The outcome of surgery in fulminant Clostridium difficile colitis

Authors

  • K. Koss,

    1. *Gastrointestinal Unit and †Academic Department of Surgery, Queen Elizabeth Hospital, University Hospital Birmingham NHS Trust, Birmingham and ‡Department of Histopathology, South Warwickshire General Hospital NHS Trust, Warwick , UK
    Search for more papers by this author
  • M. A. Clark,

    1. *Gastrointestinal Unit and †Academic Department of Surgery, Queen Elizabeth Hospital, University Hospital Birmingham NHS Trust, Birmingham and ‡Department of Histopathology, South Warwickshire General Hospital NHS Trust, Warwick , UK
    Search for more papers by this author
  • D. S. A. Sanders,

    1. *Gastrointestinal Unit and †Academic Department of Surgery, Queen Elizabeth Hospital, University Hospital Birmingham NHS Trust, Birmingham and ‡Department of Histopathology, South Warwickshire General Hospital NHS Trust, Warwick , UK
    Search for more papers by this author
  • D. Morton,

    1. *Gastrointestinal Unit and †Academic Department of Surgery, Queen Elizabeth Hospital, University Hospital Birmingham NHS Trust, Birmingham and ‡Department of Histopathology, South Warwickshire General Hospital NHS Trust, Warwick , UK
    Search for more papers by this author
  • M. R. B. Keighley,

    1. *Gastrointestinal Unit and †Academic Department of Surgery, Queen Elizabeth Hospital, University Hospital Birmingham NHS Trust, Birmingham and ‡Department of Histopathology, South Warwickshire General Hospital NHS Trust, Warwick , UK
    Search for more papers by this author
  • J. Goh

    1. *Gastrointestinal Unit and †Academic Department of Surgery, Queen Elizabeth Hospital, University Hospital Birmingham NHS Trust, Birmingham and ‡Department of Histopathology, South Warwickshire General Hospital NHS Trust, Warwick , UK
    Search for more papers by this author

Dr J Goh, Consultant Gastroenterologist, Gastrointestinal Unit, University Hospital, Birmingham NHS Foundation Trust, Raddlebarn Road, Birmingham, West Midlands B29 6JD, UK.
E-mail: Jason.Goh@uhb.nhs.uk

Abstract

Background  The clinical presentation of Clostridium difficile infection ranges from asymptomatic carriage, colitis with or without pseudomembranes, to fulminant colitis. Although not common, fulminant C. difficile colitis can result in bowel perforation and peritonitis with a high mortality rate. Colectomy is often indicated in these cases.

Methods  We retrospectively analysed the outcome of 14 patients who underwent surgery for fulminant C. difficile colitis in the period 1996–2003 in our Unit.

Results  The indications for surgery were systemic toxicity and peritonitis (n = 10), radiological and clinical evidence of progressive toxic colonic dilatation (n = 3) and progressive colonic dilatation with bowel perforation (n = 1). C. difficile infection as the cause of colitis was diagnosed pre-operatively in seven (50%) patients, six of whom underwent a total colectomy and one a right hemicolectomy. Overall mortality in our series was 35.7%. Total colectomy was associated with a lower mortality rate of 11·1% (1/9) when compared with left hemicolectomy was 100% (4/4) (P = 0·01). One patient who underwent a right hemicolectomy (on the basis of deceptively normal external appearance of the rest of the colon intra-operatively) survived after a prolonged hospital stay.

Conclusions  Early or pre-operative microbiological diagnosis of C. difficile infection can be difficult in patients with a fulminant presentation. Those patients with C. difficile colitis, who develop signs of toxicity, peritonitis or perforation, should undergo a total colectomy as the operation of choice.

Introduction

The exposure of anaerobic gut flora to antibiotics creates a microenvironment favouring the colonization and proliferation in the colon of a gram-positive anaerobic bacillus Clostridium difficile (C. difficile). Enterotoxin A and cytotoxin B, produced by C. difficile, induce an inflammatory process [1] causing a broad spectrum of clinical presentation, which may range from asymptomatic colonization, mild diarrhoea, to severe debilitating disease, with high fever, severe abdominal pain, paralytic ileus, colonic dilation (or megacolon), or even perforation [2].

Patients with severe to extremely severe disease at the time of admission, those immunosuppressed following solid organ transplantation and those who have an impaired antibody-mediated immune response to C. difficile toxins are at an increased risk of severe form of C. difficile colitis [3,4]. Prompt recognition of C. difficile colitis and aggressive therapy for the pathogen are essential for a favourable outcome [5]. Despite optimal medical treatment, some patients with C. difficile colitis fail to respond and progress on to systemic toxicity, peritonitis, or toxic colonic dilatation and bowel perforation. Careful monitoring of these patients is vital, as they require early surgical intervention [6,7].

Enteropathogenic microorganisms including C. difficile are found infrequently in patients with relapse of inflammatory bowel disease (IBD), despite intensive microbiological screening and play only a minor role in the exacerbation of IBD [8]. The presence of C. difficile in the stool is not associated with the activity of IBD, but it is related to the number of hospital admissions and antibiotic therapies in these patients [9–11]. About 10% of patients with IBD relapse have detectable C. difficile or its toxins in the stool, an incidence similar to control patients with diarrhoea [8,12]. Nevertheless, it appears that undiagnosed C. difficile colitis in patients with IBD may be responsible for failure to respond to treatment of IBD. Therefore stool culture, stool immuno-assay for C. difficile toxin A, and colonoscopy should be a part of the diagnostic work up in patients with IBD, who develop intractable diarrhoea during or after antibiotic therapy [13,14].

Current treatment options of C. difficile colitis are mainly antibiotic-based, and oral Metronidazole or oral Vancomycin are equally effective at resolving clinical symptoms. Oral metronidazole is preferred as a first line of treatment because of lowered cost and less selective pressure for vancomycin-resistant bacteria. Vancomycin is usually reserved for severely ill patients and recurrence of C. difficile colitis [15].

Emerging dietary and therapeutic interventions of pre- or probiotics aim to immunomodulate the host–pathogen interaction in a beneficial way, particularly in elderly people or individuals at risk of recurrent C. difficile associated diarrhoea [16–18].

There is a paucity of data on outcome of surgery for fulminant C. difficile colitis. Predictors of high mortality include increasing age, concurrent disease, the need for pre-operative vasopressors, and the type of surgery undertaken [19,20]. There are substantial variations in the reported mortality rates following colectomy, ranging from 38% to 80%. Total colectomy (resection of all the colon, including the sigmoid, but with preservation of the rectum and pelvic tissue planes) has the lowest mortality in the treatment of fulminant C. difficile colitis compared to more limited colonic resection [7,19].

The objective of this report is to retrospectively analyse the outcome of surgical treatment for fulminant C. difficile, and to identify potential parameters that may influence patient survival.

Patients and methods

Patients were identified from the University Hospital Birmingham NHS Trust pathology database by searching for the terms ‘pseudomembranous colitis’ or ‘antibiotic-associated colitis’ from the year 1996–2003. Other sources (such as the Colorectal Nurse Specialists' records) were also used to identify further patients.

Data collected included verification of the clinical syndrome, microbiological diagnosis, demographic data, comorbidities, laboratory indices, peri operative findings, type of surgery performed, and clinical outcome thereof.

A total of 3472 patients tested positive for C. difficile toxin A in the study period. Fourteen (0.4%) of these patients underwent a colectomy as a direct consequence of C. difficile colitis. Symptomatic C. difficile colitis was defined as > 3 bowel motions/day, and a positive stool culture or toxin A immuno-assay for C. difficile, or characteristic endoscopic findings [21].

Values were expressed in mean, median or range. Parametric variables were compared with the Fisher's Exact test for paired data. All statistical analyses were performed on Microsoft Excel 2002.

Results

Patient information

Fourteen patients were identified (5 male: 9 female) who underwent partial or total colectomy for fulminant C. difficile colitis in the study period (1996–2003). Median age was 64 years (range 30–93 years). Total colectomy with end ileostomy was performed in nine patients, left hemicolectomy in four patients, and right hemicolectomy in one patient (Table 1).

Table 1.  Details of the 14 patients, who underwent surgery for fulminant C. difficile colitis.
Patient numberPatient age (years)Antibiotics used within 2/12 prior to surgeryType of surgeryImmuno- suppressionOrgan failure prior to surgeryDiagnosis of C. difficile colitis30-days mortality
  1. Cef, Cefuroxime; Met, Metronidazole; CA, Co-amoxiclav; Cip, Ciprofloxacine; Fluc, Flucloxacillin; LHC, Left hemicolectomy; TC, Total colectomy; EI, End ileostomy; R, Renal failure; Re, Respiratory failure; C, Cardiac failure; H, Haematological failure; N, Neurological failure; Post-op, post-operative diagnosis of C. difficile colitis; Pre-op, pre-operative diagnosis of C. difficile colitis

 152Cef, MetLHCNoR, Re, H, NPost-opDied
 287Cef, MetLHCNoNoPost-opDied
 376CefLHCNoR, Re, C, HPost-opDied
 493CALHCNoR, Re, CPost-opDied
 567CipRHCNoNoPre-opSurvive
 664CipTC & EINoNoPost-opSurvive
 772Cef, AmoxTC & EINoR, RePost-opSurvive
 876Cef, Cip, MetTC & EINoR, Re, C, H, NPost-opSurvive
 930CATC & EINoNoPre-opSurvive
1050FlucTC & EINoNoPre-opSurvive
1154Cef, MetTC & EINoRPre-opSurvive
1263CefTC & EIYesCPre-opSurvive
1363CefTC & EIYesCPre-opSurvive
1447FlucTC & EINoR, Re, C, NPre-opDied

Four patients with 2–14 days (median 5·5 days) history of diarrhoea were directly admitted to the Surgical Unit with pyrexia and an abdominal pain, distension and peritonism.

Eight patients were transferred to the Gastrointestinal Unit from other units: Orthopaedic (n = 2), Cardiothoracic (n = 3), Neurosurgical (n = 2) and Gynaecology (n = 1) Units. These patients developed C. difficile colitis within 3–14 days (median 5·5 days) in the postoperative period following various procedures: total hip replacement, shoulder replacement, thoracic aortic transection repair, lung or combined heart/lung transplant, drainage of subdural haematoma, meningioma resection, total abdominal hysterectomy.

Two patients with severe active ulcerative pancolitis and positive stool samples for C. difficile toxin A were transferred from the Medical Gastrointestinal Unit (n =2) for emergency colectomy. Both of them failed to respond to standard treatment with intravenous (IV) hydrocortisone and metronidazole. Total colectomy with ileostomy formation was performed on the forth and eight day of admission, and IV/oral metronidazole was continued for seven days after surgery. In both cases histological examination of the surgical specimen confirmed severe active ulcerative pancolitis with superimposed C. difficile colitis, without pseudomembrane formation. Both of them made a good recovery.

Two of 14 patients had pre-existing diverticulosis of the colon. Two had pre-existing IBD as discussed. The remainder (71%) were not known to have pre-existing colonic disease.

All patients included in this study had a history of antibiotic use within the two months prior to presentation. Antibiotics were given for 3–14 days (median 10 days). Cefuroxime was most frequently used (in 8 patients), and others received Ciprofloxacin (3), Flucloxacillin (2), Co-amoxiclav (2), Amoxicillin (1). Metronidazole and Cefuroxime were used simultaneously in four patients who developed fulminant C. difficile colitis (Table 1).

Two organ transplant patients received immunosuppressive treatment with cyclosporin and azathioprine prior to diagnosis of fulminant C. difficile colitis. None of the patients had a previous history of C. difficile infection.

Clinical and laboratory findings

In this series all patients complained of abdominal pain, which was diffuse in 12 patients and localized to the left iliac fossa in two patients. Watery diarrhoea was reported in 11 patients with from four to 12 motions per day. Two patients had nausea and vomiting. Thirteen had abdominal distension, and nine had peritonism. All of the patients had pyrexia with recorded temperature above 38·0 °C. Eleven patients were tachycardic and tachypnoeic.

The most common finding on laboratory examination was leucocytosis with WBC from 11–35 × 109/l (median WBC 23 × 109/l), which was found in all but two patients (who were on immunosuppressive treatment). Elevated urea and creatinine were noted in seven patients. Five patients had abnormal clotting with prolonged INR > 1·5.

Nine patients included in this study had evidence of at least one organ system failure, and six had multiorgan failure prior to surgery. Six patients required intubation and mechanical ventilation prior to surgery. Vasopressors were used pre-operatively in six patients.

Supine and erect plain abdominal radiographs were performed in 11 patients. Diffuse air–fluid levels were present in six, progressively distended colon in four patients and air under the diaphragm in one patient. Abdominal CT was performed in 9 of 14 patients. Eight patients showed CT evidence of colitis of which four had pancolitis. One patient had only right-sided colon and three only left-sided colon involvement. CT findings suggesting colitis included colonic wall thickening (8 patients), nodular mucosal thickening (5 patients), the ‘accordion’ sign, which describes the appearance of oral contrast material trapped between nodular thickened folds of bowel (2 patients), pericolonic oedema (6 patients) and ascites (4 patients). One of 9 patients showed no CT evidence of colitis. The sensitivity of CT finding in detection of C difficile colitis was 87%. The clinical severity of disease did not statistically differ between patients with CT evidence of colitis and those without colitis.

The indications for surgery were systemic toxicity and clinical peritonitis in 10 patients, progressive colonic dilatation in three patients, and progressive colonic dilatation with bowel perforation in one patient.

In our study the time between onsets of symptoms and surgery varied from three to 18 days (median of 6 days). Only 7 (50%) patients had a positive stool culture or toxin A immuno-assay for C. difficile prior to colectomy. The remaining 7 patients had required emergency laparotomy prior to availability of microbiology results.

All patients diagnosed pre-operatively with fulminant C. difficile colitis underwent total colectomy (6 patients) or right hemicolectomy (1 patient). Treatment with oral metronidazole or vancomycin was commenced pre-operatively in this group.

The external surface of colon was frequently unremarkable at laparotomy, although colonic distension and oedema was present in all cases. The colon was necrotic in two patients, and perforated at the hepatic flexure in one patient.

Histological findings confirmed the diagnosis of C. difficile colitis in all patients. Two patients had pre-existing evidence of active ulcerative pancolitis and two patients had diverticulosis of the colon. None of the histological specimens had evidence of pre-existing ischaemia or vasculitis.

Outcome of surgery

The survival rate was higher in patients without multiorgan failure as compared with those with multiorgan failure (87.5%vs 33.3%). Patients that maintained haemodynamic stability without vasopressors prior to surgery had a better survival rate than those who required vasopressors pre-operatively (75%vs 50%). These differences however, did not reach statistical significance (P = 0.059 and P = 0.279 retrospectively, Fisher exact test). Interestingly, a pre-operative diagnosis of C. difficile colitis had a positive effect on patient's survival rate as compared to post operative diagnosis (85.7%vs 33.3%). This may in part be explained by the fact that once aware of a diagnosis of C difficile colitis surgeons were more likely to perform a total colectomy (85.7%) rather than a partial colectomy (14.3%). As the numbers were small this difference was not statistically significant (P = 0.122, Fisher exact test) (Table 2). A patient's; age, sex, type of hospital admission (e.g. medical or surgical), previous surgical history, past medical history, duration and type of antibiotic treatment (prior to a diagnosis of C. difficile colitis) had no apparent effect on patient's survival following surgery for fulminant C. difficile colitis.

Table 2.  The outcome of right, left and total colectomy in patients with fulminant C. difficile colitis observed within 30 days post surgery. The highest survival rate was observed in a group of patients with total colectomy. Development of multiorgan failure, pre-operative requirement of vasopressors and undiagnosed C. difficile colitis were associated with a negative surgical outcome, although P-values didn't reach statistical significance
30-day outcome →Patients survived (n = 9)Patients died (n = 5)Fisher exact test P-value
Multiorgan failure
 No7 (87.5%)1 (12.5%)0.059
 Yes2 (33.3%)4 (66.6%) 
Pre-operative use of vasopressors
 No6 (75%)2 (25%)0.279
 Yes3 (50%)3 (50%) 
Pre-operative diagnosis of C. difficile
 No3 (33.3%)4 (80%)0.122
 Yes6 (85.7%)1 (14.3%) 
Surgery:
 Right hemicolectomy10 
 Left hemicolectomy04 (100%)0.022
 Total colectomy8 (88.8%)1 (11.1%) 

Overall 30-day mortality in our series was 5/14 (35.7%); 1 (11.1%) of 9 in those who underwent a total colectomy, and in all four of those undergoing left hemicolectomy (100%). The cause of death in all cases was due to multiorgan failure. Total colectomy was therefore associated with a lower mortality compared with left hemicolectomy or with left/right hemicoloctomy (P = 0·01, P = 0.022 by Fisher's exact test)., (Table 2). One patient who underwent a right hemicolectomy survived but required 27 days of supportive treatment in the intensive care unit. The other eight patients who underwent total colectomy and survived required intensive care admission from 2 to 11 days (median 6 days). Six of the patients with total colectomy underwent further surgery at a later time, five of whom had successful formation of ileorectal anastomosis, and one who underwent completion proctectomy and formation of ileo-anal pouch for previously diagnosed ulcerative colitis.

Discussion

The incidence and severity of C. difficile colitis are rising disturbingly, and the appearance of antibiotic-resistant organisms is of major clinical concern [22–24]. Although antimicrobial use (within two months) is the most important predisposing risk factor for C. difficile infection, host and environmental factors also play important roles.

The diagnosis of C. difficile diarrhoea is often delayed. Increased awareness of C. difficile diarrhoea, and the use of C. difficile toxin A immuno-assay rather than stool culture, may reduce the time from onset of symptoms to diagnosis and the commencement of treatment [25]. Johal et al. [26] recently proposed using sigmoidoscopic and histological examinations to improve the detection rate in a subgroup of patients with pseudomembranous colitis.

Abdominal CT findings associated with C. difficile colitis include; colonic wall thickening and nodularity, pericolonic stranding and oedema, ascites and the ‘accordion’ sign [27–30]. Colonic wall thickening is the most commonly seen abnormality; it is nonspecific and can also be found in other types of colitis. Findings such as wall nodularity, the ‘accordion’ sign and ascites are less sensitive but more specific [28]. The reported sensitivity of abdominal CT imaging in detecting colonic abnormalities in patients with C difficile disease ranges between 52% and 85% with a specificity between 48% and 93%[29,31–33]. In our study the sensitivity of abdominal CT was calculated at 87%. Our study confirmed that CT changes in C. difficile colitis correlate poorly with clinical severity [31].

Experience in the surgical management of C. difficile colitis is limited. The results of colectomy for fulminant colitis are often disappointing, with reported mortality rates from 38 to 80%. The poor outcome may in part be related to the premorbid state of the subjects in the series. However, the type of surgery performed appeared critical to patient outcome [5,6].

A survey of the literature postulates that total colectomy is associated with the lowest mortality rates in C. difficile colitis. Our findings concur with published data. We found that total colectomy in patients with fulminant C. difficile colitis has the lowest mortality rate of 11.1%, compared to left hemicolectomy with a mortality rate of 100%. Although only limited conclusions can be drawn from the relatively small number in our series, there will probably never be randomized clinical trials to compare the outcome of different surgical modalities. Therefore on the basis of our observations and published series, it would seem reasonable to recommend total colectomy as the procedure of choice in patients with fulminant C. difficile colitis in spite of a deceptively normal external appearance of the colon.

The discrepancy in the outcome following different surgery modalities may be attributed to persistent sepsis with residual C. difficile infection in the colonic remnant. This remains speculative. There are currently no in vitro or ex vivo studies in the literature specifically addressing segmental or regional colonization of the human gut by C. difficile. From our and others' data, microscopic assessment of the colon intra-operatively does not provide reliable information on the extent of disease involvement.

Disappointingly, stool microbiology results were unavailable in seven of our patients prior to surgery. Four patients in this group who underwent left hemicolectomy died and three patients, who had total colectomy with end-ileostomy survived following surgery. On the other hand, positive stool culture or toxin A immuno-assay for C. difficile was available in seven patients prior to surgery and six of them underwent total colectomy and only one right hemicolectomy. Only one patient died in this group. This highlights the importance of a pre-operative diagnosis of C. difficile infection as it may influence surgical decisions about the choice of resection and result in improved survival rate post surgery.

Patients with immunosuppression are at increased risk of developing fulminant C. difficile colitis [34]. Two of our immunosuppressed patients had positive diagnosis of C. difficile prior to operation, underwent total colectomy with terminal ileostomy, and fully recovered postoperatively.

The influence of C. difficile superinfection in subjects with inflammatory bowel disease is emergingly recognized as a clinical problem but is not the subject of our study.

Fulminant C. difficile colitis remains an under-appreciated cause of death due to a combination of its non-specific clinical syndrome and a lack of general awareness of the spectrum of clinical disease. Rapid diagnosis and treatment are crucial to a positive outcome, and early surgical intervention should be considered in nonresponders to medical therapy.

Early surgical consultation, joint medical and surgical management, and timely surgery before the onset of multiple organ failure may improve survival in fulminant C. difficile colitis. Once the decision is made for surgical intervention, total colectomy with ileostomy may be the lifesaving procedure of choice.

Ancillary