Abstract
Abstract
Background:
Clostridioides difficile infection (CDI) accounts for as many as 25% of episodes of antibiotic-associated diarrhea and is the most common cause of healthcare-associated diarrhea. Rectal vancomycin irrigation is a therapy option; however, evidence is limited for its value post-colectomy. The objective of this study was to describe outcomes of patients who underwent total colectomy for fulminant C. difficile colitis and received rectal vancomycin post-operatively.
Methods:
This was a single-center retrospective chart review of adult patients who underwent total colectomy for fulminant CDI. Efficacy outcomes were all-cause in-hospital death, intensive care unit (ICU) and hospital length of stay, ventilator-free days at day 28 post-procedure, development of proctitis or pseudomembranes, need for re-initiation of CDI therapy, and normalization of infectious signs and symptoms at completion of CDI therapy. The primary safety outcome was the incidence of rectal stump blowout.
Results:
Of the 50 patients included, 38 (76%) received treatment with rectal vancomycin at the discretion of the surgeon. The Sequential Organ Failure Assessment score on the day of the procedure was higher in the rectal vancomycin group; however, this difference did not reach statistical significance. No difference was observed between the groups in the primary outcome of all-cause death. There was no significant difference between the groups for hospital length of stay, but there was a trend toward longer ICU length of stay for patients who received rectal vancomycin (9.5 days vs. 2.5 days; p = 0.05). No differences in the remaining secondary efficacy outcomes were observed. No episodes of rectal stump blowout were observed in either group.
Conclusions:
This study aimed to add to the limited data on the use of rectal vancomycin irrigation post-colectomy for toxic C. difficile colitis. Although our results do not support routine use of rectal vancomycin irrigation, they suggest that this therapy is not harmful if providers are considering its use for severe infections refractory to alternative treatment.
Clostridioides difficile * is a spore-forming, gram-positive, anaerobic bacillus that accounts for as many as 25% of episodes of antibiotic-associated diarrhea and is the most common cause of healthcare-associated diarrhea [1]. In the past, the attributable mortality rate of C. difficile infection (CDI) has been low, less than 2% of cases. In the early to mid-2000s, hospitals began to report more severe and recurrent CDI outbreaks. More recently, the 30-day mortality rate attributable directly to CDI has reached a reported 6.9%, with another 7.5% of deaths thought to be attributable indirectly to CDI. [2]
Traditional pharmacologic management of CDI includes oral or intravenous (IV) metronidazole, oral vancomycin, or fidaxomicin [1,3]. However, in severe cases that involve toxic megacolon, septic shock, perforation or impending perforation, necrotizing colitis, peritonitis, rapidly progressing disease, or disease refractory to pharmacologic therapy, surgical intervention via total abdominal colectomy is indicated [1,4]. There are limited data to guide pharmacologic therapy post-colectomy. As pharmacokinetic and anatomic limitations render other therapies ineffective, one option is the use of vancomycin irrigation for the newly constructed rectal stump. Rectal vancomycin irrigation is available at many institutions; however, the more common indication is for patients with ileus. There also are potential risks associated with its use, including colonic perforation and rectal stump blowout. Descriptions of the role of rectal vancomycin post-colectomy in the literature are limited, and there currently is no definitive recommendation for or against the use of rectal vancomycin irrigation [5,6]. Therefore, the objective of this study was to describe outcomes for patients who underwent colectomy for fulminant C. difficile colitis and received rectal vancomycin post-operatively.
Patients and Methods
Study design and patient population
This was a single-center retrospective chart review of adult patients who underwent total colectomy for CDI between January 1, 2011, and December 31, 2016. All patients were admitted to The Ohio State University Wexner Medical Center (OSUWMC), a tertiary academic referral center. Patients were identified electronically for inclusion on the basis of current procedural terminology codes for colectomy or ICD-9/ICD-10 coding for colectomy. Patients were included only if the indication for colectomy was CDI. The definition used was the presence of diarrhea or evidence of megacolon or severe ileus in correlation with a positive polymerase chain reaction (PCR) assay or evidence of pseudomembranes on surgical pathology review [1]. Patients with any type of CDI were included (healthcare facility-onset [HO]; community-associated [CA]; community-onset, healthcare facility-associated [CO-HCFA]), and the episode was defined as either initial or recurrent. Patients who did not have CDI as the indication for colectomy, or who were less than 18 or more than 89 years of age, pregnant, or incarcerated, were excluded. Informed consent was waived because of the retrospective nature of this study, which was approved by The Ohio State University Institutional Review Board.
Outcome measures
Patients who received vancomycin were compared with those patients who did not using a primary outcome of all-cause in-hospital death. Secondary efficacy outcomes were intensive care unit (ICU) and hospital length of stay (LOS), ventilator-free days at day 28 post-procedure, development of proctitis or pseudomembranes, need for re-initiation of CDI therapy after completion of therapy, and normalization of signs and symptoms associated with infection from colectomy to completion of CDI therapy. The primary safety outcome was rectal stump blowout. Normalization of infection features was defined as the normalization of white blood cell count (<11,000/mm3) and temperature (<100.4
Data collection
All data were collected retrospectively from the electronic medical record. Baseline characteristics including age, sex, admitting diagnosis, and medical history of immunosuppression (defined as ≥10 mg prednisone daily, HIV medications, or chemotherapy), transplantation or an active cancer diagnosis as documented in the provider notes was collected, as was exposure to broad-spectrum antibiotics before and after the colectomy. The PaO2:FiO2, serum bilirubin, platelet count, mean arterial blood pressure and vasopressor requirements, serum creatinine concentration, and Glasgow Coma Score were reviewed to calculate the SOFA score at the time of the colectomy. Variables related to CDI included the type of CDI acquisition (CA, HO, or CO-HCFA), C. difficile PCR results, and type and duration of CDI pharmacologic treatment before and after colectomy (including IV and oral metronidazole, oral and rectal vancomycin, and fidaxomicin). Hospital and ICU LOS data were collected, along with discharge disposition. Clinical variables, including white blood cell count, temperature, reduction or discontinuation of vasopressors, time on mechanical ventilation, and initiation of total parenteral nutrition (TPN) post-procedure were assessed through chart review and daily progress notes from the time of CDI diagnosis to the time of discontinuation of CDI therapy.
Statistical analysis
Patients who received rectal vancomycin irrigation post-colectomy for CDI were included in the “rectal vancomycin” group and were compared with those patients who did not receive rectal vancomycin irrigation (“no-vancomycin” group). All comparisons were unpaired, and all tests of significance were two-tailed with α
Results
A total of 68 colectomy cases was evaluated, and 50 patients met the criteria for inclusion. The primary reason for exclusion was surgery for a condition other than CDI (n = 14 patients). Additional exclusions were for intra-operative death (n = 2), no CDI on pathologic examination (n = 1), and a change in the procedural plan to exploratory laparotomy only (n = 1). Of the 50 patients who were included, 38 (76%) received treatment with rectal vancomycin irrigation post-colectomy. These patients received the standard dose of 500 mg every 8 hours. Each dose dwelled for 60 minutes and was administered by nursing staff via a rectal Foley catheter (18F; 30 mL balloon). No patient in either group received any other rectal irrigations. The distribution of patients receiving rectal vancomycin was relatively consistent throughout the study period, with only two patients treated in 2011 but eight to ten patients per year from 2012 to 2016. Any use of rectal vancomycin was at the discretion of the surgeon.
Baseline patient demographics were similar in the two groups; however, significantly more patients in the no-vancomycin group had an admission diagnosis of CDI or an alternative gastrointestinal cause for their symptoms (Table 1). Most patients in both groups were admitted from home and were considered to have CA-CDI, although 28.9% of patients in the rectal vancomycin group were classified as having HO-CDI. The number with a medical history of a transplant or an active cancer diagnosis was similar between groups; however, there were significantly more patients on immunosuppression in the no-vancomycin group. No differences were observed in patients receiving broad-spectrum antibiotics either before or after the procedure. The Sequential Organ Failure Assessment (SOFA) score on the day of the procedure was non-significantly higher in the rectal vancomycin group (7.0 [4–10] versus 4.5 [0.5–9.5]; p = 0.186).
Patient Demographics for Rectal Vancomycin and No-Vancomycin Groups
CDI = C. difficile infection; GI = gastrointestinal; LTACH = long-term acute care hospital; NH = nursing home; OSH, outside hospital; SD = standard deviation; SNF = skilled nursing facility; SOFA = Sequential Organ Failure Assessment.
Medications received as part of CDI therapy are summarized in Table 2 for both groups. There were no patients who received rectal vancomycin prior to colectomy. For patients who did receive rectal vancomycin, therapy was continued for a median of five days post-procedure. Alternative antibiotics used in combination included oral vancomycin, oral metronidazole, IV metronidazole, and fidaxomicin. No differences were observed in the exposure to oral vancomycin, IV metronidazole, or fidaxomicin; however, there were significantly more patients who received oral metronidazole in the no-vancomycin group. Patients who received oral vancomycin in the rectal vancomycin group received significantly more days of therapy post-procedure. There was minimal use of fidaxomicin in both groups.
Infection Therapies Administered before and after Colectomy for Rectal Vancomycin and No-Vancomycin Groupsa
Presented as median (IQR) unless otherwise noted.
No difference was observed between the groups for the primary outcome of all-cause death (Table 3). There was no significant difference between the groups in the hospital LOS, but there was a trend toward longer ICU LOS for patients who received rectal vancomycin (9.5 [5–17] days versus 2.5 [1.5–13.5] days; p = 0.05) (Table 3). All patients went to the ICU post-operatively except one in the no-vancomycin group. No differences in the remaining secondary efficacy outcomes were observed, including normalizing infection features, need for CDI treatment re-initiation, need for TPN initiation, ventilator-free days at day 28 post-procedure, and development of proctitis or pseudomembranes. All rectal stumps were stapled off and left in the pelvis. No episodes of rectal stump blowout were observed in either group.
Efficacy and Safety Outcomes for Rectal Vancomycin and No Rectal Vancomycin Groups
CDI = C. difficile infection; ICU = intensive care unit; IQR = interquartile range; LOS = length of stay; TPN = total parenteral nutrition.
Discussion
When comparing post-operative patients who received rectal vancomycin irrigations with those who did not, we found no difference in all-cause death or in any secondary endpoints. Notably, rectal vancomycin irrigation did not cause any harm, as there was no observed incidence of the primary safety outcome of rectal stump blowout. This single-center study was intended to add to the limited data surrounding the use of rectal vancomycin irrigations post-colectomy for C. difficile toxic colitis. Of the 50 patients who underwent total abdominal colectomy for CDI over a six-year period, 76% were treated post-operatively with rectal vancomycin irrigation. Calculated SOFA scores and longer duration of hospital and ICU LOS suggest that patients who received rectal vancomycin irrigation were more critically ill both pre- and post-operatively than those who did not receive this additional treatment, suggesting that irrigation therapy was reserved for sicker patients.
There have been few investigations into post-colectomy antibiotic therapy in patients with CDI and even fewer descriptions of outcomes with the use of rectal vancomycin irrigations for the fresh rectal stump. Brown et al. reported two patients who presented with rectal drainage that tested positive for the C. difficile toxin three weeks post-colectomy [5]. In both cases, symptoms were limited to the rectal stump; ileostomy contents had not changed in quantity and tested negative for C. difficile toxin. All drainage resolved after receipt of vancomycin irrigation for five to seven days. The authors were unable to discern whether the infection was an extension of the initial episode of CDI or disease recurrence, and they hypothesized that the inflammatory response was secondary to the presence of bacteria or spores retained from the time of the colectomy. Post-operatively, patient No. 1 received oral vancomycin as the sole agent, whereas patient No. 2 received four days of IV metronidazole. Oral vancomycin alone would not have reached the rectal stump, and the authors speculate that four days of IV metronidazole may not have been sufficient to eradicate any residual infection. In the current study, there were no instances of proctitis in either group; however, the median durations of IV metronidazole therapy post-operatively were seven and five days in the rectal vancomycin and no-vancomycin group, respectively. Although it may not be recommended to give patients rectal vancomycin irrigations solely for the prevention of C. difficile proctitis, the treatment may be considered in patients who have evidence of pseudomembranes at the level of the divided rectal remnant.
Whereas the Brown et al. report utilized rectal vancomycin irrigation as the sole treatment, the more common use for this therapy is in conjunction with other CDI measures, although the optimal combination remains unclear. A larger series of 100 patients who underwent total abdominal colectomy for fulminant C. difficile colitis compared four antibiotic regimens that included various combinations of IV metronidazole, oral vancomycin, and rectal vancomycin irrigation [6]. No combination was found to be superior as judged by survival, although shorter ICU LOS and more ventilator-free days were associated with the combination of IV metronidazole and oral vancomycin or IV metronidazole alone. The addition of rectal vancomycin irrigation did not impact outcomes, although it is difficult to interpret this result, as only 24 patients received rectal vancomycin irrigation. Because of these findings, the authors suggest that complications such as proctitis or pouchitis could be treated with the addition of rectal vancomycin but make no other recommendations for its use. Similarly, in the current study, we were unable to find any significant differences in clinical outcomes, although significantly, we did not find any evidence that vancomycin irrigations led to a greater incidence of rectal stump blowout.
Of note, we selected rectal stump blowout as the primary safety outcome for two reasons. Pragmatically speaking, the risk of blowout is always a concern for the clinical staff when considering any therapeutic enema or transrectal irrigation in a fresh post-operative patient, as the exact length of the rectal remnant often is not known by the nursing staff, and there is appropriate concern that transrectal administration of any medication could precipitate dehiscence of a fresh staple or suture line. Second, we recognize that gross dehiscence of the rectal stump often requires surgical re-exploration and washout or percutaneous drain placement in order to mitigate the complications of pelvic sepsis. Whereas the incidence of rectal stump blowout ranges from 3% to 6% in selected series, the morbidity of a second emergency laparotomy in a critically ill patient is significant [7,8].
This study has several limitations worth noting. First, it was retrospective and based on documentation provided in the medical record. Although the collected data existed in the form of discrete clinical events and quantitative variables, we cannot claim definitive causation between all notable clinical conditions and subsequent outcomes. The single-center design limits generalizability, as the results are specific to the patient population seen at OSUWMC. The sample was small despite collection over six years, and we were unable to stratify by severity of infection. Moreover, although the baseline characteristics were similar in the two groups, calculated SOFA scores on the day of procedure were higher in the rectal vancomycin group. We did not collect information on immunosuppressive biologics, which may limit generalizability in that population. Most confirmatory testing was done by PCR, which has known limitations of false-positive results, although patient inclusion and treatment choices were done in correlation with clinical symptoms. Finally, because of the retrospective nature of the study, we were unable to discern CDI-attributable deaths from all causes.
Conclusion
Although the results of our study do not support routine rectal vancomycin irrigation, they do provide some evidence suggesting that its use is not harmful if providers are considering it for extensive disease refractory to alternative treatment. Further investigation in a larger prospective study is warranted to help guide CDI treatment post-colectomy.
Footnotes
Author Disclosure Statement
MF was affiliated with The Department of Pharmacy at The Ohio State University Wexner Medical Center during the time of the study.
The authors have no financial disclosures or conflicts of interest to report.
Presented in part at the 2019 Society of Critical Care Medicine Annual Congress (Abstract 597), San Diego, California, February 18, 2019.
*
As is well known, this organism was long called Clostridium difficile. The name change, which arose after consideration of new 16S and ribosomal protein sequence data, began to be used by the Clinical Laboratory Sciences Institute and then by the Centers for Disease Control and Prevention. An attempt to change the name to Peptoclostridium was rejected because of the widespread use of Clostridium difficile or C. diff. in the literature. Surgical Infections is grateful to the authors of this paper for bringing the change to our attention.