Low-dose intravenous azathioprine may be effective in the management of acute fulminant colitis complicating inflammatory bowel disease
Murch Dr University Department of Paediatric Gastroenterology, Royal Free and University College Medical School, Rowland Hill Street, London NW3 2PF UK.
: The management of acute fulminant colitis unresponsive to intravenous steroids is usually surgical. However, recent evidence suggests that intravenous administration of azathioprine at very high doses may allow more rapid onset of clinical efficacy, although its use has not previously been reported in the emergency situation.
: To report the successful use of intravenous azathioprine in the management of acute fulminant colitis complicating both Crohn’s disease and ulcerative colitis.
: We initially used intravenous azathioprine because of the refusal of the family of the first patient to accept surgery following failure of conventional medical management. Importantly the azathioprine was successful at the low dose of 3 mg/kg.day, equivalent to standard oral doses. Two subsequent patients demonstrated a similar resolution. All were weaned successfully to oral azathioprine and have remained in long-term endoscopic and histological remission.
: These preliminary data suggest that low-dose intravenous azathioprine may be helpful adjunct therapy in selected cases of severe fulminant colitis. However, the need for close monitoring and daily surgical assessment remains paramount, and a formal trial of low-dose intravenous azathioprine is required before it may be more widely recommended.
Acute fulminant colitis which is unresponsive to oral medication is a life-threatening complication of inflammatory bowel disease which often requires emergency surgical intervention.1 Standard medical treatment comprises intravenous steroids with broad spectrum antibiotics. If there is no clinical improvement within 5 days, or if a toxic megacolon develops, considerations of safety dictate that colectomy must be performed as soon as possible, as delay carries significant mortality2–4.
Azathioprine, and its metabolite 6-mercaptopurine, have been widely and successfully used as steroid sparing agents in ulcerative colitis and Crohn’s disease.5–7 Clinical remission has been prolonged in both adults and children8, 9 and in Crohn’s disease resolution of fistulae has been demonstrated.5 The therapeutic effect is variable depending upon dosage, concomitant use of steroids and the duration of treatment.10–13 A meta-analysis of randomized placebo-controlled trials of azathioprine in Crohn’s disease has demonstrated clinical response which becomes significant only after 17 weeks but which increases thereafter.14 This has largely restricted its use in long-term treatment. However, Sandborn and colleagues have made the important observation that the use of a high-dose intravenous loading regime may allow much earlier efficacy,15 thus suggesting a possible extension of its use to include acute severe disease.
In light of the report of Sandborn et al.,15 and the particular circumstances of a child with toxic dilatation of the colon whose family refused colectomy, we embarked on a course of intravenous azathioprine in addition to standard treatment with intravenous steroids and antibiotics. As we were unable to check thiopurine methyltransferase (TPMT) levels before starting treatment we elected for reasons of safety to use low dose azathioprine, equivalent to standard oral dosage, rather than the large loading dose reported by Sandborn et al.15 The clinical response was sufficiently favourable that we have used the therapy electively in a further two cases, with equally favourable outcome.
This 14-year-old boy had an 8-month history of biopsy-confirmed ulcerative colitis and had been maintained in clinical remission on prednisolone and mesalazine despite having persistent mucosal inflammation. He was admitted with severe abdominal pain, profuse diarrhoea and nonbilious vomiting. Examination showed him to be pyrexial and distressed, with abdominal tenderness and guarding, while initial investigations revealed anaemia (Hb 8.5 g/dL) and evidence of systemic inflammation (white cell count (wcc) 17.1 × 109/L with neutrophils 12, platelets 675 × 109/L; C-reactive protein (CRP) 16 mg/L, erythrocyte sedimentation rate (ESR) 10 mm/h and albumin 28 g/L). Abdominal X-ray showed moderate colonic dilatation of 5 cm and mucosal thickening. No organism or toxin was isolated from his stool.
Intravenous hydrocortisone (1 mg/kg q.d.s.), metronidazole and cefotaxime resulted in a gradual improvement over the next 4 days. At this stage oral prednisolone (1 mg/kg) was substituted for the intravenous hydrocortisone. The abdominal pain and diarrhoea worsened although the abdominal X-ray did not show any further colonic distension. The white cell count increased to 26.8 × 109/L (neutrophils 24 × 109); platelets to 826 × 109/L and CRP 37 mg/L. Intravenous hydrocortisone was thus recommenced. However, there was no symptomatic resolution over the next 2 days. Urgent total colectomy was strongly recommended but refused by his parents. Intravenous azathioprine (2.6 mg/kg.day) was thus commenced as a once daily bolus over 30 min. His symptoms improved significantly over the next 3 days and his inflammatory indices reverted to normal (CRP 4 mg/L, ESR 5 mm/h, albumin 34 g/L). Conversion to oral azathioprine at equivalent dosage was followed by prompt clinical relapse with increasing neutrophilia, and was reversed by recommencing intravenous therapy. Subsequent transfer to oral azathioprine was successful 2 days later. He obtained full clinical remission within the next month and follow-up colonoscopy confirmed full mucosal healing. He has subsequently been followed-up for 33 months, and has remained in complete remission on oral azathioprine in conjunction with mesalazine and a minimal dose of prednisolone (2 mg, o.d.).
This 8-year-old girl with a 3-year history of indeterminate colitis presented with a 1-month history of worsening diarrhoea, abdominal pain, weight loss and lethargy. Examination showed erythema nodosum and investigation revealed anaemia (Hb 8.6 g/dL), systemic inflammation (wcc 18.4 × 109/L with neutrophils 15.9, CRP 144 mg/L, ESR 52 mm/h and platelets of 516 × 109/L) and albumin 31 g/L. Colonoscopy and histology demonstrated severe ileo-caecal and colonic Crohn’s disease.
Despite starting enteral nutrition therapy her abdominal pain and diarrhoea continued to worsen over the following 2 weeks, together with deterioration of inflammatory indices (wcc 20.6 × 109/L with neutrophils 18.1, platelets 752 × 109/L, CRP 202 mg/L, albumin 27 g/L). She suffered acute deterioration characterized by nonbilious vomiting, marked increase in bloody diarrhoea and abdominal pain. On examination she was apyrexial but volume depleted, with generalized abdominal tenderness. Plain abdominal X-ray demonstrated 4.5 cm colonic dilatation. Stool microscopy and culture did not reveal any pathogens. Biopsies of sigmoid colon and rectum demonstrated patchy active inflammation.
Intravenous hydrocortisone (1 mg/kg q.d.s.), metronidazole and cefotaxime were started without clinical improvement and without any change in the X-ray appearance. Daily surgical review was instituted and surgery was strongly considered after 4 days. In the light of our (enforced) experience in Case 1, intravenous azathioprine was commenced as a final medical option (2.3 mg/kg.day).
Over the next 4 days her abdominal pain decreased considerably in severity, her diarrhoeal frequency decreased and it contained considerably less blood. Repeat abdominal X-ray showed no colonic dilatation. Her inflammatory indices showed dramatic improvement (wcc 11.7 × 109/L with neutrophils 8.1, platelets 508 × 109/L, CRP 13 mg/L and ESR 18 mm/h, albumin 35 g/L). Following 5 days of intravenous azathioprine she was changed to oral azathioprine (2 mg/kg.day) and prednisolone (1 mg/kg.day).
Colonoscopy performed 2 months after starting azathioprine demonstrated very mild disease limited to the rectum and ileo-caecal valve. The histological series showed minimal patchy chronic inflammation. The prednisolone dose was gradually reduced and stopped after 4 months. She was maintained on azathioprine for 7 months at which time it was withdrawn due to a rising mean cell volume (mcv). She has now been followed for 2 years following the acute exacerbation and has remained in clinical remission throughout this time.
This 8-year-old boy with just a 4-week history of diarrhoea and weight loss was found to have Crohn’s disease extending from ileum to mid-transverse colon. He was started on an enteral feeding regime and aminosalicylate therapy (mesalazine 500 mg b.d.). However, he was re-admitted 7 days later with increasingly severe abdominal pain, vomiting which was occasionally bilious and increasingly frequent bloody diarrhoea. On examination he was markedly unwell, pyrexial and had generalized abdominal tenderness most marked in the right iliac fossa. Plain abdominal X-ray demonstrated colonic dilatation of 5.5 cm. He had an elevated wcc of 19.3 × 109/L with neutrophils of 15.4, platelets 727 × 109/L, CRP 70 mg/L and albumin 29 g/L. Treatment was started with intravenous hydrocortisone 50 mg q.d.s., intravenous metronidazole and ciprofloxicin. Over a 7-day period his clinical condition fluctuated, he continued to pass frequent bloody diarrhoea and required two blood transfusions and two albumin infusions, instituted after further decrease in serum albumin to 25 g/L. Daily abdominal X-rays demonstrated persistent colonic dilatation and his inflammatory markers remained persistently elevated.
Close surgical monitoring was initiated while intravenous azathioprine was commenced (3 mg/kg.day). Over the following 7 days he gradually improved. His abdominal pain abated, his stool form and frequency returned to normal and his abdominal X-ray returned to normal. His inflammatory indices fell (platelets 379 × 109/L, ESR 9 mm/h, CRP 4 mg/L, albumin 34 g/L). By the end of this period it had been possible to reduce his prednisolone dose from 40 mg to 30 mg and to start oral azathioprine (2.3 mg/kg). At follow-up 8 weeks after the initial diagnosis he had remained well. Repeat colonoscopy demonstrated a largely normal colon. Colonic and ileo-caecal biopsies showed marked mucosal improvement. He remains well 12 months after the acute episode and has been able to stop all steroids.
These cases suggest that low-dose intravenous azathioprine may be valuable in the emergency treatment of severe fulminant colitis complicating both ulcerative colitis and Crohn’s disease. The initial decision to use intravenous azathioprine was made in Case 1 when both patient and parents refused surgical intervention. The concept of intravenous loading was developed initially by Sandborn et al. who used a dose of 1800 mg of azathioprine, given as a continuous infusion over 36 h to adults with Crohn’s disease.15 Whereas the clinical advantage of oral azathioprine only becomes evident after ≈ 17 weeks, this regime was effective within 4 weeks.
We were reluctant to use the large dose as used by Sandborn et al. as we had no foreknowledge of the thiopurine methyltransferase (TPMT) levels in these patients and all were gravely ill. This enzyme is important in the metabolism of azathioprine allowing the formation of 6-methyl-mercaptopurine, a metabolite which appears to be biologically inactive. It has a tri-modal activity distribution. There is concern that low levels of TPMT will allow high circulating levels of the 6-thioguanine nucleotides (6-TGN). These nucleotides, acting as purine antagonists, may then cause severe bone-marrow suppression if present at high concentrations. Therefore we chose an intravenous dose in keeping with the oral dose of 2–3 mg/kg.day. We have closely monitored haematological and biochemical indices in these patients, weekly for 4 weeks and subsequently monthly. No adverse events were noted over the course of administration of intravenous azathioprine. Over long-term follow-up the only adverse event noted was an increasing mcv in Case 2. This resolved on cessation of the azathioprine. In two further cases (not presented), in which i.v. steroids and i.v. azathioprine were started simultaneously, similar resolution without side-effects was noted.
In each case initial treatment was with standard medical therapy comprising intravenous hydrocortisone and antibiotics. This was ineffective despite continuation over a period of 4–7 days. Subsequent intravenous azathioprine for a period of 5–7 days coincided with resolution of the fulminant phase of the colitis. In all cases the remission has been maintained for a considerable period, and has been confirmed histologically. In addition, steroid reduction has been possible in all three.
The duration of the intravenous azathioprine was dictated by the clinical condition of the patient. In the cases reported here it was administered for at least 5 days. This regime thus appears to have induced a sufficiently shortened response time to allow its use in the acute situation. In all cases the azathioprine administration could be converted from intravenous to oral after a minimum of 5 days and when the neutrophil count had fallen consistently below 75% of its maximum during the acute exacerbation. Conversion to the oral route before this point was unsuccessful in Case 1 (and in a subsequent unreported case). The duration of administration required for a successful outcome could not be predicted by correlation with white cell count, lymphocyte count, platelet count, CRP or ESR. Therapeutic drug monitoring may possibly allow accurate determination of the required duration of intravenous azathioprine administration, although there is a notably wide variation in effective steady-state concentrations on oral administration. Elevated erythrocyte 6-TGN levels, in adolescents with Crohn’s disease treated with 6-mercaptopurine, have been shown to correlate with efficacy16, although concerns were raised about the design of this study17.
The exact mechanism of action of azathioprine is uncertain. Myelosuppression has been postulated as its main therapeutic role; however, this has not been directly demonstrated18, 19. In our patients, remission occurred in the absence of leucopenia, although a consistent fall of the neutrophil count below 75% of its pre-treatment values did herald improvement. It is notable that the onset of action in our children at this low-dose appeared to be more rapid than reported with the high-doses used by Sandborn et al.15 This probably reflects the immunological differences between chronic severe disease and the acute changes of toxic dilatation in childhood, in which markedly elevated production of cytokines such as tumour necrosis factor-α, well above that found in chronic disease, have been noted.20 Thus the clinical effect of immunosuppressants may well be more striking in the first few days, whereas the time to complete remission may not be shorter. There may also be different effects in adults compared to children. However, this rapid onset of action implies a direct effect on the mucosal inflammatory process, possibly related to in vitro studies on peripheral blood lymphocytes from healthy subjects which demonstrated both impaired cytotoxicity and depletion of natural killer cells in response to 6-MP.21
The exact mechanism of the azathioprine-mediated improvement must remain speculative. However, we postulate that intravenously administered azathioprine may work, at least partly, in a manner distinct to that of the orally administered drug, explaining the accelerated time to clinical effect. In the metabolism of azathioprine, 6-MP is formed via a non-enzymatic reaction as the first intermediary. Because of this the two drugs have generally been considered to be interchangeable. However, several studies have demonstrated improved efficacy of azathioprine over 6-MP,22–24 suggesting a distinct effect mediated by azathioprine itself or by metabolites derived from an alternative pathway. We thus suggest that the rapidity of intravenous administration may exceed the capacity to generate 6-MP, thus allowing higher concentrations of unmetabolized azathioprine or alternative pathway metabolites.
We consider it significant that mucosal healing, both macroscopic and microscopic, occurred in two of these three patients. Thus the avoidance of surgery was not at the expense of continuing inflammation with its attendant risks of dysplasia or neoplasia. It is known that steroids may induce such clinical improvement which does not correlate with endoscopic or histological resolution as demonstrated in case 1.25, 26 In contrast, azathioprine has induced endoscopic healing of severe recurrent Crohn’s ileitis,27 and such mucosal healing is a particularly important goal for the treatment of inflammatory bowel disease in children.28
It is self-evident that definitive conclusions on the role of novel therapies such as low-dose intravenous azathioprine cannot be made from such preliminary reports, and we are currently commencing a controlled trial of its use. However, in advance of such trials we suggest that low-dose intravenous azathioprine should be considered in the management of acute fulminant colitis in selected patients whose only alternative is urgent colectomy. It is mandatory that any patient so treated is maintained under extremely close surgical and radiological supervision, and colectomy should be performed without delay in those not responding.3