Efficacy of endoscopic isotonic saline-epinephrine injection for the management of active Mallory-Weiss tears

Therapeutic endoscopy with isotonic saline-epinephrine (ISE) injection is a convenient and widely used procedure for hemostasis in upper gastrointestinal bleeding. We retrospectively evaluated 36 patients (from January 1996 to April 1999) who had been diagnosed with recent or active bleeding due to Mallory-Weiss tears in emergency endoscopic examination. The endoscopic hemostatic method with ISE injection was performed in 15 of 36 patients. The other 21 patients received conservative treatment with hemodynamic support. Patient's clinical data, laboratory data, transfusion requirements, endoscopic findings, and length of hospital stays were evaluated. Initial hemoglobin was significantly lower in the ISE group than the conservative treatment group (9.74 +/- 2.86 g/dL vs. 12.57 +/- 2.80 g/dL, respectively; p < 0.01). Mean transfusion requirements were significantly higher in the ISE group than the conservative treatment group (7.26 +/- 8.78 units vs. 2.85 +/- 6.21 units, respectively; p < 0.1). Patients in the ISE group were supposed to be having a more severe bleeding episode. Most patients achieved initial hemostasis in the ISE group and the conservative treatment group (93% and 95%, respectively). The rebleeding rate was also similar in both groups (1 in 15 in the ISE group and I in 21 in the conservative treatment group). There was no significant difference in length of hospital stay and rebleeding between these two groups (3.47 +/- 1.92 days vs. 2.47 +/- 1.47 days, respectively: p = 0.89). The endoscopic ISE injection is an inexpensive, simple, convenient therapeutic method and it can achieve initial hemostasis for active Mallory-Weiss tears.

Successful treatment with a combination of endoscopic injection and irrigation with coca cola for gastric bezoar-induced gastric outlet obstruction

We report a case of gastric bezoar-induced gastric outlet obstruction that was successfully treated with a combination of endoscopic injection and irrigation with Coca Cola. A 73-year-old diabetic woman had a history of perforated peptic ulcer and had received pyloroplasty more than 20 years previously. She had been ingesting Pho Pu Zi (Cordia dichotoma Forst. f.) as an appetizer for 1 month. She presented with epigastric pain, nausea, and vomiting. Upper gastrointestinal endoscopy, performed at a local hospital, showed 2 gastric bezoars in the stomach, and 1 of them impacted at the pylorus. She was referred to our emergency department for removal of the gastric bezoars that were suspected to be causing gastric outlet obstruction. All attempts at endoscopic removal using a polypectomy snare, biopsy forceps and Dormia basket failed. We then injected Coca Cola directly into the bezoar mass, followed by irrigation with Coca Cola. Follow-up endoscopy was performed the next day, which revealed that the gastric bezoars had dissolved spontaneously.

Role of Helicobacter pylori in cirrhotic patients with dyspepsia: a 13C-urea breath test study

The role of Helicobacter pylori in dyspeptic, cirrhotic patients remains unclear. This prospective outpatient study, conducted to assess the relationship of gastroduodenal disease and H. pylori as determined by the (13C) urea breath test, enrolled 109 consecutive cirrhotic patients with dyspepsia. All patients underwent upper-gastrointestinal endoscopy, which revealed respective prevalences of peptic ulcer, gastric ulcer, and duodenal ulcer of 41.3%, 23.9%, and 22.9%; H. pylori infection was found in 52.3%. The rate of peptic ulcer disease in the H. pylori-positive (45.6%) and -negative (36.5%) groups was not significantly different; neither was the prevalence of H. pylori in patients with or without portal hypertensive gastropathy and with or without esophageal varices. The relationship between peptic ulcer disease and H. pylori in dyspeptic patients with cirrhosis appears to be weak. Likewise, no significant relationship was evident between H. pylori and portal hypertensive gastropathy or esophageal varices. This organism may not be a major pathogenetic factor in gastroduodenal diseases in dyspeptic patients with cirrhosis.

Ectopic papilla of Vater in duodenum bulb: A hospital-based study

The major papilla of Vater can be ectopically present in the stomach, pyloric canal, duodenal bulb, and third or fourth portion of the duodenum. In this study, we determined the clinical significance of ectopic papilla of Vater by endoscopic retrograde cholangiopancreatogram (ERCP).A retrospective study was conducted by reviewing the medical records of 6133 patients receiving ERCP from 1988 to 2011. The diagnosis was confirmed if both the common bile duct (CBD) and the main pancreatic duct (PD) drained into the same opening, either by ERCP or magnetic resonance cholangiopancreatography.Eight patients with major papilla of Vater in the duodenal bulb were identified among 6133 patients receiving ERCP from 1988 to 2011, with an incidence rate of 0.13%. The mean age was 67 years and patients were predominantly male. Duodenal bulb deformity was noted in all patients and three of them had shallow gastric and/or duodenal ulcers. Hook-shaped CBD configuration was seen only in half of our cases. Three patients with CBD stones were treated successfully after endoscopic sphincterotomy or papillary balloon dilation.Ectopic orifice of papilla is a rare finding of ERCP. Opacification of both the CBD and main PD from the same opening is an essential criterion for diagnosing an ectopic papilla of Vater in the duodenal bulb.

Mycoplasma pneumonia-associated acute hepatitis in an adult patient without lung infection

Mycoplasma pneumonia is a major cause of respiratory infections in school-aged children. Most M. pneumonia infections in adults involve the respiratory tract. Extrapulmonary manifestations of M. pneumonia infection may be found in the skin, cardiovascular, neurologic and hematologic systems. Concomitant liver disease is rare in adults. Here, we report an unusual case of a patient who presented with fever and abdominal pain, but without pulmonary manifestations. The laboratory work-up demonstrated a hepatocellular pattern of acute hepatitis caused by M. pneumonia infection. Symptoms subsided and laboratory parameters improved with antibiotics treatment. Thus, this case can help raise clinicians' awareness of the possibility of M. pneumonia infection, with or without lung involvement, as a part of the evaluation of undetermined hepatitis.

Clinical predictors of large esophagogastric varices in patients with hepatocellular carcinoma

Hepatocellular carcinoma (HCC) is one of the most common malignancies worldwide, especially in Asia. Gastrointestinal bleeding due to esophagogastric variceal hemorrhage is one of the leading causes of death in HCC patients. The aim of study was to determine whether clinical variables were predictive of the presence of large esophagogastric varices (EGV) before performing endoscopy. Three hundred and four HCC patients who received endoscopy were enrolled and studied retrospectively. Univariate and stepwise logistic regression analysis were used to evaluate associations between the presence of large EGV and patient characteristics. There were 248 patients with small or no EGV and 56 patients with large EGV. The optimal critical values determined by a receiver operating characteristic curve for platelet count and albumin level were 135,000/mm3 and 3.5 g/dl, respectively. Stepwise logistic regression analysis demonstrated that splenomegaly [odds ratio (OR): 9.72; confidence interval (CI): 3.75-25.17], portal vein thrombosis (OR: 2.73; CI: 1.50-4.97), low platelet count (<135,000/mm3) (OR: 3.78; CI: 2.07-6.90) and low albumin level (<3.5 g/dl) (OR: 3.44; CI: 1.73-6.82) were significant, independent predictors for large EGV. Large EGV also could be independently predicted by Child-Pugh classification, splenomegaly (OR: 4.93; CI: 1.87-13.01), or portal vein thrombosis (OR: 2.37; CI: 1.28-4.39) while excluding the non-cirrhotic patients. In conclusion, splenomegaly, low platelet count (<135,000/mm3), and low albumin level (<3.5 g/dl) are clinical predictors to stratify HCC patients at risk of developing large EGV. Besides factors related to liver cirrhosis, portal vein thrombosis is also an important predictor for HCC patients with large EGV.

Association Between Cholangiocarcinoma and Proton Pump Inhibitors Use: A Nested Case-Control Study

Background: The present study aimed to examine the odds of cholangiocarcinoma (CCA) in patients with proton pump inhibitors (PPIs) use. Methods: A nested case-control study design was employed using data obtained from Taiwan's National Health Insurance Research Database. In total, 2,293 patients with confirmed diagnosis of CCA were identified and served as the CCA group. The CCA patients were propensity score-matched with 2,293 subjects without CCA who served as the control group. The cumulative defined daily dose (DDD) of PPIs was calculated based on the total supply in days and quantity of individual PPIs. Univariable and multivariate logistic regression models were used to determine the odds of CCA, and calculated odds ratios (ORs) and 95% confidence intervals (CI) were used to assess PPIs use and odds of CCA. Results: The overall adjusted OR of PPIs use-associated CCA was 2.58 (95% CI 2.27, 2.93). The adjusted OR of CCA by cumulative DDD dose of PPIs and CCA was analyzed and revealed those odds of CCA are associated with all types of PPIs. Conclusions: There were odds of intrahepatic and extrahepatic CCA among PPIs users. All PPIs use was associated with odds of CCA. Analyses of larger numbers of cases are needed to confirm these findings.

Association of Endoscopic Sphincterotomy or Papillary Balloon Dilatation and Biliary Cancer: A Population-Based Cohort Study

Endoscopic sphincterotomy (EST) and endoscopic papillary balloon dilatation (EPBD) have become the main therapeutic procedures in the treatment of biliary and pancreas disease. The risk of cholangiocarcinoma (CCA) is not well investigated among post-EST/EPBD patients with benign diseases, particularly in Asia population. A retrospective nationwide cohort study using data from Taiwan's National Health Insurance Research Database (from January 1, 1998 through December 31, 2010) was conducted. Among patients with history of biliary stone with cholangitis, there were 17,503 patients in the EST/EPBD cohort and 69,998 subjects in the comparison. The incidence rate ratio was calculated using the Poisson regression model. Multivariable Cox proportional hazard models, adjusted for potential confounding factors, were used to assess the risk of developing CCA associated with endoscopic EST/EPBD. The cumulative incidences of CCA in the 2 cohorts were calculated using Kaplan-Meier analyses, and differences between the survival curves of the 2 cohorts were analyzed using a log-rank test. The overall incidence of CCA in the EST/EPBD cohort was higher than in the controls (1.36 vs 7.37 per 1000 person-years, IRR = 5.40, 95% CI = 5.15-5.67), with an adjusted HR of 4.41 (95% CI = 3.86-5.04). There were no CCA occurrences among patients receiving EST over the follow-up period 3 year after EST performed. The cumulative incidence of extrahepatic CCA seemed to be little growing in patients receiving EPBD. The cumulative incidence of intrahepatic CCA was also steady increasing in patients treated with EPBD and was more than patients receiving EST 10 years after EPBD by Kaplan-Meier analysis. In the population-based cohort study, EST is not associated with a long-term risk of intrahepatic and extrahepatic CCA. The risk of CCA for EPBD needs further investigation.

Characteristics of hepatocellular carcinoma in hemodialysis patients in hepatitis B endemic area

BACKGROUND/AIMS

Hepatocellular carcinoma is one of the most common malignancies in Taiwan, a hepatitis B endemic area. In this study, we aimed to evaluate the characteristics of hepatocellular carcinoma in patients receiving hemodialysis, as this patient group is at high risk for malignancy.

METHODOLOGY

From October 1991 to September 1997, thirteen patients receiving hemodialysis and diagnosed with hepatocellular carcinoma were enrolled in this retrospective study. Patients' clinical course, laboratory data, image study and treatment were evaluated.

RESULTS

Among these 13 patients, hepatitis B virus related in 6 and hepatitis C virus related in 7. There was no statistical significance in serum levels of asparate aminotransferase, alanine aminotransferase, bilirubin and alpha-fetoprotein between hepatitis B virus and hepatitis C virus related hepatocellular carcinoma. Hepatitis B virus related hepatocellular carcinoma had a shorter mean dialysis period (29.67 +/- 22.18 vs. 87.86 +/- 79.90 months, P = 0.25) and mean duration from beginning hemodialysis to diagnosis of hepatocellular carcinoma (17.16 +/- 26.94 vs. 76.08 +/- 65.69 months, P = 0.05), but there was no statistical significance. In the area of treatment, the survival curves of the treatment (hepatic resection and/or transcatheter arterial chemoembolization) group and supportive group were compared by log-rank test and there was no statistical significance for these two groups (P = 0.69).

CONCLUSIONS

Both hepatitis B virus and hepatitis C virus are equally important for hepatocellular carcinoma in hemodialysis patients in hepatitis B endemic Taiwan. The acquisition of hepatitis B virus or hepatitis C virus might be not related to hemodialysis. Periodic screening with ultrasonography and serum alpha-fetoprotein is necessary among hemodialysis patients with evidence of hepatitis B virus or hepatitis C virus infection.

Hemoperitoneum secondary to spontaneous rupture of metastatic epidermoid carcinoma of liver: case report and review of the literature

Massive hemoperitoneum due to spontaneous rupture of liver metastasis is remarkably rare. This is of interest because ruptured hepatocellular carcinoma is the most common cause of fatal hemoperitoneum in the Far East. We describe a case of hemoperitoneum secondary to spontaneous rupture of metastatic epidermoid carcinoma of liver. A 60-year-old man with chronic hepatitis C suffered from sudden onset of abdominal pain and pallid face with a tentative diagnosis of hepatocellular carcinoma with rupture. Tumor growth with rupture was suspected based on dynamic computed tomography, but the angiography could not demonstrate the tumor lesion. We diagnosed that the tumor was a metastatic epidermoid carcinoma, of unknown origin, based on the histopathological features. Therapy is palliative rather than curative, and the prognosis is very poor.

Comparison the efficacy of intermediate dose argon plasma coagulation versus hemoclip for upper gastrointestinal non-variceal bleeding

BACKGROUND/AIMS

Endoscopic argon plasma coagulation (APC) and hemoclip were used for the treatment of bleeding peptic ulcers. There are wide ranges of hemostatic doses (power and flow) of APC used in previous studies. The aim of our study was to assess the efficacy and safety of "intermediate dose" APC compared to hemoclips for hemostasis from bleeding peptic ulcer.

METHODOLOGY

The present study was designed as a retrospective study using historical controls. One hundred and ninety-four consecutive upper GI bleeding patients with bleeding visible vessel lesions were treated with either APC or hemoclips. There are 110 patients received APC treatment and 84 patients received hemoclip hemostasis. The main outcome measurements were one week rebleeding rate, one month rebleeding rate, surgery, morality, amount of blood transfusion and durations of hospital stay.

RESULTS

There were no significant differences between the two groups in 1 week rebleeding rate (1.8% vs. 2.4%, p = 1.0), 1 month rebleeding rate (0% vs. 1.2%, p = 0.433), mortality, surgery and amount of blood transfusion (2.67 +/- 3.27 vs. 3.04 +/- 2.75 units, p = 0.322). However, the hospital stay was longer in hemoclip group (5.38 +/- 6.76 vs. 8.49 +/- 11.19 days p = 0.011).

CONCLUSIONS

APC and hemoclip are with different hemostatic mechanisms, but the hemostatic outcomes were not significantly different between the two groups. APC is an effective, safe, and easily applicable endoscopic hemostatic modality as hemoclip for patients with non-variceal bleeding.

Factors contributing to the failure of argon plasma coagulation hemostasis in patients with nonvariceal upper gastrointestinal tract bleeding

BACKGROUND/AIMS

Among endoscopic therapies, injection methods, thermal coagulation and mechanical devices are the major therapeutic modalities for upper gastrointestinal bleeding. Argon plasma coagulation (APC), non-contact electrocoagulation, is a procedure which involves thermal coagulation of tissue. In this study, our aim was to clarify the factors associated with APC treatment failure in patients with non-variceal gastrointestinal bleeding.

METHODS

We followed up 58 non-variceal upper GI bleeding patients who received endoscopic treatment in our emergency department with APC to control bleeding from November 2006 to March 2008. Patients who received APC hemostasis were followed up to check for evidence of rebleeding by clinical symptoms and signs or drops of hemoglobulin during admission or after discharge. Rebleeding was defined as bleeding from the same lesions detected by repeat endoscopy.

RESULTS

There were no significant differences between non-rebleeding and rebleeding patients in gender, age, presentation, initial systolic blood pressure, diastolic pressure, heart rate, hemoglobulin (pre-endoscopic and post-endoscopic treatment), BUN, creatinin or albumin. Patients with rebleeding had a lower platelet count than those withoutrebleeding(253.78 +/- 90.80 vs. 135.25 +/- 69.06, p = 0.020). In addition, patients with rebleeding had more comorbid disease (24/54 (44.4%) vs. 4/4 (100%), cirrhosis (3/54 (5.6%) vs. 3/4(75%), p = 0.002) and shock (4/54 (7.4%) and 2/4(50%) than those without rebleeding. There were no significant differences between non-rebleeding and rebleeding patients in ulcer size, location of ulcer, Forrest classification, blood transfusion, mean length of hospital stay, surgery or mortality.

CONCLUSIONS

APC is an effective endoscopic therapy for non-variceal gastrointestinal bleeding. The presence of shock, low platelet count and cirrhosis may be associated with treatment failure of APC in non-variceal bleeding.

The prevalence and significance of hypoalbuminemia in non-variceal upper gastrointestinal bleeding

BACKGROUND/AIMS

Hypoalbuminemia occurs in a variety of disease states and is associated with an increased rate of complications during hospitalization, resulting in an increased length of stay. However, there are no data about hypoalbuminemia in patients with non-variceal upper gastrointestinal bleeding. The purpose of this study is to evaluate the prevalence of hypoalbuminemia in patients with non-variceal upper gastrointestinal bleeding and to examine its significance in relation to severity and outcome of bleeding.

METHODOLOGY

This is a retrospective analysis of data collected prospectively on consecutive patients admitted to the Emergency Department of Taichung Veterans General Hospital with upper GI bleeding, and variceal bleeding was excluded. Hypoalbuminemia is defined as serum albumin < 3.5 g/dL. The outcome assessments in the hypoalbuminemia and normal albumin groups were compared.

RESULTS

There were three hundred and twenty-nine patients with non-variceal upper GI bleeding identified from July 2000 to January 2001. Two hundred and fifty were male, and 79 were female. Their ages ranged from 21 to 90 (64.60 +/- 14.84) years. Of these 329 patients eligible for the study, hypoalbuminemia was seen in 204 (62.0%). When compared to patients with normal serum albumin, the hypoalbuminemia group was older (66.81 +/- 13.45 vs. 60.98 +/- 16.29 years, P < 0.01), had more associated with underlying diseases (78.4% vs. 57.6%, P < 0.01), and had more leukocytosis (47.5% vs. 35.2%, P < 0.05), had lower hemoglobin (71.1% vs. 29.6% P < 0.01), and elevated BUN (85.3% vs. 72.8%, P < 0.01) at admission. In addition, these patients had longer hospital stay (6.82 +/- 9.45 vs. 2.38 +/- 3.48 days, P < 0.01), greater requirements of blood transfusion (5.76 +/- 7.43 vs. 1.38 +/- 2.20 units, P < 0.01), need of therapeutic endoscopy (41.7% vs. 16.0%, P < 0.01), with higher rebleeding rate (13.2% vs. 0%, P < 0.01), surgery rate (5.4% us. 0.8% P < 0.05), and mortality rate (9.3% vs. 0%, P < 0.01).

CONCLUSIONS

Hypoalbuminemia is common in patients with non-variceal upper GI bleeding, appears to reflect the severity of the bleeding episode, and is associated with a more complicated course.

Is idiopathic recurrent pancreatitis attributed to small stones?

Idiopathic recurrent pancreatitis remains a clinical challenge. Intraductal ultrasonography in the management of idiopathic recurrent pancreatitis may be a new strategy for undetermined causes after initial diagnostic approaches, including endoscopic retrograde cholangio-pancreatography (ERCP). However, no definite cause after ERCP should be defined under optimal settings and with experienced technique.

Alternative percutaneous approach for endoscopic inaccessible common bile duct stones

BACKGROUND/AIMS

Endoscopic retrograde cholangiopancreatography (ERCP) with sphincterotomy followed by stone extraction is a well-established and standard procedure for common bile duct (CBD) stones. An altered anatomy may be related to inaccessible papilla. Percutaneous transhepatic approach, surgery or alternative endoscopy is considered for inaccessible papilla. Percutaneous transhepatic wire-guided retrieval balloon may not be stiff enough for pushing CBD stones in some situations, such as multiple stones or impacted stones. A percutaneous transhepatic cholangiodrainage (PTCD) assisted retrieval balloon technique for increasing the power for pushing CBD stones was designed in this study.

METHODOLOGY

A PTCD tube was placed in a situation of inaccessible CBD initially. Percutaneous transhepatic approach with wire-guided balloon dilatation of the papilla assisted with a dilatation balloon. After balloon dilatation of papilla, we performed PTCD assisted retrieval balloon under wired-guided for pushing CBD stones into intestine.

RESULTS

We report on two patients with inaccessible papilla treated with papillary dilation followed by PTCD tube assisted pushing common bile duct stones.

CONCLUSION

We recommend that papillary dilatation followed by PTCD tube assisted retrieval balloon push in patients with endoscopic inaccessible common bile duct stones, particularly in multiple stones or impacted stone.

Hemolysis, elevated liver enzymes, and low platelet count (HELLP) syndrome with acute cortical blindness

The coincidence of hemolysis, elevated liver enzymes, and low platelet count (HELLP) syndrome and cortical blindness is an uncommon but very dramatic event. We describe a case of HELLP syndrome complicating with acute cortical blindness before delivery. A 27 year-old woman, gravida 1, para 0, with normal medical history, was referred to our emergency department at the 33th week of gestation due to headache, vomiting, and blurred vision. The ophthalmologic examination showed intact pupillary light reflexes and normal ophthalmoscopic findings, but no light perception in either eye. Brain computed tomography showed normal findings. HELLP syndrome and preeclampsia was diagnosed based on the findings of hypertension and proteinuria as well as laboratory data. Prompt delivery was performed in order to achieve good maternal and neonatal outcomes.

Endoscopic stenting in the treatment of bile leakage after laparoscopic cholecystectomy

BACKGROUND

Bile leakage is one of the most common complications after laparoscopic cholecystectomy surgery, and biliary decompression is a key factor in treatment. We retrospectively investigated 6 patients with bile leakage after laparoscopic cholecystectomy who were treated with endoscopic stent.

METHODS

From March 1995 to May 1999, six patients (4 men and 2 women) aged 30-64 years (mean, 51 years) with bile leakage after laparoscopic cholecystectomy were enrolled. Biliary stent (10 French, 6-10 cm) placement with (n = 4) or without (n = 2) sphincterotomy was attempted. The symptoms, results and outcomes were reviewed.

RESULTS

The interval from operation to presentation of bile leakage ranged from 1 to 10 days. Bile leakage was detected from cystic duct stump in 5 patients (83%) and from right IHD in 1 patient (17%). Plastic stent placement was successfully in all patients. Endoscopic stenting healed bile leakage successfully in 5 cases (83%). One patient required surgical correction due to persistent bile leakage. The mean duration between stent placement and cessation of bile leakage was 6.8 days (range 1 to 24 days).

CONCLUSIONS

Endoscopic stenting is a safe, rapid and effective treatment for bile leakage after laparoscopic cholecystectomy.

Application of a wire-guided side-viewing duodenoscope in total esophagectomy with colonic interposition

Therapeutic endoscopic retrograde cholangiopancreatography (ERCP) is the mainstay treatment for bile duct disease. The procedure is difficult per se, especially when a side-viewing duodenoscope is used, and when the patient has altered anatomical features, such as colonic interposition. Currently, there is no consensus on the standard approach for therapeutic ERCP in patients with total esophagectomy and colonic interposition. We describe a novel treatment design that involves the use of a side-viewing duodenoscope to perform therapeutic ERCP in patients with total esophagectomy and colonic interposition. A gastroscope was initially introduced into the interposed colon and a radio-opaque standard guidewire was advanced to a distance beyond the papilla of Vater, before the gastroscope was withdrawn. A side-viewing duodenoscope was then introduced along the guidewire under fluoroscopic guidance. After cannulation into the papilla of Vater, endoscopic retrograde cholangiography (ERC) revealed a filling defect (maximum diameter: 15 cm) at the distal portion of the common bile duct (CBD). This defect was determined to be a stone, which was successfully retrieved by a Dormia basket after complete sphincterotomy. With this treatment design, it is possible to perform therapeutic ERCP in patients with colonic interposition, thereby precluding the need for percutaneous drainage or surgery.

Clinical significance and prediction factors of gastric varices in patients with hepatocellular carcinoma

BACKGROUND/AIMS

Hepatocellular carcinoma is part of the natural history of liver cirrhosis. Gastrointestinal bleeding and hepatic failure are the leading causes of death in hepatocellular carcinoma patients. With gastrointestinal bleeding, variceal bleeding is the most prominent, and most variceal bleeding is of esophageal origin. Gastric varices bleeding is often a massive and severe bleeding episode. The role of gastric varices among patients with hepatocellular carcinoma remains to be clarified. In this study, we aimed to evaluate the prevalence, clinical significance and prediction of gastric varices in patients with hepatocellular carcinoma.

METHODOLOGY

From 1998 to 2000, we reviewed 304 patients with hepatocellular carcinoma receiving upper gastrointestinal endoscopic examinations. Patients' clinical characteristics, physical findings, laboratory data, image studies, endoscopic examinations and treatment were reviewed.

RESULTS

Among 304 patients with HCC, twenty-one (6.9%) had gastric varices among 304 patients with hepatocellular carcinoma. The location of gastric varices were the posterior wall in 12 (57%), the lesser curvature in 1 (5%), the greater curvature in 4 (19%) and the fundus in 4 (19%). Three (14%) of these 21 patients with hepatocellular carcinoma and gastric varices had clinical evidence of bleeding. One of them died due to uncontrollable bleeding. Child-Pugh classification, hepatic encephalopathy, portal vein or splenic vein dilatation, ascites, splenomegaly, albumin level, prothrombin time and platelet count were significantly different between hepatocellular carcinoma patients with gastric varices and without gastric varices under the univariate analysis. Ascites (Odds ratio: 5.45; 95% confidence interval: 2.12-14.01) and portal vein or splenic vein dilatation (Odds ratio: 4.38; 95% confidence interval: 1.77-10.86) were the two most important predictors under the stepwise logistic regression analysis.

CONCLUSIONS

The prevalence of gastric varices in patients with hepatocellular carcinoma is 6.9% and the risk of bleeding is low in this study. The Predictors of gastric varices among hepatocellular carcinoma are related to liver cirrhosis, Child-Pugh classification, hepatic encephalopathy, portal vein or splenic vein dilatation, ascites, splenomegaly, albumin level, prothrombin time and platelet count.