Octreotide for acute gastrointestinal bleeding secondary to portal hypertension in pediatric patients : experience of a tertiary center

Octreotide for acute gastrointestinal bleeding secondary to portal hypertension in pediatric patients: experience of a tertiary center Uso de octreotida na hemorragia digestiva alta secundária à hipertensão portal em pacientes pediátricos: experiência de um serviço terciário Uso de octreotide en la hemorragia digestiva alta secundaria a hipertensión portal en pacientes pediátricos: experiencia de un servicio terciario


Introduction
Extrahepatic portal vein obstruction due to thrombosis is the most common cause of non-cirrhotic portal hypertension and biliary atresia is the more prevalent cause of cirrhosis (1) .
Gastrointestinal bleeding from esophageal varices is the main cause of morbidity in patients with portal hypertension, bleeding recurrence is very high and mortality varies from 30 to 50% (2) .During acute bleeding, in addition to measures of hemodynamic support, endoscopic treatment (sclerotherapy, band ligation, tissue adhesives such cyanoacrylate) and specific drugs such as octreotide and somatostatin are used for controlling bleeding.
Octreotide is a synthetic somatostatin that has been shown to reduce splanchnic blood flow in healthy volunteers and hepatic venous pressure in cirrhotic patients (3,4) .Thus, it has been used for variceal bleeding.This drug has a marked clinical importance for patients awaiting endoscopy for esophageal varices diagnosis and treatment, for patients presenting voluminous bleedings that impair visibility during the endoscopy, and for patients under increased rebleeding risk and who require the drug hypotensive effect for some days (5) .Few studies have focused the use of octreotide in pediatric patients.
This study aimed to describe clinical data in children and adolescents with portal hypertension, with and without cirrhosis, who were treated with octreotide during acute gastrointestinal bleeding, and to evaluate eventual differences between the use or not of a bolus dose.

Method
Twenty patients were seen between March 1998 and December 2006 at the University Hospital in Faculty of Medical Sciences, Unicamp.Patients aged 7 months up to 18 years.All had portal hypertension diagnosis based on ultrasonographic criteria, presented upper gastrointestinal bleeding and received octreotide to control acute bleeding episodes.Three patients were excluded because their records were incomplete.
A retrospective and descriptive study was conducted.Patients were classified as cirrhotic or non-cirrhotic.The diagnosis of cirrhosis was based on histological data.Ultrasound was performed using a Toshiba Power Vision 6000 apparatus equipped with 3.75-MHz sectorial and 5-MHz linear transducers.The following parameters were evaluated in order to diagnose portal hypertension: splenomegaly, gallbladder wall thickness, small omentum thickness close to the venous ligament, small omentum/aorta ratio, and presence of splenorenal shunt (6)(7)(8)(9) .
The Child-Pugh score (10) was used to determine the severity of liver disease in cirrhotic patients.The following parameters were evaluated: serum albumin and bilirubin levels, prothrombin time, presence of ascites and encephalopathy.According to the routine of the service, the presence of coagulopathy and thrombocytopenia was investigated by INR measurement and platelet count.Vitamin K, fresh frozen plasma and platelets were replaced as needed.Requirement of blood transfusion was recorded.
The origin of bleeding identified by endoscopy was registered.If In the source could not be identified, the site of bleeding was classified as indeterminate.Associated endoscopic treatment was also recorded.Endoscopy was performed with a Pentax or Olympus videoendoscope.Sclerotherapy or band ligation was indicated in patients with active bleeding of esophageal varices and/or bleeding of medium and large size esophageal varices even in the absence of active bleeding.Cyanoacrylate was administered in some cases of bleeding from gastric varices.
The variables related to octreotide treatment were evaluated: dose, duration and side effects.Patients received a continuous intravenous infusion of 1-2µg/kg/h after an intravenous bolus dose (1µg/kg).The bolus administration was not universal.The criteria for the indication and duration of treatment were individually determined.The hospital stay in order to control bleeding was also recorded.
Rebleeding and treatment failure were defined as a new episode of bleeding, accompanied by a drop in hemoglobin levels during treatment or up to 15 days after the end of the last octreotide infusion.Control of glycemia, blood pressure and symptoms were performed in order to detect possible adverse effects during octreotide infusion.
Descriptive statistical analysis was performed, including measures of position and dispersion for continuous variables and frequency tables for categorical variables.Fisher's exact test was used to compare proportions.Continuous or ordinal variables were compared by Mann-Whitney test.A level of significance of 5% was adopted.All analyses were performed using the SPSS for Windows program, version 7.5.This study was approved by the Ethical Committee of the University.

Results
Twenty-six episodes of upper gastrointestinal bleeding due to portal hypertension occurred in 17 patients between March 1998 and December 2006.Patients age ranged from 0.56 to 18.9 years (mean: 8.64 and median: 7.91 years).The characteristics of the 26 upper gastrointestinal bleeding episodes are showed in Table 1.
Among the 17 patients, 8 (47.1%) were males; 6 (35.3%) had portal hypertension associated with cirrhosis, and 11 (64.7%)s presented obstruction of the extrahepatic portal vein with normal liver function (associated with congenital hepatic fibrosis in one patient and with hepatoportal sclerosis in another).
Among the 26 bleeding episodes, six occurred were observed in cirrhotic patients and 20 in non-cirrhotic patients.Propranolol was used on the occasion of bleeding as primary prophylaxis in 4/26 (15.4 %) episodes and as secondary prophylaxis in 7/26 (26.9%).
Patients undergone endoscopy in all bleeding episodes.Octreotide was already being used by the patients on the occasion of the exam in 13/26 (50%) episodes.Only 5/26 (19.2%) episodes were characterized as active bleeding at the time of endoscopy, with octreotide already being used in 2 episodes.The site of bleeding was identified in 16/26 (61.5%) episodes: esophageal varices in 9, gastric varices in 3 and hypertensive gastropathy in 4. Endoscopic treatment was performed during 15/26 (57.7%) episodes: sclerotherapy (n=6) and band ligation (n=6) for esophageal varices and A bolus dose of octreotide (ranged from 0.5 to 2.63µg/kg; mean: 1.19; median: 1µg/kg, P25/P75: 0.88/1.59)was administered during 20 of the 26 episodes.In the other cases, only a maintenance dose was infused.Some patients' weight was similar to adults weight so they received the maximum dose established by our protocol, conversely, in some cases the dose was lower than 1mcg/kg.The maintenance dose ranged from 0.44 to 1.43µg/kg/h (mean: 0.89, median: 0.92µg/kg/h, P25/P75: 0.73/1.00).The mean duration of treatment was 3.1 days (median: 3 days, P25/P75: 2.0/4.0) and the median of hospitalization was 6 days (P25/P75: 4.7/9.0).All patients received a proton pump inhibitor or H2 inhibitor during bleeding.Vitamin K replacement therapy was indicated in 16/26 (61.5%) episodes.
Hyperglycemia was the only side effect observed during 5 (22.7%) of 22 episodes.In all cases, hyperglycemia was reversed by gradual reduction of the octreotide dose without the need for insulin.In this study, 4 episodes that octreotide was administered simultaneously with glucose solution were excluded.The maximum level of blood sugar achieved was 169mg/dL.
Only 3/17 (17.6%) patients presented rebleeding.As shown in Table 2, one patient was not submitted to endoscopic treatment (patient 1).In this patient, rebleeding occurred from the esophageal varix on day 4 of treatment and was controlled by band ligation.The other two patients had undergone band ligation.One patient with Child C cirrhosis (patient number 2) maintained bleeding throughout the treatment and died.The other patient (patient number 3) presented bleeding on day 8 after the end of infusion.No significant differences in the occurrence and the determination of site of active bleeding were observed between patients who received octreotide and those who did not receive the medication before the endoscopy.In the patients who received octreotide the active bleeding was present in 2 and the site of bleeding was determined in 9.In patients that did not receive octreotide before endoscopy, the active bleeding was present in 3 and the site of bleeding was determined in 7.
Comparison of patients who received bolus and maintenance doses of octreotide and those who received only the maintenance dose showed no significant differences regarding mean transfusion volume, mean drop in hemoglobin levels, duration of octreotide treatment or hospital length of stay (Table 3).
When bleeding episodes were compared between cirrhotic and non-cirrhotic patients, a higher drop in hemoglobin levels, a higher transfusion volume and a longer duration of hospital stay were observed in cirrhotic patients, but this difference was not significant.The time of use of octreotide was similar in the two groups (Table 4).
Treatment failure was observed for 2/6 bleeding episodes in cirrhotic patients and for 1/20 episodes in non-cirrhotic ones (Fisher's exact test, p=0.12).Frequency of side effects (hyperglycemia) was lower in cirrhotic patients, but no significant difference was observed between groups (p=0.38)(Table 5).

Discussion
The frequency of gastrointestinal bleeding in children with portal hypertension is higher in those with extrahepatic portal vein obstruction than in cirrhotic patients.Portal vein obstruction is responsible for about 70% of all pediatric patients with portal hypertension (11) .In the present study, 64.7% of the patients had portal hypertension due to extra-hepatic portal vein obstruction and presented 20/26 (76.9%) episodes of gastrointestinal bleeding.Active bleeding was observed during the endoscopic exam in only 5/26 (38%) bleeding episodes analyzed at this study.Despite the small number of patients, no statistical differences were observed between episodes previously treated with octreotide and those untreated.This finding supports the hypothesis that bleeding ceases spontaneously in many cases.D'Amico et al (12) reported spontaneous cessation of bleeding in 40 to 50% of adult patients with upper gastrointestinal bleeding due to portal hypertension.
In children, Eroglu et al (13) noted the absence of active bleeding during endoscopic examination in 71% of episodes that occurred in patients under treatment with octreotide, but the percentage of cases in which bleeding cessation was spontaneous or secondary to the use of the medication could not be determined.Some studies show that bolus administration of octreotide to clinically stable cirrhotic patients causes systemic vasoconstriction, a transient decrease in the hepatic venous pressure gradient  and a significant reduction of cardiac output (3,5) .However, in the present study, the administration of a bolus dose did not modify significantly the mean transfusion volume, mean hemoglobin decline, duration of octreotide treatment or hospital stay.The only side effect detected during the administration of octreotide was hyperglycemia.In a meta-analysis on octreotide for acute esophageal variceal bleeding, Corley et al (14) observed a frequency of hyperglycemia ranging from 0 to 23%.In another meta-analysis, Bañares et al (15) compared endoscopic treatment with endoscopic treatment combined with pharmacological therapy for the control of acute variceal bleeding and found that hyperglycemia was the only adverse effect in the studies analyzed, with this side effect being significantly more frequent in the group of patients that underwent pharmacological treatment.However, the use of somatostatin and its derivatives (e.g., octreotide) was considered to be safe in view of the finding that none of the patients required discontinuation of therapy (15) .
Regarding the three patients who presented rebleeding, two had cirrhosis classified as Child C and the other underwent endoscopic treatment until the occurrence of rebleeding.Probably, the episodes of rebleeding were mainly related to the severity of the underlying disease and to the need of a definitive treatment rather than a medication effect.
Comparison of the bleeding episodes between cirrhotic and non-cirrhotic patients showed that the drop in hemoglobin levels, the blood transfusion volume, and hospital stay were slightly greater and in cirrhotic patients, although no statistical significance was observed, probably due to the small number of patients in groups .This finding is probably related to the r clinical severity of patients with cirrhosis who, in addition to impaired hepatic function, often present malnutrition and impaired general health.
Comparison of the present results with other studies is limited by the characteristics of the series, i.e., a pediatric population in which the main cause of upper gastrointestinal bleeding due to portal hypertension was extra-hepatic portal vein obstruction.Most studies evaluating octreotide were conducted on adults with cirrhosis and patients with gastrointestinal bleeding not related to portal hypertension (16) .Similar results in terms of the decline in hemoglobin levels, transfusion volume and duration of treatment were reported in a study of pediatric patients (13) .
The present report has limitations regarding its retrospective design and a small number of patients, however, it differs from international series due to the lower frequency of rebleeding, treatment failure and mortality.The effect of cirrhosis and hepatic insufficiency on the occurrence of rebleeding should be investigated in multicenter studies of patients with different demographic characteristics.Larger cohort studies are necessary to determine the benefits of combined or single treatment with octreotide in children, irrespective of the etiology of portal hypertension.
The use of octreotide in children and adolescents with portal hypertension with or without cirrhosis was considered safe.The drug seems to be effective in controlling acute bleeding in patients with portal hypertension, irrespective of etiology and the infusion schedule.

Table 1 -
Clinical data from 26 upper gastrointestinal bleeding episodes in 17 children and adolescents followed in a university hospital

Table 2 -
Site of bleeding, endoscopic treatment, etiology of portal hypertension, and period of rebleending in three rebleeding patients

Table 4 -
Blood volume transfused, decline in hemoglobin levels, time of use of octreotide and duration of admission in cirrhotic and non-cirrhotic patients *Mann-Whitney test.

Table 5 -
Distribution of the clinical data of bleeding episodes according to the presence or absence of liver cirrhosis

Table 3 -
Blood volume transfused, decline in hemoglobin levels, time of use of octreotide and duration of admission according to the administration or not of a bolus dose of the drug