Is a diverting loop ileostomy and colonic lavage an alternative to colectomy for the treatment of severe Clostridium difficile–associated disease?

Commentary

There has been a steady increase in the incidence of CDAD over the last 10 years owing in part to increased awareness and the availability of more sensitive testing. Of the more than 22 000 cases of CDAD reported last year, approximately 3%–15% would be expected to progress to the “fulminant” state, a severe form of the disease unresponsive to medical therapy.1 Surgical intervention has been on the rise, and the preferred treatment in these critically ill patients has been a colectomy and ileostomy.

Neal and colleagues2 reported on the use of diversion and colonic lavage as a minimally invasive surgical approach for the management of CDAD with excellent outcomes. The authors’ hypotheses for the success of colonic lavage and diversion were 1) a diverting loop ileostomy with colonic lavage through a minimally invasive approach poses minimal stress for the already critically ill patient and 2) the fecal stream is diverted and the luminal flora is deprived of nutrition, such that mechanical lavage and topical vancomycin administration should remove the causative bacteria and toxin with ultimate reversal of the pathologic process. Although novel to the recent era of increasingly prevalent and hypervirulent CDAD, a similar surgical approach has been advocated in the past. In 1971, Turnbull and colleagues3 described colonic decompression and diversion in the form of a skin-level colostomy and a loop ileostomy for fulminant toxic megacolon in patients with inflammatory bowel disease (IBD). The Turnbull-Blowhole colostomy was used to bridge the patient over the critical illness to a time when a definitive, elective procedure could be safely performed. However, with improved medical care of patients with IBD, refined operative approaches and extensive perioperative supportive resources, the blowhole procedure is now rarely performed.

Neal and colleagues presented the results of 42 patients prospectively enrolled and treated with colonic lavage with 8 L of warmed polyethylene glycol 3350/electrolyte solution through an antegrade catheter placed via a diverting loop ileostomy. The loop ileostomy was created laparoscopically in 83% of patients. The patients in this study were critically ill before surgery, as demonstrated by intensive care unit (ICU) admission (90%), mechanical ventilation (64%), vasopressor support (74%), mean white blood cell count of 25.4 × 10⁹/L and a mean APACHE-II score of 29.7 at the time of surgical evaluation. Furthermore, this cohort included immunocompromised (45%) and elderly patients (mean age 65 yr). The authors compared the outcomes of the prospectively followed group to a historical control group of 42 consecutive patients with similar demographic characteristics and degree of critical illness who underwent total abdominal colectomy and an end ileostomy.

The primary outcome of the study was resolution of CDAD, as documented by resolution of symptoms (which was not defined by the authors) and normalization of leukocytosis. The secondary outcomes were mortality and morbidity. The authors reported “all patients who underwent diversion and lavage had resolution of leukocytosis and clinical signs of CDAD.” However, they did not explicitly report this outcome in the colectomy cohort. There was a single patient in whom recurrent C. difficile infection occurred, but it is not clear how the authors monitored the other patients for this outcome or whether follow-up was complete in all patients. The secondary outcome of mortality was reported in both groups: 8 of 42 (19%) in the ileostomy group versus 21 of 42 (50%) in the historic colectomy group (p = 0.006).

This cohort study has numerous methodological weaknesses that could confound the outcomes and undermine the validity of the findings. First, the authors did not report important information regarding the total number of hospitalized patients with diagnosed C. difficile colitis. There is evidence that high mortality in patients with CDAD is correlated to specific strains of C. difficile that result in higher rates of surgery and worse outcomes.1 The NAP1 strain is known to be hypervirulent and has been associated with outbreaks.4 Thus, it would be important to determine whether there was a difference in strain types between the case series and historical cases. Viral strain analysis would be ideal but is often not available. Information about the proportion of patients needing ICU admission related to the overall incidence of C. difficile infection would provide a surrogate for this information, but the authors did not report this information. Recent National Surgical Quality Improvement Program and meta-analysis data suggest 30%–40% mortality in patients treated with colectomy for CDAD, although the range includes the 50% that this group reports.5,6 The authors themselves state that because this strategy is less invasive, they are using it earlier in patients with CDAD. Thus, while the 2 cohorts appear to be similar, there may be some differences in disease severity that account for the differences in outcome.

In addition, the temporal protocol change introduces bias related to cointerventions. There are few details about the perioperative management during the era in which colectomy was the procedure of choice. Were all patients treated with appropriate perioperative antibiotics? Were they treated with appropriate deep vein thrombosis prophylaxis? What were the indications for surgery? When was surgery performed in relation to the start of the patients’ symptoms? The introduction of the protocol described by the authors includes regimented perioperative management that may have contributed substantially to the survival improvement. Furthermore, we don’t know if there were patients in the colectomy era treated with segmental colectomy or other lesser surgeries. This might suggest that only patients with the most dire clinical conditions were treated with colectomy, and this might not be the best comparison cohort.

Despite these limitations, the authors present compelling data that suggest a colon-preserving alternative to colectomy may be an effective option in patients with fulminant CDAD. In addition to the mortality benefits, the authors report a 79% rate of ileostomy reversal at 6 months. This compares favourably to the reported 20% gastrointestinal continuity restoration after colectomy.7 Long-term follow-up in order to assess the rate of recurrent disease in this group of patients will be important to ensure that this approach has durability. The strategy has biologic plausibility, and further study via a randomized controlled trial is necessary before it replaces the current standard colectomy in patients with severe CDAD.