See “Variceal Bleeding and Morbidity-Considerations for Primary Prophylaxis” by Bass on page 311.
What Is Known
- In adults, acute variceal bleeding causes significant mortality and primary prophylaxis of first acute variceal bleeding is standard of care.
- In children, mortality due to acute variceal bleeding is unclear but is presumed to be low.
- In children, current expert opinion does not recommend routine primary prophylaxis.
What Is New
- There is significant morbidity associated with acute variceal bleeding in children.
- Future studies should consider morbidity events when measuring the risks and benefits of primary prophylaxis of first acute variceal bleeding in children.
Acute variceal bleeding (AVB) is a severe complication of portal hypertension in children. There has yet to be an evidence-based resolution to the debate about provision of primary prophylaxis to prevent the first variceal bleed in children with cirrhosis (1,2). Effective primary prophylaxis is available for adults with varices and is widely applied in clinical practice to help reduce mortality from variceal bleeding, which is still in the order of 10% to 20% at 6 weeks (3). Analysis of published retrospective pediatric case series, however, reveals inadequate data to provide a confident estimate of mortality associated with the first variceal bleed, but suggests that this mortality may be <1% (4–6). If this estimate is accurate, then it will be difficult to show benefit to mortality risk by the application of primary prophylactic therapy for children with varices.
The morbidity associated with the first variceal hemorrhage in children has not previously been quantified or systematically characterized to our knowledge, although has been analyzed in adults suffering from AVB (7–10). If the risk of death from the first variceal bleed is presumed to be low, the need for primary prophylaxis of the first AVB in children would largely depend on the extent of the associated morbidity. The aim of our study was to investigate the morbidity associated with AVB in children and to identify factors that contribute to this morbidity.
We conducted a retrospective cohort analysis of all consecutive patients with AVB admitted to the Hospital for Sick Children in the 15 years between May 1, 2000 and April 30, 2015. Patients were included if younger than 18 years, presenting with melena and/or hematemesis with requirement for intravenous fluid bolus and/or blood products to restore circulating volume or maintain hemoglobin, and with endoscopic confirmation of esophageal varices. We also included patients who had endoscopy performed at a referring institution and an immediate transfer was undertaken, first and subsequent episodes of AVB and patients who had AVB while already an in-patient. We excluded children with concomitant hepatic malignancy, previous transjugular intrahepatic portosystemic shunt, previous liver transplantation, patients with bleeding as a complication of an endoscopic procedure, or unavailable medical records. Recurrent bleeding episodes within 6 weeks of the first index bleed were considered rebleeds and counted as morbidity following the index bleed, not as an additional, independent episode of bleeding. Recurrent bleeding episodes after 6 weeks were treated as a new bleeding episode. We identified subjects from the operating room database and health records discharge codes. We recorded the following morbidities: length of stay, admission to intensive care unit (ICU), rebleed, failure to control bleeding, infectious complications, respiratory complications, development or worsening ascites, development or worsening encephalopathy, and acute kidney injury (AKI). We also collected laboratory variables to permit calculation of Model for End-Stage Liver Disease (if age >12 years) and pediatric end-stage liver disease (PELD, if age < 12 years). To decide whether each morbidity was present or absent, we used all available data in the chart to reach a conclusion, including objective data when available, but also more subjective comments from the attending physicians at the time. Findings in the chart were related to the definitions provided below.
Clinical care for children with portal hypertension at our institution during the study timeline did not include routine use of primary prophylaxis or a protocol for screening endoscopy. After an admission for variceal bleeding, follow-up endoscopy was usually planned at 4-week intervals until eradication of varices was achieved.
AVB was defined as hematemesis and/or melena or gastric aspirate containing blood with requirement of intravenous fluid bolus and/or blood products to restore circulating volume or maintain hemoglobin, with endoscopy confirming esophageal varices and no other cause of upper gastrointestinal bleeding (11). The endoscopy report was used to determine the presence and grade of varices and the presence of portal gastropathy. The cause of bleeding was determined from the endoscopy report and the comments of the attending physicians expressed in the patient chart at the time of the admission. Portal hypertension was present when abdominal ultrasound, computer tomography, or magnetic resonance showed either hepatic parenchymal abnormalities consistent with cirrhosis or occlusive main portal vein thrombosis, and splenomegaly and/or portosystemic collaterals, including esophageal varices. We used the Baveno recommendations to define time zero as time of admission to the first hospital the patient attended (12); the time frame for the acute bleeding episode was 120 hours (5 days) (6). Rebleeding was defined as recurrent melena or hematemesis after day 5 and within 6 weeks of the first episode resulting in hospital admission, need for blood transfusion, 3 g/dL drop in hemoglobin, or death (from any cause) (12,13). Failure to control bleed was defined as death (from any cause) or need to change therapy during the first 5 days due to one of the following criteria: fresh hematemesis or nasogastric aspiration of >2 mL/kg or >100 mL of fresh blood ≥2 hours after the start of a specific drug treatment or therapeutic endoscopy, development of hypovolemic shock, or 3 g/dL drop of hemoglobin within any 24 hours period after the initial resuscitation (2,14). The time frame defining mortality due to AVB was 6 weeks (15) and liver transplant was also recorded if it occurred within the time frame of 6 weeks. We used the Paquet classification of esophageal varices (13). Active variceal bleeding was defined as visible oozing or spurting of blood from a varix (16). Stigmata of recent bleeding were defined as the presence of adherent clot on varix, fibrin plug or white nipple on varix, or red signs on varices (cherry red spot, red wale sign, or hematocystic spots) (16).
We included any type of morbidity that happened during the admission for AVB and within 6 weeks of the bleed. Presence of ascites was based on the definition of the International Club of Ascites: grade 1, mild ascites (only detectable by ultrasound), grade 2 (moderate symmetrical distension of abdomen), grade 3 (distension of abdomen) (17). Hepatic encephalopathy was determined using the classification of the Pediatric Acute Liver Failure Study Group for children younger than 4 years (18). For older children, the West Haven classification system was used (19). Sepsis was defined according to the International Pediatric Sepsis Consensus (20), when there was evidence of an abnormality of temperature (fever or hypothermia) or age-specific abnormality of the white blood cell count and one of the following: tachycardia, bradycardia, respiratory distress, or pulmonary condition requiring mechanical ventilation associated with at least 1 positive blood culture. Occult bacteremia was defined as fever in a child that looked well and had unsuspected, occult bacteria in the blood. Localized infection was considered if there was a known source of infection/fever. Respiratory complications were considered if there was any pulmonary abnormality occurring during admission that affected the clinical course. AKI was defined according to the Kidney Disease Improving Global Outcomes guidelines (21) by any of the following: increase in serum creatinine by ≥0.3 mg/dL (≥26.5 mol/L) within 48 hours; increase in serum creatinine by ≥1.5 times baseline, which is known or presumed to have occurred within the prior 7 days; or urine volume <0.5 mL/kg/h for 6 hours.
Statistical analysis was completed using SAS 9.4 (SAS Institute Inc., Cary, NC). Descriptive statistics such as frequency count (proportions), mean (SD), and median (range) were calculated. Both univariate analysis and multiple regression were performed. A P value <0.01 was statistically significant.
The study was undertaken with approval from the Research Ethics Board of the Hospital for Sick Children, Toronto.
A total of 70 episodes of AVB were identified among 57 patients with portal hypertension. Forty-seven (67%) episodes were the patient's first AVB and 23 (33%) episodes were subsequent bleeds. Table 1 presents the baseline characteristics of the study sample. All patients were followed at our center and only 4 were referred from other institutions.
Median time from admission to endoscopy was 35 hours (range 2–288 hours) and did not change over the 5-year eras (2000–2004, 2005–2010, 2010–2015). Nineteen (28%) endoscopies were performed within 24 hours, 26 (38%) within 24 to 48 hours, and 23 (34%) after 48 hours from admission. Only a minority (3%) had active variceal bleeding on endoscopy. Endoscopic therapy was performed in 55 episodes (79%): Endoscopic band ligation in 38 (69%) and ES in 17 (31%).
Proton pump inhibitors (PPIs) or ranitidine were used in 97% of AVB episodes and octreotide in 89%. Median time to start octreotide was 4 hours (range 1–96 hours). The median dose of octreotide was 2 μg · kg−1 · h−1. Median duration of octreotide infusion was 4 days (range 1–33 days). The use of octreotide became more frequent in the most recent era (93% in 2010–2015 compared to 86% in 2000–2004, 2005–2009).
Two patients died (3% of all patients) within 6 weeks of AVB. These 2 patients had cirrhotic liver disease (1 post-Fontan, 1 idiopathic cirrhosis) and died after a first variceal bleed, 1 from failure to control bleed after banding and the other from sepsis and failure to control bleed after sclerotherapy. Therefore, the mortality associated with first AVB in cirrhotic liver disease was 8%. One patient with cirrhosis due to biliary atresia (post Kasai portoenterostomy) underwent liver transplantation within 6 weeks of the first AVB episode.
Post-AVB morbidity during admission occurred in 40 (57%) of all episodes and in 30 (64%) of first AVB (Fig. 1). Of the 70 episodes of bleeding, imaging data for the presence or absence of ascites were available in 53 episodes a median (range) of 61 days (0–512 days) before the bleeding episode. In only 9 of these episodes were the ultrasound scan performed more than 6 months before the bleeding episode. Among the remaining 17 episodes without prior ultrasound evaluations, 9 occurred in patients with portal vein thrombosis or Caroli disease in whom ascites would not be expected. Eight episodes in patients with cirrhosis had no prior ultrasound evaluation available, and in 4 of these episodes the patient had documented clinically worsening ascites on serial clinical evaluation and imaging after presentation, enabling classification of ascites as a morbidity associated with the bleeding episode. Ascites was already documented before AVB in 8 (11%) episodes with new onset ascites developing in 16 (23%) and worsening ascites in 8 (11%). Ascites was grade 2 in 63% and grade 3 in 38%. All episodes of ascites occurring or worsening after the variceal bleeding episode were treated with diuretics, apart from 1 episode in a child with congenital hepatic fibrosis in whom the grade 3 ascites improved spontaneously (Table 2).
Despite the use of antibiotics within 24 hours in 38 (54%) episodes, infection occurred in a significant percentage. Nine patients had bacteremia/occult bacteremia, 3 had sepsis, and 9 had localized infections (1 aspiration pneumonia, 1 cholangitis, 1 liver abscess, 2 respiratory infections, 2 viral enteritis, 2 urinary infections). The incidence of infection did not change between 5-year eras, although the use of prophylactic antibiotics was more frequent in the last 2 eras (86% in 2005–2009, 68% in 2010–2015) compared to 2000 to 2004 (25%). Median time to diagnosis of infection after AVB was 3 days (range 0–30 days). Fourteen (68%) patients with documented infection had received antibiotics within 24 hours from admission and in 12 a cephalosporin was part of the antibiotic regimen. In 8 episodes the antibiotic regimen had to be changed due to the nature of the infection and/or bacterial sensitivities.
Fourteen (82%) of all patients with respiratory complications needed oxygen or noninvasive ventilation support and 6 (35%) required invasive ventilation. ICU admissions occurred within a median time of 4 hours after presentation (range 0–192 hours). Median length of stay in ICU was 48 hours (7–192 hours). 86% of admissions to ICU happened in first bleeds. Rebleeding occurred in 11% at a median of 16 days (range 7–38 days) from the index episode of AVB. Patients with rebleed received a higher dose of octreotide or endoscopic therapy to control bleeding. The 3 episodes of failure to control bleed happened during the patient's first bleeding episodes. These patients underwent second endoscopy but 2 of these patients died and in 1 patient the bleed was managed with a second therapeutic endoscopy and octreotide. Encephalopathy (7%) and AKI (6%) only occurred during patients’ first bleed. All cases of encephalopathy were new onset (2 patients had grade 1, and 1 case each had grade 2, 3, and 4).
By univariate analysis, total bilirubin was predictive of the outcome “any morbidity” (Table 3). Overall morbidity rates were similar in cirrhotic and noncirrhotic portal hypertensive children and did not change during 15 years. Univariate analysis showed that cirrhotic liver disease, PELD, and total bilirubin were significantly associated with new onset or worsening ascites after AVB (Table 3). By multiple regression, PELD was associated with onset or worsening ascites (Table 3). For all other morbidities, we did not identify any statistically significant predictive factors.
Median length of stay was 7 days (quartiles 2–5, 6–7, 8–22, 23–220 days). PELD, total bilirubin, use of prophylactic antibiotics, presence of any morbidity, and number of morbidities were significantly associated with length of stay by univariate analysis with log transformed outcome (Table 4). Median length of stay of patients with “any morbidity” was longer (18 days, range 3–220) than in those with no morbidity (4 days, range 2–30, P < 0.0001). Median length of stay of patients with infection during admission was longer (19 days, range 3–220) than those with no infections (7 days, range 2–209, P < 0.0001).
Our study shows that there is a high incidence of morbidity after AVB in children. We believe that this is the first detailed characterization of the morbidity associated with AVB in children, because previous studies have mainly focused on the complications of endoscopic therapy of variceal bleeding (5,6,22). Significant morbidity occurred in 57% of our cohort, and was associated with extended length of stay in hospital, similar to adult reports (9,12,23). Morbidity was predicted by total bilirubin in univariate analysis (P = 0.001) but by no other clinical characteristics at the time of the bleed. Ascites after AVB was more common in cirrhotic liver disease with higher PELD score and total bilirubin.
Infections were identified in 21 (30%) episodes after AVB, 14 in cirrhotic liver disease. Children who received antibiotics from admission had a high incidence of infections and other morbidity, perhaps representing an association of antibiotic usage in sicker patients (eg, there was a nonsignificant trend [P = 0.04] toward a higher PELD in children developing infection). Those who received antibiotics were predominantly patients with cirrhosis (20 [53%]). Infections may occur as a contributory cause of variceal bleeding, as a consequence of bleeding and its treatment, or as a coincidental finding. It is challenging to determine in retrospect which of these options may be relevant to individual cases. We are not aware of the previously published pediatric data describing the infection rate in children admitted with AVB, but studies in adults have shown a high incidence of infection among patients with cirrhosis (8,9,23,24). Data from studies of adults show a benefit of routine antibiotic use in variceal bleeding in patients with cirrhosis, although this has not been studied to our knowledge in adults with portal vein thrombosis (13).
In our cohort, the rebleeding rate (11%, 13% in first bleeds) was lower than that reported in 2 previous pediatric studies (52%, 33%) (5,6). The definition of rebleed differs in these 2 studies. Eroglu et al (6) reported 21 children with portal hypertension with a rebleeding rate of 52%, at a median of 1 month (range: 1 day–8 months). Sokal et al described 19 children with variceal bleeding with a “short-term” rebleeding rate of 33%. Our results are comparable to a series from India in which rebleeding occurred in 13% among 263 children with cirrhosis and portal vein thrombosis (PVT) (25), although no clear definition of rebleeding was provided.
The mortality that we report (3% overall, 8% in patients with cirrhotic liver disease and first bleeds) was lower than described in previous published series. Eroglu et al (6) reported a mortality rate of 19% within 35 days of variceal bleeding episodes among North American children with liver disease of various etiologies. Two other retrospective studies describe mortality of 5% and 15% in children with biliary atresia and variceal hemorrhage (26,27). Poddar et al (25) reported mortality of 1.7% among children with AVB and PVT and 30% in children with cirrhotic liver disease. These reports are between 8 and 27 years old and mortality may have since improved thanks to improvements in the care of children with AVB.
There is a lack of evidence-based guidelines for the management of AVB in children. Our data show a significant variation in management for the children attending our institution, including numbers of patients who did not receive octreotide (11%), endoscopic therapy (21%), and antibiotics (46%). This may be partly explained by individual physician's clinical decision making in the absence of clear evidence, and changing opinions on approaches to therapy during the extended time period studied. For example, data from clinical studies and publications presenting expert opinions have not addressed whether prophylactic antibiotics should be used routinely in children with AVB, or the appropriate endoscopic therapy for small esophageal varices found after upper gastrointestinal hemorrhage. Thus, we believe there is substantial opportunity to improve the care of children with AVB through quality improvement initiatives, although much clinical research is required to provide the necessary evidence.
Our study was limited by its retrospective design and relatively small number of patients with cirrhotic liver disease and PVT. We included first bleeds and subsequent bleeds, which may be expected to differ in their morbidity and thus we have presented results in 2 ways, 1 including all bleeding events and the other for first bleeds only. Our definition of AVB was based on previous consensus (11) and on our clinical experience that a significant variceal bleed almost always requires volume expansion, but we acknowledge the difficulty of finding a definition that is highly sensitive and specific for variceal bleeding when applied to retrospective research. Furthermore, as is often necessary with retrospective clinical research, we used all available data in the patients’ charts to decide on the presence or absence of each morbidity based on the definitions provided, including objective data when available, and subjective opinions expressed by the attending physicians at the time of the admission.
In conclusion, our results demonstrate a high incidence of morbidity among children with AVB. These findings fill a gap in the literature regarding morbidity in children with AVB and add to the few available data regarding mortality risk. Currently, there is insufficient evidence to recommend the routine use of primary prophylaxis in children. We believe that future studies of the efficacy and safety of primary prophylaxis of AVB in children should consider the occurrence of morbidity as an important outcome.
1. Ling SC. Advances in the evaluation and management of children with portal hypertension. Semin Liver Dis
2. Shneider BL, Bosch J, De Franchis R. Portal hypertension in children: expert pediatric opinion on the report of the Baveno v consensus workshop on methodology of diagnosis and therapy in portal hypertension. Pediatr Transplant
3. De Franchis R. Baveno VI Faculty. Expanding consensus in portal hypertension: report of the Baveno VI Consensus Workshop: stratifying risk and individualizing care for portal hypertension. J Hepatol
4. Shneider BL, De Ville de Goyet J, Leung DH, et al. Primary prophylaxis
of variceal bleeding in children and the role of MesoRex bypass: summary of the Baveno VI Pediatric Satellite Symposium. Hepatology
5. Sokal EM, Van Hoorebeeck N, Van Obbergh L, et al. Upper gastro-intestinal tract bleeding in cirrhotic children candidates for liver transplantation. Eur J Pediatr
6. Eroglu Y, Emerick KM, Whitingon PF, et al. Octreotide therapy for control of acute gastrointestinal bleeding in children. J Pediatr Gastroenterol Nutr
7. Del Olmo JA, Peña A, Serra MA, et al. Predictors of morbidity and mortality after the first episode of upper gastrointestinal bleeding in liver cirrhosis. J Hepatol
8. Thomopoulos K, Theocharis G, Mimidis K, et al. Improved survival of patients presenting with acute variceal bleeding. Prognostic indicators of short- and long-term mortality. Dig Liver Dis
9. Chalasani N, Kahi C, Francois F, et al. Improved patient survival after acute variceal bleeding: a multicenter, cohort study. Am J Gastroenterol
10. Sorbi D, Gostout CJ, Peura D, et al. An assessment of the management of acute bleeding varices: a multicenter prospective member-based study. Am J Gastroenterol
11. De Franchis R. Developing consensus in portal hypertension. J Hepatol
12. Jairath V, Rehal S, Logan R, et al. Acute variceal haemorrhage in the United Kingdom: patient characteristics, management and outcomes in a nationwide audit. Dig Liver Dis
13. Garcia-Tsao G, Sanyal AJ, Grace ND, et al. Prevention and management of gastroesophageal varices and variceal hemorrhage in cirrhosis. Hepatology
14. De Franchis R. Revising consensus in portal hypertension: report of the Baveno V consensus workshop on methodology of diagnosis and therapy in portal hypertension. J Hepatol
15. Jalan R, Hayes PC. UK guidelines on the management of variceal haemorrhage in cirrhotic patients. British Society of Gastroenterology. Gut
16. Bambha K, Kim WR, Pedersen R, et al. Predictors of early re-bleeding and mortality after acute variceal haemorrhage in patients with cirrhosis. Gut
17. Moore KP, Wong F, Gines P, et al. The management of ascites in cirrhosis: report on the consensus conference of The International Ascites Club. Hepatology
18. Squires RH Jr, Shneider BL, Bucuvalas J, et al. Acute liver failure in children: the first 348 patients in the pediatric acute liver failure study group. J Pediatr
19. Ferenci P, Lockwood A, Mullen K, et al. Hepatic encephalopathy—definition, nomenclature, diagnosis, and quantification: final report of the working party at the 11th World Congresses of Gastroenterology, Vienna, 1998. Hepatology
20. Goldstein B, Giroir B, Randolph A. International pediatric sepsis consensus conference: definitions for sepsis and organ dysfunction in pediatrics. Pediatr Crit Care Med
21. Pan HC, Chien YS, Jenq CC, et al. Acute kidney injury classification for critically ill cirrhotic patients: a comparison of the KDIGO, AKIN, and RIFLE classifications. Sci Rep
22. Howard ER, Stringer MD, Mowat AP. Assessment of injection sclerotherapy in the management of 152 children with oesophageal varices. Br J Surg
23. Wilbur K, Sidhu K. Antimicrobial therapy in patients with acute variceal hemorrhage. Can J Gastroenterol
24. Tandon P, Abraldes JG, Keough A, et al. Risk of bacterial infection in patients with cirrhosis and acute variceal hemorrhage, based on child-pugh class, and effects of antibiotics. Clin Gastroenterol Hepatol
25. Poddar U, Thapa BR, Rao KL, et al. Etiological spectrum of esophageal varices due to portal hypertension in Indian children: is it different from the West? J Gastroenterol Hepatol
26. Stringer MD, Howard ER, Mowat AP. Endoscopic sclerotherapy in the management of esophageal varices in 61 children with biliary atresia. J Pediatr Surg
27. Van Heurn LW, Saing H, Tam PK. Portoenterostomy for biliary atresia: Long-term survival and prognosis after esophageal variceal bleeding. J Pediatr Surg