Department of Pediatrics, Section of Pediatric Gastroenterology and Nutrition, University Hospital Ghent, Belgium: and *Department of Pediatrics, St. Augustinus Veurne, Belgium
Received April 12, 1995; revision received November 21, 1996; accepted November 26, 1996.
Address correspondence and reprint requests to Dr. E. Robberecht, Department of Pediatrics, University Hospital Ghent, De Pintelaan 185, B-9000 Ghent, Belgium.
Although the lesions caused by Henoch-Schönlein purpura are confined largely to the skin, kidney, joints, and bowel, all body systems can be involved (1-6). To our knowledge, however, the septicemia we report here has never been described as a related complication.
A 10-year-old boy suffered dull abdominal pain, with increasingly frequent postprandial cramps, and swelling of both ankles with arthralgia successively over a 2-week period. When small blue spots appeared over his buttocks and lower legs, the diagnosis of Henoch-Schönlein purpura was made. Blood and urine analyses showed no abnormalities. After 2 weeks, prednisolone was started at a dose of 3 mg/kg/day in an attempt to mitigate abdominal pain. No positive reaction was initially noted. On the contrary, after 3 days he complained of more severe midepigastric pain and vomited repeatedly. Upper intestinal endoscopy revealed several small hemorrhagic ulcers in the duodenum, which were not recognized as caused by Henoch-Schönlein vasculitis but were erroneously attributed to the steroids, which were then tapered to 1 mg/kg/day. In addition, administration of ranitidine (Zantac) was started at 4 × 25 mg daily. The patient's clinical condition did not change.
Three days later melena was noted on several occasions and because the patient refused all oral food, convinced that eating provoked more severe abdominal cramps, he was referred to our institution. On arrival, he looked lethargic and had apparently lost 4 kg since the beginning of his disease. Clinical and biochemical examinations confirmed the diagnosis of Henoch-Schönlein purpura. Platelet count and coagulation studies were normal. Abdominal ultrasound revealed local thickening of the intestinal wall, as has been described elsewhere (7,8). In the duodenum, these zones could with high probability be linked to the hemorrhagic ulcers that had been seen endoscopically (9,10). Drug treatment was continued. In an attempt to avoid total parenteral nutrition and to deliver enough nutrients, constant-rate enteral nutrition through a nasogastric tube was initiated using a hydrolyzed, lactose-free formula (Pregestemil). It was well tolerated. The patient did not vomit, and even complaints of abdominal pain diminished. Melena was no longer noted; only on three occasions occult blood was present in the stools, which were otherwise normal. Blood pressure, renal ultrasonography, biochemical nephrologic parameters, and urine analyses showed no abnormalities.
Within several days, the patient's general condition ameliorated. Abdominal pain, especially the crises of cramps, disappeared progressively, and there was daily weight gain without clinical evidence of edema. On the patient's 10th day of hospitalization at our center, small normal meals could be reintroduced. While discharge was being planned 3 days later, the patient's temperature suddenly rose to 41°C fever, and he had chills, pallor, hypotension, diffuse sweating, and general malaise. On clinical examination, no etiologic abnormalities could be disclosed. He had neither central lines nor a bladder catheter. Urinanalysis showed completely normal results, but biochemical evaluation of the blood revealed serious signs of septicemia (leukocytosis of 27,600/μl with 37% bands and a C-reactive protein level of 18 mg/dl). After hemoculture, intravenous antibiotic therapy with ceftriaxone (100 mg/kg/day) was started. Fever decreased within the following 12 h, and biochemical and clinical normalization followed in the next few days. Enterobacter cloacae grew from several hemocultures. Because this organism proved to be sensitive to ceftriaxone, this treatment was continued for 7 more days. Meanwhile, the boy recovered further. All other medications were discontinued. The further course of his illness was uncomplicated. Intravenous antibiotics were given for a total of 10 days, and after a further safety margin of 3 days, he was able to be discharged. A normal, unrestricted oral intake could progressively be achieved so that he regained his original weight after 3 weeks. Follow-up examinations 6 months later showed no sequelae. No dermal, articular, intestinal, or nephrologic relapses during the interim period were mentioned.
The history and clinical signs of this case are undoubtedly typical for mild Henoch-Schönlein purpura (1-6). The acute severe septicemia, however, is very particular and has, to our knowledge, never been reported before. This is a surprising fact because the necrotic ulcers provide an almost direct connection between bowel lumen and circulation. These ulcers are caused by vasculitis (9,11) and are found endoscopically or ultrasonographically (7-10) on many sides of the intestinal wall from the duodenum to the ileum. It can be compared with the situation in necrotizing enterocolitis of the neonate, in whom sepsis is a frequent complication. The findings of both E. cloacae, a typically intestinal germ (12), and the intestinal necrotic lesions support a hypothesis of an enterogenic origin for this septicemia. The dissemination of the germ was probably not favored by the endoscopy, because it was performed 17 days earlier. It could be speculated that their development was facilitated by the reduction of gastric acid caused by ranitidin and by the immunosuppressive activity of corticosteroids. A relation is uncertain, however, because both drugs are mostly used without infectious complications. In the presented case, the endoscopically observed hemorrhagic ulcers in the duodenum were mistaken for side effects of the steroids and not recognized as a typical manifestation of Henoch-Schönlein (10). Retrospectively, therefore, there is no grounded reason for administering ranitidine. Because the value of steroids in the treatment of Henoch-Schönlein is debated (6,13), the present complication, although exceptional, warns against their administration in patients with Henoch-Schönlein purpura.
1. Lanzkowsky S, Lanzkowsky L, Lanzkowky P. Henoch-Schönlein purpura. Pediatr Rev
2. Amitai Y, Gillis D, Wasserman D, Kochman RH. Henoch-Schönlein purpura in infants. Pediatrics
3. Tapson KM. Henoch Schonlein purpura. Am Fam Physician
4. Barajas de Frutos D, Pedrero Vera J, Bravo Mancheno B, et al. Manifestaciones infrecuentes en la purpura de Schönlein-Henoch. An Esp Pediatr
5. Causey Al, Woodall BM, Wahl NG, et al. Henoch-Schönlein: four cases and a review. J Emerg Med
6. Szer IS. Henoch-Schönlein purpura. Curr Opin Rheumatol
7. Connolly B, O'Holpin D. Sonographic evaluation of the abdomen in Henoch Schönlein purpura. Clin Radiol
8. Kagimoto S. Duodenal findings on ultrasounds in children with Schönlein Henoch purpura and gastrointestinal symptoms. J Pediatr Gastroenterol Nutr
9. Sasaki K, Nukuda Y, Maxida T, et al. Endoscopically and histologically documented gastrointestinal lesions in an adult patient with Henoch-Schönlein purpura. Endoscopy
10. Kato S, Shibuya H, Naganuma H, et al. Gastrointestinal endoscopy in Henoch Schonlein purpura. Eur J Pediatr
11. Agha FP, Nostrant TT, Keren DF. Leucocytoclastic vasculitis (hypersensitivity angiitis) of the small bowel presenting with severe gastrointestinal Hemorrhage. Am J Gastroenterol
12. Andresen J, Asmar BI, Dajani AS. Increasing Enterobacter
bacteremia in pediatric patients. Pediatr Infect Dis J
13. Bodemer C. Vascularites allergiques chez l'enfant. Ann Pediatr Paris