The efficacy of intracolonic vancomycin for severe Clostridium difficile colitis: a case series

We present a retrospective study of 127 patients describing the utility of treating severe CDI using intracolonic vancomycin. To our knowledge this is the second largest case series published to date [6, 9‐11] (Table 4).

In this study of patients with moderate to severe CDI, those treated with intracolonic vancomycin were critically ill since more than 90 % were admitted into the ICU and the average length of hospital stay was almost 40 days. This group had a higher incidence of toxic megacolon, which is a serious and life-threatening condition associated with a high mortality rate and systemic toxicity [16, 17]. Additionally this group underwent a greater number of colectomies, a therapy that is often reserved for severe refractory disease [17]. Despite having a greater number of Clostridium difficile-associated severe complications and a higher severity score (5.69 vs 5.02), the group that received adjunctive intracolonic vancomycin maintained a similar mortality rate as the group that received standard treatment without intracolonic vancomycin (38.5 vs 38.6 %). The severity score index that had been previously derived and validated in a VA population would have predicted a mortality closer to 50 % [13, 14]. This may suggest a protective role of intracolonic vancomycin.

In our study, patients treated with intracolonic vancomycin were also more likely to be infected or colonized with MRSA and VRE. The predominance of VRE and MRSA in the group that received intracolonic vancomycin may be demonstrative of excessive antibiotic use predisposing to severe CDI [18]. Antibiotic overuse is a major cause of severe CDI [18, 19]. Non-essential antibiotics can eliminate protective colonic flora and predispose to colonization and infection by virulent strains of Clostridium difficile [20]. VRE and MRSA are established markers of antimicrobial overuse and patients diagnosed with VRE are at greater risk for developing recurrent CDI [21]. Another explanation for our observation is that CDI has been shown to be a common coinfection in patients diagnosed with VRE secondary to the use vancomycin and metronidazole and prior use of metronidazole [22]. There have been a number of other hypotheses for this relationship, such as the sharing of common risk factors between the two groups [23], treatment with broad spectrum antibiotics for VRE, and use of vancomycin for the treatment of CDI.

Current practice guidelines for the initial management of CDI recommend stratifying patients into mild, moderate and severe CDI [1‐3, 24]. However, no single classification has been validated in a broad population or achieved national acceptance for this purpose [25, 26]. We chose to use the severity score index developed by Velazquez-Gomez [14] and validated by Toro [13] because these studies were done in a VA population similar to ours. However, this index would have predicted an approximate 50 % and 100 % mortality in moderate and severe CDI groups respectively, which is in contrast to the 39 % observed for both groups in our study. One explanation for this discrepancy is that the original prediction rule may not be applicable to other populations since it was derived from a small sample of 101 patients and validated in another small sample of 54 patients, both from the same single institution [27]. Another possibility is that because the intracolonic vancomycin group was critically ill with more comorbidities, their physicians were more likely to prescribe intracolonic vancomycin secondary to non-response from oral antibiotics. This confounding by indication is a well known bias of retrospective studies.

Current guidelines vary in their recommendations for the initial treatment of severe CDI and the role of intracolonic vancomycin. The Society for Healthcare Epidemiology of America (SHEA) and Infectious Diseases Society of America (IDSA) [2] recommend, besides oral vancomycin and IV metronidazole, “if complete ileus consider (italics added by the authors) adding rectal installation of vancomycin.” Gastrointestinal guidelines [1] recommend for CDI patients complicated with ileus or toxic colitis and/or significant abdominal distention, the “treatment of choice” is oral and per rectum vancomycin plus IV metronidazole. The European Society of Clinical Microbiology and Infectious Diseases [3] recommend one of 3 regimens for “severe, and/or complicated or refractory CDI”: IV metronidazole plus per rectum vancomycin or IV metronidazole plus oral vancomycin or IV tigecycline. The use of rectal vancomycin is based on low quality evidence, mostly case series and expert opinion [1‐3]. In our series, many of the cases occurred before the publication of the current guidelines. As with most other case series described in the literature (see Table 4), the dosage of intracolonic vancomycin varied from 0.25 to 1g every 6 h with the most common dosage being 500 mg every 6 h. The latter dose and frequency reflect current practice guidelines recommendations, although this is based on low quality evidence [1, 3]. The duration of use in our study was also quite variable ranging from 1 to 34 days with a median of 7 days. The authors were unable to find the duration of intracolonic vancomycin use in most other published case series listed in Table 4. Current guidelines also vary with their recommendations for duration of use ranging from 10 days [3] to “until the patient improves” [1], but in both situations this is also based on low quality evidence.

Finally, other potential therapies for CDI not used in our case series include the use of intravenous immunoglobulin (IVIG) and Fecal Microbiota Transplantation (FMT). In recent years there has been a renewed focus on the use of Fecal Microbiota Transplantation (FMT) for the treatment of recurrent CDI. The current data on FMT for the treatment of severe CDI is limited only to case reports or small case series [28, 29]. It remains to be seen how FMT will fit into the treatment of severe CDI.

Compared to other case series, the strengths of our study include the consecutive identification of all patients with CDI based on our hospital’s infection control surveillance during the time period studied, systematic collection of information from charts using a standardized form, application of a single severity index that stratified patients into comparable risk groups, and complete follow-up to at least 90 days after the diagnosis of CDI was made to record complications and outcomes.

The limitations of our study include those inherent to most retrospective studies including selection bias, recall bias, and confounding by indication. The population studied comes from a single institution, focusing on a veteran population, which may also limit this study’s application to more diverse populations. As noted in Table 3, although the severity score indices between the two groups were not statistically significant, the baseline therapy between the two groups differed significantly. The non-intracolonic vancomycin group received more oral metronidazole compared with the intracolonic vancomycin group, which received more IV metrondizaole and oral vancomycin. Since this was a retrospective study over 10 years, the “standard” of care evolved and primary care providers varied in their treatment approaches which has the potential to confound the study results. As noted in Table 4, the clinical indications for intracolonic vancomycin, doses and duration of uses, and outcomes vary significantly in this study and among other reported cases series in the literature. Data differentiating PCR-positive, toxin-negative patients from PCR-positive, toxin-positive patients were not collected and the specific strain of CDI for each case was also not available. Finally, the standardized information form recorded only all-cause 90 day mortality without a specified cause of death.