Skip to main content
. 2010 Jan 8;9(4):695–704. doi: 10.1074/mcp.M900234-MCP200

Table II. Univariate and multivariate Cox regression analyses of overall survival since the start of gemcitabine therapy (n = 304).

Factors except sex are regarded as continuous variables. A forward stepwise selection based on Akaike's information criterion was used to select parameters for multivariate analysis. p values of <0.050 are shown in bold. AST, aspartate aminotransferase; ALT, alanine aminotransferase; ALP, alkaline phosphatase.

Univariate analysis
Multivariate analysis
Hazard ratioa (95% CI) p Hazard ratioa (95% CI) p
Age (years) 0.99 (0.98–1.01) 0.380
Female sex (vs. male) 1.07 (0.83–1.38) 0.610
ECOG performance status 1.49 (1.22–1.80) <0.001 1.36 (1.11–1.67) 0.003
Body surface area (m2) 0.70 (0.33–1.50) 0.360
Leukocytes 1.08 (1.05–1.11) <0.0001 1.04 (1.00–1.08) 0.066
Platelets 1.07 (0.90–1.28) 0.450
Hemoglobin (g/dl) 0.93 (0.85–1.01) 0.098
Albumin (g/dl) 0.61 (0.45–0.82) 0.001
Creatinine (mg/dl) 1.13 (0.60–2.14) 0.700
AST (IU/liter) 1.01 (1.00–1.01) <0.001
ALT (IU/liter) 1.00 (1.00–1.01) 0.033
ALP 1.09 (1.06–1.11) <0.0001 1.07 (1.05–1.10) <0.0001
α1-Antitrypsinb 5.92 (3.09–11.37) <0.0001 3.66 (1.89–7.11) 0.0001
α1-Antichymotrypsinb 11.60 (2.69–50.01) 0.001
Clinical stage IVac (vs. IVb) 1.10 (0.85–1.38) 0.453

a Hazard ratios are per 1,000/mm3 increase for leukocytes, per 10 × 104/mm3 increase for platelets, and per 100 units/liter increase for ALP. Hazard ratios for other continuous variables are per 1 unit increase for each variable.

b Logarithmic variable determined by reverse-phase protein microarray.

c According to Ref. 23.