Association Between Deep Sedation and Risk of Post-colonoscopy Colorectal Cancer: Swedish Register Data 2007-2016

Post-colonoscopy colorectal cancer (PCCRC) refers to cancer diagnosed subsequent to a colonoscopy where no malignancy was initially detected. Established risk factors include inflammatory bowel disease, prior colorectal cancer (CRC) diagnosis, or a history of polypectomy. Additionally, younger age and female sex are associated with increased risk, with PCCRC cases more commonly located on the right side.1Burr N.E. et al.BMJ. 2019; 367: l6090Crossref PubMed Scopus (96) Google Scholar,2Forsberg A. et al.Clin Gastroenterol Hepatol. 2020; 18: 2724-2733.e3Abstract Full Text Full Text PDF PubMed Scopus (22) Google Scholar Causes of PCCRC can be divided into missed lesions, incompletely resected lesions, or lesions that were detected but not resected.3Anderson R. et al.Gastroenterology. 2020; 158: 1287-1299.e2Abstract Full Text Full Text PDF PubMed Scopus (117) Google Scholar The proportion of colonoscopies performed under deep sedation varies considerably worldwide. In Sweden, only a small percentage (<2%) of colonoscopies are performed under deep sedation,4Forsberg A. et al.Scand J Gastroenterol. 2017; 52: 1042-1047PubMed Google Scholar which is a substantially lower frequency compared with the United States.5Krigel A. et al.Clin Gastroenterol Hepatol. 2019; 17: 2489-2496Abstract Full Text Full Text PDF PubMed Google Scholar Being able to change the body position of the patient during the colonoscopy may facilitate the detection of bowel lesions,6Lee S.W. et al.Am J Gastroenterol. 2016; 111: 63-69Crossref PubMed Scopus (0) Google Scholar although this maneuver is more challenging to execute under deep sedation. Nonetheless, only a limited number of studies, notably from China, have explored the impact of deep sedation on polyp miss rates: Lee et al reported a significantly higher adenoma miss rate during deep-sedated compared with unsedated colonoscopy,7Sui Y. et al.Eur J Intern Med. 2023; 110: 48-53Abstract Full Text Full Text PDF Scopus (0) Google Scholar whereas Dong et al reached the opposite conclusion.8Dong H. et al.Scand J Gastroenterol. 2021; 56: 484-491Crossref Scopus (9) Google Scholar We hypothesize that the risk of PCCRC increases when colonoscopy is conducted under deep sedation and that this risk varies depending on the lesion's location, with a greater likelihood of missing lesions on the left side. In a register-based study, all colonoscopies conducted in Sweden between 2007 and 2016 on individuals aged 18 and above were included, utilizing surgical procedure codes for colonoscopy and colonoscopy with biopsy or polypectomy in colon (UJF32, UJF35, JFA15). The local Ethics Committee in Stockholm approved the study (2015/690-31/2). Individuals with a prior colectomy were excluded. Individuals with CRC diagnoses (C18, C19, or C20) from 2007 to 2019 were identified. CRCs displaying histology other than invasive adenocarcinoma, appendiceal cancer (C181), and anal cancers were excluded. The procedure code ZXH60 defined deep sedation, which is synonymous with deep propofol sedation in Sweden. The International Classification of Diseases, 10th revision, codes defined the anatomical location of the tumor, recorded as right-sided, left-sided (distal to the splenic flexure), or undefined. The registers are briefly described in the Supplementary Appendix, and methods for retrieval of data are previously reported.2Forsberg A. et al.Clin Gastroenterol Hepatol. 2020; 18: 2724-2733.e3Abstract Full Text Full Text PDF PubMed Scopus (22) Google Scholar The CRC was defined as PCCRC-3yr when detected within 6 to 36 months after undergoing a colonoscopy examination that was negative for cancer (a false negative [FN] examination). The CRC was defined as "detected" (dCRC) if it was registered within 6 months of a colonoscopy; this was assigned as a true positive (TP) examination. The World Endoscopy Organization method calculated the PCCRC rate: [{FN / (FN + TP)} × 100] %.9Rutter M.D. et al.Gastroenterology. 2018; 155: 909-925.e3Abstract Full Text Full Text PDF PubMed Scopus (226) Google Scholar Multiple Poisson regression calculated the relative risks (RRs) for PCCRC, where the occurrence of a PCCRC or a dCRC was defined as the outcome in the analysis, and death was the only censoring event. First, deep sedation was evaluated as the explanatory variable adjusted for sex, age, country of birth, comorbidity (inflammatory bowel disease, diabetes, ischemic heart disease, heart failure, chronic obstructive pulmonary disease, prior CRC, and polypectomy), polypectomy at colonoscopy, tumor stage and location, and time period. Second, an interaction term between deep sedation and location of cancer was added to this model to investigate if location influenced the risk of PCCRC during sedation. A total of 648,819 colonoscopies performed (53.6% in women) in 469,431 individuals (55.1% women) between January 1, 2007, and December 31, 2016, were evaluated. The mean age at the initial colonoscopy was 60.1 ± 17.2 years, with men averaging 60.4 years and women 59.7 years. The number of FN colonoscopies was 1799, and the number of TP colonoscopies was 22,665. Thus, the overall PCCRC-3yr rate was 7.35% for the 10-year period. The PCCRC rate for deep sedation according to the World Endoscopy Organization method was 13.2%. Descriptives are given in Supplementary Table 1. The index colonoscopy was performed under deep sedation in 434 cases of CRC, and of these, 64 (14.8 %) were PCCRCs. Descriptive data regarding sex, age, detailed location, stage, comorbidity, and time period for the PCCRCs and the dCRCs respectively, stratified on whether the index colonoscopy was performed under deep sedation or not, are given in Table 1. In the deep sedation group, 67% were men, compared with 51% in the group without deep sedation. Correspondingly, higher rates for PCCRC were observed for left-sided location and location in the right flexure.Table 1Descriptive Data for PCCRCs Diagnosed During Colonoscopy, Performed With and Without Deep SedationWith deep sedationaThree of 64 individuals (5%) had the diagnosis familial gastrointestinal cancer (International Classification of Diseases code Z800). (n = 64)Without deep sedation (n = 1735)NRate, %nRate, %Sex Men426787851 Women223385749Age group, years <30711111 ≥30–4035473 ≥40–5058664 ≥50–604616810 ≥60–70213343225 ≥70–80182864437 ≥806936721IBD (UC or CD) Yes142221412 No5078152188Location Cecum91435721 Ascending colon101630017 Right flexure711955 Transverse colon4617410 Left flexure12493 Descending colon46634 Sigmoid colon172729217 Rectum121938021 Unknown location00251.4 Total right-sided304792653.4 Total left-sided345378445.2Stage I172743125 II132041749 III111744626 IV7111619 Unknown162528016Prior CRC Yes101622013 No5484151587Prior polypectomy Yes111723413 No5383150187Time period 2007–2011213383548 2012–2016436790052 Total641001735100Note: Total CRCs under deep sedation (N = 434), of these PCCRCs (n = 64; 14.8%).CD, Crohn's disease; CRC, colorectal cancer; IBD, inflammatory bowel disease; PCCRC, post-colonoscopy colorectal cancer; UC, ulcerative colitis.a Three of 64 individuals (5%) had the diagnosis familial gastrointestinal cancer (International Classification of Diseases code Z800). Open table in a new tab Note: Total CRCs under deep sedation (N = 434), of these PCCRCs (n = 64; 14.8%). CD, Crohn's disease; CRC, colorectal cancer; IBD, inflammatory bowel disease; PCCRC, post-colonoscopy colorectal cancer; UC, ulcerative colitis. The multivariable RR of PCCRC in deep sedation was 1.36 (95% confidence interval [CI], 1.10–1.69) (Supplementary Table 2). PCCRC was relatively more often right-sided compared with dCRC (RR left-sided vs right-sided PCCRC, 0.72; 95% CI, 0.66–0.79). There was a statistically significant interaction between location and deep sedation, so that PCCRC was more commonly seen for left-sided than right-sided cancers if the colonoscopy was performed under deep sedation (RR, 1.58; 95% CI, 1.05–2.39; P = .029). This study represents the first investigation into the effect of deep sedation on PCCRC. The result support our hypotheses that the risk of PCCRC is increased when the colonoscopy is performed under deep sedation and that the risk depends on the location of the cancer, with a greater likelihood of missing a left-sided cancer under deep sedation. Deep sedation is used sparingly in Sweden, and mainly in severe cases of irritable bowel syndrome, anxiety, previous trauma, and cognitive impairments (ie, conditions not associated with an increased risk of CRC). The majority of colonoscopies in Sweden are performed by gastroenterologists, including those under deep sedations. One possible explanation for our observations could be the challenge of implementing dynamic position changes during deep sedation, a maneuver demonstrated to enhance the adenoma detection rate.6Lee S.W. et al.Am J Gastroenterol. 2016; 111: 63-69Crossref PubMed Scopus (0) Google Scholar,10Facciorusso A. et al.Clin Gastroenterol Hepatol. 2019; 17: 2439-2454.e25Abstract Full Text Full Text PDF PubMed Google Scholar The left colon is often easier to inflate and examine in right lateral position of the patient, and the right flexure often needs examination in left lateral position to be properly examined.6Lee S.W. et al.Am J Gastroenterol. 2016; 111: 63-69Crossref PubMed Scopus (0) Google Scholar In conclusion, deep sedation increases the risk of PCCRC and, specifically, cancer located on the left side, suggesting a need for heightened vigilance towards the left colon when the colonoscopy is performed under deep sedation. Download .docx (.06 MB) Help with docx files Supplementary Table 1 Download .docx (.04 MB) Help with docx files Supplementary Table 2 Download .docx (.03 MB) Help with docx files Supplementary Appendix