Table 3. CRP and prognosis.
Univariable | Multivariable* | |||||
Survival from Cancer Diagnosis Date | Hazard Ratio | 95% CI | p-value | Hazard Ratio | 95% CI | p-value |
CRP (time dependent)¥ | 1.005 | (1.005, 1.006) | < .0001 | 1.003 | (1.003, 1.004) | < .0001 |
Maximum CRP (continuous) | 1.004 | (1.004, 1.005) | < .0001 | 1.001 | (1.000, 1.002) | 0.005 |
Maximum CRP (categorized: high v normal) | 2.76 | (2.44, 3.11) | < .0001 | 1.57 | (1.34, 1.84) | < .0001 |
TWBC (time dependent)§ | 1.06 | (1.05, 1.07) | < .0001 | 1.02 | (1.004, 1.04) | 0.01 |
Maximum TWBC (continuous) | 1.06 | (1.05, 1.07) | < .0001 | 1.01 | (0.998, 1.02) | 0.10 |
Modified Glasgow Prognostic Score (mGPS)** | ||||||
Maximum CRP ≤ 10 mg/L | ref | ref | ||||
Maximum CRP > 10 mg/L & Albumin ≥ 35 g/L | 1.68 | (1.44, 1.95) | < .0001 | 1.43 | (1.22, 1.68) | < .0001 |
Maximum CRP > 10 mg/L & Albumin < 35 g/L | 4.71 | (4.14, 5.36) | < .0001 | 2.42 | (2.06, 2.85) | < .0001 |
¥ 10mg/L u = increase § K/μL increase |
||||||
Univariable | Multivariable* | |||||
Survival from Maximum CRP Test Date | Hazard Ratio | 95% CI | p-value | Hazard Ratio | 95% CI | p-value |
CRP (time dependent)¥ | 1.003 | (1.002, 1.004) | < .0001 | 1.001 | (1.000, 1.003) | 0.03 |
Maximum CRP (continuous) | 1.004 | (1.004, 1.005) | < .0001 | 1.001 | (1.000, 1.002) | 0.005 |
Maximum CRP (categorized: high v normal) | 2.65 | (2.35, 3.00) | < .0001 | 1.50 | (1.23, 1.76) | < .0001 |
TWBC (time dependent)§ | 1.02 | (0.99, 1.05) | 0.08 | 1.00 | (0.97, 1.03) | 0.98 |
Maximum TWBC (continuous) | 1.05 | (1.04, 1.06) | < .0001 | 1.01 | (0.999, 1.02) | 0.07 |
modified Glasgow Prognostic Score (mGPS)** | ||||||
CRP ≤ 10 mg/L | ref | < .0001 | ref | < .0001 | ||
CRP > 10 mg/L & Albumin ≥ 35 g/L | 1.69 | (1.46, 1.97) | 1.39 | (1.18, 1.63) | ||
CRP > 10 mg/L & Albumin < 35 g/L | 4.18 | (3.68, 4.76) | 2.31 | (1.96, 2.71) |
¥ 10mg/L u = increase
§ K/μL increase
* Each multivariable model adjusted for the following: patient age at diagnosis, patient gender, WBC or CRP, hemoglobin, albumin, BMI, primary cancer site (respiratory, genital, digestive, urinary, breast, brain), liver metastases, number of metastatic sites, comorbidities (liver, cardiac, inflammatory bowel disease, rheumatic arthritis, venous thromboembolism)
** CRP and albumin removed from multivariable model to avoid multicollinearity