Loop Ileostomy With Colonic Lavage for Fulminant Clostridium Difficile Colitis

PROTOCOL

1. Study purpose and rationale
2. Specific Aim and Hypothesis: The objective of this study is to determine if loop ileostomy and colonic irrigation reduces 30 day all-cause mortality compared to the current standard of care for fulminant Clostridium difficile colitis (FCDC), total abdominal colectomy with end ileostomy. We hypothesize that loop ileostomy and colonic irrigation (ileostomy) will significantly reduce 30 day all-cause mortality compared to total abdominal colectomy with end ileostomy (TAC). Furthermore, we hypothesize that this intervention will reduce 30-day morbidity, 90 day all-cause mortality and improve gastrointestinal restoration rates compared to TAC. ii. Significance: With the emergence of a hypervirulent toxin-producing NAP1/027 strain of C. difficile in 2003, C.difficile colitis has become an increasingly common problem (1). Up to 10% of patients with C.difficile infection progress to FCDC with a mortality of 36-80% (1-3). The standard of care for FCDC is a timely TAC; however despite this intervention mortality rates remain high (38-55%) (2,3). The standard of care for FCDC is a TAC. This includes end ileostomy, which becomes permanent in over 65% of patients as reversal necessitates an extensive operation (5). Neal et al. proposed a new surgical approach for FCDC which includes creation of a loop ileostomy, intraoperative colonic lavage with warmed polyethylene glycol via the ileostomy and postoperative antegrade instillation of vancomycin flushes into the diseased colon via the ileostomy. In a single-institution series, this alternative approach for 42 patients was compared to 42 historical controls that underwent TAC (4). While demographics and severity of illness in the control and intervention groups were comparable, the 30 day mortality was significantly decreased in ileostomy compared to TAC groups (19% vs. 50%). To date, this is the only report of this new alternative procedure for FCDC. Neal et al.'s work is a single center, single surgeon series of loop ileostomy cases compared to a historical series of patients treated with TAC. To further investigate their promising results and innovative treatment, we propose a multi-center national (Canadian) prospective cohort study for the evaluation of this intervention compared to controls.
3. Description of study population
4. Basic study design: This is a national, multi-center prospective cohort study to evaluate loop ileostomy and colonic lavage to the standard of care, total abdominal colectomy with end ileostomy. This study will consist of a prospective registry (from 2014 to 2017) with 2 prospective arms: the loop ileostomy group and the control (TAC) group. In addition, we will concurrently perform a retrospective chart review at all participating sites from (from 2010 to 2014) to serve as a third group of historical controls. Study population and assembly of patients: Subjects will be drawn from 20 participating adult, tertiary Canadian hospitals. All patients with Clostridium difficile infection who are being evaluated for surgery are eligible for this study. Subject identification will be carried out by the attending surgeon on call, eligibility will be evaluated, and patients will be approached for study participation. Consent for the study will be obtained from the subject or a legally authorized representative (the legal representative must be a court-appointment mandatory, curator or tutor). Vulnerable patients with pre-existing incapacity to consent and who do not have a legally authorized representative will be excluded. In contrast, some subjects may have a late presentation of FCDC and therefore may lack the physical and mental capacity to provide an independent informed consent. At this time, the consent for the study will be obtained by the next of kin (next of kin will be defined as in Civil Code of Québec). When and if the subject recovers, consent will be obtained by him/her at that time. Enrolment into the trial will be in accordance with the Civil Code of Québec (articles 15, 20, 21) (6) QC notation. As this disease process and surgery only occur on an emergency basis, the surgeon on-call (pre-determined by each hospital's call schedule) will assess the patient and he/she will invite the patient to enrol in the study. The decision to undergo loop ileostomy or TAC will be left up to the discretion of the surgeon and the patient. In either case, the patient will be will be given the option to enrol in the study either in the investigational arm (loop ileostomy) or the control arm (TAC). We will concurrently perform a retrospective chart review at all participating sites from 01/2010 to 01/2014 to serve as a historical comparative control group (180 patients estimated). These patients will be identified from hospital databases by discharge and procedure codes. Chart review will be conducted using the same inclusion and exclusion criterias used for the prospective patients. The stable nature and incidence of C.difficile infection in Canada since 2009 and the well-established standard surgical approach (TAC) for FCDC (6) justify the comparison to a historical control.
5. Sample size and how it was determined We estimate that 60 colectomies for FCDC will be performed per year (at 20 institutions) (5). Neal et al. report an absolute risk reduction of 31% in 30 day mortality. However, given that this was a single surgeon experience and that our current 30-day mortality is 38% (5) (in contrast to 50% for TAC in Neal et al.'s series), we anticipate an absolute risk reduction of 22%. For this reduction, with a type I error rate of 0.05, a power of 80%, a 1:1 control to intervention ratio, 63 patients are required in each arm, and are estimated to be accrued over 2 years. We will also cluster by surgeon and institution, however as each surgeon and institution are only expected to contribute very few (approximately 1-3) cases, we are estimating a small intraclass correlation coefficient and design effect close to 1.
6. Design and description of methodology
7. Exposures and their measurement: The exposure is the treatment strategy. The intervention being evaluated is loop ileostomy with intra-operative colonic lavage. Ileostomy and colonic lavage is not considered standard of care at our institutions. Patients managed with ileostomy and colonic lavage who do not respond as expected may be converted to TAC, intra- or post-operatively, at the discretion of the surgeon. These patients will be analyzed in the ileostomy group (similar to an intention to treat analysis) since this was the original treatment strategy assigned. A subgroup analysis of these patients who had to be converted to TAC, and their impact on the overall results, will also be examined. ii. Confounders and their measurement: Fortunately, FCDC is managed on an emergency basis only; thus the surgeon on-call (based on the on-call schedule) is the one who will care for the patient and the patient does not choose their surgeon, nor is the surgeon choosing their patient. This eliminates any referral bias, that is, surgeon assignment can be considered random and there is no bias for different surgeons to be treating different types of patients. All general and colorectal surgeons are equally qualified to make the clinical decision to operate for FCDC (inclusion criteria for the study) and to perform the both proposed operations, TAC and loop ileostomy. However, individual surgeon bias may exist towards one procedure. This will be examined by a short qualitative questionnaire requesting the rationale for the choice of operation. Due to the infrequency of the operation for FCDC and the need to recruit multiple participating surgeons, each surgeon may only contribute 1-3 patients. Variability in individual and institution clinical practice will be accounted for by clustering by surgeon and institution in the analysis. Institution and surgeon variability may introduce may reflect differences in technical skill, hospital infrastructure, degree of adoption of early surgical consultation for FCDC, and operating room availability. Due to the severity of FCDC requiring surgery and the critically ill nature that patients reach prior to surgical consultation, the patient population entered into the trial will be relatively homogenous. This remarkable homogeneity within the patient population presenting for surgical management of FCDC was observed in our hospitals (2,5) and in Neal et al.'s study (4). Nonetheless, we will anticipate and account for confounding patient factors in the analysis including older age, immunosuppression, elevated pre-op white blood cell count, lactate, intubation and vasopressors (2,3,5). Earlier time to surgery is associated with improved survival and may also be a confounder as surgeons may be more inclined to offer loop ileostomy and lavage earlier than they would offer TAC given its minimally invasive nature. Similarly due to the new intervention's less invasive nature, older patients with co-morbidities may be more inclined to consent to surgery than they would be if offered TAC; introducing confounding by indication.
8. Definition of end-points The primary outcome is 30 day all-cause mortality. This binary outcome is the most suitable and reported outcome for FCDC. The secondary outcomes are: 1) 30 day morbidity, a composite binary outcome to include all Clavien-Dindo grade 2-5 (serious) complications (6) which also includes death (Clavien grade 5); 2) 90 day all-cause mortality (binary) in order to capture those who survive the first month but have a long and complicated hospitalization with poor outcome; 3) restoration of gastrointestinal continuity by ileostomy closure (binary) which serves as a surrogate for excellent recovery; 4) length of hospital stay, a censored outcome, using time to discharge as the outcome censoring by death (patients who die during hospitalization will contribute towards length of stay for the days that they survived).
9. Measurements and study instruments (including questionnaires, data collection forms, etc) Prospective data collection will be performed during the hospitalization daily for 30 days, and then patients will be followed at 3, 6 and 12 months post-operatively with a phone follow-up . If these time points coincide with a visit to the surgeon's office, the data can be collected at that time. The patient will not be required to do any extra tests or hospital visits if they participate in the study
10. Data analysis plan The distribution of each measured variable will be assessed graphically and with diagnostics. Descriptive variables and confounders will be compared between ileostomy and TAC groups using two sample t-tests, Wilcoxon rank sum, and fisher exact tests for normal, right-skewed and categorical variables, respectively. A logistic regression model predicting exposure (TAC vs. ileostomy) will be created with potential confounders outlined above (surgeon and patient factors) and a propensity score will be created for each individual. A logistic regression model will then be fit for the main outcome, with the main predictor being intervention, and the propensity score included as a second co-variate to control for confounding. Alternatively, we may choose to match on the propensity score. The impact of ileostomy on mortality will be expressed as an odds ratio (95% CI). Secondary analyses (considered exploratory) will compare 30 day morbidity, 90 day all-cause mortality, length of hospital stay and rate of GI restoration for ileostomy and TAC. A subgroup analysis will be performed for ileostomy patients who were converted to TAC. The historical control arm will be analyzed separately and should they be similar to both groups will be compared to them.
11. Recruitment procedures including copies of advertisements;

Subject identification will be carried out by the attending surgeon on call, eligibility will be evaluated, and patients will be approached for study participation. A consent for the study will be obtained from the subject or a legally authorized representative. As this disease process and surgery only occur on an emergency basis, the surgeon on-call (pre-determined by each hospital's call schedule) will assess the patient and he/she will invite the patient to enrol in the study. The decision to undergo loop ileostomy or TAC will be left up to the discretion of the surgeon and the patient. In either case, the patient will be will be given the option to enrol in the study either in the investigational arm (loop ileostomy) or the control arm (TAC). The patient will be given a consent form as soon as they are identified as a potential candidate by the surgeon-on-call. Once the patient and/or the legal representative have signed the consent form, the patient will be considered accrued to the study.

Details on confidentiality

Patient data will be linked to a hospital ID number and a study ID number. Once the patient has completed the follow-up period, their hospital ID number will be removed such that they are de-identified.

The data will be maintained in a secure web-based registry (Redcap). This will be used only by the research team and will always be accessed from a password protected hospital computer.

The data will be kept in a de-identified manner for 5 years following study completion.

The study will be conducted according to ethical principles stated in the Declaration of Helsinki (2013), ethics approval will be obtained before initiating study. Also, consent forms will take into consideration the well-being, free-will and respect of the participants, including respect of privacy.