Introduction: Split dose preparation (SDP) has been shown to have a superior bowel cleansing effect compared day before preparation (DBP) for screening colonoscopy in previous trials. Most of these studies reported bowel cleanliness as primary outcome. Compared to assessment of bowel cleansing which inherently has certain degree of subjective component, adenoma detection rate (ADR) is more objective and relevant parameter when it comes to colonoscopy for cancer screening. Unlike bowel cleanliness, increased ADR has been directly linked to colon cancer survival rate. A recent study estimated that higher adenoma detection rate is inversely related to Colon Cancer mortality. A previously done meta-analysis which compared ADR as a secondary outcome of interest included two trials and did not show a statistically significant difference between the two preparation methods. In the light of new available evidence, we wanted to compare ADR between the two groups.
Methods: On May 17,2017 we conducted a systematic search of Medline, EMBASE, Web of Science Collection databases and the Cochrane Central Register of Controlled Trials for abstracts and fully published studies comparing day before versus split dose colonoscopy regimens. Both fully published randomized clinical trials and abstracts of unpublished trials were included in our meta-analysis. Studies which were non-randomized, included pediatric population or compared different preparation regimens in two groups were excluded. The primary outcome of interest was comparison of adenoma detection rate (ADR) between the two groups. Secondary outcome of interest was advanced adenoma detection rate between split dose and day before preparation. Using Mantel-Haenszal method, a fixed effect model was used to calculate weighted odds ratio (OR). Duval and Tweedie's trim and fill test was used to detect any publication bias. RevMan 5.3 and CMA version 3.0 was used for statistical analyses.
Results: Five fully published randomized clinical trials and one abstract of randomized trial met inclusion criteria of our analysis. Using Mantel-Haenszel method, a fixed effect model was used to calculate the weighted effect size. Our analysis included a total of 2033 patients. Split dose preparation showed significantly higher adenoma detection rate (ADR) compared to day before group (odds ratio [OR], 1.58; 95% confidence interval, 1.32- 1.90). Tests for statistical heterogeneity did not show any significant heterogeneity P-value=0.52 (I2=0 %). After adjustment for publication bias using Trim and Fill test, our analysis showed adjusted OR of 1.56; 95% confidence interval, 1.30-1.88. In secondary analysis, split dose preparation had significantly higher advanced adenoma detection rate compared to day before preparation (OR 1.46; 95% confidence interval, 1.14-1.89).
Conclusion: Based on our meta-analysis, split dose colonoscopy preparation has significantly higher adenoma detection rate (ADR) compared to traditional day before preparation.