Introduction

Acute variceal bleeding (AVB) is a medical emergency that requires urgent evaluation. It is a complication of portal hypertension and, despite the improvements in the management of these patients, mortality is still high. According to the best available data, all patients should be risk stratified according to the validated clinical scores MELD and Child-Pugh.

Aims & Methods

To access the predictive capacity of PALBI and CLIF-C AD scores and to compare them with the currently recommended MELD and Child-Pugh scores for the following outcomes: in-hospital mortality, mortality at 6 weeks and rebleeding.
Retrospective review of all cirrhotic patients admitted for AVB between January 2016 and December 2022 in a Gastroenterology department in Portugal. The area under the receiver operating characteristics curve (AUROC) was calculated for all scores and compared with DeLong test.

Results

100 patients included, average age 61.87±14.68years, 77% male. Alcohol was the most frequent aetiology (n=56).
All patients were hemodynamically resuscitated and more than 90% underwent endoscopic evalutation within 12hours from admission. 30% of patients had active bleeding at the time of upper GI endoscopy. 90% of patients had esophageal varices (GOV-1) bleeding treated with endoscopic band ligation. 10% of patients had acute gastric (cardiofundal) variceal (GOV2, IGV1) bleeding treated with cyanoacrylate injection. 5% of patients had persistent bleeding and thus a Sengstaken-Blakemore tube was placed. Rebleeding occurred in 11% of patients.
Median time of admission was 6 days (1-30 days). In-hospital mortality was 28% and mortality at 6 weeks was 19%.
The average scores were MELD: 15.98 ± 6.89 points; Child-Pugh 8.65 ± 2.10 points; PALBI -1.71 ± 0.47 points; CLIF-C AD 58.48 ± 11.42 points.
The AUROC of PALBI for in-hospital mortality (AUROC=0.742) and for mortality at 6 weeks (AUROC=0.616) was similar to the traditional scores MELD (AUROC=0.747 and AUROC=0.698) and Child-Pugh (AUROC=0.794 and AUROC=0.673). PALBI was also a good predictor for rebleeding (AUROC=0.795), superior to MELD (AUROC=0.726) and inferior to Child-Pugh (AUROC=0.801). Comparison of the AUROCs of the three scores showed no statistically significant differences.
Regarding CLIF-C AD, it was a good predictor for in-hospital mortality (AUROC=0.708) and mortality at 6 weeks (AUROC=0.706), with no statistically significant differences compared to MELD and Child-Pugh scores. The predictive capacity for CLIF-C AD for rebleeding was not satisfactory (AUROC=0.552).

Conclusion

In addition to liver-specific scores, the predictive performance and clinical utility of PALBI and CLIF-AD C scores for identifying patients with AVB at higher risk for in-hospital mortality, mortality at 6 weeks and rebleeding is very good.
PALBI score deserves recognition, not only because of it is very good predictive capacity for all outcomes, but also because it does not take into account variables whose evaluation has some degree of subjectivity, like hepatic encephalopathy and ascites.