538 The Association of Cecal Withdrawal Time With Adenoma Detection Rate (ADR) and Sessile Serrated Adenoma/Polyp Detection Rate (SDR) in Average Risk Screening Colonoscopies

Abstract

INTRODUCTION: The ASGE, ACG & AGA in the Multi Society Task Force of Colorectal Cancer recommended quality benchmarks in 2017 to enhance the quality of screening colonoscopies. Cecal withdrawal time of ≥6 minutes (m) was recommended to achieve an adenoma detection rate (ADR) of ≥25%. The primary objective was to assess the relationship between withdrawal time & ADR in the USA using a large national database. The secondary objective was to assess the polyp detection rate (PDR) & sessile serrated adenoma/polyp detection rate (SDR). METHODS: The Clinical Outcomes Research Initiative (CORI) V4 database was queried for screening colonoscopies from 2009 to 2014. PDR, ADR & SDR were measured for each minute of withdrawal time. The detection rates were defined as the number of procedures bearing at least one lesion of interest (polyp for PDR, or adenoma for ADR or sessile serrated adenoma/polyp for SDR) divided by the total number of colonoscopies meeting the inclusion criteria. Inclusion criteria was age ≥18 years, average risk screening colonoscopies, performance years 2009 to 2014, successful cecal intubation, documentation of withdrawal time & post polypectomy pathology submission to CORI at a rate of ≥90%. SAS 9.4 was used. RESULTS: 23242 colonoscopies met the inclusion criteria. PDR increased with longer cecal withdrawal times in a linear fashion (Figure 1). The rise in PDR was very similar in pattern to the rise in ADR. The ratio of PDR to ADR decreased as the withdrawal time increased (2.7 at 6 m, 1.8 at 10 m, 1.5 at 13 m & 1.2 at 22 m). ADR increased with longer withdrawal times in a linear fashion (Figure 2). ADR was 8.1% at 6 m, 25.4% at 10 m & 42.1% at 13 m. The ADR continued to rise in association with longer withdrawal times reliably reaching 75.5% at 26 m. ADR continued to rise beyond 26 m but it was less reliable due to the number of procedures being < 100 per withdrawal minute beyond 26 m. SDR increased with longer withdrawal times in an exponential fashion (Figure 3). It was 0.6% at 6 m, 1.3% at 10 m, 2.6% at 13 m & 9.0% at 22 m. It seemed to reach a plateau after 22 m & SDR was less reliable past that point due to <100 procedures per withdrawal minute. CONCLUSION: PDR & ADR are likely to be higher with longer withdrawal times in a linear fashion. SDR is likely to be higher with longer withdrawal times in an exponential fashion until 22 m after which it tends to plateau. The optimal withdrawal time for reaching the benchmark of ADR ≥25% may be as high as 10 m when assessed at a national level.