This was a prospective, single-centre, operator- and assessor-blinded randomised clinical trial. The trial complied with the Declaration of Helsinki, written informed consent was obtained from the patients and the study protocol was approved by the Ningbo First Hospital Institutional Ethics Committee (2019-R065). The study was registered at Clinical Trials.gov (NCT 04,247,386) on 30/1/2020.
Outpatients aged 18–75 years old, who required colonoscopies and were willing to participate in this study between December 2019 and August 2020 were consecutively enrolled. The following exclusion criteria were used: (1) patients with constipation; (2) patients with severe liver, kidney and heart dysfunction; (3) patients with a history of poor intestinal preparation; (4) pregnant/lactating women; (5) patients who were allergic or intolerant to any type of research drug; (6) patients with severe gastrointestinal diseases, such as intestinal obstructions or perforations, toxic colitis and toxic megacolon; (7) patients with a history of inflammatory bowel disease; (8) patients with significant electrolyte anomalies, including abnormalities in phosphorus, sodium, potassium, calcium, chloride and magnesium levels; (9) patients with a history of colorectal resections; and (10) patients with diabetes. After receiving a full explanation of the study, all of the patients were provided with a detailed oral education and written informed consent before the recruitment.
All of the patients were instructed to adhere to a 1-day low-residue diet before the colonoscopy [9,10,11]. Patients who were allocated into the PEG + Mizone group (experimental group) received 180 g PEG with 1.2 L clear water plus 1.8 L Mizone liquids, whereas the patients in the PEG + water group (control group) received 180 g PEG with 3 L clear water for the bowel preparations. They were instructed to ingest one-third of the solution (60 g PEG + 1 L Mizone liquids plus clear water/clear water) between 9 and 10 pm on the day before the colonoscopy, after which they would ingest the remaining two-thirds of the solution (120 g PEG + 2 L Mizone liquids plus clear water/clear water) at 4–6 h before the colonoscopy. Intestinal cleanliness was evaluated according to the video by a specific assessor of outcomes, other information (such as patient age, height, weight, advent events and willingness) was collected by another assessor during the colonoscope, and operator was only responsible for colonoscopy. Partial blinding of the participants was applied (blinding of two assessors of outcomes, blinding of the operator, but patients in this trial were nonblinded).
All of the procedures were performed with the use of the same model of high-definition video colonoscope (CF- H290 or CF- HQ290 video colonoscope, Olympus Co, Tokyo, Japan) by a single experienced endoscopist, and all of the patients underwent the colonoscopies with air.
PEG powder (Hengkang Zhengqing, Jiangxi Hengkang Pharmaceutical Co. Ltd., China) was packed into a box that included 3 bags of reagents A, B and C. Each bag of reagent A consisted of 1.68 g of sodium bicarbonate and 0.74 g of potassium chloride. Each bag of reagent B contained 5.68 g of sodium sulphate and 1.46 g of sodium chloride. Each bag of reagent C contained 60 g of PEG 4000. A total of three boxes of PEG powder and 3 L clear water were used to produce 3 L PEG solution, whereas a total of three boxes of PEG powder with 1.2 L clear water and 1.8 L Mizone (Danone China Food & Beverage) were used to produce 3 L PEG-Mizone solution.
The primary endpoint in this study was the degree of colonic cleanliness in the PEG + water group and the PEG + Mizone group. Colonic cleanliness was measured according to the Ottawa Bowel Preparation Scale (OBPS) . The OBPS consists of two parts, colon segment (right: cecum, ascending; mid: transverse, descending; and rectosigmoid) and fluid quantity. Colon segment was scored on a four-point scale (0–4), fluid quantity was scored on a three-point scale (0–2). Two parts scores were summed to yield a total score (range, 0–14, higher is worse). With the use of the OBPS, bowel preparation was considered to be “adequate” when the total score was less than 7. Secondary endpoints included the patients’ tolerabilities of the bowel preparations (including palatability, repeat willingness to undergo another preparation and the rates of completeness of the administrated liquids) and the rate of adverse events (including electrolyte disorders that required clinical treatment, nausea, vomiting, abdominal pain and bloating). Each patient completed a form regarding their palatability for the bowel preparation solution. The form was scored by using a modified 5-point Likert scale for ease of use (scores 5–1: very good, good, neutral, bad or very bad, respectively), and palatability was considered to be “good” when the score was greater than 3. All of the patients received follow-ups via telephone within 3 days to assess the frequency of adverse postoperative events. Only the electrolyte disorders that required clinical treatments were regarded as being adverse events.
The noninferiority test was used to calculate the sample size. The rate of adequate bowel preparation (Boston bowel preparation scale ≥ 6) of 92.5% has been reported in an article in 2019 . By comparison, the rate of adequate bowel preparation (Ottawa bowel preparation scale < 7) of 83% has been reported in another article in 2011 . The absolute risk difference between the two researches was 9.5%, we hypothesised that PEG was 92.5% efficacious for both groups in achieving an adequate score, the non-inferiority margins for PEG + Mizone compared with PEG + water were defined at 9.5% in absolute risk difference (RD), in order to ensure that the “adquete rate” for the PEG + Mizone group would exceed 83%. Based on a 2-sided significance level of 0.025 and a power of 80%, the estimated sample size was determined to be 121 patients in each arm. Additionally, we assumed that approximately 10% of the patients would eventually be excluded from the analysis set; thus, a total sample size of 270 patients would be needed for the study.
Randomisation and masking
A simple randomisation strategy was used, and the patients were randomly assigned at a 1:1 ratio to the two treatment groups by using Stata (version 13.0, Stata Corp LP, College Station, TX). Subsequently, the generated randomised sequence with a serial number was assigned to an opaque, sequentially numbered envelope by a staff member who was unaffiliated with the study. The allocation table was concealed from the operators. Intestinal scoring was performed by a proficient independent observer, according to the video.
Statistical analyses were performed by using SPSS (version 21, SPSS, Chicago, IL, USA). All of the included patients were enrolled in the intention to treat (ITT) analysis. The primary endpoint was also analysed according to the per-protocol (PP) analysis principle. In regard to the efficacy of the bowel cleansing, if the lower bound of the 95% confidence interval (95% CI) of the RD was greater than − 9.5%, then the noninferiority of the PEG + Mizone group could be concluded. We analysed the categorical outcomes by using Fisher’s exact test and compared the continuous outcomes by using the Mann–Whitney U test. Results with p-values < 0.05 (two-sided) were considered to indicate significance.