Diagnostic value of the urea-to-creatinine ratio for gastrointestinal bleeding source: influence of renal function

Abstract

Background: Gastrointestinal (GI) bleeding is frequent and clinically critical, especially in patients with renal dysfunction. Localization of the source is relevant for treatment and outcomes. The urea-to-creatinine ratio (UCR) has been proposed as a tool to differentiate between upper and lower GI bleeding. However, its diagnostic utility across varying degrees of renal impairment remains unclear.

Methods: From January 2021 to October 2023, all patients with suspected GI bleeding who underwent endoscopy were retrospectively analyzed in a single center. Patients without confirmed bleeding were excluded. Patients were stratified by renal function: normal renal function, chronic kidney disease (CKD), acute kidney injury (AKI), and AKI with pre-existing CKD (acute-on-chronic). We assessed the discriminatory power of the UCR to distinguish upper from lower GI bleeding within each group.

Results: A total of 849 patients (mean age 66.9 ± 18.2 years) were included. Approximately two-thirds of the patients (n = 544; 64.1%) were male. CKD was present in 321 (37.8%) and AKI in 354 (41.7%) patients; 199 (56.2%) of those with AKI had pre-existing CKD. Upper GI bleeding occurred more frequently in patients with AKI (75.1%) and CKD (72.9%) than in those with normal renal function (p < 0.0001 and p = 0.0003, respectively); both groups also required transfusions more frequently (p < 0.0001 for both). UCR values were significantly higher in upper vs. lower GI bleeding in patients with normal renal function (67.5 vs. 42.5; p < 0.0001), but this difference was reduced in CKD (49.3 vs. 41.0; p = 0.0103) and not significant in AKI (53.1 vs. 47.7; p = 0.09). The diagnostic performance of UCR was best in patients with normal renal function (AUROC 0.69) and markedly impaired in CKD (AUROC 0.56) and AKI (AUROC 0.54).

Conclusion: The diagnostic value of UCR to localize GI bleeding depends strongly on renal function. While the UCR may help to identify the bleeding site in patients with normal kidney function, impaired renal function hampers the reliability of the UCR, especially in patients with AKI. As renal dysfunction is common in patients with GI bleeding, kidney function must be taken into account when interpreting UCR values.

Supplementary Information: The online version contains supplementary material available at 10.1186/s12882-025-04382-y.

Keywords: Acute kidney injury (AKI); Chronic kidney disease (CKD); Endoscopy; Gastrointestinal (GI) bleeding; Renal dysfunction; Urea-to-creatinine ratio (UCR).

Fig. 1

Study design. Gastrointestinal bleeding was confirmed in 849 patients after review of electronic in-patient records and endoscopic findings

Fig. 2

Venn diagram of the study population according to kidney function. Of the 354 patients with AKI, 199 also had CKD. CKD without AKI was present in 122 patients, while 373 patients had normal kidney function (i.e., neither AKI nor CKD)

Fig. 3

Occurrence of site of GI bleeding and transfusion in relation to renal impairment. (A and B) Localization of bleeding in patients with CKD (A) and AKI (B). Upper GI bleeding is more common in patients with CKD and AKI (p = 0.0003 and p < 0.0001, respectively). (C) Risk of upper GI bleeding. Both patients with AKI and patients with CKD have an increased risk of upper GI bleeding compared to those with normal kidney function (OR 2.16; CI 1.56 to 2.95 and OR 2.16; CI 1.60 to 2.91, respectively). (D) Transfusion requirement. Patients with CKD were significantly more likely to require transfusions (63.9% vs. 47.5%; **** = p < 0.0001). (E) Probability of transfusion according to renal involvement. Both CKD and AKI patients are more likely to require transfusion during their hospital stay if they presented with GI bleeding (OR 3.00; CI 2.21 to 4.08 and OR 2.40; CI 1.80 to 3.19, respectively)

Fig. 4

UCR for differentiation of upper and lower GI bleeding according to renal function. (A) UCR in the overall cohort. Upper GI bleeding showed a significantly higher UCR (**** = p < 0.0001). A UCR (red threshold line) above 75 or 100 was highly sensitive for upper GI bleeding (93.5% and 98.3%), but poorly specific (23.8% and 12.4%). (B) UCR in patients without CKD or AKI. The UCR for upper GI bleeding was also significantly higher in patients without AKI and CKD (**** = p < 0.0001). (C) UCR in patients with CKD. In patients with CKD, there was only slightly but significantly higher in UCR for upper GI bleeding (* = p < 0.05). (D) UCR in patients with AKI. There were no significant differences in UCR between upper and lower GI bleeding in patients with AKI (ns, p = 0.09). (E) UCR in patients with acute on chronic kidney injury. In patients with acute on chronic kidney injury, there was a weakly significant increase in the UCR. (F) Receiver operating characteristics (ROCs) curves for all subgroups. The largest areas under the ROC curves (AUROCs) were found in those without CKD and AKI, whereas the AUROCs were lower in those with AKI (0.69 vs. 0.54; p < 0.0001). Thick dashed line represents median, thin dashed lines represent the 25th and 75th percentiles, respectively