Scoring system for prediction of mortality after endoscopic ligation in esophageal variceal bleeding

Abstract

Liver cirrhosis (LC) is the final stage of chronic liver disease. It is classified as compensated or decompensated cirrhosis based on the presence or absence of jaundice, ascites, encephalopathy, and gastrointestinal bleeding including esophageal varices (EVs). The transition rate from the asymptomatic compensatory period to the symptomatic decompensated period has been reported to be 5–7%/year and ascites is often an early symptom of decompensation. Portal hypertension (PH) is defined as the clinical state in which portal pressure is elevated due to some etiology. Moreover, clinically significant portal hypertension (CSPH) is defined as a decompensated LC state, and the diagnostic criterion is the hepatic venous pressure gradient (HVPG) ≥10 mmHg. Furthermore, severe PH, defined as an HVPG ≥12 mmHg, is a risk factor for EVs bleeding.¹ EVs bleeding is a major concern in patients with LC that constitutes a serious decompensating event with high mortality.² It is very important to decrease the recurrence rate of EVs in patients with LC, because liver transplantation is limited in Japan compared with that in Western countries. Endoscopic variceal ligation (EVL) is the standard method for the treatment of EVs bleeding. However, complications such as liver failure, renal failure, infection, and rebleeding may occur after EVL. The incidence of rebleeding is about 60%¹ and the in-hospital mortality rate is reported as between 15% and 20%.³

For predicting mortality, HVPG measurement is also a helpful technique, but this measurement is invasive and requires the clinical experience of a physician. Recently, the usefulness of liver or spleen stiffness measured by transient elastography (TE) is often reported,⁴ and CSPH is highly suspected when TE in the liver is >15 kPa.² In the Baveno VII workshop consensus, it is also stated that splenic TE ≥50 kPa has the risk of CSPH and splenic TE ≤40 kPa is a low probability of high-risk varices.² However, TE is expensive ultrasonography and not widely used in general hospitals. Against this background, as a method to easily determine prognosis from blood examination results, Xavier et al.⁵ showed that the albumin–bilirubin index was a good predictor of mortality during hospitalization or within 30 days (area under the curve 0.81, P < 0.01) in 111 patients with LC complicated with acute upper gastrointestinal bleeding. International guidelines, such as the Baveno VII workshop consensus, on managing variceal bleeding, recommend using prognostic scores like the Child–Pugh and Model for End-stage Liver Disease (MELD).^{1, 2} In these guidelines, Child–Pugh class C and MELD score ≥18 are defined as risk factors for rebleeding. However, unfortunately, these scores do not specifically focus on estimating mortality following hemostasis.

About the conversion from JCS to GCS, there is a report showing the approach method with a concordance rate of 80.3%, a 95% confidence interval ranging from 77.4% to 82.9%, a relative concordance rate permitting a one-category deviation of 93.2%, and a confidence interval between 91.2% and 94.8%.⁷ Regarding HCC with portal venous invasion and background, Ichita et al. stated that HCC and a history of liver cirrhosis did not emerge as significant predictors in the analysis. This means that these factors do not affect prognosis and may be due to the small number of HCC with portal venous invasion. However, the background of liver cirrhosis and the amount of alcohol would be better clarified in a further study. Regarding the endoscopic findings of EVs, as written in their limitation section in their article, this information could potentially contribute to mortality rates. Assessing these factors could lead to the development of a more useful scoring system. Especially, the presence of a red color sign, the form of EVs, and the bleeding pattern are directly related to life prognosis.⁸ In Japan, EVL has been developed to the crowding method and the bi-monthly method,⁹ and the rate of EVs rebleeding is much lower than in other countries. Therefore, further studies are needed to confirm whether this score will be applicable to Western countries. Moreover, regarding EVs, information on whether it is a new case or a recurrent case is very important for prognosis, because recurrence cases have a worse prognosis. The previous treatment method is also an important point. It has also been reported that additional consolidation treatment with argon plasma coagulation after initial treatment of EVs improves the prognosis.¹⁰ It seems possible to create a more accurate scoring system by adding endoscopic findings, a history of EVs treatment, and the background of liver cirrhosis in addition to the information from DPC. This HOPE-EVL score shown by Ichita et al. is a simple, clinically relevant, and easy-to-calculate scoring system. It is expected to prospectively collect data and ensure accurate recording of these findings in future research projects.

Authors declare no conflict of interest for this article.

None.