Albumin level and risk of major bleeding in patients with atrial fibrillation on direct oral anticoagulants

Figures

• 

  Figure 1. Study flowchart. The patients were divided into three tertiles based on their serum albumin levels: the first tertile consisted of patients with serum albumin levels lower than 3.7 g/dL (lower tertile, N = 860), the second tertile included those with serum albumin levels between 3.7 and 4.1 g/dL (middle tertile, N = 835), and the third tertile comprised patients with serum albumin levels of 4.1 g/dL or higher (higher tertile, N = 828).

• 

  Figure 2. A lower albumin level is associated with a bleeding risk for AF patients using DOACs. (A) The Kaplan-Meier survival curves illustrate the survival probability over time stratified by albumin tertiles. The tertiles are categorized as Higher tertile (solid line), Middle tertile (dashed line), and Lower tertile (dotted line). (B) This figure illustrates the relationship between albumin levels (Alb, g/dL) and the relative hazard ratios. The impact of albumin levels was analyzed using a Cox proportional hazards model with a linear term for albumin. The covariates for Cox proportional hazards model were age, female sex, body mass index, diabetes mellitus, hypertension, history of bleeding, history of stroke, malignant neoplasm, haemoglobin levels, creatinine clearance, C-reactive protein, liver dysfunction, antiplatelet therapy, and NSAIDs use. The reference level of albumin was set at 4.0 g/dL. The solid lines represent the estimated hazard ratios, while the shaded grey areas indicate the 95% CIs. Relative hazard ratio is presented on a loge (natural logarithm) scale. Alb, albumin; HR, hazard ratio.

• 

  Figure 3. Improvement in predictive performance of bleeding risk scores for 1-year major bleeding with the inclusion of albumin. This figure shows the ROC curves for the ORBIT, DOAC, and HAS-BLED scores for 1-year major bleeding, comparing their predictive accuracy alone (dashed lines) vs. when combined with albumin (solid lines). IDI, Integrated discrimination improvement; NRI, Net reclassification improvement.

• 

  Figure 4. Effect of albumin levels on major bleeding in four different DOACs. This figure illustrates the relationship between albumin levels (Alb, g/dL) and the relative hazard ratios for four different DOACs. The impact of albumin levels on bleeding risk for each DOAC was analyzed using a Cox proportional hazards model with a linear term for albumin. The covariates for Cox proportional hazards model were age, female sex, body mass index, diabetes mellitus, hypertension, history of bleeding, history of stroke, malignant neoplasm, haemoglobin levels, creatinine clearance, C-reactive protein, liver dysfunction, antiplatelet therapy, and NSAIDs use. The reference level of albumin was set at 4.0 g/dL. The solid lines represent the estimated hazard ratios, while the shaded grey areas indicate the 95% CIs. Relative hazard ratio is presented on a loge (natural logarithm) scale. Alb, albumin; HR, hazard ratio; DOAC, direct oral anticoagulant.