Abdominal Pain

A 10-year-old girl presented with a 10-day history of severe postprandial epigastric abdominal pain, lasting 60 minutes, relieved by emesis and associated with a 4.5-kg weight loss. Initial abdominal ultrasound and computed tomography suggested possible superior mesenteric vein thrombosis, and enoxaparin was started. The patient continued to have recurrent abdominal pain requiring hospitalisation twice during the following year, despite repeat imaging demonstrating clear perfusion of the superior mesenteric vein. Abdominal pain and vomiting escalated, resulting in food aversion and further weight loss, mandating rehospitalisation 16 months after initial presentation. Multiple ultrasounds and magnetic resonance angiography of the abdomen were performed, clearly confirming the absence of intravascular thrombus. An upper gastrointestinal study showed normal anatomy with no evidence of superior mesenteric artery syndrome. Upper and lower endoscopies were normal. Because of worsening emesis, magnetic resonance imaging (MRI) of the head was completed. Incidental finding of upper cervical spinal lesion prompted further spinal imaging. Intramedullary lesions at the levels C1–2 to C4–5, T5, and T7 were discovered (Fig. 1).FIGURE 1: Spine magnetic resonance imaging showing 3 distinct enhancing intramedullary lesions.A broad differential included granulomatous, inflammatory, infectious, and malignant etiology. Screening for rheumatic and metabolic disease, including porphyria, was negative. Serum and cerebrospinal fluid angiotensin-converting enzyme were normal. Both tuberculin skin test and culture of cerebrospinal fluid for acid-fast bacilli were negative. Serology was negative for Bartonella, Borellia, Helicobacter, arbovirus, and human immunodeficiency virus, as was polymerase chain reaction for Mycoplasma and Chlamydia pneumoniae. Lumbar puncture showed normal opening pressure, cell count, biochemistry, electrophoresis, and cytology. The patient remained bedridden, dependent on total parental nutrition, with persistent episodic pain despite a morphine infusion. A trial of gabapentin was initiated, and within 36 hours, the patient had significantly reduced pain, improved effect, and was ambulatory. Total parental nutrition and morphine were successfully weaned. Diagnostically, there remained a question as to the optimal time to consider biopsy. Despite evaluations by 3 paediatric and 1 adult neurologist, the patient demonstrated no neurologic evidence of spinal cord lesions. The patient was discharged symptom free while taking gabapentin. Follow-up spinal MRI showed subtle increase in 2 of the enhancing lesions. Despite the absence of neurologic symptoms, the family opted to pursue biopsy. Pathology revealed pilocytic astrocytoma (WHO grade I). The patient presently has mild and improving deficits post biopsy and is being treated for her tumour with vinblastine. Gabapentin has been continued and her abdominal pain has not recurred. DISCUSSION The association of abdominal pain and spinal cord pathology is not a new phenomenon. Before the advent of penicillin, when syphilis posed significant disease burden, tabetic gastric crises were well described. Patients presented with recurrent paroxysms of severe abdominal pain as a manifestation of leptominingeal inflammation and spinal cord atrophy (1). Paediatric tumours of the spinal cord commonly present with pain over the affected area. Although abdominal pain is a common paediatric complaint, it is rarely an isolated presenting symptom of a spinal cord tumour. A low index of suspicion can result in significant diagnostic delay and misdiagnosis of alternative abdominal pathology. Five previous reports have described abdominal pain as a presenting symptom of spinal lesions in children. These patients demonstrated involvement of the spinal cord down to at least T8 (2–5), and misdiagnosis included irritable bowel syndrome, constipation, and functional abdominal pain. All diagnoses were reconsidered with worsening abdominal pain and the appearance of focal neurological findings (Table 1). Our patient was initially misdiagnosed with superior mesenteric vein thrombus, until repeat imaging refuted this diagnosis. Following the finding of spinal cord lesions, thorough and systemic elimination of causes other than primary tumour was considered necessary before biopsy in this neurologically asymptomatic patient. There were no signs of spinal cord pathology, such as the thoracic scoliosis, gait abnormalities, or lower limb neurological findings described in previous cases (2–5). Ultimately, the decision to biopsy was made at the preference of the family, highlighting the potential frustration for both patient and family when symptoms persist and diagnosis remains evasive.TABLE 1: Summary of published literature on paediatric spinal cord tumours presenting with recurrent abdominal painThe cervical and thoracic levels involved in our patient presented a further diagnostic challenge. Abdominal pain secondary to spinal pathology is traditionally a result of involvement of nerve tracts supplying the abdominal wall, T8 through T12. As previously suggested in the literature (3,5), we suggest the consideration of referred pain, and specific to our case, from more cranial spinal cord involvement, much akin to tabetic gastric crises. We consider the rapid response to the GABA analogue gabapentin to be further evidence of the neuropathic origin of the abdominal pain. Our patient experienced an extremely unusual clinical presentation that required 4 independent admissions to hospital during a period of 18 months to establish a diagnosis and provide comprehensive care. The lack of connection between abdominal pain and spinal cord lesions at the level presented caused significant hesitation in consideration of primary tumour as a unifying diagnosis. Following unyielding systemic evaluation, little evidence was found to support the ideal time to pursue tissue samples, which, in this case, proved essential for providing comprehensive care.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.