While the etiology of Henoch-Schonlein purpura (HSP) remains unclear, at least 50% of children with HSP have a preceding upper respiratory tract infection. Various viruses, such as Epstein-Barr virus, adenovirus, and parvovirus, and bacteria, such as streptococci and mycoplasma pneumoniae, have been linked with this condition (1). We report here an instructive case of HSP associated with Clostridium difficile infection and a severe necrotizing enteritis with pseudomembrane formation. CASE REPORT A previously well 3-year-old Chinese girl presented with a 1-month history of intermittent fever, abdominal pain, and vomiting. The abdominal pain and vomiting were typically worse about half an hour after meals. Vomiting was not bilious and there was neither hematemesis nor diarrhea. Two episodes of rectal bleeding occurred ≈3 weeks after the onset of illness. There were no musculoskeletal complaints. Her appetite was poor and she had lost weight. During this period, she had been seen by a number of general practitioners and had received a few courses of antibiotics, including amoxycillin and cotrimoxazole, without improvement. She was finally referred to the Paediatric Department at University Hospital, Kuala Lumpur. Examination on admission revealed a child who was obviously extremely distressed about her abdominal pain. The abdomen was slightly tender on palpation, but it was soft, with no guarding or rebound tenderness. There were no other major physical abnormalities found. Laboratory investigations showed a raised white count of 47.9 × 109/L, a platelet count of 870 × 109/L, and a normal erythrocyte sedimentation rate of 10 mm/h. The hemoglobin and serum electrolyte levels were normal. Initial blood and urine cultures were sterile. Urinalysis did not reveal any proteinuria, hematuria, or pyuria. A barium meal suggested mucosal thickening in the third segment of the duodenum and the proximal jejunum. An upper gastrointestinal endoscopy showed inflammation of the duodenal mucosa. Colonoscopic examination was normal. A laparotomy was performed in view of the history of severe abdominal pain and vomiting after meals. Multiple segments of inflamed jejunum were found, with fibrinous exudate on the serosa (Fig. 1). These segments were red, thickened, and of a leathery consistency. A 5-cm segment of inflamed jejunum was resected and an end-to-end anastomosis performed. Pseudomembranes could be seen on the mucosal aspect of the resected bowel (Fig. 2). The ileum and colon were normal. Histopathology of the resected segment revealed heavy transmural infiltration of the intestinal wall by acute inflammatory cells, mainly neutrophils. Large areas of mucosal ulceration were present (Fig. 3) and these were associated with heavy exudates of fibrin and neutrophils, forming pseudomembranes. Intervening nonulcerated mucosa appeared as islands surrounded by inflammatory exudate (Fig. 4). No granulomas or epithelial dysplasia was present. There was no evidence of vasculitis. After laparotomy, the child was given intravenous ceftriaxone, gentamicin, and metronidazole and supported with parenteral nutrition for 2 weeks to rest the bowel. On the second day after surgery, the child developed purpuric spots in the lower limbs up to the buttocks. There was also evidence of proteinuria and hematuria. At about this time Clostridium difficile toxin was detected in the stools by enzyme immunoassay. The organism was also cultured from the stools. The child improved on parenteral nutrition, intravenous metronidazole, and vancomycin followed by oral cholestyramine. Two weeks after the laparotomy, the child developed septicemic shock caused by an Enterobacter species, which responded to imipenem, intravenous fluids, and inotropic support. She improved and was discharged on the twenty-eighth day after operation. She appeared to be fully recovered at follow-up 2 weeks after discharge. DISCUSSION Abdominal manifestations of HSP are well-known (2) and include abdominal pain, gastrointestinal bleeding, nausea, vomiting, intussusception, and perforation. Antibiotic-associated pseudomembranous colitis, although rare, has also been reported in association with HSP by Narchi et al. (3). In their report, there were two cases of HSP with pseudomembranous colitis, but in only one of them could Clostridium difficile be detected. While the finding of Clostridium difficile in the feces of an adult is usually associated with symptoms, positive identification of the organism in the stool can occur asymptomatically in ≈40% of healthy neonates and infants (4). Nonetheless, if this is associated with detection of toxin and intestinal inflammation, it is probably wise to assume that the organism has a causative role and to treat accordingly. It is noteworthy that in this case there was no evidence of intestinal vasculitis as may be found in a Henoch-Schonlein enteritis, although the absence of demonstrable vasculitis does not exclude the diagnosis. On the other hand, it may be that the pathology here was due to a pseudomembranous inflammatory process, in which vasculitis is less prominent a feature, rather than a typical Henoch-Schonlein enteritis. In Narchi and co-workers' first patient, too, there was a lack of evidence of vasculitis in the bowel. It is also interesting that there was no colonic involvement to suggest classical pseudomembranous colitis. Nevertheless, the detection of Clostridium difficile and its toxin in the stool of a child with necrotizing enteritis and pseudomembrane formation suggests that a process similar to pseudomembranous colitis might be occurring in the small intestine. This was also felt to be a strong indication for the use of anticlostridial therapy. It remains debatable whether HSP along with multiple antibiotic usage had led to colonization by Clostridium difficile and the development of “pseudomembranous enteritis” or whether Clostridium difficile infection had, in fact, predisposed to the development of HSP. In any event, the presence of abdominal pain, vomiting, gastrointestinal bleeding, and thrombocytosis should make one suspect HSP as a differential diagnosis even in the absence of joint involvement. Acute arthritis, affecting most frequently the knees and the ankles, occurs in 60 to 84% of patients (1). The development of classical purpuric lesions may occur late, as demonstrated by this case. In general, palpable nonthrombocytopenic purpura, which is a prerequisite for the diagnosis of HSP, is a presenting sign in only 50% of cases (1). In most reports the typical rash, though not the presenting sign, appears within the first week (5,6). Lind et al. reported the case of a 15-year-old boy who developed HSP associated with Campylobacter jejuni enteritis in which a hemorrhagic rash was noted only 1 week after initial presentation (7). The recommended treatment of C. difficile enteritis is oral vancomycin. Metronidazole by either the intravenous or the oral route is an alternative. In the present case, the diagnosis of C. difficile infection was made in the immediate postoperative period, when the child was suffering from an extended period of ileus that resulted in an inability to tolerate oral medication. In such a situation, intravenous metronidazole alone is usually recommended. In this patient, intravenous vancomycin was also given but the extent to which it played a part in the organism's response is unclear. After the course of antibiotics and the resolution of ileus, oral cholestyramine was given, with the aim of preventing relapses, based on the regime recommended by Roy et al. (8). Cholestyramine is thought to act by binding the toxin of C. difficile and some clinical success has been reported with its use (9). Nevertheless, the effective prevention of relapses is still a difficult problem. Current evidence of the efficacy of cholestyramine is inconclusive. Cholestyramine can be unpleasant for a patient to take and many centers do not use it. We also wish to emphasize the point that indiscriminate use of antibiotics should be strongly discouraged in HSP. Finally, this case demonstrates that C. difficile-related pseudomembranous intestinal inflammation may affect the proximal small bowel while sparing the ileum and colon. Acknowledgment: The authors thank their colleagues in the Departments of Paediatrics, Surgery, and Pathology for their assistance in the management of this case.FIG. 1: . Inflamed jejunum with fibrinous exudate on the serosa.FIG. 2: . The mucosal aspect of resected jejunum showing the presence of pseudomembranes.FIG. 3: . Photomicrograph of resected jejunum showing mucosal ulceration associated with heavy fibrinous and acute inflammatory exudation. Hematoxylin and eosin; original magnification, ×150.FIG. 4: . One of numerous islands of intact mucosa covered by a pseudomembrane of acute inflammatory exudate. Hematoxylin and eosin; original magnification, ×150.
No takes yet. Share an insight, caveat, or question.
Boey et al. (1997) studied this question.