Article Text

Download PDFPDF

Gas gangrene after colonoscopy
  1. N F Jamieson,
  2. C P Willoughby
  1. Gastroenterology Unit, Basildon Hospital, Nether Mayne, Basildon, Essex SS16 5NL, UK
  1. Dr Willoughby


A case of spontaneous clostridial myonecrosis developing shortly after diagnostic colonoscopy is described. The prime underlying factor proved to be an unsuspected colonic cancer, developing in a patient with pre-existing ulcerative colitis and sclerosing cholangitis.

  • gas gangrene
  • colonoscopy

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Gas gangrene was originally described as a complication of battlefield wounds and other penetrating injuries. Spontaneous clostridial myonecrosis can occur in patients with predisposing conditions such as diabetes, leukaemia, and other neoplasms.1 We describe a case occurring shortly after a diagnostic colonoscopy, but the apparent causative association between the two events proved incorrect.

Case report

A 48 year old man was admitted to hospital as a day case for colonoscopy. He had a 25 year history of extensive ulcerative colitis and was also found to have primary sclerosing cholangitis. This was diagnosed by endoscopic retrograde cholangiography arranged after persistently abnormal liver function tests were noted in 1984 (fig 1). Previous regular surveillance colonoscopies had shown low grade inflammation to the caecal pole but no dysplastic features on examination of biopsy specimens; the latest colonoscopy had been two and a half years before the present admission.

Figure 1

Endoscopic retrograde cholangiography film showing typical changes of sclerosing cholangitis in the patient.

The current examination had been booked because of two attacks of abdominal pain and constipation. Full blood count and erythrocyte sedimentation rate were normal when checked in the outpatient department. Colonoscopy was carried out under intravenous sedation given through a plastic cannula in the right antecubital fossa.

The colonoscope was passed to the hepatic flexure, but the right side of the colon could not be seen adequately. Multiple biopsy specimens were taken from the large bowel mucosa for dysplasia screening. As there was quite a lot of blood in the proximal transverse colon and a caecal lesion could not be excluded on this incomplete examination, a barium enema was arranged to evaluate the proximal large bowel.

Two days after the colonoscopy the patient was readmitted to hospital with pain and swelling in the right arm for the previous 12 hours. He was feverish and unwell. The right upper arm was markedly swollen, with mottling of the skin and some blister formation at the elbow. The initial differential diagnosis was of a cellulitis or axillary vein thrombosis. The latter was excluded by a Doppler ultrasound scan, while a plain x ray revealed gas in the tissue planes (fig 2).Clostridium septicum was isolated from multiple blood cultures.

Figure 2

Plain x ray of the right arm, showing gas in the tissue planes (arrows).

An urgent surgical debridement of the upper arm was performed and repeated four times during the next week. High dose intravenous penicillin was given, together with intravenous metronidazole and ofloxacin.

Ten days after his admission the patient had a large rectal bleed followed by increasing abdominal pain and distension. A plain abdominal film showed dilated small bowel loops suggestive of obstruction. At laparotomy, a perforated caecal carcinoma was found. The liver showed macronodular cirrhotic changes. A right hemicolectomy was performed, but the patient developed disseminated intravascular coagulopathy and hepatorenal failure, leading to death eight days after the abdominal surgery.


This patient's admission with gas gangrene two days after colonoscopy caused understandable concern to the gastroenterology unit, particularly as the indwelling cannula for the intravenous sedation had been placed close to the site of the later development of myonecrosis. As matters evolved, it became evident that the condition was likely to be secondary to the underlying colorectal cancer. Spontaneous myonecrosis caused by Clostridium septicumis associated with colonic cancer in about one third of cases.2 The organism probably gains access to the blood stream through mucosal defects in the large bowel caused by an ulcerating tumour, and may be particularly likely to infect normal areas of muscle, as it is more aerotolerant thanC perfringens. 3 Release of exotoxins results in haemolysis, vascular injury, cytolysis, and inhibition of neutrophil function, leading to massive tissue necrosis. Treatment involves aggressive surgical debridement or amputation in cases of limb involvement, together with systemic antibiotic treatment. When available, hyperbaric oxygen may help by inhibiting clostridial growth and toxin production, but there are no controlled studies of the clinical value of such treatment.4 Even with aggressive therapy, the prognosis of spontaneous myonecrosis is poor, with a mortality of 60–100%.

In this case, the underlying cancer developed on a background of long term extensive ulcerative colitis and primary sclerosing cholangitis. Ulcerative colitis is a well recognised antecedent of colorectal cancer. Since the description of a close relation between colonic neoplasia and mucosal dysplastic changes in ulcerative colitis,5 and the wide availability of colonoscopy in general hospitals, it has become conventional to undertake regular surveillance endoscopy in patients with colitis thought to be at particular risk of cancer development. Although some series have claimed that colonoscopic surveillance reduces the cancer mortality,6-8 a critical analysis of the data available implies that such programmes may be largely ineffective in the detection of tumours at an early and curable stage, while involving significant utilisation of resources for a very limited reward.9 10 Nevertheless, 92% of endoscopy units in the United Kingdom still perform surveillance colonoscopy for ulcerative colitis, in the perhaps optimistic expectation of reducing cancer deaths.11 The chance of developing colonic dysplasia or neoplasia appears to be particularly increased if a colitic patient also has primary sclerosing cholangitis.12 13 As a result, close endoscopic surveillance with multiple biopsy sampling—probably at yearly intervals—has been advised in such cases to try to reduce the mortality from colorectal tumours.14 15 In our patient, domestic circumstances made him a rather poor complier with follow up, and the interval between colonoscopies was arguably longer than ideal.

At necropsy examination, the patient had established liver cirrhosis related to his primary sclerosing cholangitis. Circulatory changes in the liver allowing easier access of colonic organisms to the systemic vasculature may have been an additional risk factor for gas gangrene development; myonecrosis caused by a gas forming coliform has been described in a case of decompensated liver disease and spontaneous bacterial peritonitis.16

Although apparently precipitated by the endoscopic procedure, this case of spontaneous clostridial myonecrosis was almost certainly an example of the known association between that condition and bowel cancer. Patients with long term ulcerative colitis and primary sclerosing cholangitis may require careful surveillance for tumour development, but this report describes a unique presentation of the problem.