Abstract
Abstract 1368
Poster Board I-390
Results from the REACH randomised trial showed that patients with relapsed or refractory CLL who received rituximab in combination with fludarabine-cyclophosphamide (R-FC) remained in progression free ten months (median) longer than patients receiving FC alone (30.6 vs. 20.6 months). Comparison of the amount of time a patient can expect before relapse without debilitating toxicities is a valuable aid in choosing between treatments.
To evaluate whether relapsed or refractory CLL patients treated with the combination of R-FC would experience longer time in a better health state compared with patients on FC alone.
The Q-TWiST was based on 25.3 months median follow-up data from the REACH trial (Robak et al., 2008) where 276 patients were treated with R-FC and 276 patients with FC alone. Because of shorter follow up time in the FC arm, the data was truncated at 52.53 months; the longest follow up in the shortest PFS curve of the comparator to exclude follow-up time bias in favor of R-FC. Patients with event times greater than 52.53 months were censored at the truncation point.
The area under survival curves was partitioned into health states: (1) Relapse (REL) - area between overall and progression free survival curves; (2) Toxicity (TOX) - time period with all treatment-related adverse events from start of study treatment to up to 28 days following last dose and/or progression; and (3) Time without disease symptoms or treatment toxicity (TWiST), defined as the area between the PFS and TOX curves. Utility weights were applied to each health state to reflect the quality of life value relative to time in TWiST. Each utility weight ranged from 0 to 1, where 0 represents a state as bad as death, 1 represents a state as good as TWiST. For example, a utility weight of 0·5 for REL would indicate that 2 month of time after relapse is valued the same as 1 month in the TWiST state. Mean differences and 95% confidence intervals (CI) were calculated for each health state.
R-FC patients gained a mean of 6.38 months TWiST (95% CI, 3.92-8.93, p<0.0001), spent a mean of 4.82 months less time in relapse (95% CI, 1.40-8.43 p<0.0009) compared with patients treated with FC, without a significant increase in the burden of toxicity (mean difference 0.06 months (95% CI, 0.39-0.50, p=0.404). With utility coefficients of 1.0 for all health states, the unadjusted mean difference in survival between R-FC and FC was 1.50 months (95% CI, 0.84-3.69, p=0.401).
Using the utility of 0.618 for REL derived by Hancock et al. (2002), an assumed utility of 0.618 for TOX, and a utility of 1.0 for TWiST, R-FC patients experienced a mean of 3.45 months longer quality-adjusted survival compared with FC (95% CI, 1.69-5.14, p<0.0001). The key driver for these results is the substantial (40.9%) reduction in the time spent in the relapsed health state due to the addition of rituximab to FC.
The threshold analyses showed that, all utility combinations for TOX and REL (0.1-0.9) with a TWiST utility of 1 resulted in a statistically significant (p<0.002) gain in Q-TWiST for R-FC patients. A second sensitivity analysis was conducted using 0.80 for TWiST and varying the REL and TOX utilities from 0.1 to 0.9. Each of these utility combinations resulted in a significant Q-TWiST outcome for R-FC relative to FC (p ≤0.039). For utilities TOX and REL combinations greater than 0.6 there was a trend towards significance. However, such combinations are clinical extremes and unlikely to present in clinical practice.
The result of the Q-TWiST analysis on the REACH study data showed that, relapsed or refractory CLL patients receiving R-FC, spend less time with disease symptoms with no significant increase in treatment toxicity compared with patients who received FC alone.
Robak: F Hoffmann-La Roche: Honoraria. Oertel: Hoffmann La Roche Ltd.: Employment, Equity Ownership. Carr: F.Hoffmann-La Roche Ltd.: Employment.
Author notes
Asterisk with author names denotes non-ASH members.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal