This 17-year-old girl was admitted to the intensive care unit because of increasing respiratory distress over the previous two weeks. Two weeks prior to admission, she started developing progressive respiratory distress and swelling in her right leg. She had difficulty climbing stairs and increasing exercise intolerance. There was a vague history of intermittent fevers without rash, cough, vomiting, diarrhea, or other symptoms. At an outside hospital, her oxygen saturation was 57% in room air, and a limited echocardiogram revealed a moderate pericardia! effusion. Due to deterioration in her respiratory status, she underwent endotracheal intubation and was transferred.
Her past medical history is quite remarkable. At birth, she was thought to have a hemangioma over her left thigh, but throughout the next several years, the lesion inexorably grew, involving more of her thigh and leading to remarkable hypertrophy of the entire left lower extremity. Eventually the diagnosis of an arteriovenous malformation (AVM) was made. When she was about 6 years old and during attempted embolization of the AVM, she developed a large pulmonary embolus and sustained a respiratory arrest during the procedure. Although the pulmonary embolus was lysed and she was successfully resuscitated, it was decided that the risk of repeat embolism was very high, and her entire left lower extremity was amputated. Despite the surgery, the residual AVM in her pelvis and in her stump continued to grow, eventually leading to high output congestive heart failure. She has had repeated attempts of embolization of portions of the AVM.
On exam on admission, she was intubated and sedated. Her temperature was 39° C, pulse 138, respiratory rate 22 synchronized with the ventilator, blood pressure 120/40. HEENT exam was unremarkable. The precordium was extremely hyperdynamic. The point of maximum intensity was in the posterior axillary line. There was a prominent left ventricular impulse. A palpable thrill was present at the left lower sternal border. A grade IV/VI systolic ejection murmur was heard at the left lower sternal border and a grade IV/IV diastolic flow murmur was heard at the apex. There was a loud S3 gallop. The lungs were clear. The liver was palpable 4 cm below the right costal margin. There was a prominent vascular malformation noted on her left leg stump and in the left pelvic wall. The remainder of the physical examination was unremarkable.
Laboratory evaluation: electrolytes normal, save for BUN 54 mg/dL and creatinine of 1.1 mg/dL; hemoglobin 8.5 g/dL, white blood count 10,000/mm3 with 86% neutrophils, 7% lymphocytes; platelet count 75,000/mm3; prothrombin time 19 seconds (elevated), partial thromboplastin time 28 seconds, fibrinogen 354 mg/dL, and D-dimers 9.8 MgAnL (elevated). Urinalysis had large blood, moderate bilirubin, 2+ protein, and too-numerous-to-count red blood cells in the sediment. Cultures of the blood and urine were negative. A chest x-ray showed massive cardiomegaly with diffuse rightsided interstitial lung markings and air space opacities consistent with congestive heart failure.
Echocardiography revealed a large amount of pericardia! fluid. The following day, repeat echocardiography revealed collapse of the right atrial wall with greater than 10% variation in the mitral valve flow. The righi and left ventricles appeared underfilled when compared with the previous echocardiogram; these signs together were felt to be consistent with cardiac tamponade. She underwent pericardiocentesis, which drained 650 cc of frankly bloody pericardial fluid. The Gram stain of the fluid was negative, albumin 2.9 g/L, and glucose 37 mg/dL. Cell count of the fluid revealed 4 million red blood cells and 145,000 white blood cells/mm3 with 77% neutrophils. Enteroviral PCR testing was negative.
Robert Listernick, MD, moderator: How do you embolize an AVM?
Jim Donaldson, MD, pediatrie radiologist: Any number of different substances can be injected into a vascular malformation to promote clotting including balloons, alcohol, metallic coils, and glue. As I understand what happened to this child, the degree of high output congestive heart failure was so great at the time that it was felt that the only way to gain some control was to amputate the extremity and eliminate as much of the AVM as possible. Most of the tumor that could be seen radiographically at that time was in the extremity rather than the pelvis.
Guy Randolph, MD, pediatrìe cardiologist: I have been following her for some time. Her high output failure has been stable over the past 5 to 7 years; whatever is going on now seems to be a new process. She has been undergoing repeated embolizations of varying parts of the AVM in order to control the congestive heart failure. Whether these have been successful is unclear, but she had been relatively stable until recently.
Dr. Listernick: Had she had normal blood counts during this time?
Dr. Randolph: Absolutely. The anemia and thrombocytopenia are new.
Dr. Listernick: Let's backtrack a bit. Can we review the classification of vascular malformations?
Anthony Mancini, MD, pediatric dermatologist: Vascular lesions can be divided into two main categories - malformations and tumors. Tumors are true proliferative neoplasms, the most classic of which is the infantile hemangioma. Other tumors that would fall into that category would include things like Kaposiform hemangioendothelioma, tufted angioma, hemangiopericytoma, and pyogenic granuloma.
Malformations are not tumors in that they are not proliferative. Malformations are collections of dilated or excess blood vessels, usually present at birth. They are named for the predominant vessel within the lesion, whether it's a capillary, vein, or lymphatic vessel. They may have combinations of blood vessels, such as an AVM or a capillary-lymphatic-venous malformation. Although they're relatively static, they may gradually increase in size and may result in increasing complications as the patient ages.
Examples of capillary malformations include the salmon patch, also known as "stork's bite" or "angel's kiss," and the port-wine stain. Venous malformations tend to be blue and may be isolated or in combination with other vessels such as an AVM. AVMs can be locally destructive or give rise to significant rightto-Ieft shunting, as in this child. Vascular malformations may be associated with a low level chronic coagulopathy due to consumption of coagulation factors and platelets within the lesion. I should also say that the term "lymphangioma" is a misnomer; these lesions are in reality lymphatic malformations, not neoplasms. The cystic hygroma should be properly termed a macrocystic lymphatic malformation.
Dr. Listernick: So is it generally easy to distinguish between a malformation and a tumor?
Dr. Mancini: No, not always. For instance, it may be difficult to distinguish between a venous malformation and a deep infantile hemangioma. Hemangiomas are fairly characteristic in that they're usually not present at birth, although now we know up to 10% may be. They classically present at 2 to 4 weeks old, and grow over the first year of life (known as the "proliferative" phase). After 12 to 1 8 months, they begin the process of spontaneous resolution (called the "involution" phase), and this phase is quite gradual, often taking 5 to 10 years to be complete. Superficial hemangiomas are bright red, hence the traditional name "strawberry hemangioma." Deep hemangiomas are bluish lesions deeper in the skin, which may be more difficult to distinguish from malformations. Usually, after 3 to 6 months of observation, we can distinguish between the two by the growth pattern. Occasionally, we rely on radiologie methods to help distinguish between the two, but even the radiologie distinction can Tie very challenging. Magnetic resonance imaging is the imaging modality of choice. AVMs are unique in that they tend to have a purple flush, look a little telangiectatic, and often have a palpable thrill or audible bruit.
Dr. Listernick: I'm a little confused. This child's AVM clearly grew, yet it's not a "tumor" in your classification scheme.
Dr. Mancini: It is confusing. It's splitting hairs, but there was no actual cell proliferation. Rather, the "growth" most likely is due to hypertrophy of underlying tissues and increasing blood flow within the vascular lesion.
Dr. Listernick: How does the term "Kasabach-Merritt syndrome" fit into your classification scheme?
Dr. Mancini: Kasabach-Merritt phenomenon refers to the combination of a vascular neoplasm with consumptive coagulopathy and thrombocytopenia. Contrary to popular belief, this does not occur within classic hemangiomas of infancy, but rather occurs with Kaposiform hemangioendothelioma or tufted angioma. These lesions are congenital, grow extensively, and may be locally destructive. Kasabach-Merritt phenomenon does not typically occur within vascular malformations, although a low-level consumption coagulopathy may be present, as previously mentioned.
Dr. Donaldson: The imaging of all these lesions is not clear cut, and it can be very difficult to distinguish between them radiographically.
Dr. Listernick: Is there any medical treatment for AVMs beyond embolization?
Dr. Mancini: Interferon has been used for corticosteroid-resistant hemangiomas that are life-threatening or lead to functional impairment, such as loss of vision. I'm not aware of any proven use for steroids or interferon in the treatment of AVMs.
Dr. Listernick: How do you treat high-output congestive heart failure?
Dr. Randolph: If you can't control the inciting lesion, such as the AVM, this can be exceedingly difficult to treat. The best we can do is use diuretics to treat her symptomatically. Functionally, she was ambulating using crutches and going to high school. Although her cardiac function had been stable for some years, there had been a progressive decline over the past few months.
Dr. Listernick: If we were to remove the offending AVM at this point, would you expect to see return of cardiac function to normal.
Dr. Randolph: Given the longterm heart failure and the remarkable degree of cardiac dilatation, I think that a return to normal function would be exceedingly unlikely now.
Dr. Listernick: Moving forward, she clearly has a large pericardia! effusion, a new finding. Could she have acute viral pericarditis?
Ellen Chadwick, MD, pediatrie infectious disease specialist: The three most likely causes of infectious pericarditis are bacterial, viral, and mycobacterial. Tuberculous pericarditis is extremely uncommon, accounting for less than 1 % of cases. It generally presents in an extremely indolent manner with progressive shortness of breath and exercise intolerance. The best way to make a diagnosis of tuberculous pericarditis is to perform a pericardial biopsy. Viral pericarditis is by far the most common scenario, usually presenting with chest pain or signs of cardiac tamponade in an otherwise healthy appearing child, resolving spontaneously. Purulent pericarditis, most commonly caused by Staphylococcus aureus, should be suspected in any child with pericarditis who appears toxic with a high fever; the mortality is greater than 80% unless there is prompt drainage of the pericardial pus. Prior to the Hemophilus influenzae type b vaccine, that organism was the most frequent cause of purulent pericarditis.
Dr. Listernick: How is the diagnosis of viral pericarditis made?
Dr. Chadwick: The most common diagnostic tool is serologie testing, using paired sera drawn at least 2 weeks apart. Enteroviral PCR may be used on pericardial fluid, if it's not bloody. Alternately, one can try to isolate enterovirus from the pharynx or stool.
Dr. Listernick: How would you approach this child from an infectious disease point of view?
Dr. Chadwick: The pericardial fluid is difficult to interpret, and there is no clear etiology at present. Interventions that have been attempted in acute symptomatic pericarditis have included the use of intravenous gammaglobulin or corticosteroids. Most infectious disease physicians would be wary of using corticosteroids in the face of an active viral infection. Of course, there are no good prospective data that support either of these therapies.
Dr. Listemick: Ultimately, over 1 L of bloody pericardial fluid was drained from the pericardial space. She received both intravenous immunoglobulin and large doses of methylprednisolone with little effect. She had persistent fevers and reaccumulation of fluid, leading to placement of a pericardial drain. Computerized tomography did not reveal any direct connection between the AVM and the pericardium. Other diagnostic considerations included neoplastic or rheumatologic pericarditis.
Michael Miller, MD, pediatrie rheumatologist: Certainly, we need to think of systemic lupus erythematosus (SLE) in this situation. In order to confirm the diagnosis of SLE, a patient should have at least 4 of the 11 criteria listed in the Sidebar (see page 180). There are certainly other vasculitides that might lead to pericarditis, such as sarcoidosis, systemic juvenile idiopathic arthritis, or Wegener's granulomatosis. However, we have no other evidence to support any of these. Her test for antinuclear antibodies was negative, which strongly speaks against SLE, and the anti-neutrophil cytopiasmic antibody (ANCA) was negative, mitigating against Wegener's granulomatosis as a diagnosis; levels of both C3 and C4 were very low. Given that this child was extremely ill and that 500 to 700 cc of pericardial fluid was being drained daily, the decision was made to start high dose corticosteroids, 30 mg/ kg/day of methylprednisolone to a maximum of 1 g/day.
Dr. Listemick: There certainly were other findings that suggested a systemic disease such as a vasculitis, including thrombocytopenia, hematuria, azotemia, and significantly depressed complement levels. Initially, she responded to the steroids with decreasing pericardial fluid drainage. However, after this initial response, she became increasingly ill and eventually died. Although it was never proven, it was felt that she had an undiagnosed collagen-vascular disease. Thank you, everybody.