Treatment of Chronic Enteritis in Glycogen Storage Disease Type IB with Granulocyte Colny-Stimulating Factor

Glycogen storage disease (GSD) type Ib is a rare inherited disorder of carbohydrate metabolism. It results from a deficiency in the glucose-6-phosphate transport system in the microsomes (1-3). Patients typically have symptoms similar to those with the more common GSD type Ia, i.e., hypoglycemia, lactic acidosis, adiposity, hepatomegaly, and growth retardation. In addition, the GSD type Ib disorder is characterized by neutropenia and neutrophil dysfunction (4-7), predisposing affected individuals to pyogenic infections (8). In 1986, Roe et al. described two adolescent boys with GSD type Ib and associated Crohn's disease-like enteritis (9). Since then there have been other reports of a similar association (10-12). Because chronic mucosal and bowel inflammation is a prevalent problem in neutropenic patients (13-16), it was suggested by Roe et al. that the neutropenia and neutrophil dysfunction in GSD type Ib may predispose these patients to inflammatory bowel disease. The administration of granulocyte-macrophage colony-stimulating factor and granulocyte colony-stimulating factor (G-CSF) to their two patients resulted in a prompt normalization of their neutrophil counts. The correction of neutropenia with G-CSF in patients with GSD type Ib who are without inflammatory bowel disease has subsequently been reported by many observers (17-25). The remarkable finding in Roe et al.'s patients was that in addition to the correction of neutropenia, they experienced a complete resolution in their bowel symptoms and showed radiographic and laboratory evidence of improvement of bowel inflammation (26). A similarly favorable response to treatment with G-CSF was subsequently reported in a 9-year-old girl with GSD type Ib and ulcerative colitis by another group (27). We present here the case of a man with GSD type Ib and associated inflammatory bowel disease who responded to administration of G-CSF with an increase in the absolute neutrophil count (ANC) and resolution of bowel complaints. During treatment with G-CSF, the patient's course was complicated by splenic peliosis and hypersplenism resulting in pancytopenia, which was treated with splenectomy. Postoperatively the patient's hematocrit and platelet count normalized, but he continued to require G-CSF at a lower dose. The patient improved remarkably over a 12-month period. The results of our trial further implicate neutropenia in the pathogenesis of the inflammatory bowel disease seen in GSD type Ib patients. CASE REPORT The patient was an 18-year-old man who had been evaluated at 11 months of age for seizures and hypoglycemia. The diagnosis of GSD type Ib was confirmed by liver biopsy and the characteristic metabolic and hematologic abnormalities. Throughout childhood, he had had chronic infections, recurrent oral ulcers, and perianal abscesses. His glycemic control was achieved with six cornstarch feedings each day (1-2 g/kg by nasogastric tube) and a diet low in simple sugars, but like many patients with GSD, he remained obese (height 10th percentile, weight 90th percentile). In August 1990, he was seen for a 4-week history of decreased appetite, Left Upper Quadrant (LUQ) pain, emesis, and diarrhea. A barium enema (BE) showed narrowing and “thumbprinting” in the ileocecal region, and a colonoscopy showed acute inflammatory changes. The stool was positive for Clostridium difficile toxin. He was diagnosed with presumptive pseudomembranous colitis, and his symptoms cleared with oral vancomycin. During his hospitalization, multiple splenic lesions were noted on abdominal ultrasound. An ultrasound-guided needle biopsy was performed to rule out abscesses, but the lesions suggested hemangiomas on histopathologic examination. The patient was lost to follow-up until he returned in December 1992 with a 2-week history of decreased appetite, a 4-kg weight loss since the previous hospitalization, feeling of fullness, sharp crampy LUQ abdominal pain, perianal fissures, and increased stool frequency. Examination of the stool was negative for blood or infectious organisms. A BE found a stricture of the terminal ileum and cecum (Fig. 1), and an upper-gastrointestinal (GI) series with small-bowel follow-through showed marked delay in gastric emptying and confirmed a stricture in the terminal ileum, cecum, and ascending colon. Esophagogastroduodenoscopy showed endoscopic and histologic changes consistent with reflux esophagitis, but the stomach and proximal small intestine were normal. A colonoscopy showed a normal appearance and histology from the transverse colon to the rectum, but the ascending colon could not be intubated. An abdominal computerized tomogram (CT) showed bowel-wall thickening, narrowed lumen, and perivisceral inflammatory changes at the ileocecal area. Additionally, the CT showed an increased number of low-density splenic lesions and two similar new lesions in the liver. A laparoscopic wedge liver biopsy showed fatty infiltration and increased glycogen stores and contained an endothelial-lined blood space consistent with the diagnosis of peliosis hepatis. Special stains (Warthin-Starry, GMS, and AFB) were negative for microorganisms. During laparoscopy, the splenic surface evidenced multiple areas that showed signs of hemangiomata, which were not biopsies. The splenic biopsy obtained 2 years earlier, however, was reviewed, and the specimen showed changes suggestive of peliosis. A Warthin-Starry stain performed on this biopsy was also negative for microorganisms. Serum taken for Rochalimaea henselae and R. quintana antibody titers to rule out bacillary angiomatosis was negative. Given that the patient's clinical and radiographic abdominal findings appeared to suggest the presence of Crohn-like enteritis, as previously reported in GSD type Ib patients, we considered experimental treatment with G-CSF in January 1993. The family declined to pursue this option at that time owing to the need for frequent diagnostic procedures. The patient was then started on 500 mg olsalazine orally twice daily, 250 mg metronidazole orally three times daily, and nasogastric feedings with a low residue enteral formula (2,000 kcal per day). He continued to lose weight, however (5 kg over the next 8 months), and had persistent anorexia, nausea, nonbloody diarrhea, abdominal pain, fatigue, and recurrent skin infections. In August 1993 he began to complain of left shoulder pain. Findings included anemia [hemoglobin 8.5 g/dl (1.3 mM/L)] and marked splenomegaly, with a spleen tip 11 cm below the costal margin. Another abdominal CT showed apparent resolution of the liver lesions, stable peliosis of the spleen, and increasing hepatosplenomegaly. A bone marrow aspirate showed myeloid hyperplasia and dysplastic erythropoietic and megakaryocytic maturation. Consent to initiate G-CSF treatment in an effort to relieve the patient's chronic gastrointestinal symptoms was granted in September 1993. The clinical investigation protocol for G-CSF (Neupogen) administration was approved by the hospital's institutional review board and by the USAF Surgeon General's Clinical Investigation Committee. The initial G-CSF dosage was 5 mcg/kg/day Subcutaneous (SQ), for a total dose of 300 mcg. Olsalazine and metronidazole were discontinued after G-CSF therapy was started. Weekly weight checks and physical examinations were Performed, and records of oral intake and stool pattern were reviewed. Initial and follow-up laboratory evaluation included a complete blood count with manual differential, serum electrolytes and chemistry panels, erythrocyte sedimentation rate, urinalysis, and stool alpha-1-antitrypsin. Gastrointestinal imaging and endoscopic studies were repeated 3 months and 12 months after starting therapy. The patient's neutrophil count normalized within 2 days of treatment with G-CSF. His pretreatment ANC ranged from 0.3 to 0.9 × 109 cells/L. The initial G-CSF dose of 300 mcg per day maintained the ANC above 2 × 109 cells/L. The G-CSF was temporarily stopped after 13 days because of symptomatic thrombocytopenia (platelet count of 38 × 109 cell/L), with petechiae noted at the injection sites. The platelet count increased during this period, and G-CSF administration was restarted. Symptomatic thrombocytopenia did not recur, but platelet counts remained in the 40-50 × 109 cells/L range. The patient was hospitalized five times during the first 6 months of therapy. Two of the admissions were for self-limitted episodes of acute gastroenteritis. He was admitted in November 1993 to rule out a splenic infarct while having complaints of LUQ abdominal pain. A splenectomy was subsequently performed for persistent pain and a progression of pancytopenia, felt to stem from splenic sequestration. Following splenectomy, lower doses of G-CSF (210 or 150 mcg/day) were required to maintain the ANC between 2 and 7 × 109 cells/L. The thrombocytopenia resolved within a week after surgery, and the anemia gradually resolved over the next 2 months [current hemoglobin level 11.4 g/dl (1.77 mM/L)]. Follow-up bone marrow examinations in November 1993 and February 1994 remained stable, without evidence of aplasia or myeloproliferative changes. The patient was readmitted in December 1993 for urokinase and heparin therapy for portal vein thrombosis, which was identified after surgery. Before G-CSF treatment the patient's erythrocyte sedimentation rate was 90 mm per hour. This value decreased to 46 within 2 weeks of treatment and to 25 after the first month and continued to gradually normalize to a level of 3 mm per hour after 12 months. The stool alpha-1-antitrypsin level declined from 5.4 mg per gram of stool at baseline to 3.7 mg after 3 months (normal 0-2.9 mg). At 8 months into treatment the level was 6.3. The stool alpha-1-antitrypsin level was immediately repeated, and a result of 0.3 was obtained. There were no significant changes in the patient's blood chemistries except for periodic three- to fivefold elevations of the serum aminotransferase levels since December 1993, 1 month after two blood transfusions. Hepatitis serology and polymerase chain reaction for hepatitis C virus RNA have been negative. Three months after initiating therapy, the findings noted in the upper GI series and BE showed only a modest improvement from baseline. The radiographic studies 12 months into treatment showed an increase in luminal filling in the terminal ileum, although a small segment of the ascending colon remained strictured (Fig. 2). A follow-up abdominal CT showed no changes in the appearance of the intestines. A colonoscopy 3 months after treatment was again unsuccessful in reaching the proximal colon, but the distal colonic appearance and histology remained normal. Follow-up esophagogastroduodenoscopy at that time was normal. Since the time treatment with G-CSF was initiated and especially following splenectomy, the patient's gastrointestinal symptoms have gradually subsided. He did require supplemental nocturnal enteral feedings for 1 month following splenectomy. His anorexia, nausea, and abdominal pain resolved thereafter, and he presently eats his usual diet and cornstarch feedings. His stools became formed within the first month of treatment, and the oral and perianal inflammation have not recurred. Beginning 3 months after therapy the patient reported a marked increase in energy, and he has since regained 5 kg of weight. The patient was so debilitated before therapy that he often required ambulation with a wheelchair. Now he actively plays football and tennis (currently playing as a reserve fullback on a varsity football team) and works part-time. The patient was continued on G-CSF after protocol termination (12 months) and remains well at the time of this writing. His only hospitalizations since January 1994 have been for episodes of hypoglycemia associated with dietary indiscretions. DISCUSSION Treatment with G-CSF in this patient with GSD type Ib and Crohn-like enteritis resulted in rapid correction of his neutropenia, followed by resolution of all preexisting gastrointestinal symptoms. The patient's energy level and quality of life improved dramatically. It seems that the absolute number of neutrophils is more important than the functional abnormalities in determining the susceptibility of patients with GSD type Ib to infection (22), and possibly to enteric inflammation. However, the possibility of a role for neutrophil dysfunction in the pathogenesis of Crohn's disease has been raised, considering that in addition to patients with GSD type Ib, subjects with congenital neutropenia (14,28) and chronic granulomatous disease (13,29) have also been reported to have a Crohn-like enteritis. Several abnormalities in neutrophil function in Crohn's disease have already been identified (30-36). Notwithstanding, the effects of G-CSF treatment on neutrophil function remain controversial. Some authors have reported improvements in neutrophil motility (20) or enhanced production of oxygen radicals (19,25) while others have not (26,27). Other treatment methods have been tried in attempts to increase neutrophil counts in patients with GSD type Ib. Although trials using lithium have been unsuccessful (37-39), splenectomy with protacaval transposition has been shown to correct neutropenia in one patient (40). The improvement in neutrophil counts after splenectomy, also documented in our subject, suggests that the spleen may play a role in the pathogenesis of neutropenia in some of these patients, but it does not entirely account for their neutropenic state. Along with the recognized motility abnormalities in the neutrophils of these patients (7), which may impair mobilization from the bone marrow, splenic sequestration of neutrophils may be important. Although hepatic peliosis has been described in two adult patients with GSD type Ib (41,42), the splenic peliosis found in our patient has not been reported in this condition to our knowledge. Other possible explanations for this finding, such as bacillary angiomatosis (43) and steroid use, were not present here. We suggest that splenic peliosis and sequestration may be an unrecognized complication of GSD type Ib and that it may indicate splenectomy in some patients, realizing that the neutropenia may not completely normalize after the procedure. In conclusion, the clinical response of our subject to treatment with G-CSF supports the speculation that neutrophil deficiency may be implicated in the pathogenesis of the Crohn-like enteritis found in patients with GSD type Ib. This is now the fourth subject that we know of with GSD type Ib and chronic enteritis or colitis who has responded favorably to G-CSF treatment. Based on this limited experience (T. F. Roe et al., personal communication), it appears that treatment must be continued indefinitely to sustain clinical improvement in these patients. In addition to the chronic bowel inflammation, our subject had other previously reported features, such as hepatic peliosis and bone marrow dysplasia, but the splenic peliosis is a new finding that may be associated with hypersplenism and may further complicate the clinical course in patients with GSD type Ib.FIG. 1.: The terminal ileum is markedly narrowed (large white arrow). The cecum is also narrowed and irregular. The appendix is seen extending from the cecal tip (small white arrow). (The dense overlying ring is due to compression paddle).FIG. 2.: Small-bowel contrast study shows improvement in the appearance of the terminal ileum and cecum (large white arrow). There is some residual narrowing suggestive of scarring.