This 5-week-old boy was admitted with a 1-day history of increased work of breathing, rhinorrhea, and fever. Although there was no history of forceful coughing, he had one episode of slight pink-tinged sputum. There was no history of cyanosis, vomiting, diarrhea, or foreign body aspiration. He had been eating normally. He was the 6-pound, 5 -ounce product of a full-term pregnancy to a 16-year-old GlPl female by a spontaneous vaginal delivery. There were no perinatal problems.
On physical examination, he was fussy, but consolable, and in no respiratory distress. Temperature was 38.7° C, pulse 140, respiratory rate 40, and blood pressure 80 by palpation. Weight, length, and head circumference were in the 50th percentile. HEENT exam was unremarkable. There were mild intermittent stridor and mild intercostal retractions with occasional grunting without wheezes or crackles. Sl and S2 were normal without murmurs. Abdomen was soft and nontender without masses or organomegaly. Deep tendon reflexes were normal and symmetric; his tone was normal. The remainder of the examination was unremarkable.
Laboratory evaluation revealed a normal blood count, urinalysis, and lumbar puncture. Blood, urine, and cerebrospinal fluid cultures were negative, and a chest x-ray was obtained the night of admission. Chest x-ray on the day of admission demonstrated a pneumomediastinum (Figure 1, see page 237).
Robert Listernick, MD, moderator: First impressions?
Mary Nevin, MD, pediatric pulmonologist: Pneumomediastinum in newborn infants is not uncommon and is often asymptomatic. It may be associated with birth trauma, neonatal respiratory distress syndrome, meconium aspiration, or mechanical ventilation. It's very unusual in a slightly older infant. I don't believe that the present pneumomediastinum can be attributed to a perinatal event. Certainly, in older children, it's most commonly associated with a bronchospastic event.
Dan Leonhardt, MD, child abuse physician: Pneumomediastinum has been reported due to penetrating esophageal trauma from a finger or foreign body. However, it also has been associated with a direct blow to the neck or suffocation.
Dr. Listernick: Can you comment on the pink-tinged sputum?
Dr. Nevin: Putting together the pneumomediastinum and hemoptysis, you could postulate that the child had an esophageal perforation leading to both. In a 5-week-old, this would almost certainly be the result of child abuse. More commonly, bloody sputum in an infant could be the result of esophageal bleeding from a small tear in the esophageal wall or from erosive esophagitis. Other causes of hemoptysis in neonates include bleeding disorders, airway hemangiomas or arteriovenous malformations, or idiopathic pulmonary hemorrhage (IPH). IPH has been reported in clusters of inner city infants, predominantly male and African-American. Most were term and previously healthy; all had a rapid onset of severe disease leading to mechanical ventilation. Although Stachybotrys atra, a toxic mold, was linked epidemiologically to some cases, its role is uncertain.
Stan Shulman, MD, pediatric infectious disease physician: One thing for sure, bloody sputum in a 5 -week-old is not the result of an infectious disease.
Dr. Listernick: Moving on, he continued to exhibit choking and coughing while feeding. An esophagram was performed which showed considerable nasopharyngeal reflux and tracheal aspiration. There had been no complaint of previous feeding problems by the family. An unusual indentation of the esophagus just below the level of the pharynx was seen prompting computerized tomography of the chest which was normal; no perforation or leaking of contrast was seen. Finally, bedside laryngoscopy was performed and was normal. Because of ongoing poor feeding, video fluoroscopy was performed which revealed poor cricopharyngeal closure and reflux of liquid into the trachea. It was recommended that he receive nasogastric feedings at home.
Figure 1. X-ray from day of ad mission illustrates pneumomediastinum.
Dr. Leonhardt: How easy is it to miss a small esophageal perforation?
Dr. Nevin: It is certainly possible depending on who's performing the procedure and whether rigid or flexible endoscope is being used. Even in the best of hands, I've known of tracheoesophageal fistulae that have been missed.
Josh Goldstein, MD, pediatric neurologist: It's hard to make much sense out of this as there had been no previous history of swallowing dysfunction and he was not failing to thrive. A common cause of acquired swallowing dysfunction and food aversion may be gastroesophageal reflux. We also see a number of children who have isolated swallowing dysfunction as a result of a developmentally immature nervous system. Generally, this improves over the first year of life. Assuming that the rest of the neurologic examination was normal as stated, this scenario doesn't make a lot of sense. There's been no mention of any other cranial nerve abnormalities.
Dr. Nevin: My impression is that food aversion and swallowing dysfunction to this degree is unusual at 5 weeks of age.
Dr. Listernick: Should we perform neuroimaging based on the known information?
Dr. Goldstein: If a child is neurologically normal and only has isolated swallowing dysfunction, neuroimaging can be deferred as long as the child is followed closely. If the swallowing doesn't improve during the first year of life or if other neurologic signs develop, magnetic resonance imaging (MRI) of the brain and posterior fossa should be obtained.
Figure 2. Left vertebral artery dissection.
Dr. Listernick: He was discharged home, and 7 weeks later he was seen in the emergency department with a 2-day history of lethargy and poor intake without fever or vomiting. His mother reported that his eyes had been crossed for the previous two days. On exam, he was lethargic with dysconjugate gaze. Temperature was 38.6 C, pulse 150, respiratory rate 38, blood pressure 92 by palpation. Weight was now in the 25th percentile, but length and head circumference remained in the 50th percentile. The anterior fontanel was bulging. Pupils were equal, round, and reactive to light. There were bilateral sixth nerve palsies. On fundoscopy, there were bilateral retinal hemorrhages, Lungs were clear. Sl and S2 were normal without murmurs. Abdomen was soft and nontender without masses or organomegaly. Deep tendon reflexes were normal. There was significant truncal hypotonia. The other cranial nerves were normal.
Dr. Goldstein: This neurologic exam is easy to interpret. The bulging fontanel, bilateral sixth nerve palsies, and the new-onset axial hypotonia all point to the presence of increased intracranial pressure. This is probably not a long-standing phenomenon as the head circumference hasn't crossed percentiles.
Delilah Burrowes, MD, pediatric neuroradiologist: The initial study was a noncontrast enhanced computed tomography (CT) scan, which demonstrated hemorrhage in the perimedullary subarachnoid space, a small amount of blood in the lateral ventricles, and enlargement of the ventricles. Etiologies of intracranial hemorrhage at this age include nonaccidental trauma or bleeding from a vascular malformation or tumor.
Dr. Listernick: Is this shaken baby syndrome (SBS)?
Dr. Leonhardt: The constellation of neuroradiologic findings is somewhat unusual for SBS. Typically, we see subdural hemorrhage, although subarachnoid hemorrhage can be seen in SBS, and is a frequent finding in children undergoing autopsy. We also don't commonly see intraventricular blood and acute hydrocephalus. However, nonaccidental trauma still needs to be high on the list.
Dr. Listernick: What about the presence of retinal hemorrhages?
Dr. Leonhardt: The presence of retinal hemorrhages does not in and of itself confirm the diagnosis of child abuse. The description of the retinal hemorrhages is critical. Retinal hemorrhages can be located in different areas of the retina - preretinal, which are on top of the retina; intraretinal, within the layers of the retina; and subretinal, underneath the retina. Multiple retinal hemorrhages located in these different areas of the retina and extend all the way out to the edge of the eye are virtually diagnostic for abusive head trauma. We can see retinal hemorrhages from increased intracranial pressure and in a small minority of cases of accidental trauma; however, they are usually few in number and located in the posterior pole of the eye. This child's hemorrhages were extensive and grouped in the periphery of the eyes. They were certainly highly suggestive of child abuse, even in the absence of the other clinical information.
Dimitrios Nikas, MD, pediatric neurosurgeon: From a neurosurgical standpoint, he had acute hydrocephalus and intraventricular hemorrhage that required immediate drainage. We found that he had extremely elevated intracranial pressure; the blood was evacuated and an extraventricular drain was left in place. Even though child abuse was high on the list, we felt that the location of the blood was unusual and were concerned about a different etiology such as a tumor or vascular lesion so further imaging was obtained.
Dr. Burrowes: Magnetic resonance imaging (MRI) of the brain demonstrated the hemorrhage but no other clear abnormalities. Magnetic resonance angiography (MRA) was also inconclusive; MRA may miss a small vascular malformation or aneurysm. MRI of the spine was performed in case there was a spinal vascular malformation or tumor which bled and tracked into the brain. Finally, CT angiography (CTA) was performed, as it is more sensitive than MRI or MRA in detecting small vascular lesions. Often, CTA will provide as detailed images as conventional angiography without exposing the child to the risks inherent in performing an angiogram. In this child, we demonstrated the left vertebral artery to have a narrowed segment just distal to the branching of the posterior inferior cerebellar artery. Angiography demonstrated a dissection of the vertebral artery at this level, which was embolized to prevent further hemorrhage (Figure 2, see page 238).
Dr. Listernick: What are the possible etiologies of this dissection?
Dr. Nikas: Vertebral artery dissections are rare in pediatrics. Most that have been described occur in the neck outside of the cranial vault as a result of either direct trauma or shearing whiplash-like forces. To have such a dissection inside the skull is even more unusual because the intracranial vessels are relatively protected against these forces. I'm not aware of any cases secondary to child abuse.
Dr. Listernick: When I heard this story I wondered whether she could have a collagen disorder such as Ehlers-Danlos type IV, the autosomal dominant "vascular type," which is characterized by spontaneous rupture of bowel and large arteries as well as the other manifestations of EhlersDanlos, such as fragile skin, easy bruisability, joint hypermobility, and "cigarette-paper scars." However, the genetics consultant didn't believe that there were sufficient phenotypic features to pursue this diagnosis.
Although the whole story was strange, nonaccidental trauma was highly suspected. How did you proceed?
Dr. Leonhardt: First, we try to get as much information as we can and to fully define the extent of the injuries. We interviewed each parent separately and tried to assess the home environment. We also searched the database to see if there had been any previous reports to the Department of Children and Family Services. X-rays of the entire skeleton were obtained because fractures are found in 33% to 50% of cases of abusive head trauma. As it turns out, he had bilateral bucket-handle fractures of the distal tibia, metaphyseal lesions that are nearly pathognomonic for child abuse.
Dr. Listernick: How do they occur?
Dr. Leonhardt: These fractures are actually the result of small microfractures along the metaphysis from torsional or shearing forces, such as grabbing the leg and twisting it or actually shaking the baby violently.
Dr. Listernick: What has been his hospital course?
Dr. Nikas: Unfortunately, he developed hydrocephalus and required placement of a ventriculoperitoneal shunt. The Department of Children and Family Services took protective custody.
Dr. Listernick: For the moment, we have to assume that the pneumomediastinum was the result of esophageal or retropharyngeal penetrating trauma that healed on its own. Thank you, everybody.