Azathioprine in the management of enteropathy in cystic fibrosis

Introduction

While respiratory complications remain the main cause of morbidity and mortality in patients with cystic fibrosis (CF), gastrointestinal (GI) problems are also a significant cause of morbidity. CF research has concentrated on the inflammation in the airways, but there is now increasing evidence of inflammation in the GI tract as well, leading some clinicians to coin the term CF enteropathy. The cause of this inflammation and the best therapeutic options are still to be determined. We describe a case of a 16-year-old boy with CF and enteropathy which was treated with azathioprine.

DECLARATIONS

Competing interests

None declared

Funding

None

Ethical approval

Not applicable

Guarantor

HLT

Contributorship

All authors contributed equally

Case report

Our patient was diagnosed with CF (genotype G542X/621 + 1G > T) at the age of 5 months following investigations for recurrent respiratory infections and failure to thrive. After the initiation of medical therapy, he remained generally well until the age of 10 years when his growth started to falter, with his height and weight dropping from the 25th to the 5th centile.

Investigations for CF-related diabetes were negative – his blood sugar profile was normal and he had a normal oral glucose tolerance test. His appetite deteriorated and he started having problems with abdominal pain. He received considerable input from the dietitians who made sure his creon dose was optimal and tried him on a range of nutritional supplements. Aged 11 years, he had a gastrostomy inserted and was commenced on overnight feeds. His regular medication included ranitidine, omeprazole and domperidone. He was chronically infected with Pseudomonas aeruginosa and was on long-term nebulized colomycin alternating monthly with tobramycin, nebulized rhDNase, oral azithromycin, flucloxacillin and three-monthly intravenous antibiotics.

At the age of 12 years, his GI symptoms worsened, with nausea and vomiting, poor appetite and increased abdominal pain. He reported persistent effortless vomiting of food eaten several hours beforehand, suggestive of delayed gastric emptying.

Further investigations performed included an upper GI contrast study, coeliac screen and mixed food specific IgE, which were all negative. A pH study was negative for gastroesophageal reflux. However, a gastric emptying study (milk scan) revealed very slow gastric emptying, with approximately 80% of the gastric contents still within the stomach after 1 h. Delayed images acquired at 2 h still showed persisting activity within the stomach with very little activity in the small bowel (Figure 1). This slow gastric motility was confirmed on an electrogastrogram.

Figure 1

Gastric emptying scan showing marked delay in gastric emptying, with persisting activity within the stomach and very little activity in the small bowel after 2 hours

He proceeded to have an upper GI endoscopy, which showed erythema of the gastric mucosa. The histology from the antral biopsy showed a marked increase in the lamina propria of plasma cells. This included plasma cell clustering, indicative of moderate chronic inactive gastritis (Figure 2). No evidence of active neutrophilic gastritis or Helico-bacter was seen. The small bowel biopsy showed an increase in the lamina propria inflammatory cell density including eosinophils (Figure 3).

Figure 2

Gastric biopsy showing marked increase in lamina propria plasma cells including plasma cell clustering, indicative of chronic inactive gastritis (in colour online)

Figure 3

Duodenal biopsy showing an increase in lamina propria inflammatory cell density including eosinophils (in colour online)

A course of prednisolone was commenced that led to a complete cessation of his symptoms, but when weaned, his symptoms recurred. To help reduce his reliance on steroids, his gastrostomy was changed to a gastrojejunostomy to bypass the problem of delayed gastric emptying and he was started on E028 elemental feeds. After several failed attempts to wean the prednisolone over a period of 4 months and a DEXA scan showing a Z score of –1.97, suggestive of potential bone complications, a steroid-sparing agent, azathioprine, was started. On this combined approach, he was successfully weaned off steroids.

He has been on maintenance azathioprine for over 2 years and has had one relapse a year after its introduction, when he was admitted with feed intolerance and severe abdominal pain. He also had a concurrent respiratory exacerbation with increased cough and haemoptysis. This resolved on a combination of intravenous antibiotics and pulsed methylprednisolone.

Presently his BMI is 15.8 compared to 14.7, when he first started azathioprine. He is eating more during the day and tolerating 1000 mL of Peptamen, a whey-based hydrolysate feed, overnight. His symptoms of nausea and abdominal pain have improved, though there are still significant concerns with regards to his appetite and growth. An upper GI endoscopy was recently repeated (3 years after the initial one). This showed that his gastric antral mucosa was now normal, though the small bowel mucosa still showed a mild increase in eosinophils. Colonoscopy showed a capacious looking colon with normal histology. Since commencing azathioprine, he has had no increase in the frequency of pulmonary exacerbations.

Discussion

CF transmembrane regulator (CFTR) is expressed in epithelial cells in the GI tract. It has multiple functions, being much more than just an ion channel regulating chloride secretion and sodium absorption. In view of the marked inflammation in the CF airway, it has long been postulated that there is likely to be a similar chronic inflammatory response in the CF gut. Previous studies from the 1960s showed a thick mucus layer, but normal histology in duodenal biopsies from CF children,^1,2 However, with the advent of newer laboratory techniques, there is now increasing evidence for the presence of inflammation in the CF GI tract.

The GI phenotype of CF mice is similar to that in humans, making them a good model for the study of the GI aspects of the disease. Using DNA microarray analysis, it has been shown that there is upregulation of genes involved in inflammation, particularly those associated with innate immunity in the CF mouse intestine. Histologically, there are increased numbers of mast cells and neutrophils in the small intestine of these mice. ³

The concept of immune activation in the CF GI tract is supported by the findings of whole gut lavage performed in clinically well CF children without any GI complications when compared to healthy controls – in the CF children, there was an increased intestinal output of albumin, IgG, IgM, eosinophil cationic protein, neutrophil elas-tase, IL-1? and IL-8. ⁴ Non-invasive markers of intestinal inflammation, faecal calprotectin and rectal nitric oxide (NO) production have also been shown to be raised in paediatric CF patients; with levels of rectal NO production being comparable to children with inflammatory bowel disease. ⁵

Capsule endoscopy examination of the small intestine of CF patients (age 10 years and above) has shown evidence of inflammation in their small intestine. Changes seen ranged from areas of oedema, erythema, villous blunting and denudation of mucosa to frank ulceration. Changes were also seen in some pancreatic sufficient patients. ⁶ CF patients with meconium ileus and distal ileal obstruction syndrome have also been shown to have lymphocytic leimyositis and myen-teric ganglionitis. ⁷

In our patient, it initially did seem surprising, given the refractory nature of his symptoms, that the inflammation seen in the small intestine biopsy was not more severe. However, this is similar to the findings of Raia et al. who, while reporting an increased expression of ICAM1, CD25, IL-2 and IFNΓ in lamina propria mono-nuclear cells of the CF small intestine, found no gross morphological changes. ⁸ Furthermore, it needs to be acknowledged that a small biopsy may not necessarily be representative of the whole bowel and varying degrees of severity may be present throughout the bowel.

The cause of this GI inflammation in CF is still unknown. Several hypotheses have been postulated: the CF intestine could be inherently proinflammatory due to the underlying defect in CFTR. Alternatively, the inflammation could be a result of the chronic administration of pancreatic enzymes, increased permeability, small bowel bacterial overgrowth, or a combination of these factors.

We postulate that our patient's delayed gastric emptying was secondary to chronic inflammation, as intestinal inflammation has been shown to have an adverse effect on GI motility.^9,10 Targeting chronic inflammation in the gut is likely to explain the resolution of our patient's symptoms following the introduction of steroids and azathioprine. Azathioprine is a prodrug for the purine antagonist mercaptopurine. It is an anti-metabolite, often used for immunosuppression in the treatment of autoimmune diseases. ¹¹ It has been used in CF mainly in the context of immuno-suppression post lung transplantation. In GI disease, azathioprine has been used with good effect in inflammatory bowel disease. There are data showing it is effective in maintaining mucosal healing and histological remission in steroid dependent Crohn's disease.^12–14 In paediatric Crohn's patients, the early use of immuno-modulators such as azathioprine is associated with reduced corticosteroid exposure and possibly fewer hospital admissions. ¹⁵

Conclusion

In conclusion, in a carefully selected subset of patients with severe GI symptoms and evidence of GI inflammation on biopsy, azathioprine could potentially be a useful steroid-sparing agent. Further research into the causes of intestinal inflammation in CF and effective therapies is warranted.