**Background:** Esophageal variceal hemorrhage typically arises from portal hypertension due to liver cirrhosis. However, a rare nonhepatic cause is cardiac-derived varices, which can develop in patients with congenital heart disease (CHD) and pulmonary venous obstruction (PVO) due to abnormal collateral vessels shunting blood from the bronchopulmonary circulation to the esophageal veins. Evidence-based guidelines for managing varices secondary to cirrhosis do not apply to this population, and the natural history and optimal management of cardiac-derived varices are poorly understood. This report describes two pediatric cases to highlight the clinical presentation, diagnostic approach, and management challenges.
**Methods:** The authors present two case reports from a single institution. Case 1: A 10-year-old child with complex CHD (double inlet left ventricle, complete atrioventricular septal defect, malpositioned great arteries, hypoplastic right ventricle, post-surgical bidirectional cavopulmonary connection [BCPC] and resection of left atrial cortriatriatum) presented with acute hematemesis and a hemoglobin drop from 21 to 18.9 g/dL. He had no evidence of liver disease (normal transaminases, platelet count 239 × 10^9/L, no splenomegaly on exam or ultrasound) and was not on anticoagulation. Urgent upper gastrointestinal endoscopy revealed grade 3 esophageal varices (EV) with active bleeding. Endoscopic variceal ligation (EVL) was performed, but within minutes he developed moderate pulmonary hemorrhage through the endotracheal tube, which resolved spontaneously. He had recurrent hematemesis 5 days later, managed conservatively with packed cell transfusions, proton pump inhibitor (PPI), and octreotide, with resolution. Computer tomography angiogram (CTA) showed PVO with extensive collateral vessels to the mainstem bronchus and paraesophageal veins. Further endoscopic reassessment and prophylactic EVL were deferred due to concern that obliterating decompressing collaterals might worsen pulmonary hemorrhage. He remained on PPI alone with no further variceal hemorrhage in 8 months of follow-up. Case 2: An 18-month-old child with postoperative total anomalous pulmonary venous drainage, persistent PVO, and pulmonary hypertension, on antiplatelet therapy, presented with first-episode hematemesis. Hemoglobin was 13.2 g/dL, platelet count 279 × 10^9/L, liver enzymes and coagulation screen normal, no hepatosplenomegaly, and Doppler ultrasound showed normal hepatopedal portal vein flow. Urgent endoscopy showed grade 2 EV without active bleeding; no endoscopic intervention was performed. CTA and cardiac catheterization confirmed PVO and collateralization between pulmonary and esophageal veins. Hepatic venous pressure gradient was low-normal at 1 mmHg, ruling out portal hypertension. Attempts to dilate and stent the left upper pulmonary vein were unsuccessful. She had no further variceal hemorrhage in 2 years of follow-up on PPI and diuretics.
**Key Results:** Both patients had cardiac-derived EV due to PVO and collateral vessel formation, without any evidence of liver disease or portal hypertension. In Case 1, EVL for active bleeding was complicated by immediate pulmonary hemorrhage, likely due to redirection of blood flow into bronchopulmonary shunts. Conservative management (PPI, octreotide, transfusions) controlled the recurrent bleed. In Case 2, no endoscopic intervention was performed, and the patient remained bleed-free for 2 years. Neither patient received prophylactic EVL, and both had favorable short- to medium-term outcomes (8 months and 2 years follow-up, respectively).
**Clinical Implications:** Cardiac-derived EV hemorrhage is a rare but important complication in children with CHD and long-standing PVO. The pathophysiology involves collateral vessels that decompress the pulmonary circulation but predispose to variceal formation. Standard management for cirrhotic varices—particularly EVL—may be dangerous in these patients because obliterating the varices can worsen pulmonary hypertension and cause pulmonary hemorrhage, as observed in Case 1. The authors propose that EVL should be reserved only for acute, significant hemorrhage refractory to medical therapy. In the absence of active bleeding, surveillance endoscopy and prophylactic EVL are not recommended. For refractory or recurrent bleeds, definitive interventions such as pneumonectomy or endovascular stenting/embolization of collaterals may be considered. This report underscores the need for a tailored, multidisciplinary approach involving cardiology, gastroenterology, and interventional radiology, and highlights the importance of recognizing nonhepatic causes of variceal bleeding in patients with complex CHD.