A three-state Markov Model was created considering the transition states of first line treatment. The model describes the current clinical practice, and reflects the natural course of the disease according to the treatment option (Fig 1). The model development, costs analysis and other assumptions were done according with the ISPOR consolidated Health Economic Evaluation Reporting Standards (CHEERS)(12)
The model used a cycle length of 3 months and 10 years as time horizon. The Brazilian public healt system willingness to pay is considered as perspective. Future costs and utilities had an annual discount rate of 5%. The Markov Model was developed and calculated with the Tree Age Pro Healthcare® software (Treeage, Williamstown, MA, USA).
All patients started in the “On treatment” state. They could either remain on this state or change to the states “Progression” or “Death”. The state “Progression” included all events from first relapse or disease progression after the treatment.
The Cost-effectiveness of R-FC over FC was assessed using an incremental cost-effectiveness ratio (ICER). The transition probabilities were obtained from the CLL8 trial progression-free survival (PFS) and overall survival (OS) curves, up to a follow up period of 3 years. Survival beyond the trial follow up period was projected according to the previous described methodology.
2.2 Target population
Patient related data, survival probabilities and clinical outcomes were obtained from the clinical study CLL8 trial (NCT00281918), a global prospective, open label, randomized phase 3 trial(Hallek et al., 2010). The data reported in this publication was used to estimate the long-terms costs relative to the adding of rituximab in the first-line treatment of CLL.
A total of 817 patients diagnosed with CLL and treatment naïve were enrolled and in the trial. The study population had a mean age of 61 years, and 74% were male. Patients enrolled were in Binet stage C (31%) or with confirmed active disease in stages A or B (68%).
With a median observation time of 3 years after the randomisation, CLL8 trial results demonstrated a significant benefit in R-FC treatment arm related with the progession free-survival (PFS) and overall survival (OS). Table 1 summarizes the clinical data extracted from CLL8.
Table 1
CLL8 clinical data: R-FC versus FC
|
R-FC versus FC
|
Progression free survival (PFS)
|
HR= 0,56 – 95% CI [0,46 – 0,69]
|
Overall survival (OS)
|
HR= 0,67 – 95% CI [0,48 – 0,92]
|
2.3 Treatment
CLL8 study design randomly assigned participants to receive either R-FC or FC treatment regimens as described in the trial protocol. Patients were scheduled to receive 6 cycles of the regimen to which they were randomized. Treatment consisted of six 28-day intravenous administration of 25mg/m2/day of fludarabine and 250 mg/m2/day of cyclophosphamide on the first 3 days of each cycle. For the R-FC protocol, rituximab was administered intravenously in addition to FC as 375 mg/m2 before the FC infusion on day 1 of the first cycle and 500 mg/m2 on day 1 of cycles 2-6.
2.4 Treatment algorithm
Although several therapeutic options are currently available for the treatment of first line, treatment naïve CLL patients, there is no formal treatment consensus on the Brazil public health system. An algorithm for the treatment of CLL was developed on the basis of information available in the following guidelines: the International Workshop on CLL (iWCLL) 2018(7), the European Society for Medical Oncology (ESMO)(13), US National Comprehensive Cancer Network (NCCN), clinical guidelines in Oncology: CLL/Small Lymphocytic Lymphoma(14), Brazilian Group of CLL(15).
Using these sources, a preliminary treatment algorithm was developed, this process was them submitted to the review of hematologists with extensive practice in treating CLL in Brazil. The final algorithm that closely reflects the treatment of CLL considering the Brazilian public health experience and the predicted use of rituximab is shown in Fig 2.
2.5 Costs
The incremental costs were calculated considering only costs driven by the difference in the treatment arms, the association of rituximab to the FC protocol. The direct costs include drug costs from both protocols, adverse event costs, drug administration costs, and costs with follow-up visits, and laboratory tests (clinical examinations, blood exams. Other costs like administrative costs were not included in the model because available study data for these parameters is scarce and they have low representativity.
The assessment of drug prices and costs was obtained from the Brazilian Health Ministry website (SIGTAP – Health procedures and drugs management system), which provides information related to the value paid by the Brazilian Public Health system (SUS) for procedures and medications. Specifically, for CLL there are 2 procedures available for first and second line treatment (nominated APAC’s). They were used as main reference for the comparison of incremental costs and the current value allocated for first line CLL treatment.
2.6 Treatment costs
The model followed the protocol described on CLL8 trial and summarized briefly in the treatment section. The body surface (m2) considered was 1.8 according to the Brazilian populational average of height and weight, according with the Brazilian Institute of Demographics and Statistics (IBGE).
The costs associated with salvage treatment were calculated based on the available guidelines and literature. The treatment costs associated with the R-FC protocol and FC were obtained in Brazilian Reais (BRL) and converted to US Dolars (USD), using a currency exchange rate of 5,39 BRL to 1 USD (currency exchange rates from November 9th 2020).
2.7 Adverse events costs
To define the costs of adverse events associated with the R-FC protocol, we considered the incidence of adverse events published in the CLL8 trial. The incidence of events reported was significant higher on patients treated with R-FC compared with FC. Among the events reported, grade 3 or 4 adverse events were significant higher in patients treated with rituximab. According to the study protocol, granulocyte-stimulating factor (G-CSF), was administered to patients that presented this adverse event. Hospitalizations costs related to adverse events were disregarded as the treatment of neutropenia usually does not require hospital admission.
2.8 Sensitivity analysis
Probabilistic sensitivity analysis were performed to determine the magnitude of change in the ICER associated with a range of variations in four most critical parameters. We carried out tornado diagrams to test the model uncertainty when applying variations on the overall survival hazard ratio, the rituximab cost per milligram, the progression-free survival hazard ratio, and 3-month toxicity costs.