Robert Listernick, MD, is Attending Physician, Ann & Robert H. Lurie Children’s Hospital of Chicago; and Professor of Pediatrics, Northwestern University, Feinberg School of Medicine, Division of Academic General Pediatrics. Address correspondence to Robert Listernick, MD, via e-mail:
Robert Listernick, MD, and colleagues discuss hard-to-diagnose cases.
A now 30-month-old girl was initially seen at 21 months of age for failure to thrive. She was described as having been a “poor eater” since age 6 months, when she began eating pureed foods. She would “spit them out of her mouth,” never swallowing. Before that time, she had been drinking breast milk and formula from the bottle without problems. At age 11 months, she began vomiting at random times throughout the day. A speech therapy evaluation revealed normal motor coordination, suck, and swallow.
Her doctor performed laboratory testing, most of which was normal except for free T4 of 0.61 ng/dL (with the lower limit of normal being 0.78 ng/dL). Thyroid-stimulating hormone (TSH) was normal. This prompted referral to an endocrinologist who did further testing and concluded that there was no endocrinopathy. At age 26 months, she was seen by a gastroenterologist who noted that she ate very few solid foods and drank only 14 ounces of milk per day in addition to water. Complete blood count, serum chemistries, tissue transglutaminase, and barium study and endoscopy of the upper gastrointestinal tract were normal. It was recommended that she enter a feeding program.
Instead of entering the feeding program after developing persistent daily vomiting, her parents took her for a third opinion at an integrative medicine program, where gastrostomy tube placement was recommended.
Following that visit, she was seen for a fourth opinion by a gastroenterologist who recommended “caloric enrichment, hypoallergenic formula, anti-reflux measures, and testing for Williams syndrome due to some minor dysmorphic features.” A gastrostomy tube was placed at an outside hospital prior to referral here.
On review of systems, she wasn’t eating much solid food. She had multiple episodes of emesis each day with normal stools. Her birth history and family histories were unremarkable. Her development was normal.
On exam at the present time, her weight had fallen substantially, to below the 5th percentile, and length and head circumference were in the 10th percentile. Her physical exam was normal save for decreased muscle mass.
Robert Listernick, MD, moderator: In reviewing her growth chart, it appears that she has actually lost about 1 pound over the past 6 months, and her length fell from the 25th percentile to the 10th percentile. As a start, do infants who have poor weight gain (failure to thrive) ever have an endocrinopathy?
Donald Zimmerman, MD, pediatric endocrinologist: In the infant/toddler age group, endocrinopathies causing poor weight gain are very rare. Hypothyroidism causes poor linear growth and excessive weight gain. Hyperthyroidism is rare in this age group, but affected children should have clear signs of hyperthyroidism. In general, we’re the wrong consultants in this clinical setting.
Dr. Listernick: What do you think about the recommendation for gastrostomy tube placement in this child?
Barbara Bayldon, MD, academic general pediatrician: That’s a decision I never enter into lightly. First, I’d like to document that the child isn’t taking in sufficient calories. If that’s the case, I’d like to better understand why. It is extremely important to distinguish among such diverse problems as food refusal due to sensory integration disorder, anorexia due to systemic illnesses, and neurologic dysfunction. Although supplemental calories are extremely important in each of these scenarios, the approach as well as the need for gastrostomy placement varies greatly.
Dr. Listernick: Does this child have gastroesophageal reflux disease (GERD) leading to poor weight gain? Vomiting is a prominent symptom here.
Jeff Brown, MD, pediatric gastroenterologist: Vomiting to this degree starting at 11 months of age is not GERD. If the etiology were gastrointestinal, I’d be much more suspicious of a food-related trigger leading to eosinophilic esophagitis. The most common food trigger would be cow’s milk if the child had not been previously exposed. Another food trigger leading to this history could be gluten and celiac disease. The pattern of vomiting is important to know. In general, in both celiac disease and eosinophilic esophagitis, vomiting should be temporally related to ingestion of the offending agent.
Dr. Listernick: Can you say a few words about distinguishing eosinophilic esophagitis from gastroesophageal reflux histologically?
Dr. Brown: First, gastroesophageal reflux is rarely diagnosed by visual endoscopy alone in children, whereas in adults we can easily see erosive esophagitis. Histology is important in children. Second, rather than only taking a biopsy in the distal esophagus as we had in the past, we now perform biopsy on both the distal and mid-esophagus in order to have a better idea of the distribution of inflammation. This helps in distinguishing eosinophilic esophagitis from gastroesophageal reflux. Finally, these children should receive therapy with a protein pump inhibitor for at least 4 weeks prior to endoscopy; this often eliminates any vestiges of reflux disease and helps provide a more accurate diagnosis of eosinophilic esophagitis.
Dr. Listernick: Do you want to comment on gastrostomy placement?
Dr. Brown: I would never do that until it was clear to me that the child could gain weight on supplemental feeding via nasogastric tube first.
Dr. Listernick: More history was obtained when she first was seen by our endocrinologist. When specifically asked, the mother reported the child had excessive thirst. “It really got bad this summer; some days she was drinking up to 60 ounces of fluid, especially water.” She does not have a clear preference for ice cold water and does not wake up in the night to drink. The mother believes that the child frequently makes up excuses to drink such as that she has the hiccups or wants to brush her teeth so that she can suck the water out of the toothbrush. She drinks water from the bathtub faucet. Overnight she is in a diaper and often soaks all her clothes. So, guess what’s going on!
Dr. Zimmerman: Seems like a slam dunk that she has diabetes insipidus (DI), doesn’t it? Of course it helps to ask the right questions once DI is suspected. For instance, one of the clinical clues to the presence of DI is that the individual has a strong preference for very cold fluids. If DI is suspected, you shouldn’t let the normal serum electrolytes fool you. As long as the person has a normal thirst mechanism and access to free water, the serum sodium can stay in the normal range. In addition, you should always check an electrolyte panel because hypokalemia and hypercalcemia can both mimic DI.
Dr. Listernick: Once the diagnosis of DI was suspected in the face of her chronic vomiting history, neuroimaging was obtained.
Tord Alden, MD, pediatric neurosurgeon: We see a large lobulated, mixed-density suprasellar mass with multiple cystic components.
Dr. Listernick: Differential diagnosis?
Dr. Alden: When I first heard about this patient before I heard the full report, I believed that she had a hypothalamic glioma. However, computerized tomography revealed speckled calcifications, which, coupled with the cysts, clearly pointed to this being a craniopharyngioma.
Jason Fangusaro, MD, pediatric neuro-oncologist: Although this clearly seems to be a craniopharyngioma, central nervous system germ cell tumors can present exactly in this fashion with vomiting and DI; they also may have calcifications.
Dr. Listernick: What about that initially low T4?
Dr. Zimmerman: She might have had sick euthyroid syndrome due to her malnutrition. More likely, she has secondary hypothyroidism due to the tumor’s mass effect on the pituitary gland. In order to distinguish between the two, it’s imperative to measure T4 by the dialysis method. This is particularly true in children younger than age 3 years who are dependent on thyroxine for brain development.
Dr. Alden: I should point out that this is how brain tumors present in infants and young children—vomiting and failure to thrive.
Dr. Listernick: She went to the operating room. What is a realistic surgical goal?
Dr. Alden: Obviously you’d like to remove the entire tumor because she is very young and not a candidate for radiation therapy. Although craniopharyngiomas are benign tumors, they can behave in a very aggressive manner. This type of tumor is often wrapped around the carotid artery or stuck underneath the optic nerve, making complete removal impossible. Sometimes, the vast majority of the tumor is cystic and we can drain the cyst without being aggressive so as to maintain pituitary function.
Dr. Listernick: Can we see the histology?
Nitin Wadhwani, MD, pediatric pathologist: Craniopharyngiomas are classified as grade I epithelial tumors. On higher power, we see squamous-lined tissue with wet keratin. Within the tumor, we see entrapped normal pituitary tissue. There are two distinct pathologic subtypes: 1) the adamantinomatous subtype, in which we see wet keratin, calcifications, and complex cysts; and 2) the papillary subtype, which lacks these features. The literature is not clear as to whether there are prognostic differences between the two subtypes. Although they arise from remnants of Rathke’s pouch, they are true neoplasms. Adamantinomatous tumors have mutations in the beta-catenin gene that are not seen in the papillary subtype.
Dr. Fangusaro: Craniopharyngiomas are great examples of why we neuro-oncologists dislike the term “benign.” It tends to give people a false sense of security. Although they don’t metastasize, they can be one of the most aggressive, quality-of-life–affecting pediatric brain tumors, with profound endocrinologic and neuropsychological long-term effects as well as seizures.
Dr. Listernick: How do they present?
Dr. Fangusaro: Many present as it did in this infant, with vomiting, poor weight gain, or endocrinologic abnormalities such as DI or growth hormone deficiency. As many as 50% of patients have ophthalmologic abnormalities as the presenting sign.
Dr. Listernick: Treatment?
Dr. Fangusaro: The best treatment is gross total resection, although somewhere between 20% and 50% of these patients have a recurrence. For older patients who have had a subtotal resection, radiation is the standard treatment. Currently, an international study is looking at the role of radiation in those patients who had a subtotal resection; patients older than age 5 years are being randomized to radiation therapy, or to observation and radiation therapy only if progression occurs. Late effects, such as neurocognitive dysfunction and secondary malignancies, are more likely to occur in young children who receive radiation therapy, so we are doing our best to avoid it. Other agents have been tried, such as intracavitary infusions of bleomycin in those tumors that have large cystic components. Presently, there’s an open study through the Pediatric Brain Tumor Consortium that is using pegylated interferon-alpha.
Dr. Brown: Hindsight is 20/20, and not all vomiting infants need magnetic resonance imaging—you’d bankrupt the system. If the child is vomiting and doesn’t respond to anti-reflux therapy and does not have the pathologic findings of eosinophilic esophagitis on endoscopy, one needs to look elsewhere for the source. Of course, the presence of an unusual history, such as the DI history in this patient, or physical findings, such as ophthalmologic abnormalities, should be huge red flags that something more ominous is going on.
Dr. Listernick: Any other endocrinologic observations?
Dr. Zimmerman: Other than DI, suprasellar tumors present in a characteristic endocrinologic sequence as the tumor grows, with progressive abnormalities of anterior pituitary function. First, one generally sees growth hormone deficiency followed by gonadotropin deficiency, TSH deficiency, and finally adrenocorticotropic hormone deficiency. An interesting clinical pearl is that children who present with panhypopituitarism may not have polyuria and polydipsia until we start giving them their needed hydrocortisone, which will unmask DI.
Dr. Listernick: Thank you, everyone.