Tailored Bowel Preparation Guided By Boston Stool Form Scale And A Development Of Predictive Model For Inadequate Bowel Preparation

Part Ⅰ:Randomized controlled trial:standard versus supplemental bowel preparation in patients with Bristol stool form 1 and 2Background and AimsColonoscopy is the standard approach for evaluating the entire colon currently.And it is an effective screening tool for colorectal cancer and a therapeutic tool for polyp removal.However,inadequate bowel preparation can result in failed detection of prevalent neoplastic lesions and has been linked to an increased risk of procedural adverse events,lower adenoma detection rates(ADRs),longer procedural time,and so on.The Bristol stool form scale(BSFS),developed and validated by Kenneth W.Heaton et al,has been widely applied in both clinical practice and research.According to the shape and consistency,BSFS divides human stool into 7 different types.In clinical practice,Bristol stool form is easy to be identified and can predict the quality of bowel preparation.Studies have demonstrated that Bristol stool form 1 and 2 is an important predictor of inadequate bowel preparation.However,how to improve bowel preparation quality for patients with Bristol stool form 1 and 2?There is lacking of clinical trials focusing on this question.The aim of the study was to evaluate the efficacy of a supplemental preparation in bowel cleansing quality among patients with Bristol stool form 1 and 2,as well as its feasibility.MethodsThis study was performed in 3 Chinese tertiary hospitals,containing Shandong University Qilu Hospital,Shandong University Qianfoshan Hospital and Binzhou People’s Hospital.One independent investigator from each center interviewed patients and recorded their main Bristol stool forms of last 7 days.Baseline demographic and clinical characteristics of all patients were recorded when they met inclusion criteria and not met the exclusion creteria.Patients with Bristol stool form 1 and 2 randomly received either 2 L polyethylene glycol electrolytes powder(PEG-ELP)(standard preparation,group A)or 10 mg bisacodyl plus 2 L PEG-ELP(supplemental preparation,group B).When every two patients were enrolled into either group A or group B,the next patient with Bristol stool form 3 to 7 was assigned to group C,receiving 2 L PEG-ELP(standard preparation)for bowel preparation.The competitive enrollment was used in this study.We educated patients the method to do bowel cleansing by verbal and phamlet.Colonoscopists,nurses and investigators who eavaluated bowel cleansing qualities were all blinded to any details of bowel preparation.In colonoscopy procedures,blinded investigators evaluated the quality of each bowl preparation according to the Boston Bowel Preparation Scale(BBPS).The primary endpoint is the rate of adequate bowel reparation for the whole colon.The adequate bowel preparation rate for separate colon segments,the polyp detection rate(PDR),tolerability,acceptability,sleeping quality and compliance were evaluated as secondary endpoints.The Likert sacle was used to eavaluate subjective indicators in secondary endpoints quantitatively.Meanwhile,we performed a multivariate binary logistic regression to evaluate factors associated with inadequate bowel preparation.ResultsFrom January to October 2015,806 eligible outpatients were assessed for inclusion and 700 patients were randomized.In intention to treat analysis,the adequate bowel preparation rates of group B and C were both significantly higher than group A(80.8%vs.56.2%,pAB<0.001;77.7%vs.56.2%,pAC<0.001).There was no significant difference between group B and C in adequate bowel preparation rate(80.8%vs.77.7%,pBC= 0.427).In per-protocol analysis,both group B and C showed significantly higher adequate bowel preparation rate than group A(88.7%vs.61.2%,PAB<0.001;85.0%vs.61.2%,pAC<0.001).The adequate bowel preparation rates were similar between group B and C(88.7%vs.85.0%,pBC= 0.316).For each separate colon segment,the adequate bowel preparation rate in group B and C was significantly higher than group A(p<0.001).The PDR in group B was significantly higher than group A(43.2%vs.25.7%,p<0.001),and similar with group C(43.2%vs.37.6%,pBC= 0.277).Ease of taking scores of preparation were not significantly different among 3 groups(p= 0.801).The rate of willingness to repeat the same preparation was significantly lower in group A than that in group B(pAB= 0.002)and C(PAC= 0.001).Sleeping quality was not significantly different among 3 groups(p=0.779).The compliance among 3 groups was not significantly different(p>0.05).The majority of patients started preparation regimens at correct time(p= 0.938)and abided by correct diet restriction(p= 0.843).Multivariate analysis showed age>60 years(OR= 1.63),diabetes(OR= 2.00),standard preparation(OR= 6.81),poor compliance to product instruction(OR= 3.67),frequency of defecation<3 times/week(OR= 1.78)and Bristol stool form 1-2(OR= 2.54)were significant factors for poor bowel preparation.ConclusionsThe supplemental preparation,10 mg bisacodyl plus 2 L PEG-ELP,can significantly improve adequate bowel preparation rate both in whole colon and each colon segment.Meanwhile,it can also significantly improve both total and each colon segment BBPS scores.Additionally,the supplemental preparation can significantly improve PDR and acceptability,without impairing tolerability,sleeping quality and compliance in patients with Bristol stool form 1 and 2.SignificanceThis study demonstrated that the supplemental preparation could significantly improve the quality of bowel preparation.In clinical practice,it is simple to indentify the Bristol stool forms of patients according to BSFS.For patients with Bristol stool form 1 and 2,the supplemental preparation should be prescribed;for patients with Bristol stool form 3 to 7,standard preparation can be used.By using tailored bowel preparation according to the Bristol stool form,the adequate bowel preparation rate may be improved.Therefore,BSFS guided tailored bowel preparation might be an easy and efficient approach for bowel preparation before colonoscopy.Part II:A predictive model for inadequate bowel preparation before colonoscopy:development and validationBackground and aimsColonoscopy is currently the predominant approach to prevent colorectal cancer(CRC)by detecting and removal precancerous lesions.The effectiveness of colonoscopy in colorectal cancer screening relies on the quality of bowel preparation.Adequate bowel preparation is critical for sufficient adenoma detection rates(ADRs).Besides CRC screening and surveillance,colonoscopy is commonly used for the diagnostic assessment of symptoms and other positive CRC screening tests.Inadequate bowel preparation is demonstrated to be associated with lower ADR,increased procedure time,higher interval cancer rates,more cost and shorter intervals between examinations.Unfortunately,multiple studies report up to one-third of all colonoscopies has an inadequate bowel preparation.There are many risk factors being discovered,such as chronic constipation,diabetes,inpatients,old age and so on.Additionally,several predictive models have also been developed by teams from different western countries.Patients with high risk of inadequate bowel preparation can be identified more accurately guided by inadequate bowel preparation predictive model.However,the current models are inconsistant with each other,and the procedures and methods have certain limitations.The aim of the study was 1)to identify risk factors of inadequate bowel preparation in chinese population;2)to develop and validate a predictive model to distinguish patients at risk for inadequate bowel preparation who may benefit from supplemental bowel preparation regimens and improve the overall quality of bowel preparation finally.MethodsA prospective cohort was established at endoscopy center in Qilu Hospital of Shandong University between April and December 2016.Before study initiation,3 investigators were educated by the Boston Bowel Preparation Scale Educational Program(BBPSEP)online and performed a calibration exercise on 30 colonoscopies according to BBPS,to achieve a satisfactory level of inter-observer agreement in the assessment of bowel preparation quality.Before colonoscopies,demographic data were collected prospectively and patient clinical data were recorded retrospectively.The same day 2 L polyethylene glycol electrolytes powder(PEG-ELP)regimen was used to conduct bowel preparation.The Boston bowel preparation scale(BBPS)was adopted to assess bowel cleansing quality which was classified as adequate or inadequate.The whole cohort was randomly divided into the development cohort(three quarters)and the validation cohort(one quarter).Univariate analysis and multivariate Logistic regression analysis were used to identify predictors associated with inadequate preparation in the development cohort to develop a prediction model.We used the regression coefficients of the remaining risk factors to assign integer points for the predictive model.The validation cohort was used to validate and evaluate the discriminative power of the predictive model.ResultsA total of 1504 colonoscopies were included in the prospective cohort,of which 346(23.0%)had an inadequate bowel preparation.1128 patients were assigned into the development cohort and 256(22.7%)patients yielded inadequate bowel preparation.Independent predictors were recognized as body mass index(BMI)≥30 kg/m2,age≥60 years,use of tricyclic antidepressants,diabetes mellitus,chronic constipation,American Society of Anesthesiologists Physical Status Classification System score≥3,inpatients,current hospitalization and history of colorectal surgery.The predictive model built with these predictors demonstrated an area under the curve of 0.715 in the development cohort and 0.683 in the validation cohort.According to the predictive model,a predictive scale with a range of 0 to 20 scores was developed.In the validation cohort,a cutoff of total scores≥2 could predict inadequate bowel preparation with a sensitivity,specificity,positive predictive value and negative predictive value of 57.0%(95%CI 46.3%-67.2%),74.9%(95%CI 69.4%-79.9%),42.7%(95%CI 33.9%-51.9%)and 84.1%(95%CI 79.0%-88.4%).ConclusionsWe identified predictors associated with inadequate bowel preparation and developed a validated predictive model with good accuracy.In the future,randomized controlled studies should be performed to further validate the predictive model in identifying subjects with an increased risk of inadequate bowel preparation.SignificanceIn this study,we developed a validated,easy-to-use predictive model that can be used to identify subjects with an increased risk of inadequate bowel preparation with good accuracy.It can be applied to clinical works to identify patients tend to yield inadequate bowel preparation.By prescribe more intensive regimens(4 L PEG-ELP regimen,2 L PEG-ELP plus bisacodyl regimen)to these subjects,the overall rates of adequate bowel preparation is hoped to be improved apparently.