Red cell distribution width in anemic patients undergoing transcatheter aortic valve implantation

Figure 1 Regression analysis for risk factors associated with 30-d mortality. A: Results from the classification tree with significant node-splits and distribution of patients. Life-threatening bleeding (P < 0.001) and acute kidney injury (P = 0.002) were found to be statistically relevant risk factors for 30-d mortality; B: Logistic regression with all covariables which were supposed to be associated with 30-d mortality. Forest plot with odds ratios and 95% confidence intervals (logarithmic scale); C: Logistic regression with the best selected covariables using AIC. Life-threatening bleeding (P < 0.001), acute kidney injury post procedure (P = 0.018) and RDW (P = 0.044) were found to be statistically relevant risk factors for 30-dmortality. AIC: Akaike information criterion; AKI stage post: Acute kidney injury stage I-III post; CAD: Coronary artery disease; Cavk: Cerebroarterial vascular disease; COPD: Chronic obstructive pulmonary disease; DM: Diabetes mellitus; ES log: Logistic EuroSCORE; LT bleeding: Life-threatening bleeding; NYHA: New York Heart Association; pAVK: Peripheral vascular disease; RDW: Red cell distribution width.

Figure 2 Regression analysis for risk factors associated with anemia. A: Classification tree with significant node-splits and distribution of patients with anemia. A creatinine > 1.1 mg/dL (P < 0.001) and age > 83 (P = 0.027) were found to be statistically relevant risk factors for anemia; B: Logistic regression with all covariables which were supposed to be associated with anemia. Forest plot with odds ratios and 95%CI (logarithmic scale); C: Logistic regression with the best selected covariables using AIC. A creatinine > 1.1 mg/dL (P < 0.001) and age > 83 (P = 0.001) were found to be statistically relevant risk factors for anemia. a: Years; AIC: Akaike information criterion; CAD: Coronary artery disease; CKD: Chronic kidney disease; Crea pre: Creatinine (mg/dL) preoperative; DM: Diabetes mellitus; LVEF: Left ventricular ejection fraction (%).

Figure 3 Histogram of the distribution of hemoglobin and red cell distribution width levels. A: Histogram of the distribution of hemoglobin levels. Vertical lines at 12 g/dL and 13 g/dL for population based cut-off points for women and men according to WHO definition of anemia; B: Histogram of RDW levels. Hb: Hemoglobin; RDW: Red cell distribution width.

Figure 4 Anemia and one-year mortality. A: One-year survival curves of patients with and without anemia (P = 0.0594); B: One-year survival curves of patients grouped according to their hemoglobin level above or below mean Hb level of 11.9 g/dL; C: One-year survival curves of patients grouped according to their hemoglobin level above or below cut-off point of 9.7 g/dL.

Figure 5 Red cell distribution width and mortality. A: Classification tree for 30-d mortality with significant node split at RDW 14% (P = 0.021); B: Thirty-day (P < 0.01) and one-year mortality (P < 0.001) of patients grouped according to the presence of RDW ≤ 14% or > 14%. RDW: Red cell distribution width.

Figure 6 Correlation of hemoglobin with red cell distribution width. RDW and hemoglobin showed a significantly negative correlation (-0.36; 95%CI: -0.45, -0.27; P < 0.001). Hb: Hemoglobin; RDW: Red cell distribution width.

Figure 7 Distribution of red cell distribution width and survival of anemic patients according to their red cell distribution width. A: Distribution of RDW in patients without anemia; B: Distribution of RDW in patients with anemia; C: Survival curves of patients with anemia grouped to their RDW above or below cut off point of 14% (P = 0.013). RDW: Red cell distribution width.