Prognostic risk stratification using C-reactive protein, albumin, and associated inflammatory biomarkers in patients with advanced cancer in palliative care

To evaluate the prognostic value of C-reactive protein (CRP), albumin, CRP/albumin ratio (CAR), and modified Glasgow Prognostic Score (mGPS) at different thresholds in patients with advanced cancer in palliative care. Prospective cohort study with patients evaluated at a palliative care unit in Brazil between July 2016 and March 2020. We included patients ≥ 20 years old, both sexes, able to provide the necessary information or accompanied by someone able to do so, and Karnofsky Performance Status ≥ 30 %. The exclusion criteria were the absence of laboratory data and previous diagnosis of autoimmune and infectious diseases. The thresholds analyzed were: CRP < 5 vs. 5-10 vs. > 10 mg/L, albumin < 2.4 vs. 2.4-2.9 vs. 3.0-3.5 vs. > 3.5 g/dL; CAR 2.0, and mGPS equal to 0 vs. 1 vs. 2. Kaplan-Meier curves and Cox regression models (with hazard ratios [HR] and 95% confidence interval [CI]) were used to evaluate prognostic value, and the concordance statistic (C-statistic) was used to evaluate the predictive accuracy of these thresholds to predict death within 90 days. A total of 1,877 patients were included. Median overall survival was 51 (19;124) days and decreased in line with the deterioration of the inflammatory biomarkers. According to the Cox regression models, HR increased as the thresholds worsened (CRP: 1.74 [95% CI, 1.50-2.02] to 2.30 [95% CI, 2.00-2.64]; albumin: 1.77 [95% CI, 1.52-2.07] to 2.60 [95% CI, 2.15-3.14]; CAR: 1.47 [95% CI, 1.21-1.77] to 2.35 [95% CI, 2.05-2.69]; mGPS: 1.78 [95% CI, 1.40-2.23] to 1.89 [95% CI, 1.65-2.15]). All the inflammatory biomarkers evaluated showed discriminatory accuracy for predicting death (C-statistic >0.70), with CAR as the best parameter (C-statistic: 0.80). Our results suggest that CRP, albumin, CAR, and mGPS can be used as clinically meaningful biomarkers to stratify patients with advanced cancer in palliative care according to the severity of these indicators.