Evaluation of a Modified Bowel Preparation Regimen in Cirrhotic Patients Undergoing Colonoscopy

BACKGROUND AND RATIONALE

Optimal diagnostic yield in colonoscopy is critically dependent on the quality of bowel preparation. Inadequate bowel preparation, observed in up to 25% of colonoscopies, adversely impacts procedural performance and diagnostic accuracy, with significantly reduced detection rates of adenomas and advanced adenomas, increased risk of undetected colorectal cancer, longer procedure times, and need for repeat colonoscopy.

Cirrhosis has been consistently identified as an independent predictor of inadequate bowel preparation, with studies reporting suboptimal cleansing in 29.8-49% of cirrhotic patients undergoing colonoscopy. Proposed mechanisms include impaired gastrointestinal motility related to autonomic dysfunction, metabolic derangements, small intestinal bacterial overgrowth, and increased bacterial translocation, serving as a potential trigger for several complications associated with chronic liver disease.

Suboptimal bowel preparation may be particularly detrimental in patients with cirrhosis as chronic liver diseases may place patients at increased risk of colorectal cancer, making adequate bowel preparation essential to maximize the efficacy of screening with colonoscopy.

Studies that have reported on modified strategies for individuals with previous inadequate bowel preparation include prolonged low-fibre diets, the addition of promotility agents and/or the use of high-volume bowel preparation regimens. Recently, the US Multi-Society Task Force (USMSTF) developed a consensus statement addressing bowel preparation regimens for individuals at high risk for inadequate bowel preparation, suggesting a split-dose high volume PEG formulation plus 15 mg bisacodyl the afternoon before the colonoscopy, along with a low-fibre diet 2 to 3 days before colonoscopy, changing to clear-liquid diet the day before colonoscopy.

To date, no randomized controlled trials have established a superior or specifically tailored bowel preparation regimen for patients with cirrhosis. Therefore, the aim of this study is to evaluate whether an intensified bowel preparation protocol improves bowel cleansing quality in adult patients with cirrhosis undergoing colonoscopy.

STUDY DESIGN, RANDOMIZATION AND BLINDING

This is a multicentre prospective, randomized, single-blind controlled clinical trial conducted at the endoscopy unit of the Gastroenterology Department of Portuguese centers. The coordinator centre is Gastroenterology Department of the Hospital do Divino Espírito Santo of Ponta Delgada. This trial was designed with the participation of four centres in Portugal. The addition of new participating sites during the recruitment period is permitted if it occurs before 50% of the total planned sample has been enrolled.

A total of 252 participants will be enrolled and randomized to one of two arms: control group (regimen A) will receive the standard bowel preparation without any additional interventions (2-L PEG regimen combined with a one-day low-fibre diet and transition to clear liquids the afternoon before colonoscopy), whereas the intervention group (regimen B) will receive the same 2-L PEG regimen in combination with the assigned adjunctive measures: (1) 15 mg bisacodyl the afternoon before the colonoscopy, (2) follow a 3-day low-fibre diet before the procedure and (3) clear liquids for the entire day before colonoscopy.

Randomization will be performed centrally using Research Electronic Data Capture (REDCap) in a 1:1 allocation ratio. Data entry will be performed online through REDCap.

Blinding of the endoscopist will be strictly enforced. Before entering the endoscopy suite, the patient will be instructed by the nurse department not to reveal to the gastroenterologist team the regimen assigned. Participants will not be blinded to the intervention.

COLONOSCOPY AND BOWEL PREPARATION ASSESSMENT

Colonoscopies will be performed in the morning sessions, according to local standard operating procedures by board-certified gastroenterologists and supervised fellows in training. Assessment of the degree of bowel preparation will be made according to the Boston Bowel Preparation Scale (BBPS), with a total BBPS <6 or a BBPS <2 in any segment being defined as inappropriate.

Prior to enrolling the first participant, all endoscopists at each participating site must complete a formal BBPS calibration exercise. This consists of the independent scoring of a standardised set of colonoscopy video recordings with pre-established reference scores, provided by the coordinating centre. Certification requires a weighted kappa coefficient of ≥0.70 relative to the reference scores. Endoscopists who do not meet this threshold must undergo additional training and repeat the exercise before enrolling participants.

To ensure complete outcome data, participants who do not attend their scheduled colonoscopy will be contacted to document the reason. If non-attendance is unrelated to bowel preparation, the procedure may be rescheduled using the originally assigned regimen to preserve study allocation and limit dropouts.

STATISTICAL CONSIDERATIONS

Sample size was calculated based on the following assumptions. The expected adequacy rate in the control arm (standard 2L split-dose PEG) was set at 70%, consistent with the BBPS distribution reported by Gow-Lee et al. (2024) in 732 cirrhotic patients (mean BBPS 7.3 ± 1.8), from which an adequacy rate of approximately 70-72% can be derived. This estimate is further supported by Anam et al. (2016), who reported inadequate preparation in 48% of cirrhotic patients using standard regimens.

The expected adequacy rate in the intervention arm (split-dose 2L PEG + 15 mg bisacodyl + 3-day low-fibre diet) was set at 87.5%, corresponding to an absolute improvement of 17.5 percentage points (25% relative improvement). This assumption is grounded in three independent lines of evidence: (1) in patients with chronic constipation - the closest available model for cirrhosis-related dysmotility - randomized evidence supports the inclusion of bisacodyl as an adjunct to bowel preparation regimens, providing a mechanistic rationale for its use in populations with impaired intestinal motility; (2) the USMSTF 2025 consensus explicitly recommends the combination of split-dose 4L PEG + 15 mg bisacodyl + extended low-fibre diet for patients at high risk of inadequate preparation, including those with cirrhosis, based on the principle that each component contributes additively to cleansing efficacy; (3) the delta of 17.5 percentage points is deliberately conservative relative to effect sizes observed in analogous high-risk populations (29 percentage points in constipated patients), reflecting the uncertainty inherent in extrapolating to a cirrhotic population for which no RCT data exist.

Assuming a two-sided alpha of 0.05 and 90% statistical power, 113 patients per arm are required. After adjustment for an anticipated 10% dropout rate - a total of 252 patients (126 per arm) will be enrolled.

Descriptive statistics will be presented as mean (standard deviation), median (interquartile range), or proportions as appropriate. The primary analysis will be conducted using a modified intention-to-treat approach, including all randomized patients who initiated bowel preparation and underwent colonoscopy with bowel preparation assessment, with additional per-protocol analyses performed as sensitivity analyses to assess the robustness of the findings. The primary outcome will be compared between groups using multivariable logistic regression adjusted for pre-specified covariates including study centre and cirrhosis severity variables. Continuous variables will be analyzed using unpaired t-test or Wilcoxon rank-sum test as appropriate, and categorical variables using chi-square or Fisher's exact test. Pre-specified secondary analyses will include per-protocol analysis and subgroup analyses according to cirrhosis etiology, Child-Pugh class, MELD 3.0 score, and portal hypertension-related complications. Statistical significance will be defined as a two-sided p-value <0.05.

SAFETY MONITORING AND INTERIM SAFETY ANALYSIS

An interim safety analysis will be performed after approximately 50% of the planned sample has been enrolled. An independent Data Safety Monitoring Board (DSMB), composed of two gastroenterologists and one biostatistician not otherwise involved in the trial, will review serious adverse events and procedure-related complications in both study arms. Safety outcomes reviewed will include hepatic decompensation, severe hepatic encephalopathy, bowel preparation-related hospitalization, and procedure-related complications. The trial may be suspended pending DSMB review if predefined safety thresholds are exceeded or if a significant between-group difference in serious adverse events is identified. The DSMB will also evaluate participant dropout rates and overall study safety throughout the trial.

CLINICAL RELEVANCE

This study aims to generate prospective evidence supporting optimized bowel preparation strategies for cirrhotic patients undergoing colonoscopy.

Inclusion criteria

• Established diagnosis of cirrhosis, scheduled for elective outpatient total colonoscopy
• Age ≥ 18 years
• Ability to follow verbal and written instructions in Portuguese

Exclusion criteria

• Urgent procedures
• Colonoscopies not intended to reach the caecum
• History of any colonic surgery
• Absolute contraindication to bowel preparation or colonoscopy
• Active hepatic encephalopathy (West Haven grade ≥2) at the time of enrolment
• Refractory ascites, defined as ascites unresponsive to maximum diuretic therapy or requiring repeated large-volume paracentesis
• Severe hyponatraemia (serum sodium <125 mEq/L) at the time of enrolment
• Subject refusal or inability to comprehend the trial