Multivariate analyses of standard prognostic variables associated with clinical outcomes
| Variable . | Overall survival . | Event-free survival . | Complete response . | Duration of complete response . | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| n/N (%) . | HR (95% CI) . | P* . | n/N (%) . | HR (95% CI) . | P* . | n/N (%) . | HR (95% CI) . | P* . | n/N (%) . | HR (95% CI) . | P* . | |
| Randomization to thalidomide | 309/634 (49%) | 1.03 (0.75, 1.41) | 0.872 | 312/642 (49%) | 0.67 (0.55, 0.83) | <.001 | 303/618 (49%) | 1.59 (1.28, 1.97) | <.001 | — | — | — |
| Cytogenetic abnormalities | 188/634 (30%) | 2.51 (1.80, 3.49) | <.001 | 194/642 (30%) | 1.64 (1.32, 2.03) | <.001 | — | — | — | 82/340 (24%) | 1.90 (1.35, 2.69) | <.001 |
| Thalidomide and presence of CA | 89/634 (14%) | 0.58 (0.35, 0.95) | 0.032 | — | — | — | — | — | — | — | — | — |
| Albumin < 3.5 g/dL | — | — | — | 117/642 (18%) | 1.35 (1.04, 1.75) | 0.023 | — | — | — | — | — | — |
| β2-Microglobulin > 5.5 mg/L | 115/634 (18%) | 1.61 (1.21, 2.16) | 0.001 | 116/642 (18%) | 1.47 (1.13, 1.90) | 0.004 | — | — | — | 57/340 (17%) | 1.89 (1.30, 2.75) | <.001 |
| Lambda light chain | — | — | — | — | — | — | 233/618 (38%) | 1.27 (1.02, 1.58) | 0.035 | — | — | — |
| IgA isotype | — | — | — | 154/642 (24%) | 1.34 (1.06, 1.70) | 0.014 | — | — | — | — | — | — |
| Lactic dehydrogenase ≥ 190 U/L | 197/634 (31%) | 1.52 (1.16, 1.98) | 0.002 | 198/642 (31%) | 1.42 (1.13, 1.77) | 0.002 | — | — | — | — | — | — |
| C-reactive protein ≥ 8.0 mg/L | — | — | — | — | — | — | 249/618 (40%) | 1.35 (1.08, 1.68) | 0.007 | — | — | — |
| Variable . | Overall survival . | Event-free survival . | Complete response . | Duration of complete response . | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| n/N (%) . | HR (95% CI) . | P* . | n/N (%) . | HR (95% CI) . | P* . | n/N (%) . | HR (95% CI) . | P* . | n/N (%) . | HR (95% CI) . | P* . | |
| Randomization to thalidomide | 309/634 (49%) | 1.03 (0.75, 1.41) | 0.872 | 312/642 (49%) | 0.67 (0.55, 0.83) | <.001 | 303/618 (49%) | 1.59 (1.28, 1.97) | <.001 | — | — | — |
| Cytogenetic abnormalities | 188/634 (30%) | 2.51 (1.80, 3.49) | <.001 | 194/642 (30%) | 1.64 (1.32, 2.03) | <.001 | — | — | — | 82/340 (24%) | 1.90 (1.35, 2.69) | <.001 |
| Thalidomide and presence of CA | 89/634 (14%) | 0.58 (0.35, 0.95) | 0.032 | — | — | — | — | — | — | — | — | — |
| Albumin < 3.5 g/dL | — | — | — | 117/642 (18%) | 1.35 (1.04, 1.75) | 0.023 | — | — | — | — | — | — |
| β2-Microglobulin > 5.5 mg/L | 115/634 (18%) | 1.61 (1.21, 2.16) | 0.001 | 116/642 (18%) | 1.47 (1.13, 1.90) | 0.004 | — | — | — | 57/340 (17%) | 1.89 (1.30, 2.75) | <.001 |
| Lambda light chain | — | — | — | — | — | — | 233/618 (38%) | 1.27 (1.02, 1.58) | 0.035 | — | — | — |
| IgA isotype | — | — | — | 154/642 (24%) | 1.34 (1.06, 1.70) | 0.014 | — | — | — | — | — | — |
| Lactic dehydrogenase ≥ 190 U/L | 197/634 (31%) | 1.52 (1.16, 1.98) | 0.002 | 198/642 (31%) | 1.42 (1.13, 1.77) | 0.002 | — | — | — | — | — | — |
| C-reactive protein ≥ 8.0 mg/L | — | — | — | — | — | — | 249/618 (40%) | 1.35 (1.08, 1.68) | 0.007 | — | — | — |
The significant interaction between thalidomide and cytogenetic abnormalities (CA; HR = 0.58; P = .032) means that the effect of thalidomide is different for patients with and without CA. In patients without CA, thalidomide has no effect on survival (HR = 1.03; P = .87). In patients with CA, the effect of thalidomide is a significant reduction in the risk of death of 41% (HR = 0.59; P = .008), which was also found to be true in an analysis done in the subset of patients who had baseline gene array data available (HR = 0.40; P = .001; data not shown).
HR indicates hazard ratio; CI, confidence interval; and —, not applicable.
P from Wald χ 2 test in Cox regression. Multivariate results not statistically significant at the .05 level. All univariate P values reported regardless of significance. Multivariate model uses stepwise selection with entry level 0.1 and variable remains if meets the 0.05 level. A multivariate P value greater than .05 indicates variable forced into model with significant variables chosen using stepwise selection.