To evaluate the effect of high dose chemotherapy (HDC) followed by autologous hematopoietic stem cell transplantation (ASCT) in comparison to conventional pharmacological therapy in the treatment of patients with refractory, progressively erosive rheumatoid arthritis (RA).
Decision analysis using a Markov model with a 5.5 year time horizon. Probabilities of transition towards 5 different health states, ranging from 70% improvement to death, were derived from published case reports, patient series, and expert panels. Quality of life (QOL) estimates were obtained from 2 RA clinical trials. Patients were hypothetical cohorts of 50-year-old female patients with progressively erosive, active RA, who failed treatment with methotrexate, combination therapy, and tumor necrosis factor blocking agents. Interventions were HDC + ASCT versus conventional pharmacological treatment with a (combination) therapy of disease modifying antirheumatic drugs. As main outcome measures, we included the number of quality adjusted life years (QALY) after HDC + ASCT compared to conventional therapy. Sensitivity analysis was performed to investigate the influence of treatment related mortality (TRM) and the influence of QOL during HDC + ASCT, and to assess the minimal desired effectiveness of HDC + ASCT for a given TRM of 1% and 10%.
HDC + ASCT and conventional pharmacological treatment were equally effective in the base-case analysis (3.48 vs 3.46 QALY). A TRM of less than 3.3% favored HDC + ASCT as the preferred treatment. The analysis showed that when TRM was set at 1%, a relatively short period of efficacy was sufficient to remain the preferred strategy, whereas a TRM of 10% would require a sustained response for several years.
This model predicted equally favorable effects of HDC + ASCT and conventional therapy in the treatment of refractory RA in the base-case. The minor differences in terms of QALY seem to indicate that clinical decision making should be guided by patient preferences. However, better clinical efficacy might be achieved by adaptation of the treatment regimen of HDC + ASCT and patient selection. The model supports the need for randomized clinical trials and may contribute to an optimal study design.