Diffuse Fibrosis of the Colon Complicating Cystic Fibrosis

Abstract

Cystic fibrosis (CF) is an inherited disease that primarily compromises pulmonary and pancreatic function but may also affect other secreting epithelial tissues (1). More than 85% of patients with CF also suffer from pancreatic insufficiency, resulting in malabsorption and nutritional growth failure (2). Pancreatic enzyme replacement therapy has been the primary treatment for malabsorption in patients with CF. Capsule preparations that deliver high concentrations of enzyme were developed to improve patient compliance and have subsequently been shown to improve energy retention (3). In earlier reports, the gastrointestinal complications of CF were described chiefly as complications of intestinal obstruction secondary to meconium ileus in infants or inspissated stool in children (4,5). Recently, however, patients with CF have presented with obstructive symptoms related to strictures of the proximal colon associated with marked submucosal fibrosis (6-11). We report three children with CF, two of whom presented with bloody diarrhea and a third with blood-streaked loose stools and abdominal pain. Each patient had diffuse stenosis, fibrosis, and shortening of the entire colon, with sparing of the small intestine. CASE REPORTS Patient 1 A 31-month-old female infant with CF (genotype, homozygous delta F508) presented to the gastroenterology service with a 1-month history of bloody diarrhea without abdominal pain (Table 1). Although she had been treated with oral Vancomycin for Clostridium difficile-positive stools, the symptoms had not resolved. Physical examination showed her height and weight at the 10th percentile; she had clear lungs and a soft, tender abdomen. She had undergone enzyme replacement therapy for the previous 6 months in the form of Pancrease MT 25 (>11,000 IU lipase/kg/meal). Flexible proctoscopy revealed an erythematous and friable mucosa with shallow ulcerations, representing acute colitis both endoscopically and histologically (Fig. 1). A subsequent colonoscopy demonstrated diffuse erythema and friable mucosa with scattered shallow ulcerations. At 40 cm from the anal verge, the colonic lumen abruptly narrowed, preventing advancement of the colonoscope, which had an external diameter of 9.5 mm. A barium enema revealed a short, tubular colon with diffuse loss of haustrations (Fig. 2a). Despite treatment with sulfasalazine and corticosteroids for suspected inflammatory bowel disease, diarrhea and weight loss continued. A subsequent barium enema demonstrated further contraction of the colon, with fixed stenosis of the distal terminal ileum compatible with ileitis (Fig. 2b). Because she developed obstructive symptoms, a colectomy was done. During surgery, the colon was noted to be markedly shortened and displaced toward the pelvis. Gross pathologic examination revealed a markedly shortened colon measuring 33 cm (normal for height, ≈100 cm) with a small amount of attached terminal ileum. The colonic circumference varied from 3.4 cm proximally to 5.4 cm distally; the wall thickness was 0.6 cm at its maximal dimension. The submucosa was white throughout the colon (Fig. 3). Normal mucosal folds were absent, and there were focal mucosal ulcerations. The serosa was normal. Histologically, the colonic architecture was preserved, with only occasional crypt branching and dropout. Rare superficial ulcerations consisted of thinning of the lamina propria with neutrophilic infiltration and overlying regenerative epithelium. The lamina propria of the colon was expanded by amorphous eosinophilic material, but there was a relative paucity of inflammatory cells (Fig. 4). Extensive submucosal fibrosis extended into a disrupted muscularis mucosae. Metaplastic fat was abundant. The muscularis propria was normal, but fibrosis was present in the serosal fat. In contrast, the terminal ileum was normal both grossly and microscopically. Sections of the ileocecal valve demonstrated extensive submucosal fibrosis on the colonic side of the valve, but without the extensive fibrosis in the ileum. When we reviewed the prior colonic biopsies, we observed that the bland, eosinophilic appearance of the lamina propria noted earlier in the colectomy was still present, but to a lesser extent. Patient 2 Patient 2, a 6-year, 9-month-old girl with CF (unknown genotype), presented to the gastroenterology clinic with a 6-month history of vague abdominal pain relieved partially by antacids (Table 1). She was suffering from posttussive emesis, a markedly decreased appetite, and loose stools with flecks of blood. She had been receiving Ultrase MT 24 (>14,500 IU lipase/kg/meal) for 6 months. This patient had a history of meconium ileus, complicated by necrotizing enterocolitis 2 weeks later. Six weeks prior to her current presentation, she had undergone an appendectomy, following a pathologic diagnosis of acute appendicitis. A barium enema had revealed a diffusely short colon with loss of haustrations and linear extravasation arising from the right colon. After recovery, her enzyme dose was increased to the current level of > 14,500 IU lipase/kg/meal. Her physical examination was significant for weight below the fifth percentile and height just above the fifth percentile. Her abdomen was diffusely tender to palpation; tenderness was more pronounced in the periumbilical and epigastric area. Stool was positive for Clostridium difficile using a latex agglutination test, so Vancomycin therapy was begun. Colonoscopy revealed diffusely erythematous mucosa without ulceration. There was obstruction to colonoscope advancement 30 cm from the anal verge. A biopsy yielded results similar to those for patient 1, with the addition of eosinophils and eosinophilic crypt abscesses. The lamina propria showed the homogeneous expansion also noted in patient 1. After reduction of the enzyme dose, the child gained weight, but 4 months later she returned with abdominal distension. A hypaque enema demonstrated severe tubular stenosis of the entire colon proximal to the rectosigmoid and a narrow terminal ileum, with high-grade obstruction of the proximal small bowel. A colectomy was done. The pathologic findings were similar to those seen in patient 1. The colonic length was 26 cm (normal, ≈100 cm); the circumference was uniformly stenotic, measuring 2 cm (Fig. 1). Five centimeters of attached ileum appeared normal. There was diffuse submucosal fibrosis but no ulceration. In addition, the ganglion cells in Auerbach's plexus were increased—a finding not seen in the colectomy specimen from patient 1. Patient 3 Patient 3 was a 22-month-old male infant with CF (genotype unknown) who presented to an outside institution with bloody diarrhea and failure to thrive (Table 1). Enzyme replacement therapy for the past 4 months consisted of Ultrase 24 (> 12,000 IU lipase/kg/meal). Because of poor weight gain, his enzymes were changed to Ultrase 30 (> 14,500 IU lipase/kg/meal). A sigmoidoscopy disclosed diffuse erythema with some ulceration. Because the bloody diarrhea continued, he was referred to our gastroenterology service for further evaluation. His weight was at the fifth percentile and height was at the 25th percentile. His abdomen was soft and nontender. A barium enema demonstrated diffuse shortening and narrowing of the colon with loss of normal haustrations, as well as a contracted cecum with fixed narrowing of the most terminal ileum and mild dilatation of the proximal ileum. There was relative sparing of the rectosigmoid. Colonoscopy revealed a narrow area in the cecum at the ileocecal valve. The colonic mucosa was edematous and friable, and there were linear ulcerations throughout. Proximally, there were deep ulcers with pseudopolyp formation. Biopsies of the colon showed expansion of the lamina propria by homogeneous eosinophilic material, similar to that seen in patients 1 and 2. The cellularity was normal and colitis was not present. A decrease in enzyme supplementation resulted in some improvement in his diarrhea, although he continued to have intermittent bleeding. Therapy with sulfasalazine (40 mg/kg/day) was started. To date, stools are negative for occult blood and he is gaining weight. The sulfasalazine is now being tapered. DISCUSSION Even though respiratory symptoms continue to be the leading cause of morbidity and mortality in patients with CF, malnutrition is also a major concern. Approximately 50% of patients with CF have a weight or height below the 10th percentile for their age-sex group (4). For more than 3 decades, replacement therapy with pancreatic enzymes has been used to treat chronic malabsorption (12). In general, use of these preparations results in decreased steatorrhea and improved growth. Recently, concentrated forms of enzyme replacement have been developed with the aim of increasing patient compliance, leading to improvement of fat absorption and energy balance. Studies of high-dose preparations have shown a reduction in fecal fat output and good patient acceptance of the newer products (2,13). Initially, the patients in the current report were suspected of having inflammatory bowel disease, due to clinical and radiographic changes compatible with that diagnosis. Colitis, consistent with inflammatory bowel disease, has been described in a few patients with CF (5). All three of our patients demonstrated a diffuse colonic abnormality both endoscopically and radiographically, with loss of haustrations and a shortened tubular and rigid appearance. There was relative sparing of the rectosigmoid but no skip lesions. The appearance was reminiscent of the “lead pipe colon” characteristically seen in patients with chronic ulcerative colitis. The lack of histologic changes characteristic of inflammatory bowel disease, along with the striking submucosal fibrosis and expansion of the lamina propria found in these patients, indicates that another process was responsible for the clinical and pathologic findings. Several case reports and small series, mostly from Europe, have described patients with CF and stenosis of the colon (6-11). The patients from England presented with distal intestinal obstruction syndrome and had submucosal fibrosis similar to our cases, but there was involvement of only the proximal colon (6-10). One article from Denmark discussed six patients who presented with watery diarrhea instead of symptoms of obstruction (11). In these cases, as in ours, there was diffuse colonic involvement by the fibrotic process. In most cases, the patients had been taking elevated doses of highstrength pancreatic enzymes, as were our patients. However, three patients from the United States suspected of having colonic fibrosis were taking standard strength enzyme microspheres (9). It is not currently known what role pancreatic enzyme supplements play in the genesis of the colonic fibrosis described in this report. Factors such as the heterogeneity of length of enzyme ingestion prior to symptoms, type of capsule administered, and maximum dosage taken have made evaluation difficult (9). We believe that the pathologic changes in the patients presented in this report are related to their regimen of high doses of pancreatic enzymes, particularly the protease component of the preparation. One striking finding in both colectomy specimens of our patients was the relative lack of involvement of the terminal ileum by the fibrotic process, which we believe may provide insight into the pathogenesis of the lesion. The physiologic process of digestion normally takes place in the small intestine, and pancreatic proteases have markedly diminished activity as they reach the terminal ileum (14). Because digestion occurs there, the small intestine would be expected to have endogenous protection from digestion by luminal proteases. In contrast, the colon would not be expected to contain such defense mechanisms. Support for this hypothesis can be extrapolated from the distribution of α-1-antitrypsin messenger RNA, which is selectively expressed in the enterocytes of the small intestine and not the colon of mice transgenic for the gene for human α-1-antitrypsin (16). Similar studies have not been done for pancreatic secretory trypsin inhibitor, the substance thought to have the greatest antiprotease effect against pancreatic enzymes; however, an analogy may exist. Protease digestion of the colon secondary to excess protease activity from the large doses of enzymes given would damage cell membranes in the colon, thus resulting in a histological picture similar to that seen here. Experimental studies are needed to test this hypothesis. In addition, epidemiologic data should also help us to establish the true relationship between pancreatic enzyme dosage and fibrosis. Clostridium difficile was detected in patients 1 and 2 but not in patient 3. Patients with CF have a high asymptomatic carrier rate of this organism; and there was no evidence endoscopically, radiographically, or pathologically that this organism was a significant pathogen in these cases. The focal acute changes seen in the biopsy specimens of patients 1 and 2 might have been related to Clostridium difficile or stasis; alternatively, they may have been the result of damage by other factors leading to colonic fibrosis. Although the ulcers may have been the source of the melena, diffuse colonic fibrosis is the primary pathologic change responsible for the morbidity seen in our patients. These patients presented a confusing picture both endoscopically and radiographically. Because the colon was markedly shortened, the terminal ileum was displaced distally and thus misinterpreted as a colonic stricture by the endoscopist. The radiographic findings of patient 1 were initially thought to represent inflammatory bowel disease, a diagnosis not supported by subsequent pathologic findings. In summary, we have presented three patients with diffuse fibrosis of the colon who not only developed symptoms of obstruction, but also presented with bloody diarrhea. Although experimental data are lacking, the circumstantial evidence suggests that high doses of pancreatic enzymes are responsible for the fibrotic process. When such characteristic radiologic and pathologic changes are noted, the team of medical specialists caring for children with CF should maintain a high level of clinical suspicion for this new complication of the disease.FIG. 1.: Patient 1. The colonic architecture is preserved, with occasional acute inflammatory cells in the lamina propria (arrow). There is a subtle expansion of the lamina propria, imparting a homogeneous appearance (*). (H&E, 1 cm bar = 0.1 mm)FIG. 2.: Patient 1. A: A prone view of barium enema shows a shortened, featureless colon with diffuse loss of haustrations. B: A barium enema 5 months later reveals marked progression of colonic narrowing with fixed stenosis of the terminal ileum.FIG. 3.: Patient 1. The colon is opened longitudinally and the walls reflected to display a flattened mucosal surface and a thickened, fibrotic wall. The striking white band (arrow) corresponds to extensive submucosal fibrosis. (1 cm bar = 1.1 mm)FIG. 4.: Patient 1. A trichrome stain of the colon highlights the dense submucosal fibrosis, which contains metaplastic fat. The fibrosis has disrupted the muscularis mucosae (arrow). The lamina propria is paucicellular and has been expanded by amorphous material (photo and inset). (large photo: 1 cm bar = 0.2 mm; inset: 1 cm bar =.05 mm)