Gastrointestinal Bleeding after Hemolytic Uremic Syndrome: A Report of Drug-Associated Granulation Tissue Polyps.

A previously healthy 4-year-old girl was admitted for bloody diarrhea, hemolytic anemia, thrombocytopenia, and acute kidney injury. She and her sibling had visited a local petting zoo with endemic cases of Shiga toxin-producing Escherichia coli hemolytic uremic syndrome. Two days after admission, she developed anuric renal failure and needed peritoneal dialysis. Her course was further complicated by bacterial and fungal peritonitis, requiring broad-spectrum anti-infective therapy and transition to hemodialysis. She underwent emergent total colectomy with ileostomy creation for colonic necrosis and perforation. Pathology ultimately revealed cytomegalovirus (CMV) colitis with concurrent plasma CMV viral load of 61,095 IU/mL. She was treated with ganciclovir, with undetectable CMV viral load 5 days later. Hemodialysis was stopped 7 weeks after admission as urine output improved. Sevelamer was started as a phosphate binder and sodium polystyrene sulfonate (SPS) was added to chelate potassium. Several months later, she underwent ileostomy takedown with ileorectal anastomosis and gastrostomy tube placement to facilitate enteral nutrition. Almost a month after reanastomosis, she developed painless rectal bleeding. Hemoglobin remained stable throughout this period. Because of the concern for SPS-related colitis, SPS was discontinued, and Renastart (Vitaflo, Alexandria, VA) a formula low in potassium, protein, calcium, and phosphorus, was initiated. Hematochezia continued and colonoscopy was performed 5 weeks after ileorectal anastomosis, showing a polypoid lesion with ulceration at the junction between small bowel and remaining large intestine (Fig. 1A). This ulceration was thought to represent an anastomotic ulcer (Fig. 1B). Sulfasalazine was started with minimal improvement. Repeat colonoscopy was performed 4 months later and showed 2 friable polyps at the anastomotic site, with resolution of the ulcerated area (Fig. 1C). The polyps were removed and sent to pathology. Hematochezia resolved soon after polypectomy, with no further need for sulfasalazine.FIGURE 1: A, Polypoid lesion (arrow) and shallow ulceration at the site of anastomosis between the small and large intestine. B, Anastomotic ulcer, close-up view (arrow). No active bleeding was visualized at this site at the time of procedure. C, Follow-up colonoscopy with resolution of the anastomotic ulcer but continued presence of discrete, friable, polypoid lesions.Histologic sections of the polyps demonstrated prominent granulation tissue characterized by inflamed, edematous stroma with numerous thin-walled vessels. Embedded within the stroma of the polyps were 2 distinct crystalline deposits, 1 purple and cracked—consistent with SPS (Fig. 2), and 1 with both yellow-pink and eosinophilic fragments—consistent with sevelamer (Fig. 3).FIGURE 2: Hematoxylin and eosin (H&E) stain. Sodium polystyrene sulfonate appears violet.FIGURE 3: Hematoxylin and eosin (H&E) stain. Deeply eosinophilic fragments with irregularly spaced cracks consistent with sevelamer.DISCUSSION Certain medications have been associated with gastrointestinal (GI) mucosal injury, and characteristic histopathologic features help distinguish them from others that may represent incidental findings. Sevelamer is classically 2-toned, pink to yellow/brown on hematoxylin and eosin stain. In ulcerated or necrotic tissue, the crystals can take on a brown or deeply eosinophilic hue as seen in this case (Fig. 3). Sevelamer has irregularly spaced cracks that give it a “fish-scale” appearance. This feature is shared with SPS, although cracks associated with SPS are often narrower and intersect with each other. SPS is violet on hematoxylin and eosin stain, a further distinguishing feature. Drug-related GI injuries are well described in the medical literature, but can go unrecognized, especially in the pediatric population. Among such injuries, concomitant use of SPS and sorbitol can cause intestinal necrosis, leading to fatal outcomes (1). This prompted a black box warning on rectally administered SPS, and the removal of 70% sorbitol containing SPS from the market by the Food and Drug Administration in 2006. Despite these interventions, there continue to be multiple reports linking SPS-without-sorbitol to colonic ulceration, necrosis, and ischemia (2). The exact mechanism by which SPS-in-sorbitol causes GI necrosis is unclear. Sorbitol was initially identified as the culprit for ischemic enterocolitis, presumably via a mechanism of osmotic ischemia or prostaglandin effect (3). Interestingly, cases of intestinal necrosis have been reported even in patients receiving “sorbitol-free” SPS enemas (4). A systematic review of case reports found that 91% of patients with SPS-induced colonic perforation had a history of kidney disease. Although this observation may be confounded by indication, patients with renal failure may be more vulnerable to injury (5). Elevated levels of renin seen in renal failure can activate angiotensin II, leading to splanchnic vasoconstriction and mesenteric ischemia. Although the association of SPS with ulceration is well described, reactive polyp formation is unusual. Non-neoplastic inflammatory polyps can form in the colon as a response to increases in C-reactive protein, C4, and procollagen peptide (6). They are often seen at sites of surgical anastomosis, as in our patient. It is possible that transient ischemia induced by SPS led to mucosal injury, resulting in granulation tissue polyp formation (7). Injury due to sevelamer, an orally administered ion-exchange resin used to lower phosphate levels, was also considered in our patient. It too has been implicated in mucosal injury in the GI tract. The mechanism of sevelamer injury in the GI tract is unknown, but it is thought that sevelamer may predispose an individual to dysmotility. Our patient underwent total colectomy before the development of rectal bleeding, which may have affected her baseline GI motility. Sevelamer use may have exacerbated this, leading to prolonged local retention of both sevelamer and SPS crystals, with subsequent mucosal injury. In summary, we describe a unique case of drug-induced GI tract injury in a complex pediatric patient with prior total colectomy and chronic kidney disease. To our knowledge, this is the first report of reactive polyp formation associated with these medications in children. Characteristic crystals identified on histology suggest that SPS and sevelamer, commonly used medications in patients with kidney injury, played a role in the development of the reactive polyps and ulceration. Further data are required to better understand the mechanism of injury and the effects of these medications on GI motility.