Univariate Cox proportional hazards analysis of leukemia-free survival for all participating patients (n = 320) and patients in CR1 (n = 261)
Variable . | Hazard ratio . | 95% CI . | P* . |
|---|---|---|---|
| All patients | |||
| Age, y (≤ 60 vs > 60)† | 0.66 | (0.50-0.87) | < .01 |
| Sex (men vs women) | 1.32 | (1.00-1.74) | .05 |
| FAB class: M2/M3/M4 (no vs yes)‡ | 1.38 | (1.05-1.81) | .02 |
| Karyotype (MRC): adverse (no vs yes)‡ | 0.48 | (0.28-0.84) | .01 |
| Percentage blast cells (≤ 15% vs > 15%) | 0.57 | (0.33-0.99) | < .05 |
| High-dose cytarabine therapy (no vs yes) | 1.41 | (1.04-1.92) | .03 |
| Months from current CR to random assignment (≤ 6 vs > 6) | 1.51 | (1.08-2.11) | .02 |
| Patients in CR1 | |||
| Age, y (≤ 60 vs > 60)† | 0.61 | (0.45-0.83) | < .01 |
| FAB class: M2/M3/M4 (no vs yes)‡ | 1.49 | (1.10-2.02) | .01 |
| Karyotype (MRC): adverse (no vs yes)‡ | 0.48 | (0.27-0.84) | .01 |
| History of antecedent hematologic disorder (no vs yes) | 0.65 | (0.44-0.97) | .03 |
| High-dose cytarabine therapy (no vs yes) | 1.550 | (1.10-2.18) | .01 |
| Months from current CR to random assignment (≤ 6 vs > 6) | 1.57 | (1.09-2.27) | .02 |
Variable . | Hazard ratio . | 95% CI . | P* . |
|---|---|---|---|
| All patients | |||
| Age, y (≤ 60 vs > 60)† | 0.66 | (0.50-0.87) | < .01 |
| Sex (men vs women) | 1.32 | (1.00-1.74) | .05 |
| FAB class: M2/M3/M4 (no vs yes)‡ | 1.38 | (1.05-1.81) | .02 |
| Karyotype (MRC): adverse (no vs yes)‡ | 0.48 | (0.28-0.84) | .01 |
| Percentage blast cells (≤ 15% vs > 15%) | 0.57 | (0.33-0.99) | < .05 |
| High-dose cytarabine therapy (no vs yes) | 1.41 | (1.04-1.92) | .03 |
| Months from current CR to random assignment (≤ 6 vs > 6) | 1.51 | (1.08-2.11) | .02 |
| Patients in CR1 | |||
| Age, y (≤ 60 vs > 60)† | 0.61 | (0.45-0.83) | < .01 |
| FAB class: M2/M3/M4 (no vs yes)‡ | 1.49 | (1.10-2.02) | .01 |
| Karyotype (MRC): adverse (no vs yes)‡ | 0.48 | (0.27-0.84) | .01 |
| History of antecedent hematologic disorder (no vs yes) | 0.65 | (0.44-0.97) | .03 |
| High-dose cytarabine therapy (no vs yes) | 1.550 | (1.10-2.18) | .01 |
| Months from current CR to random assignment (≤ 6 vs > 6) | 1.57 | (1.09-2.27) | .02 |
P values were calculated using the Wald test and stratified by country and, if applicable, CR stratum. Variables with a P value for the hazard ratio ≤ 0.1 were retained from the backward selection model.
Of the dichotomized groups, the second group is always the baseline.
If any of the subgroups of FAB classes or leukemic karyotypes was retained by the selection procedure, then all groups were retained in the final model.