Abstract
Introduction
In cirrhotic patients with esophageal variceal bleeding (EVB), most guidelines recommend endoscopy within 12 h. However, the optimal timing for the highest treatment efficacy remains controversial, and current studies provide inconsistent results. This study aims to evaluate the optimal timing by comparing clinical outcomes between early endoscopy (< 12 h) and delayed endoscopy (≥12 h) groups.
Methods
A retrospective study was conducted on 265 cirrhotic patients with EVB who underwent endoscopic variceal ligation (EVL) from January 2022 to December 2024. Patients were divided into early and delayed endoscopy groups based on the timing of endoscopic intervention since admission. The primary outcome was the 5-day treatment failure.
Results
Among the 265 eligible patients, 109 (41.1%) underwent early endoscopy, while 156 (58.9%) underwent delayed endoscopy. There were no statistically significant differences between the early and delayed endoscopy groups regarding 5-day treatment failure (10.1% vs. 9%, p = 0.759), in-hospital mortality (10.1% vs. 7.1%, p = 0.377), the need for ICU care (13.8% vs. 10.9%, p = 0.481), total blood transfusion units (p = 0.585), hemoglobin decrease requiring transfusion (p = 0.31), and the length of hospital stay (p = 0.35). Similarly, after Propensity Score Matching, no significant differences were observed between the two groups regarding those outcomes. Univariate and multivariate logistic regression analysis identified low systolic blood pressure (OR = 0.959, CI: 0.927–0.992), elevated serum creatinine levels (OR = 1.026, CI: 1.004–1.047), and Child-Pugh class C (OR = 7.516, CI: 1.538–36.721) at admission were independently associated with an increased risk of in-hospital mortality.
Conclusion
The timing of endoscopy (before or after 12 h) does not significantly affect clinical outcomes in EVB patients. Endoscopy timing should depend on the patient’s condition and clinical judgement. Key predictors of in-hospital mortality include low systolic blood pressure, high serum creatinine levels, and Child-Pugh class C at admission.
Keywords: Cirrhosis, Gastrointestinal bleeding, Esophageal varices, Timing of endoscopy
Introduction
Upper gastrointestinal (GI) bleeding is a serious and common medical emergency in daily clinical practice, which can lead to unpredictable consequences [1]. In patients with cirrhosis, the most common cause of upper GI bleeding is the rupture of esophageal varices due to portal hypertension. This condition occurs when blood flow in the portal vein is obstructed by structural changes in the cirrhotic liver, resulting in elevated portal pressure and the dilation and rupture of esophageal varices [2]. Nowadays despite advances in management, treatment, and prevention of complications in cirrhosis, the incidence of bleeding from ruptured esophageal varices remains high, ranging from 20 to 40%, with a six-week mortality of 10–20% [3, 4]. Hence, this is a serious complication, directly threatening the patient’s life if not intervened promptly. Currently, overall management of esophageal variceal bleeding (EVB) mainly focuses on key measures such as initial resuscitation to stabilise hemodynamics, the use of vasoconstrictive drugs to reduce portal pressure (octreotide, somatostatin, terlipressin), prophylactic antibiotics, and urgent endoscopic intervention [5]. These approaches not only control bleeding but also contribute significantly to improving long-term outcomes in cirrhotic patients.
It is undeniable that endoscopic variceal ligation (EVL) is the mainstay treatment for EVB, however, the optimal timing for endoscopy remains a topic of debate in present guidelines and research. Most existing guidelines, including the Baveno VII consensus, the American Association for the Study of Liver Diseases (AASLD) guideline, the European Society of Gastrointestinal Endoscopy (ESGE) guideline, and the Belgian guideline, suggest that endoscopy should be performed within 12 h of symptom onset in cirrhotic patients with EVB [5, 6, 7, 8]. In contrast, the UK guideline recommends that in case of severe variceal bleeding and hemodynamic instability, endoscopy should be performed immediately after resuscitation. For others, it should be conducted within 24 h after hospitalisation [9]. Meanwhile, Chinese guideline proposes that endoscopy should be conducted within 12–24 h after the presentation of variceal bleeding [10]. Moreover, these conflicting recommendations appear to rely more on expert consensus than on solid clinical evidence.
Similarly, previous observational studies according to endoscopic timing in cirrhosis with AVB have also yielded inconsistent and contradictory results. Recently, two retrospective studies from China reported no statistically significant differences in the rates of treatment failure and mortality between early endoscopy (< 12 h) and delayed endoscopy (≥12 h) [11], as well as between before and after 6 h since admission [12]. A retrospective study from South Korea also showed comparable results when comparing early endoscopy (< 12 h) and delayed endoscopy (≥ 12 h), with no significant differences in six-week rebleeding rates or mortality [13]. Nevertheless, a separate study from Taiwan involving 311 patients with acute variceal bleeding found that delaying endoscopy > 15 h after admission was an independent risk factor for in-hospital mortality [14]. This issue becomes even more complex with two recent systematic reviews and meta-analyses reporting conflicting findings. While Jung DH et al. [15] concluded that the timing of endoscopy (before or after 12 h) was not related to mortality or rebleeding rates in cirrhotic patients with EVB, Bai et al. [16] suggested that early endoscopy within 12 h might improve survival rates, though it did not provide significant benefits in preventing rebleeding.
In Vietnam, there are currently no specific recommendations from the Vietnamese Gastroenterology or Endoscopy Societies, nor any research studies addressing this issue. As a result, the decision on when to perform endoscopy mainly depends on the clinical judgment and experience of individual physicians. Therefore, we conducted this study intending to shed light on the optimal timing for endoscopic intervention in cirrhotic patients with EVB by comparing outcomes between the early endoscopy group (< 12 h) and the delayed endoscopy group (≥ 12 h). In addition, we also identify independent risk factors predicting in-hospital mortality in cirrhotic patients with EVB after endoscopy treatment.
Materials and methods
Patients cohort
We retrospectively included all cirrhotic patients admitted to Da Nang Hospital for esophageal variceal bleeding from our electronic inpatient database between January 2022 and December 2024, following the return of medical activities to normal after the COVID-19 pandemic. Cirrhotic patients with EVB were divided into 2 different groups based on the timing of endoscopy after being admitted to the emergency department: the early endoscopy group (< 12 h) and the delayed endoscopy group (≥ 12 h).
The inclusion criteria were as follows (1) aged 18 years and older; (2) Diagnosed with cirrhosis either through a prior history of cirrhosis or based on a combination of clinical symptoms, biochemical tests, and imaging results; (3) Admitted for upper GI bleeding (hematemesis, melena, hematochezia) due to EVB, confirmed by endoscopy and treated with EVL. Exclusion criteria were as follows (1) not undergo endoscopy or have contraindications for the procedure; (2) upper GI bleeding due to other causes (peptic ulcer bleeding, gastritis, portal hypertensive gastropathy bleeding, etc.), additionally, gastric varices were also excluded as our facility did not have histoacryl sclerotherapy at the time of the study; (3) endoscopy had been undergone at other hospitals within 7 day prior to admission.
Management
Upon arrival at the emergency department with suspected variceal bleeding, emergency physicians consulted with the on-duty gastroenterologist to assess the feasibility of performing the endoscopic intervention. The procedure was generally carried out once the patient had been intensively resuscitated and their hemodynamics had been stabilized. The decision to proceed with endoscopy, as well as the timing, was based on the clinical evaluation of the on-duty gastroenterologist.
Endoscopic variceal ligation (EVL) is performed by an experienced endoscopist with at least 3 years of experience in therapeutic endoscopy, using advanced endoscopic systems, including the Olympus Elvis X1, CV 190, and CV 170, all featuring straight scopes and large working channels. The procedure also involves suction devices capable of holding multiple rubber bands for ligation.
Before the procedure, both the patient and their family were thoroughly informed about the endoscopy process, including its risks and benefits, and they signed a written consent form for the procedure.
In terms of treatment, all patients are managed according to current ESGE and Baveno VII guidelines [6, 7]. Adequate resuscitation with fluids and blood is initiated to ensure the proper tissue perfusion, along with prophylactic antibiotics and a vasoactive medicine with Octreotide (50 µg intravenous bolus, followed by a continuous infusion of 50 µg/hour for 2–5 days) immediately upon admission. Blood transfusions are given when hemoglobin levels fall to ≤ 70 g/L, with the goal of maintaining levels between 7 and 9 g/dL.
Data collection
We identified cirrhosis patients with EVB by reviewing the electronic inpatient database of Da Nang Hospital during the study period. Patient eligibility for inclusion was determined through a comprehensive evaluation, including medical history, clinical presentation, laboratory findings, and endoscopic results during hospitalisation. If a patient was hospitalised multiple times for variceal bleeding during the study period, only the first admission was included. We gathered information on patient demographics, medical history, comorbidities, and key laboratory results recorded at the time of admission. The Child-Pugh score, MELD score, MELD-Na, and Glasgow-Blatchford score were calculated based on the patient’s medical records.
MELD score = 9.57 × loge (creatinine) + 3.78 × loge (total bilirubin) + 11.2 × loge (INR) + 6.43 [17].
MELD-Na = MELD–Na – [0.025 × MELD × (140-Na)] + 140 with parameters Na < 125 calculated as 125 and Na > 140 calculated at 140 [18].
Outcomes
The primary outcome was the 5-day treatment failure. The secondary outcomes included in-hospital mortality, need for ICU care, total blood units transfused, hemoglobin decrease requiring transfusion, and length of hospital stay.
Definitions
The endoscopic duration was defined as the time interval between admission to the emergency department and the commencement of the endoscopy procedure. An endoscopy performed within 12 h was classified as an early endoscopy, while one conducted after 12 h was classified as a delayed endoscopy.
5-day treatment failure is defined as the inability to control bleeding or the recurrence of bleeding within the first 5 days (Baveno VII) [7]. In-hospital mortality is defined as death occurring within the hospital during treatment due to the current bleeding episode or severe liver disease. Need for ICU care refers to the requirement for intensive care unit admission due to hemodynamic instability, coma, or respiratory failure following an endoscopic procedure.
Statistical analysis
Qualitative variables were presented as numbers and percentages (%), while quantitative variables were presented as median with interquartile range (IQR).
To assess the differences between the early endoscopy and delayed endoscopy groups, we applied the nonparametric Mann-Whitney U test for quantitative variables and the Chi-square test or Fisher’s exact test for qualitative variables.
Univariate and multivariate logistic regression analyses were performed to identify independent risk factors predicting in-hospital mortality in patients with EVB.
The Propensity Score Matching (PSM) method with a 1:1 ratio and a matching tolerance of 0.02 was used to compare outcomes between the two groups of endoscopy time. The goal of this method was to minimize bias and ensure comparability of baseline, clinical, and laboratory characteristics between the two groups. Matching factors included: age, gender, history of bleeding, etiology of cirrhosis, liver cancer, comorbidities, symptoms, heart rate, systolic blood pressure, hemoglobin levels, platelet count, liver function (AST, ALT, total bilirubin), albumin, urea, creatinine, sodium, INR, prothrombin time, Child-Pugh score, MELD, MELD-Na, and Glasgow-Blatchford score.
All p-values were two-tailed, and a p-value less than 0.05 was considered statistically significant.
All statistical analyses were performed using SPSS software, version 29.0.
Results
Patients
At first, there were a total of 415 cirrhotic patients with symptoms of upper gastrointestinal bleeding. After applying the inclusion and exclusion criteria, 265 eligible patients with cirrhosis admitted for EVB were included in this study (Fig. 1). Table 1 outlines the baseline characteristics of the patients enrolled. The median age was 55 years (IQR: 50–64), and 231 (87.2%) patients were male. Of the patients, 32.5% (86) reported a history of variceal bleeding before this study. Alcohol abuse was the predominant etiology of cirrhosis (53.2%), followed by chronic hepatitis B infection (18.9%). Liver cancer accounted for 15.5% (41) of the cases. The majority of patients (79.3%) were admitted with hematemesis as the primary symptom.
Fig. 1.
Patient selection flowchart
Table 1.
Baseline features of cirrhotic patients with EVB in early (< 12 h) and delayed (≥12 h) endoscopy before and after PSM
| Before PSM | After PSM | |||||
|---|---|---|---|---|---|---|
| Early endoscopy (< 12 h) n = 109 |
Delayed endoscopy (≥12 h) n = 156 |
p | Early endoscopy (< 12 h) n = 90 |
Delayed endoscopy (≥12 h) n = 90 |
p | |
| Age (years), median (IQR) | 55 (50–62) | 56 (50–65) | 0.383 | 55 (49–62) | 56 (50–65) | 0.284 |
| Sex (Male) | 96 (88.1%) | 135 (86.5%) | 0.713 | 78 (86.7%) | 77 (85.6%) | 0.829 |
| Bleeding history, n (%) | 36 (33%) | 50 (32.1%) | 0.867 | 30 (33.3%) | 28 (31.1%) | 0.75 |
| Etiology, n (%) | 0.795 | 0.903 | ||||
|
HBV HCV Alcohol HBV + alcohol HCV + alcohol Others |
19 (17.4%) 8 (7.3%) 56 (51.4%) 10 (9.2%) 10 (9.2%) 1 (0.9%) |
31 (19.9%) 11 (7.1%) 85 (54.5%) 18 (11.5%) 11 (7.1%) 0 |
13 (14.4%) 6 (6.7%) 50 (55.6%) 11 (12.2%) 9 (10%) 1 (1.1%) |
16 (17.8%) 7 (7.8%) 47 (52.2%) 12 (13.3%) 8 (8.9%) 0 |
||
| Liver cancer, n (%) | 20 (18.3%) | 21 (13.5%) | 0.279 | 11 (12.2%) | 16 (17.8%) | 0.297 |
| Comorbidity, n (%) | 34 (31.2%) | 54 (34.6%) | 0.560 | 27 (30.0%) | 29 (32.2%) | 0.747 |
| Symptom, n (%) | ||||||
|
Hematemesis Hematochezia Melena |
95 (87.2%) 11 (10.1%) 75 (68.8%) |
115 (73.7%) 13 (8.3%) 112 (71.8%) |
0.008 0.624 0.600 |
77 (85.6%) 9 (10.0%) 60 (66.7%) |
79 (87.8%) 7 (7.8%) 59 (65.6%) |
0.661 0.6 0.875 |
| Heart rate (beats per minute), median (IQR) | 100 (89–108) | 90 (85–100) | < 0.001 | 95 (88–106) | 93 (87–102) | 0.326 |
| Systolic blood pressure (mmHg) |
110 (100–120) |
110 (100–120) |
0.326 |
110 (100–120) |
110 (100–120) |
0.829 |
| Hemoglobin (g/l), median (IQR) | 82 (66–98) | 85 (68.5-105.5) | 0.183 | 84 (68–102) | 86 (68–105) | 0.496 |
| Platelet (109/L), median (IQR) | 97 (66–151) | 97.5 (72.5–135) | 1 | 97.5 (67–150) | 102 (73–138) | 0.653 |
| AST (U/L), median (IQR) | 73 (43.3-129.5) | 70.4 (47–126) | 0.825 | 73 (42-127.4) | 66 (50–104) | 0.828 |
| ALT (U/L), median (IQR) | 34.2 (22.1–59) | 33.4 (24.1–47.7) | 0.754 | 33.8 (21-57.9) | 33.3 (26–50) | 0.838 |
| Total Bilirubin (µmol/L), median (IQR) |
35.3 (20.4–57.7) |
35.5 (21.4–55.6) |
0.957 |
35.3 (20.4–53) |
28.6 (20.4–47.1) |
0.493 |
| Albumin (g/L), median (IQR) | 28 (24–31) | 27.9 (23-30.5) | 0.839 | 28 (24-31.1) | 28 (24-31.1) | 0.493 |
| Urea (mmol/L), median (IQR) | 6.9 (5–10) | 6.7 (4.8–9.7) | 0.593 | 6.7 (4.9–10) | 6.7 (5-9.1) | 0.819 |
| Creatinine (µmol/L), median (IQR) | 71 (62.5–88) | 71.3 (58.9–86) | 0.616 | 70 (62.6–87) | 70.7 (58–82) | 0.357 |
| Na+ (mmol/L), median (IQR) |
133.9 (131.5 -136.7) |
134.5 (131.3–137) |
0.691 |
133.9 (131.5-136.7) |
134.5 (130.8-137.4) |
0.498 |
| INR, median (IQR) | 1.4 (1.2–1.6) | 1.3 (1.2–1.6) | 0.338 | 1.4 (1.2–1.6) | 1.3 (1.2–1.5) | 0.201 |
| PT (%), median (IQR) | 15.4 (13.2–18.2) | 14.8 (13.1–18.5) | 0.444 | 15.4 (13.2–18) | 14.6 (12.8–17.9) | 0.255 |
| Child-Pugh, median (IQR) | 8 (7–9) | 8 (7–10) | 0.389 | 8 (7–9) | 8 (7–9) | 0.322 |
| MELD, median (IQR) | 13 (11–16) | 13 (10–16) | 0.496 | 13 (11–16) | 11.5 (9–16) | 0.146 |
| MELD-Na, median (IQR) | 17 (15–21) | 18 (14–21) | 0.861 | 17.5 (15–21) | 18 (14–20) | 0.486 |
| GBS, median (IQR) | 10 (9–12) | 9 (7–12) | 0.162 | 10 (8–12) | 10 (7–12) | 0.699 |
Abreviation: PSM, propensity score matching; HBV, hepatitis B virus; HCV, hepatitis C virus; AST, aspartate aminotransferase; ALT, alanine aminotransferase; INR, international normalized ratio; PT, prothrombin time; MELD, model for end-stage liver disease; GBS, Glasgow-Blatchford score
Upon admission, patients presented with an elevated heart rate with a median of 92 beats per minute, and a median systolic blood pressure was 110 mmHg. In terms of the severity of liver disease, more than 80% of cirrhotic patients had a Child-Pugh score of 7 or higher (Child B or C). In addition, the majority of admitted patients had MELD and MELD-Na scores exceeding 10, with prevalence rates of 73.6% and 94.7%, respectively. At the same time, their median Glasgow-Blatchford Score (GBS) was also ≥ 7, indicating a high risk of bleeding and the need for urgent endoscopy intervention.
In comparing the two groups, patients who underwent early endoscopy (< 12 h) had a significantly higher incidence of hematemesis and heart rate at admission compared to those who received delayed endoscopy (≥ 12 h) (p = 0.008 and p = < 0.001). However, no significant differences were observed between the two groups regarding other baseline characteristics with p > 0.05. After 1:1 PSM, 90 patients from each group were analysed, and no significant differences were found between the two groups regarding all baseline features.
Influence of endoscopic timing on clinical outcomes
In the overall analysis, within the first five days, the incidence of treatment failure was lower in the delayed endoscopy group (≥ 12 h) compared to the early endoscopy group (< 12 h) (9.0% vs. 10.1%) (Fig. 2). After propensity score matching (PSM), the findings remained unchanged (6.7% vs. 11.1%). However, the differences were not statistically significant (p = 0.759 and p = 0.295).
Fig. 2.
Treatment failure within the first 5 days of patients receiving early (< 12 h) and delayed (≥12 h) endoscopy, before and after PSM
The overall in-hospital mortality rate and the need for ICU care were 8.3% (22 patients) and 12.1% (32 patients), respectively (Figs. 3 and 4). Similarly to the 5-day treatment failure, although the early endoscopy group (< 12 h) showed higher mortality and ICU admission rates than the delayed endoscopy group (≥ 12 h), the differences remained statistically insignificant both before and after PSM (p > 0.05).
Fig. 3.
In-hospital mortality rate in patients receiving early (< 12 h) and delayed (≥12 h) endoscopy, before and after PSM
Fig. 4.
The need for ICU care in patients receiving early (< 12 h) and delayed (≥12 h) endoscopy, before and after PSM
The two endoscopy groups also showed comparable results with no notable variation regarding the length of hospital stay, total blood units transfused, and hemoglobin decrease requiring transfusion (Table 2).
Table 2.
Clinical outcomes regarding hospital duration and total blood units transfused in patients receiving early (< 12 h) and delayed (≥12 h) endoscopy, before and after PSM
| Before PSM | After PSM | |||||
|---|---|---|---|---|---|---|
| Early endoscopy (< 12 h) n = 109 |
Delayed endoscopy (≥12 h) n = 156 |
p | Early endoscopy (< 12 h) n = 90 |
Delayed endoscopy (≥12 h) n = 90 |
p | |
| Length of hospital stay (days) | 6 (5–8) | 7 (5–8) | 0.35 | 6 (5–8) | 7 (6–8) | 0.187 |
| Hemoglobin decrease requiring transfusion |
71 (65.1%) |
92 (59%) |
0.31 |
58 (64.4%) |
53 (58.9%) |
0.443 |
| Total blood units transfused | 1 (0–3) | 1 (0–3) | 0.585 | 1 (0–3) | 1 (0–3) | 0.688 |
Abbreviation: PSM, propensity score matching
Predictors for in-hospital mortality
The logistic regression model was applied to identify patients’ characteristics as well as liver disease severity at baseline that predicted death during hospitalization. The result is presented in Table 3. Both univariate and multivariate analyses revealed that a low level of systolic blood pressure at admission was associated with an increased risk of in-hospital mortality (OR = 0.959, CI: 0.927–0.992; p = 0.014). Furthermore, we also found that the risk of in-hospital mortality was also considerably increased among patients with high creatinine levels (OR = 1.026, CI: 1.004–1.047; p = 0.018) and Child-pugh class C at admission (OR = 7.516, CI: 1.538–36.721; p = 0.013).
Table 3.
Predictors for in-hospital mortality
| Logistic regression | ||
|---|---|---|
| Univariate | Multivariate | |
| Age (years) | p = 0.951 | |
| Gender- Male | p = 0.586 | |
| Liver cancer | p = 0.118 | |
| Comorbidity | p = 0.885 | |
| Heart rate (beats per minute) | p = 0.019 | 0.422 |
| Systolic blood pressure (mmHg) | p < 0.001 | OR = 0.959 (CI: 0.927–0.992) p = 0.014 |
| Hemoglobin (g/L) | p = 0.244 | |
| Albumin (g/L) | p = 0.085 | |
| Urea (mmol/L) | p = 0.037 | 0.854 |
| Creatinine (µmol/L) | p = 0.001 | OR = 1.026 (CI: 1.004–1.047) p = 0.018 |
| INR | p = 0.015 | 0.561 |
| Child Pugh C | p < 0.001 | OR = 7.516 (CI: 1.538–36.721) p = 0.013 |
| MELD | p < 0.001 | 0.201 |
| MELD-Na | p < 0.001 | 0.342 |
| GBS | p = 0.002 | 0.541 |
| Days of hospital stay > 7 | p = 0.281 | |
Abreviation: INR, international normalized ratio; MELD, model for end-stage liver disease; GBS, Glasgow-Blatchford score
Survivor vs. non-survivor comparison during hospitalization
Baseline characteristics and clinical parameters were compared between survivors (n = 243) and non-survivors (n = 22), as shown in Table 4. No significant differences were observed in age, sex, comorbidities, endoscopy timing, or liver cancer status (p > 0.05).
Table 4.
Comparison of baseline clinical and laboratory characteristics between in-hospital survivors and non-survivors
| Alive patients n = 243 |
Deceased patients n = 22 |
p | |
|---|---|---|---|
| Endoscopy time | 0.377 | ||
|
< 12 h ≥12 h |
98 (40.3%) 145 (59.7%) |
11 (50%) 11 (50%) |
|
| Age (years), median (IQR) | 55 (50–64) | 54.5 (51.5–62.3) | 0.928 |
| Sex (Male) | 221 (86.8%) | 20 (90.9%) | 0.75 |
| Bleeding history, n (%) | 79 (32.5%) | 7 (31.8%) | 0.947 |
| Liver cancer, n (%) | 35 (14.4%) | 6 (27.3%) | 0.124 |
| Comorbidity, n (%) | 81 (33.3%) | 7 (31.8%) | 0.885 |
| Heart rate (beats per minute), median (IQR) | 92 (85–102) | 105 (89.8-110.5) | 0.007 |
| Systolic blood pressure (mmHg) | 110 (100–120) | 100 (77.5–110) | 0.001 |
| Hemoglobin (g/l), median (IQR) | 84 (68–104) | 75 (66-87.5) | 0.185 |
| Platelet (109/L), median (IQR) | 96 (67–137) | 109.5 (84.8-153.5) | 0.229 |
| AST (U/L), median (IQR) | 69 (44.4-123.8) | 88.8 (69.9-200.9) | 0.037 |
| ALT (U/L), median (IQR) | 33.4 (22.6–50) | 47.5 (27.9–88.8) | 0.062 |
| Total Bilirubin (µmol/L), median (IQR) | 34.9 (20.4–55.4) | 43.4 (30.5–79.9) | 0.051 |
| Albumin (g/L), median (IQR) | 28 (24–31) | 27 (21.4–28) | 0.094 |
| Urea (mmol/L), median (IQR) | 6.6 (4.9–9.4) | 8.1 (5.3–11.7) | 0.111 |
| Creatinine (µmol/L), median (IQR) | 70 (59-84.7) | 97.3 (81.1-122.5) | < 0.001 |
| Na+ (mmol/L), median (IQR) | 134.4 (131.6-136.9) | 131.7 (128.1-135.2) | 0.015 |
| INR, median (IQR) | 1.35 (1.19–1.61) | 1.49 (1.28–1.89) | 0.051 |
| PT (%), median (IQR) | 15.2 (13.1–18.2) | 16.7 (13.5–21.0) | 0.106 |
| Child-Pugh, median (IQR) | 8 (7–9) | 10 (9–11) | < 0.001 |
| MELD, median (IQR) | 13 (10–16) | 16.5 (13–20) | 0.001 |
| MELD-Na, median (IQR) | 17 (14–20) | 22 (18–27) | < 0.001 |
| GBS, median (IQR) | 9(8–12) | 12.5 (10–13) | 0.003 |
Abreviation: AST, aspartate aminotransferase; ALT, alanine aminotransferase; INR, international normalized ratio; PT, prothrombin time; MELD, model for end-stage liver disease; GBS, Glasgow-Blatchford score
However, non-survivors had significantly higher heart rates (p = 0.007) and lower systolic blood pressure (p = 0.001) at admission. Laboratory findings showed worse liver and kidney function in non-survivors, with higher AST (p = 0.037), creatinine (p < 0.001), and lower sodium (p = 0.015). Severity scores, including Child-Pugh, MELD, MELD-Na, and GBS, were also significantly higher in the deceased group (all p < 0.01).
Discussion
The optimal timing for endoscopy in cirrhotic patients with EVB has been a longstanding subject of debate. While most guidelines have suggested performing endoscopy within 12 h since arrival at the hospital [5, 6, 7, 8], there remains a lack of robust clinical evidence to firmly support this recommendation. Furthermore, these guidelines seem to rely more on “expert opinion” rather than solid clinical data. Here, our data did not support the above recommendation and found no significant benefit of early endoscopy (< 12 h) on the outcomes of cirrhotic patients with EVB. Specifically, we concluded that the endoscopy timing did not correlate with short-term outcomes regarding 5-day treatment failure, in-hospital mortality, need for ICU, total blood units transfused, hemoglobin decrease requiring transfusion, and hospital duration in these patients.
In this study, the baseline characteristics of all eligible patients showed some variation due to the complexity of real-world clinical settings, posing challenges for further analysis. To correct this imbalance, we applied propensity score matching (PSM) using a 1:1 nearest-neighbour method to pair patients from the two distinct endoscopy groups. After matching, no substantial differences were found in the baseline characteristics between the 90 matched pairs of patients. Despite this, we also observed no significant correlation between the timing of endoscopy (before or after 12 h) and the clinical outcomes of patients with EVB.
Aside from the ongoing debate regarding selecting the timing of endoscopy in acute variceal bleeding, the definition of endoscopy time is also heterogeneous across different studies and guidelines. This inconsistency underscores the challenges in assessing the effectiveness of endoscopic procedures. While some studies classify endoscopy performed within 12 h of admission as urgent and after 12 h as non-urgent [13, 19, 20], others use terms like early endoscopy for procedures conducted within 12 h and delayed endoscopy for those performed after 12 h [2, 15, 16, 21]. Moreover, a study from China even considered procedures done within 6 h as urgent and those done between 6 and 24 h as early endoscopy [12]. Here, we followed the most common classification: endoscopy performed < 12 h as early and ≥ 12 h as delayed.
In our research, despite no significant differences, patients experiencing delayed endoscopy (≥ 12 h) appeared to have lower risks of treatment failure, in-hospital mortality, and ICU care than those with early endoscopy (< 12 h). This challenges the conventional belief that early endoscopy could improve outcomes by enabling faster hemostasis [22]. However, performing endoscopy too early may disrupt essential resuscitation efforts, increase the risk of ischemic complications, and shorten the action duration of vasoactive drugs or antibiotics before endoscopy [11, 23]. Moreover, patients who undergo early endoscopy tend to have more active bleeding, which means there is often a larger volume of blood and stomach contents present. This can obscure the source of bleeding and make it more difficult to obtain a clear view during the procedure, giving rise to technical challenges as well as complications such as aspiration or perforation [24], possibly resulting in worse outcomes. On the other hand, delayed endoscopy may be safer, providing a clearer visual field, especially after vasoactive drugs have had time to reduce portal pressure [13].
Although contrary to guideline recommendations, our data is consistent with recent cohort studies [11, 12, 13]. A recent retrospective study on 534 cirrhotic patients with acute variceal bleeding by Peng et al. also reported no significant association between the timing of endoscopy and clinical outcomes, when comparing treatment failure rates and in-hospital mortality between different endoscopy timing groups (before and after 12 h, before and after 24 h, before and after 48 h) [11]. Similarly, another multi-centre study with a larger patient population also yielded similar results, indicating that conducting endoscopy within 6–12 h or up to 24 h after presentation led to comparable outcomes in terms of treatment failure, death, and ICU care [12]. In a meta-analysis of 5 studies, Jung et al. [15] concluded that the timing of endoscopy (before and after 12 h) did not significantly impact mortality or rebleeding rates in cirrhotic patients with variceal bleeding. Therefore, selecting an appropriate time for endoscopy based on the patient’s condition is more important than performing an emergency endoscopy.
In Vietnam, there are currently no official guidelines as well as research on the optimal timing of endoscopy and this study represents the first attempt to bridge this gap. Clinical decisions regarding endoscopy timing are primarily guided by physician experience and recommendations from major gastroenterology societies. Although our facility is equipped with adequate resuscitation and intensive care resources, challenges remain when compared to healthcare systems in developed countries. Factors such as patient overload can impact the ability to provide optimal treatment and resuscitation care, potentially influencing clinical outcomes. This may explain why the overall treatment failure and in-hospital mortality rates in our data were higher than those reported in other studies [11, 12, 19]. By doing so, our findings contribute to the ongoing discussion on endoscopy timing, offering valuable perspectives tailored to Vietnam’s healthcare conditions. We hope this research will serve as a foundation for future guidelines, ultimately improving patient care and outcomes in this setting.
Variceal bleeding is a severe complication in cirrhotic patients with high mortality despite advances in treatment [25]. In our study, the in-hospital mortality rate reached 8.3%, emphasising the ongoing clinical challenge. Identifying predictors of mortality is crucial for early risk assessment, optimizing treatment strategies, and improving patient survival. Through logistic regression analysis, we identified three independent risk factors for in-hospital mortality. First, low systolic blood pressure (SBP) at admission was associated with an increased risk of in-hospital mortality in EVB patients (OR = 0.959, CI: 0.927–0.992), aligning with Chirapongsathorn et al. [26], who also found an inverse relationship between SBP and mortality risk. Low SBP may signal inadequate compensation for blood loss, increasing the risk of organ failure and mortality. Second, hepatorenal syndrome (HSR) not only has high morbidity and mortality rates but also significantly increases the burden on healthcare resources [27, 28]. In EVB, acute blood loss worsens renal perfusion, triggering HRS. Several studies have linked elevated serum creatinine levels with short-term mortality [29, 30], and our study also identified increased serum creatinine as an independent predictor of in-hospital mortality (OR = 1.026, CI: 1.004–1.047). Lastly, Child-Pugh type C was strongly associated with in-hospital mortality (OR = 7.516, CI: 1.538–36.721), consistent with prior studies [29, 31]. These findings emphasise the importance of Child-Pugh not only for risk stratification but also for guiding treatment strategies and enabling early and timely interventions. Unlike the Child-Pugh score, the MELD score did not show significant predictive value in in-hospital death through multivariate analysis. While both scores are widely recognized for predicting outcomes in cirrhotic patients [32, 33], the Child-Pugh score includes clinical parameters, which may provide a more comprehensive assessment of acute decompensation in patients with EVB. In contrast, the MELD score, which relies primarily on laboratory values, may not fully capture the complex clinical deterioration that occurs in this patient population.
In addition to the multivariate regression analysis, our comparison between survivors and non-survivors also revealed that non-survivors presented with worse hemodynamic and biochemical profiles. Specifically, higher heart rate, lower systolic blood pressure, elevated AST and creatinine levels, hyponatremia, and significantly higher severity scores (Child-Pugh, MELD, MELD-Na, and GBS) were all associated with in-hospital mortality. These findings are consistent with the logistic regression results and further emphasize the importance of early identification and aggressive management of high-risk patients with poor liver and kidney function. Recognizing these factors at admission may assist clinicians in prioritizing care and improving outcomes in patients with EVB.
We are aware of several limitations in our study. First, as a single-centre retrospective study, selection bias and incomplete treatment details may have influenced our findings. Although our findings revealed numerically higher rates of 5-day treatment failure, in-hospital mortality, and ICU admissions in the early endoscopy group, these differences did not reach statistical significance. However, this lack of significance may be partly attributed to the limited sample size, which might reduce the statistical power to detect clinically meaningful differences. Larger prospective or multi-centre studies are needed to provide a more comprehensive evaluation of endoscopy timing in cirrhotic patients with variceal bleeding. While a randomized controlled trial would offer the highest level of evidence, ethical concerns make it challenging to conduct, as delaying endoscopy in high-risk patients could compromise their outcomes. Additionally, we excluded patients with gastric variceal bleeding due to the unavailability of histoacryl injections at our hospital during the study period. This may affect the generalizability of our results compared to studies that included such cases. Finally, due to limitations in electronic medical records, our analysis was restricted to short-term outcomes in the two endoscopy groups. Long-term follow-up studies are essential for a more thorough assessment of treatment efficacy and a better comparison of outcomes across different endoscopy timing strategies.
Conclusion
The timing of endoscopy (before or after 12 h) did not significantly impact clinical outcomes in patients with variceal bleeding. Therefore, the decision on endoscopy timing should be guided by the patient’s condition and the clinical judgment of the physician. Independent risk factors for in-hospital mortality included low systolic blood pressure, elevated serum creatinine, and Child-Pugh type C at admission.
Acknowledgements
Not applicable.
Abbreviations
- GI
Gastrointestinal
- EVB
Esophageal variceal bleeding
- EVL
Endoscopic variceal ligation
- AASLD
Association for the Study of Liver Diseases
- ESGE
European Society of Gastrointestinal Endoscopy
- UK
United Kingdom
- ICU
Intensive care unit
- IQR
Interquartile range
- OR
Odds ratio
- PSM
Propensity Score Matching
- HBV
Hepatitis B virus
- HCV
Hepatitis C virus
- AST
Aspartate aminotransferase
- ALT
Alanine transaminase
- INR
International normalised ratio
- PT
Prothrombin time
- MELD
Model for end-stage liver disease
- GBS
Glasgow-Blatchford score
- HRS
Hepatorenal syndrome
Author contributions
K.P.N. and T.H.D. were involved in the study concept and design. K.P.N. and X.V.N. were involved in data curation. K.P.N. was involved in statistical analysis and wrote the main manuscript text. T.H.D. was involved in the supervision of the study. All authors reviewed the manuscript.
Funding
None.
Data availability
The datasets used and/or analysed during this research are available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
This retrospective study was conducted in accordance with the principles of the Declaration of Helsinki. The study protocol was reviewed and approved by the Da Nang Hospital’s Internal Medicine Council. Given the retrospective nature of the study and the use of anonymized patient data, the requirement for informed consent was waived by the Ethics Committee of Da Nang Hospital.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Stanley AJ, Laine L. Management of acute upper Gastrointestinal bleeding. BMJ. 2019;364:l536. [DOI] [PubMed] [Google Scholar]
- 2.Qi X, et al. Timing of endoscopy in cirrhotic patients with acute variceal bleeding: protocol of a multicenter randomized controlled trial. Th Adv Gastroenterol. 2024;17:17562848241295452. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Garcia-Tsao G, Bosch J. Management of varices and variceal hemorrhage in cirrhosis. N Engl J Med. 2010;362(9):823–32. [DOI] [PubMed] [Google Scholar]
- 4.de Franchis R. Expanding consensus in portal hypertension: report of the Baveno VI consensus workshop: stratifying risk and individualizing care for portal hypertension. J Hepatol. 2015;63(3):743–52. [DOI] [PubMed] [Google Scholar]
- 5.Garcia-Tsao G, et al. Portal hypertensive bleeding in cirrhosis: risk stratification, diagnosis, and management: 2016 practice guidance by the American association for the study of liver diseases. Hepatology. 2017;65(1):310–35. [DOI] [PubMed] [Google Scholar]
- 6.Gralnek IM, et al. Endoscopic diagnosis and management of esophagogastric variceal hemorrhage: European society of Gastrointestinal endoscopy (ESGE) guideline. Endoscopy. 2022;54(11):1094–120. [DOI] [PubMed] [Google Scholar]
- 7.de Franchis R, et al. Baveno VII - Renewing consensus in portal hypertension. J Hepatol. 2022;76(4):959–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Colle I, et al. Upper Gastrointestinal tract bleeding management: Belgian guidelines for adults and children. Acta Gastroenterol Belg. 2011;74(1):45–66. [PubMed] [Google Scholar]
- 9.Tripathi D, et al. U.K. Guidelines on the management of variceal haemorrhage in cirrhotic patients. Gut. 2015;64(11):1680–704. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.[Guidelines on the management of esophagogastric variceal bleeding in cirrhotic portal hypertension]. Zhonghua Gan Zang Bing Za Zhi, 2022. 30(10): pp. 1029–1043. [DOI] [PMC free article] [PubMed]
- 11.Peng M, et al. Timing of endoscopy in patients with cirrhosis and acute variceal bleeding: a single-center retrospective study. BMC Gastroenterol. 2023;23(1):219. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Zhang W et al. Timing of endoscopy for acute variceal bleeding in patients with cirrhosis (CHESS1905): A nationwide cohort study. Hepatol Commun, 2023. 7(5). [DOI] [PMC free article] [PubMed]
- 13.Yoo JJ, et al. Timing of upper Gastrointestinal endoscopy does not influence short-term outcomes in patients with acute variceal bleeding. World J Gastroenterol. 2018;24(44):5025–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hsu YC, et al. Delayed endoscopy as a risk factor for in-hospital mortality in cirrhotic patients with acute variceal hemorrhage. J Gastroenterol Hepatol. 2009;24(7):1294–9. [DOI] [PubMed] [Google Scholar]
- 15.Jung DH, et al. Optimal endoscopy timing in patients with acute variceal bleeding: A systematic review and meta-analysis. Sci Rep. 2020;10(1):4046. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Bai Z, et al. Outcomes of early versus delayed endoscopy in cirrhotic patients with acute variceal bleeding: a systematic review with meta-analysis. Eur J Gastroenterol Hepatol. 2021;33(1S):e868–76. [DOI] [PubMed] [Google Scholar]
- 17.Singal AK, Kamath PS. Model for End-stage liver disease. J Clin Exp Hepatol. 2013;3(1):50–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kim WR, et al. Hyponatremia and mortality among patients on the liver-transplant waiting list. N Engl J Med. 2008;359(10):1018–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wu K, et al. Analysis of the timing of endoscopic treatment for esophagogastric variceal bleeding in cirrhosis. Front Med (Lausanne). 2022;9:1036491. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Huh CW, et al. Optimal endoscopy timing according to the severity of underlying liver disease in patients with acute variceal bleeding. Dig Liver Dis. 2019;51(7):993–8. [DOI] [PubMed] [Google Scholar]
- 21.Chen PH, et al. Delayed endoscopy increases re-bleeding and mortality in patients with hematemesis and active esophageal variceal bleeding: a cohort study. J Hepatol. 2012;57(6):1207–13. [DOI] [PubMed] [Google Scholar]
- 22.Esmat G, Hamza IM, Abbas BE. Outcome of patients undergoing early versus late endoscopic intervention for acute oesophageal variceal bleeding at Air Force Hospital, Egypt. A population based cohort study. Open Journal of Gastroenterology, 2014. 2014.
- 23.Baradarian R, et al. Early intensive resuscitation of patients with upper Gastrointestinal bleeding decreases mortality. Am J Gastroenterol. 2004;99(4):619–22. [DOI] [PubMed] [Google Scholar]
- 24.Merola E, Michielan A, de Pretis G. Optimal timing of endoscopy for acute upper Gastrointestinal bleeding: a systematic review and meta-analysis. Intern Emerg Med. 2021;16(5):1331–40. [DOI] [PubMed] [Google Scholar]
- 25.Augustin S, et al. Expanding the Baveno VI criteria for the screening of varices in patients with compensated advanced chronic liver disease. Hepatology. 2017;66(6):1980–8. [DOI] [PubMed] [Google Scholar]
- 26.Chirapongsathorn S, Akkarachinores K, Chaiprasert A. Development and validation of prognostic model to predict mortality among cirrhotic patients with acute variceal bleeding: A retrospective study. JGH Open. 2021;5(6):658–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Kaewput W, et al. In-hospital mortality of hepatorenal syndrome in the united States: nationwide inpatient sample. World J Gastroenterol. 2021;27(45):7831–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Simonetto DA, Gines P, Kamath PS. Hepatorenal syndrome: pathophysiology, diagnosis, and management. BMJ. 2020;370:m2687. [DOI] [PubMed] [Google Scholar]
- 29.Mandal AK, et al. Factors predicting mortality of acute variceal bleeding in liver cirrhosis. JNMA J Nepal Med Assoc. 2018;56(209):493–6. [PMC free article] [PubMed] [Google Scholar]
- 30.Altamirano J, et al. Predicting 6-week mortality after acute variceal bleeding: role of classification and regression tree analysis. Ann Hepatol. 2009;8(4):308–15. [PubMed] [Google Scholar]
- 31.Martín-Flores UI, Aguayo-Villaseñor JF, Contreras-Marín ME. 6-Week mortality predictors in patients with acute variceal bleeding from the Western National medical center of the Mexican social security Institute. Ann Hepatol, 2022. 27.
- 32.Ricardo Flores-Rendón Á, et al. Model for end stage of liver disease (MELD) is better than the Child-Pugh score for predicting in-hospital mortality related to esophageal variceal bleeding. Ann Hepatol. 2008;7(3):230–4. [PubMed] [Google Scholar]
- 33.Buckholz A et al. MELD, MELD 3.0, versus child score to predict mortality after acute variceal hemorrhage: A multicenter US cohort. Hepatol Commun, 2023. 7(10). [DOI] [PMC free article] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets used and/or analysed during this research are available from the corresponding author upon reasonable request.