Cyclic vomiting syndrome is a disorder of many causes. The major challenge is to label the disorder as idiopathic or primary only when all possible etiologies, particularly those that can be specifically treated, have been ruled out. In retrospect, our patient had subtle prenatal and early postnatal radiological findings supportive of a urological cause for his symptoms, the importance of which was initially missed. It would probably be wise to include an abdominal ultrasound examination (looking for hydronephrosis) in the “must-do” battery of tests for the investigation of all cases of recurrent cyclic vomiting without apparent cause. Otherwise, it seems likely that many cases of ureteropelvic junction obstruction will be missed, cases in which surgical intervention can be curative.
Cyclic vomiting syndrome is a common disorder with many causes. In most cases, no gastroenterological cause, such as intermittent bowel obstruction, is found. In such cases, the vomiting is often deemed primary or idiopathic and is frequently assumed to be a migraine variant, so-called abdominal migraine.
Of the many definable causes of cyclic vomiting, whether gastroenterological, neurologic, metabolic, mitochondrial, or endocrinological, it is particularly important to search for causes that are specifically treatable. Most such secondary cases if specifically treated usually have a better outcome than those that are primary or idiopathic, where the only intervention is abortive, namely symptomatic treatment of the child’s vomiting and frequently accompanying abdominal pain.
We present a case of refractory cyclic vomiting for which a very extensive investigation initially failed to uncover a specific cause. A misdiagnosis of migraine variant resulted and led to antimigrainous therapy that first was reassuring, for the child appeared to respond. His symptoms soon returned, however. Further investigative studies were belatedly pursued, the cause of the child’s cyclic vomiting was finally found and definitive treatment led to complete resolution of his symptoms.
A 4-year 1-month-old boy began to have episodes of recurrent vomiting at the age of 14 months. The episodes initially occurred only in the evening, between 5 and 10 pm. There were no obvious triggers. At onset, the child became quiet, refused to talk, and appeared tired. Once his vomiting started, it would continue to recur at intervals of 10-15 minutes, for durations varying from 3 hours to 2 days. During the episodes, he became pale and clammy, with sunken eyes. The episodes would leave the child dehydrated, sometimes severely so. The vomiting was sometimes accompanied by a low-grade fever and occasionally by diarrhea.
Over the next 14 months, the child had a total of 9 episodes of cyclic vomiting, with intervals of 4-8 weeks between attacks. Blood studies, including serum chemistries, liver function tests, serum amylase and lipase levels and, and blood ketone levels had normal findings. Urinalysis occasionally showed 1+ ketones. In a few of the episodes, the vomiting was bilious, and in these cases, abdominal plain-film radiographs were taken; they too showed normal findings, with no evidence of intestinal obstruction. With 8 of the 9 episodes, the child was hospitalized overnight for treatment with intravenous fluids, ondansetron, and lorazepam. The one time he was not hospitalized, he was successfully managed at home with rectal administration of diazepam and ondansetron.
The child had been born at term, with a birth weight of 3.6 kg (40 percentile). Prenatal sonogram at 18 weeks showed bilateral hydronephrosis. Follow-up ultrasound studies after birth and at 2 weeks showed only mild left hydronephrosis, and at age 1 year, the hydronephrosis had cleared. There were no other antenatal concerns or perinatal complications.
Developmentally, the child had early mild gross motor delay; for example, he was slow to crawl. Aided by several months of early intervention therapy, he quickly caught up. His language development and cognitive skills were at all times normal. In addition, between 2 and 6 months, the child had plagiocephaly associated with a left torticollis that resolved quickly with an orthotic helmet.
The family history included car sickness in the child’s mother when she was a child, which later resolved. In adulthood, she had longstanding intermittent severe abdominal pain with nausea and vomiting. Her symptoms, initially thought to be “abdominal migraine,” were later determined to be from chronic cholecystitis; all symptoms resolved after a cholecystectomy. The child’s 70-year-old maternal grandmother has a history of car sickness. There was a history of migraine headaches in the child’s paternal great grandmother, paternal grandmother, and a paternal aunt. There was no family history of seizures.
The child’s medical history was remarkable for poor appetite and poor weight gain between 6 and 12 months of age (when his weight percentiles dropped from 25%-50% at 4-6 months to 3% by 12 months of age). He had multiple upper respiratory tract and ear infections during that period. Tympanostomy tubes were placed at the age of 20 months. Because of his poor appetite and poor weight gain, at 14 months of age, he was prescribed a high-calorie diet and an appetite stimulant, cyproheptadine, which helped. Cystic fibrosis was ruled out by negative results on 2 sweat chloride tests.
The child’s history of recurrent stereotypical episodes of vomiting, with complete resolution of symptoms and return to baseline between episodes, prompted in-depth investigations by several specialists to search for an underlying cause. Upper gastrointestinal endoscopy showed only a hiatal hernia. Biopsies from the esophagus, duodenum, and stomach had normal findings. He was administered lansoprazole, a proton pump inhibitor, but without benefit. Metabolic and mitochondrial laboratory tests showed a normal blood ammonia, carnitine, and acylcarnitine profile; plasma amino acids, urine organic acids, blood creatine kinase and serum lactate levels were normal.
Early in his course, during acute presentations of cyclic vomiting at the ages of 7 and 14 months, the child had a low blood glucose level with 1+ urine ketones, expected in a child with a low glycogen reserve and in a fasting catabolic state, so-called ketotic hypoglycemia. Those abnormalities reversed as the child grew older and gained weight.
Most of the child’s endocrinological studies showed normal findings, although he had elevated levels of blood adrenocorticotropic hormone (726 pg/mL; normal <46 pg/mL) and blood cortisol (40.2 mcg/dL; normal 5-25 mcg/dL) during acute attacks. This was initially thought to be due to glucocorticoid resistance, but when repeated during an asymptomatic period, those levels had normalized; then it was concluded that the elevated levels were most likely induced by stress. Insulin-like growth factor-1 and insulin growth factor–binding protein-3 levels, which were initially low, also normalized with weight gain.
Neurologic examination had shown no papilledema, “sun-setting,” abducens paresis or other signs suggesting elevated intracranial pressure. A brain magnetic resonance imaging scan showed no intracranial abnormalities such as a tumor or hydrocephalus. An electroencephalogram showed normal findings, with no epileptiform activity or slowing.
At the age of 25 months, during the child’s sixth presentation to the emergency room with recurrent vomiting, a renal sonogram was obtained, given his history of prenatal hydronephrosis. It disclosed a new left hydronephrosis, moderate in degree without dilatation of the distal left ureter; the proximal ureter was not visualized. These findings suggested left ureteropelvic junction (UPJ) obstruction. The right kidney and ureter were normal. A follow-up dynamic renal scintigraphy (MAG3) with diuretic renography was performed next day. The right kidney showed normal spontaneous drainage (cortical transit time was 3-4 minutes; renal time activity curve showed a time to peak of 4 minutes, with a 20-minute residual of 22% of the peak activity). The hydronephrotic left kidney had markedly delayed spontaneous drainage (cortical transit time was 4 minutes; renal time activity curve showed a time to peak of 13 minutes, with a 20-minute residual of 86% of the peak activity). However, after furosemide administration, the left kidney washout half-time was 3 minutes, with a residual of 5% at 30 minutes, indicating no obstruction during diuresis. A computed tomography angiogram of the abdomen and pelvis with contrast did not show a crossing renal vessel compressing the ureter at the left UPJ or any other anatomical abnormality at that site. A follow-up renal sonogram, repeated during an asymptomatic period at the age of 27 months, showed improvement in the left-sided hydronephrosis. Urology concluded that there was insufficient clinical or radiological evidence of left UPJ obstruction (causing the left hydronephrosis) to warrant surgical exploration.
As extensive investigations had failed to identify a specific cause for the child’s episodes of recurrent vomiting, it was concluded that this was probably a migraine variant, cyclic vomiting or abdominal migraine; that diagnosis was supported by the family history of both migraine and car sickness. The child was administered 4 mg/kg/d dose of verapamil orally, in 2 divided daily doses. His initial response was very encouraging, for he remained free of vomiting for 4 months—his longest previous vomiting-free period had never been greater than 2 months. Then, after 4 vomiting-clear months, the child’s vomiting returned, and for the first time, it occurred in the morning. In addition to the vomiting, he complained of left flank pain, with left flank tenderness to palpation. In retrospect, the mother reported that for the previous year, during his episodes of recurrent vomiting, the child seemed to be experiencing left-sided abdominal or flank pain. Repeat blood tests showed mildly elevated serum transaminases; the other blood chemistries, including blood urea nitrogen, serum creatinine, serum amylase and lipase levels, were normal. Urinalysis showed 3+ blood and 120 red blood cells per high-power field for the first time, in addition to 1+ ketones and 1+ protein. A renal sonogram showed mild left hydronephrosis, now with a mild left hydroureter involving the proximal and middle sections. No renal or ureteral calculi were visualized.
A week later, repeat dynamic renal scintigraphy (MAG3) with diuretic renography showed a hydronephrotic left kidney with markedly delayed spontaneous drainage. The left kidney showed a cortical transit time of 5 minutes; the renal time activity curve showed a time to peak of 12.5 minutes, with a 20-minute residual of 93% of the peak activity; yet, after furosemide administration, the left kidney washout half-time was 4 minutes, with a 30-minute residual of 5%; that is, prompt drainage with nonobstructive parameters, little changed from the prior scan. The right kidney showed normal spontaneous drainage (cortical transit time was 3 minutes; the renal time activity curve showed a time to peak of 3.3 minutes, with a 20-minute residual of 27% of the peak activity). After furosemide administration, there was rapid and essentially complete washout of tracer.
The child had another episode of cyclic vomiting accompanied by severe left flank pain 2 months later, which was similar to his most previous episode. A repeat renal sonogram showed a left moderate hydronephrosis, with proximal and middle left hydroureter, greater than seen on the most recent previous sonogram..