Cirrhosis complicated with acute esophageal variceal bleeding is a common critical illness, responsible for the high morbidity and mortality in patients with cirrhosis[6–8]. Both EVL and sclerotherapy can safely and effectively treat esophageal variceal bleeding, but EVL is more advantageous in reducing the number of endoscopic treatments and complications[8, 12, 13]. Current guidances recommend that EVL be the first choice for emergency hemostasis if there are no contraindications[6–8, 14]. Blood transfusion is an another important aspect of treatment in patients with EVB, but for patients with upper gastrointestinal bleeding, especially patients with EVB, excessive or rapid blood transfusion or infusion may increase portal pressure and induce rebleeding due to deteriorated pre-existing portal hypertension. Blood transfusion is required for hemoglobin below 70 g/L. The Baveno-VII consensus recommended blood transfusion to maintain hemoglobin at 70 g/L∼80 g/L[6]. The European Society of Gastrointestinal Endoscopy Guidelines released in 2015 also suggested a target hemoglobin of 70 g/L ~ 90 g/L, and the Chinese guideline recommended that it be maintained at 60 g/L ~ 70 g/L[8]. The RBC transfusion threshold in these guidelines is Hb ≤ 70 g/L, which is based on the high-quality randomised controlled trial study of Villanueva et al.[15]. A meta-analysis study by Ayodele et al. suggest that for patients with acute upper gastrointestinal bleeding, use of a restrictive transfusion strategy is associated with a reduction in mortality and rebleeding[16]. Gautham et al.[17]conducted a single center, prospective, noninferiority, randomized controlled trial. In this study, the mortality, the occurrence of rebleeding, and average length of hospital stay in the restrictive transfusion group and the liberal transfusion group during the 45 day follow-up were similar, indicating that the restrictive transfusion strategy is as safe and effective as the liberal transfusion. However, most of the existing restrictive transfusion strategies[15, 17–21] study the prognostic impact on non-variceal patients. Considering that low hemoglobin levels may affect the healing of post-ligation ulcers and domestic blood product resources are scarce, it is necessary to explore whether restrictive transfusion will affect the safety of patients with esophageal varices after EVL. Therefore, this study conducted a related discussion.
In the present study, there were 10 patients with early rebleeding after EVL, which was lower than the results of previous studies. There were 7 patients (8.54%) in the RBC transfusion group and 3 patients (1.83%) in the non RBC transfusion group, which was in keeping with previous report[20]. Considering that the Child-Pugh class B, the presence of ascites, low Hb level, and the proportion of plasma and albumin transfusion were greater in the RBC transfusion group, the general condition of patients was worse than that of the non transfusion group. The proportion of patients with acute hemorrhage in the RBC transfusion group was relatively large, while preventive treatment was mainly used in the non RBC transfusion group, which was consistent with previous literature[21], and also one of the reasons for the difference in early rebleeding between the two groups. EVB can lead to the reduction of hemoglobin and the loss of coagulation factors, and patients with cirrhosis have poor liver function and insufficient synthesis of coagulation factors, thus increasing the risk of rebleeding. The results of this study also showed that there were significant differences between the RBC transfusion group and the non transfusion group in the length of hospital stay and hospitalization costs, which may be attributed to the increase in blood transfusion costs, and the increase in hospital stay due to severe illness in the RBC transfusion group. Although univariate analysis of early rebleeding after EVL suggested that the type and amount of blood transfusion were risk factors, the correlations were not observed in the results of multivariate analysis and further research is needed.
Previous studies have shown that hemoglobin level is an important predictor of early rebleeding after EVL[10, 11, 22]. As for the impact of blood transfusion on the safety after EVL, this study also further analyzed the impact of RBC transfusion to maintain different postoperative Hb levels (Hb < 70 g/L, 70 g/L ≤ Hb < 90 g/L, and Hb ≥ 90 g/L) on early rebleeding after EVL. The early rebleeding rates of the three groups were 8.57%, 8.51% and 1.83%, which revealed that cirrhotic patients with lower Hb levels are more likely to rebleed. It may be related to the fact that the high hemoglobin level of patients after ligation is conducive to the healing of post-ligation ulcers, but it cannot be ruled out that liver function classification, the presence of ascites and low albumin level affect the results at the same time, which is worthy of further prospective studies in the future. However, pairwise comparison analysis could not prove the difference in rebleeding rate between the groups. Although this study found this relationship between Hb and postoperative rebleeding rate, yet we did not analyze what range of hemoglobin level is the most appropriate, which is worth exploring in prospective randomized controlled trials.
In this study, there were significant differences between the two groups in varicose vein diameter, postoperative Hb, INR, serum albumin, RBC transfusion or not, and the amount of plasma or RBC transfused after ligation. Previous studies have found that the severity of liver cirrhosis is closely related to the prognosis[23–26]. The higher the Child-Pugh grade, the greater the risk of early rebleeding, especially the Child-Pugh class C, but this study failed to confirm this relationship. Considering the multiple and correlation between the factors, combined with the clinical practice, varicose vein diameter, postoperative Hb, INR, RBC transfusion or not, and preoperative albumin were finally included for further binary logistic regression analysis. The results showed that postoperative hemoglobin and preoperative albumin were independent influencing factors of rebleeding after EVL, which may be related to the reduction in albumin and Hb affecting the healing of ulcers after EVL.
The current guidelines recommend albumin as the volume expander in cirrhotic patients with ascites[5, 27]. Indeed, albumin has demonstrated to improve effective hypovolemia, prevent circulatory dysfunction after paracentesis, improve the outcomes in patients suffering spontaneous bacterial peritonitis (SBP) or hepatorenal syndrome (HRS)[28–31]. However, the role of albumin in the management of EVB in cirrhotic patients has not yet been elucidated. Zhou et al.[32]conducted a multicenter retrospective study and found that the albumin concentration > 31.5 g/L was a protective factor for early rebleeding after EVL. A retrospective study by Wang et al.[33]showed that in patients with hypoalbuminemia, albumin infusion can reduce the risk of in-hospital rebleeding, and in patients with Child-Pugh class C, albumin infusion can reduce the number of in-hospital deaths. The results in this study also showed that rebleeding after EVL may be related to a low albumin level. Given that preoperative blood transfusion may increase the risk of rebleeding, it needs to be considered that the low albumin level should be corrected within 3 days after EVL before the formation of ligation ulcers, so as to promote wound healing and improve the hemostatic effect.
According to the results of regression analysis, to further evaluate the relationship between hemoglobin, albumin and early rebleeding after EVL for cirrhotic patients with esophageal varices, this study also constructed a nomogram prediction model, which gives each independent influencing factor the corresponding score, calculates the risk prediction value of early rebleeding according to the total score, and converts the complex regression equation into a visual graph, which is more intuitive and operable. It shows that the lower albumin and hemoglobin levels, the higher the incidence of early rebleeding in patients with cirrhosis.
In the study, 3 patients were treated with balloon tamponade. According to the medical records, one person died. On the fifth day after ligation, the patient developed severe hematemesis, hematochezia, and hemodynamic instability, and finally died after being remedied by balloon tamponade. The mortality rate was not analyzed in this study because the situation of some patients discharged automatically was unknown.
There are some limitations in this study. Firstly, this study was a unicentric clinical review, with a small sample size and certain bias. Secondly, because of the consensus and ethical principles, the control group without RBC transfusion, which should receive RBC transfusion to maintain Hb ≥ 90 g/L. Therefore, in order to reduce rebleeding events and save blood products, prophylactic esophageal variceal ligation should be paid more attention to perform. Thirdly, this study did not record whether the patient had variceal bleeding in the past and endoscopic manifestations (whether there was high-risk erythema, etc.) during ligation, which may affect postoperative rebleeding. Finally, in actual clinical practice, the hemoglobin and albumin levels of cirrhotic patients with esophageal varices rarely reach the high values shown in the nomogram model, and this study did not verify and evaluate the prediction model, which needs further improvement in future research.
In conclusion, the higher the hemoglobin level of cirrhotic patients with esophageal varices after EVL, the lower the incidence of early rebleeding, and restrictive blood transfusion appeared to increase the risk of early rebleeding after EVL, which needs be further confirmed in future research.