Anastomotic dehiscence following retroflexion in surveillance colonoscopy: a case report

Abstract

Anastomotic dehiscence following colonoscopy for routine surveillance after anterior resection for colorectal cancer is unreported in the English literature. It is a potentially fatal complication requiring awareness, quick recognition and management. We present the case of a 45-year-old woman who presented 12 hours after a routine follow-up colonoscopy with peritonitis due to anastomotic rupture diagnosed on computed tomography. The patient was taken to theatre for emergency laparotomy and formation of an end colostomy. Her postoperative recovery and follow-up were optimal.

Background

Routine colonoscopic surveillance is internationally recommended for all patients who have undergone any bowel surgery for cancer including rectal cancer.¹ Retroflexion of the colonoscope in the rectum is a mandatory recommended step during colonoscopy to ensure that the low rectum and anus are examined fully and to allow detection of lesions that could otherwise be missed by forward viewing alone.² The technique involves the endoscopist deflecting the tip of the colonoscope up to 180 degrees, allowing better visualisation of the rectal canal.³ Perforation during retroflexion is rare, with one study quoting a perforation rate of 0.01%.⁴ All reported cases of rectal perforation to date are in patients who have an intact rectum.⁴ Complications following colonoscopy after low anterior resection have not been reported to date; this is the first reported case of bowel perforation subsequent to retroflexion during colonoscopy for routine post-cancer surgery follow-up.

Case history

A 45-year-old woman initially presented with a partially obstructing low rectal cancer, 5cm from the anal verge, subsequently staged as a T4aN2b tumour by magnetic resonance imaging of the rectum and pelvis. Laparoscopy and defunctioning loop ileostomy was created for impending complete obstruction, following which she was offered long-course chemoradiation therapy to downstage the cancer, to which she showed good response with no evidence of distant metastasis. Three months later, the patient underwent laparoscopic low anterior resection with curative intent. The postoperative period was uneventful with no evidence of anastomotic leak or collection. Subsequently her loop ileostomy was closed. She had good postoperative recovery and opened her bowels regularly with good continence.

The patient was followed up with annual surveillance computed tomography (CT) and, nearly two years following her anterior resection, she underwent her first surveillance colonoscopy. During the procedure there was an aborted attempt at performing retroflexion of the colonoscope in the narrow neorectum. Following the colonoscopy the patient complained of passing metal clips per rectum and was admitted to hospital within 12 hours with features of shock and peritonitis. After adequate resuscitation CT was arranged, which showed evidence of dehiscence of rectal anastomosis with pneumoperitoneum (Fig 1). At laparotomy, fluid, pus and free intestinal contents were seen in the abdomen, and a complete disruption of the low colorectal anastomosis was noted, which required Hartmann’s procedure and formation of an end colostomy.

Figure 1.

Computed tomography image showing evidence of perforation of rectal anastomosis (highlighted on the image) with pneumoperitoneum.

Postoperatively, the patient developed pelvic collection, which was drained with ultrasound guidance. The patient was discharged home from hospital on the 25th postoperative day, with continuing input from the colorectal nurse specialists and outpatient follow-up with the colorectal surgeon.

At five weeks post-laparotomy and Hartmann’s procedure, the patient had a healthy, functioning stoma with two small midline wounds, which were healing well. It was decided that the stoma would not be reversed given the technical challenge and high risk of reversal.

Discussion

Rectal retroflexion has been added to be one of the key performance indicators during lower gastrointestinal endoscopy, resulting in an increasing number of patients undergoing the manoeuvre. Although the step is often painful for the patient and there have been reports of de novo perforations, the complication rate is low and the potential for detecting lesions undetectable by straight viewing justify its use.⁴ However, the role of rectal retroflexion during surveillance colonoscopy in patients following cancer surgery, specifically rectal cancer surgery, remains undefined. These patients would normally have had a full index colonoscopy with retroflexion prior to surgery.

Given that the full and intact rectum is no longer retained in these patients, retroflexion during surveillance colonoscopy would singularly depend on the pliability and the ability of the neorectum to accommodate the retroflexed colonoscope. Studies have shown that the neorectum has a reduction in the compliance and reservoir capacity.⁵ This could perhaps have been the underlying cause for the complete dehiscence of the anastomosis in our patient, even at 23 months after surgery. Moreover, the neoadjuvant chemoradiotherapy that the patient received prior to the anterior resection may have been a risk factor for the anastomotic dehiscence in this patient, although it is not possible to substantiate this possibility, and further research is required.

Rectal cancers are common in the UK and the majority of these cases are operated with a primary resection and anastomosis. Routine postoperative colonoscopic surveillance for up to five years is routine for these patients. There is no national or international guidance regarding specific risks that the colonoscopy poses in these postoperative patients other than the routine risks of colonoscopy. We highlight the potential for a life-changing complication of disruption of the colorectal anastomosis during the retroflexion step of colonoscopy in these patients and advice caution when carrying out this step. We also highlight the fact that this step could potentially be avoided in these patients. The diagnostic yield from this recommended step will be potentially negligible given that they would have had a full index colonoscopy prior to surgery.

In summary, clinicians must be aware of this potentially life-changing complication of a recommended routine procedure after cancer resection.