Stepwise forward multivariable Cox regression analysis model for the time to first treatment in 104 CLL patients
| Variable . | Hazard ratio . | CI (95%) . | P . |
|---|---|---|---|
| Step 1 | |||
| Disease stage at diagnosis | 3.26 | 1.8–5.8 | <.001 |
| Step 2 | |||
| Disease stage at diagnosis | 2.95 | 1.7–5.3 | <.001 |
| ZAP-70 status | 3.09 | 1.7–5.8 | <.001 |
| Step 3 | |||
| Disease stage at diagnosis | 2.46 | 1.3–4.5 | .004 |
| ZAP-70 status | 3.03 | 1.6–5.7 | .001 |
| BCL2 status, −938AA vs AC/CC | 1.90 | 1.1–3.5 | .034 |
| Variable . | Hazard ratio . | CI (95%) . | P . |
|---|---|---|---|
| Step 1 | |||
| Disease stage at diagnosis | 3.26 | 1.8–5.8 | <.001 |
| Step 2 | |||
| Disease stage at diagnosis | 2.95 | 1.7–5.3 | <.001 |
| ZAP-70 status | 3.09 | 1.7–5.8 | <.001 |
| Step 3 | |||
| Disease stage at diagnosis | 2.46 | 1.3–4.5 | .004 |
| ZAP-70 status | 3.03 | 1.6–5.7 | .001 |
| BCL2 status, −938AA vs AC/CC | 1.90 | 1.1–3.5 | .034 |
The cutoff levels used in the analysis were as follows: age, 60 years; disease stage at diagnosis, Binet stage A versus B and C; ZAP-70 or CD38 status, 20% positive cells. Because of missing data in a substantial proportion of patients, the prognostic marker thymidine kinase and genomic aberrations (Table 1) were not included in the analysis.
CI indicates confidence interval.