Table 9

Multivariable Cox regression analysis of relapse-free survival in patients with primary breast cancer showing the impact of adding the marker (PMN-E) to a base model of recognized prognostic variables [59].

Factor

HR (95% CI)

P


Base model

Age and menopausal status combined

0.005

Age premenopausala

0.68 (0.55 to 0.85)

Age postmenopausala

0.96 (0.84 to 1.09)

Post- versus premenopausal

1.83 (1.27 to 2.46)

Tumor size

<0.001

2 cm to 5 cm versus ≤2 cm

1.69 (1.36 to 2.10)

>5 cm versus ≤2 cm

2.31 (1.73 to 3.10)

Nodal status

<0.001

N1-3 versus N0

1.66 (1.30 to 2.11)

N>3 versus N0

2.75 (2.18 to 3.47)

ER (positive versus negative)b

0.87 (0.68 to 1.11)

0.25

PgR (positive versus negative)b

0.76 (0.61 to 0.95)

0.02


Additions to base model

+PMN-E (high versus low)c

1.45 (1.10 to 1.89)

0.01

+PMN-E (continuous)d

1.06 (0.98 to 1.14)

0.13


aAge in decades for pre- and postmenopausal patients; bPositive, ≥10 fmol/mg protein; negative <10 ng/mg protein; cHigh, >36.4 ng/mg protein; low, ≤36.4 ng/mg protein; dLog-transformed variable. CI: confidence interval; ER: estrogen receptor; PgR: progesterone receptor.

Altman et al. BMC Medicine 2012 10:51   doi:10.1186/1741-7015-10-51

Open Data