P095 - GASTROSCOPY-INDUCED HEMODYNAMIC INSTABILITY IN A CHILD WITH A BERLIN HEART EXCOR AND ACQUIRED VON WILLEBRAND DISEASE.

Case background: On average, 20 to 30% of children awaiting heart transplantation are currently supported by a ventricular assist device (VAD) as a bridge to transplant.¹ The Berlin Heart (BH) EXCOR is a pneumatically driven, pulsatile pump with a fixed-volume chamber. Cannulation typically involves an inflow cannula placed in the left ventricular apex or left atrium, and an outflow cannula positioned in the ascending aorta. As an extracorporeal system with an external drive console, it is suitable for infants as little as 3 kg.¹ ² (Fig. 1)
Bleeding is a frequent complication in this population, due to the unique features of developmental hemostasis in children and the need for combined anticoagulation and antiplatelet therapy to prevent device thrombosis. Additionally, high shear stress from the BH can cause a loss of high-molecular-weight von Willebrand factor (vWF) multimers, leading to an imbalance in hemostasis. This condition is known as acquired von Willebrand syndrome (AVWS).³
This case highlights the hemodynamic challenges of performing a gastroscopy in a small child supported by a Left VAD and illustrates the diagnostic complexity of AVWS in this setting.

Management Challenges: The patient was a 9-month-old, 7-kg girl with a prenatal diagnosis of atrioventricular block secondary to maternal auto-immune disease. She also had a history of gastroesophageal reflux. Despite progressive pacemaker upgrade to a three-chamber DDD device and escalating medical therapy, her cardiac function declined. She was referred to our center during an episode of acute on chronic heart failure, for transplant assessment and mechanical assist. She underwent implantation of a 15 mL BH LVAD.

Postoperative right heart support included nitric oxide, epinephrine, dobutamine, and milrinone, with rapid wean to a maintenance regime of milrinone and sildenafil.BH settings were 75 bpm, resulting in a cardiac index of 2.9 L/min/m². Echocardiography showed preserved right ventricular function on milrinone (1 mcg/kg/min) and effective left ventricular decompression.

Following ACTION network harmonization and institutional protocol, antithrombotic therapy was initiated with bivalirudin and aspirin.

Ten days post-implantation, she developed melena despite therapeutic anticoagulation (APTT 70–90 seconds) and gastric protection with famotidine. Management included pantoprazole, low-dose epinephrine, blood transfusions, and bedside gastroscopy under general anesthesia. Induction was achieved with midazolam, ketamine, and rocuronium; maintenance was provided with remifentanil and propofol infusions, along with norepinephrine support. Intubation was uneventful, and the patient was maintained on mechanical ventilation.

During gastric insufflation, the patient had systemic hypertension, abdominal distension and decreased BH filling. Intermittent pauses in insufflation allowed for CO₂ release, improving preload and restoring device filling. The gastroscopy revealed diffuse gastritis but no treatable lesion.

Persistent bleeding in the following weeks prompted multiple investigations. Initial vWF testing showed a VWF activity of 1.06 IU/mL and a VWF antigen level of 1.19 IU/mL, with an activity-to-antigen ratio of 0.89. Multimer analysis showed a normal distribution of VWF multimers. CT angiography suggested possible jejunal angiodysplasia, though no active bleeding was identified. Persistent hemorrhage forced a reduction in anticoagulation targets (APTT 50–70 seconds), with subsequent fibrin accumulation. Repeat vWF testing showed a decreased activity-to-antigen ratio of 0.56. A few days later, subsequent multimer analysis confirmed the loss of high-molecular-weight multimers, establishing the diagnosis of acquired von Willebrand syndrome. Targeted treatment with Humate-P (loading dose of 50 U/kg followed by 25 U/kg every 24 hours) resolved bleeding and transfusion requirements.

Fifteen days later, the patient underwent heart transplantation with favorable outcomes.

Discussion
Hemodynamics in patients supported by the BH-LVAD depends on the interaction between pump function, native cardiac performance, and systemic and pulmonary vascular resistances. Preload to the left-sided VAD is influenced by right ventricular (RV) output, which often requires inotropic support. The main anesthetic goals in BH patients are to preserve circulating volume (given the fixed VAD output), support RV function, and ensure effective device performance.⁴ In this case, fluid status was optimized with crystalloids and blood products prior to induction of anesthesia. Milrinone was maintained to support RV contractility, and norepinephrine infusion was initiated to counteract the effects of induction. Special attention was paid to avoid triggers of pulmonary hypertension: hypercapnia, hypoxia, cold stress, pain, and acidosis. Device filling was monitored continuously throughout the procedure.

Gastric insufflation during gastroscopy increased intra-abdominal pressure, impairing venous return and reducing preload to the BH. Intermittent desufflation allowed for rapid improvement in preload and hemodynamic stability. To minimize this risk, CO₂ is preferred over air for insufflation due to its rapid absorption and elimination. Using the lowest effective insufflation volume is also advised.⁵

Acquired von Willebrand syndrome is a recognized complication in pediatric patients with Berlin Heart support. Gossai et al. reported AVWS-compatible laboratory abnormalities in 10 of 19 patients.⁶ Diagnosis, particularly of type 2A AVWS, can be challenging. Standard assays such as VWF antigen, ristocetin cofactor activity, and collagen-binding activity may seem normal or elevated. Although an activity-to-antigen ratio below 0.7 is suggestive, this abnormality may not be present in all cases. In some instances, the loss of high-molecular-weight multimers may be the only detectable laboratory abnormality.⁷ ⁸ This case highlights the importance of maintaining a high level of suspicion despite initially normal findings. Serial testing demonstrated a progressive decline in the activity-to-antigen ratio and ultimately confirmed multimer loss, which led to the diagnosis.

Targeted treatment with Humate-P, a plasma-derived concentrate of VWF and factor VIII, achieved rapid bleeding control and resolved transfusion need.

Conclusion

This case illustrates the impact of gastroscopy on the delicate physiological balance of a pediatric patient supported by a Berlin Heart EXCOR LVAD. Gastric insufflation can significantly impair venous return and device preload, compromising pump performance and systemic circulation. Anticipating and mitigating these effects through careful anesthetic planning and minimizing the volume of CO₂ insufflation are critical to ensuring patient safety.

The diagnosis of acquired von Willebrand syndrome remains particularly challenging in this setting. While initial laboratory tests may seem deceptively normal, serial assessments, including VWF multimer analysis, may be required for definitive diagnosis. This case underscores the importance of maintaining a high index of suspicion for AVWS in Berlin Heart patients with unexplained gastrointestinal bleeding and highlights the efficacy of targeted therapy with Humate-P in achieving hemostatic control.