Table 2.
Univariate analysis | Multivariate analysis | ||||||
Number of patients | Hazard ratio per 10 IDS units for p21 | 95% confidence interval | P | Hazard ratio per 10 IDS units for p21 | 95% confidence interval | P | |
Overall survival | |||||||
p21 | |||||||
p21 nucleus | 64 | 0.96 | 0.77, 1.19 | 0.70 | 0.96 | 0.75, 1.22 | 0.73 |
p21 cytoplasm | 63 | 1.12* | 1.01, 1.25 | 0.03* | 1.15* | 1.03, 1.28 | 0.01* |
HER-2 (2+/3+ versus 0/1+) | 63 | 2.26* | 1.00, 5.27 | 0.05* | 3.14* | 1.15, 8.64 | 0.03* |
Relapse-free survival | |||||||
p21 | |||||||
p21 nucleus | 64 | 1.01 | 0.84, 1.22 | 0.89 | 1.03 | 0.84, 1.25 | 0.80 |
p21 cytoplasm | 63 | 1.12* | 1.01, 1.24 | 0.03* | 1.15* | 1.02, 1.28 | 0.02* |
HER-2 (2+/3+ versus 0/1+) | 63 | 2.60* | 1.19, 5.71 | 0.02* | 4.08* | 1.60, 10.39 | 0.003* |
Hazard ratios, confidence intervals and P values are given for the results of both the univariate analyses and the multivariate analyses. For univariate and multivariate analyses, the hazard ratio is given for nuclear and cytoplasmic p21 WAF1/CIP1 per 10 unit increase in intensity distribution score (IDS), along with the 95% confidence interval. The hazard ratio and confidence interval for HER-2 relates to 2+/3+ versus 0/1+. Patients with ductal, lobular or 'mixed' cancer were included where the marker data was available. The multivariate analysis is adjusted for the Nottingham Prognostic Index (NPI) (nodes, grade and size) and for treatment (tamoxifen/chemotherapy/no treatment), with the NPI based on imputed grades and tumour sizes where the relevant data were missing as described in the statistical methods. *Relationship reached significance (P < 0.05).