A total of 120 patients were included (60 patients in each group; mean age: 72 years; 42.5% women). The demographic, clinical, and analytical characteristics of the sample and the bivariate analyses based on the presence of active bleeding are shown in Table 1.

Table 1: Table 1: Characteristics of the patients included in the study. Quantitative variables are expressed as mean and standard deviation (X ± SD). Qualitative variables are expressed as absolute (n) and relative frequencies (%). Difference between the value of the parameter in the baseline analysis (1) and pre-CT angiography (2) (negative values indicate an increase in the parameter between the two). INR: International Normalized Ratio. INR3: Percentage of patients with increased INR, defined as INR >2 and INR >3 for patients without and with anticoagulant treatment, respectively. aPTT: Activated Partial Thromboplastin Time ratio. Hb: Serum hemoglobin (g/dL). Platelet count: Platelet count (number of platelets * 10^6/mL). SBP: Systolic Blood Pressure (mmHg). DBP: Diastolic Blood Pressure (mmHg). MAP: Mean Arterial Pressure (mmHg). HR: Heart Rate (beats/minute). *Significant value.

The univariate logistic regression analysis for the dependent variable "active bleeding" is summarized in Table 2. The variables that showed the highest odds ratios (OR) were the difference between previous INR and immediately prior to CTA (OR = 0.235; 95% CI: 0.099-0.558), tachycardia (OR = 3.5; 95% CI: 1.556-7.874), and hemodynamic instability (OR = 4.571; 95% CI: 2.032-10.284).

Table 2: Table 2: Univariate logistic regression analysis for the variables that were significant in the bivariate analyses. OR: crude odds ratio. ^Difference between the value of the parameter in the baseline analysis (1) and prior to CTA (2).

The multivariate logistic regression analysis, which included the last three variables mentioned, was statistically significant, with adjusted OR values of 0.232, 2.807, and 4.326, respectively (Table 3).

Table 3: Table 3: Multivariate logistic regression analysis for the variables that showed higher odds ratios in the univariate logistic regression. ORa: adjusted odds ratio.

The ROC curve analysis showed the following results: for the "INR difference" variable, the AUC value was 0.736 (95% CI: 0.646-0.825; p < 0.001); for the "tachycardia" variable, the AUC value was 0.633 (95% CI: 0.533-0.733; p = 0.012); and for the "hemodynamic instability" variable, the AUC value was 0.667 (95% CI: 0.569-0.764; p = 0.02). Finally, the ROC analysis of the combined multivariate model showed an AUC of 0.813 (95% CI: 0.738-0.887; p < 0.001) (Figure 4).

Fig 4: Figure 4: ROC curves for the variables included in the multivariate model (difference between baseline and prior to CTA INR, tachycardia, and hemodynamic instability).

In 9 of the 60 patients in the case group, a venous origin of active bleeding was identified. The results of the bivariate analyses based on the dependent variable "type of bleeding" (arterial vs. venous) are shown in Table 4. The importance of differentiating between arterial or venous bleeding lies in its subsequent management, as patients who show extravasation in the arterial phase will undergo diagnostic and therapeutic angiography.

Table 4: Table 4: Bivariate Analysis for the Dependent Variable "Type of Active Bleeding”. Quantitative variables are expressed as mean and standard deviation (X ± SD). Qualitative variables are expressed as absolute frequencies (n) and relative frequencies (%). ^Difference between the parameter value in the first baseline analysis (1) and pre-angioCT (2) (negative values indicate an increase in the parameter between the two). INR: International Normalized Ratio. INR3: Percentage of patients with increased INR, defined as INR > 2 and INR > 3 for patients without and with anticoagulant treatment, respectively. aPTT: Activated Partial Thromboplastin Time Ratio. Hb: Serum Hemoglobin (g/dL). Platelet count: Platelet count (number of platelets × 10^6/mL). SBP: Systolic Blood Pressure (mmHg). DBP: Diastolic Blood Pressure (mmHg). MAP: Mean Arterial Pressure (mmHg). HR: Heart Rate (beats/minute). ✞p-value for the Student's t-test for independent samples in the case of quantitative variables and for Fisher's exact test for qualitative variables (due to a low count in at least one of the cells in the contingency table). *Significant value.

Several studies agree that hemodynamic instability is the main variable in which active contrast extravasation is significantly more likely. However, other studies have not identified significant differences regarding these parameters. Due to the intermittent, transient, and sometimes self-limiting nature of LGIB the diagnostic yield of CTA may be limited. Furthermore, contrast extravasation that is only observed in the venous phase (portal and/or delayed) could correspond to a true venous bleeding or, less frequently, to intermittent, transient, or slow-flow arterial bleeding.