- Email: [email protected]
The Relationship Between Distal Findings and Proximal Neoplasia in Colorectal Cancer Screening: A Systematic Review and Meta-Analysis
AGA Abstracts
factors, including age, sex, cancer stage, grade of differentiation, subsite, alcohol use, smoking status, physical activity, body mass index, regular use of aspirin, and intake of total fat, folate, calcium, and vitamin D. Results: We observed 731 deaths, of which 145 were classified as colorectal cancer-specific deaths over a median of 10 years of follow-up. High fiber intake after diagnosis was associated with lower risk of colorectal cancer-specific mortality (P for trend=0.04) and all-cause mortality (P for trend=0.01). Compared to patients in the lowest quartile, those in the highest quartile had 54% lower risk of colorectal cancerspecific mortality (HR, 0.46; 95% CI, 0.24-0.87) and 28% lower risk of all-cause mortality (HR, 0.72; 95% CI, 0.54-0.96). We did not detect any statistically significant nonlinear relationship by restricted cubic spline analysis. The results remained essentially unchanged after further adjusting for dietary glycemic or insulin load. For different fiber sources, vegetable fiber was associated with lower colorectal cancer-specific mortality (comparing extreme quartiles: HR, 0.54, 95% CI, 0.30-0.97), whereas no association was found for fiber from cereals or fruits. Conclusion: Among patients with colorectal cancer, increasing intake of fiber may reduce the risk of all-cause and colorectal cancer-specific mortality. Table 1. Post-diagnostic total fiber intake and mortality among colorectal cancer patients (n=1,487)*
Abbreviation: CI, confidence interval; HR, hazard ratio. *Post-diagnostic intake was assessed at least one year after diagnosis to minimize the influence of active treatment. †Cox proportional hazards regression model stratified by age groups at diagnosis (<60, 60-64, 65-69, 70-74, and ≥75 years), sex, and cancer stage (I, II, III, and unspecified), with additional adjustment for age at diagnosis (continuous). ‡Further adjusted for tumor grade of differentiation (1-3 and unspecified), subsite (proximal colon, distal colon, rectum and unspecified), pre-diagnostic fiber intake (in quartiles), alcohol consumption (<0.15, 0.15-1.9, 2.0-7.4, ≥7.5 g/d), pack-years of smoking (0, 1-15, 16-25, 26-45, >45), BMI (<23, 23-24.9, 2527.4, 27.5-29.9, ≥30 kg/m2), physical activity (women: <5, 5-11.4, 11.5-21.9, ≥22 METhours/week; men: <7, 7-14.9, 15-24.9, ≥25 MET-hours/week), regular use of aspirin (yes or no). §Further adjusted dietary intake of total fat, folate, calcium, and vitamin D.
Figure 1. Dose-response relationship between post-diagnostic fiber intake and colorectal cancer-specific mortality (A) and all-cause mortality (B) among colorectal cancer patients. Dashed lines represent the 95% confidence intervals of the hazard ratio (HR). Multivariable model was adjusted for the same set of covariates as in Table 1. No spline variable was selected. P for linearity = 0.04 for colorectal cancer-specific mortality, and 0.005 for allcause mortality.
83 THE RELATIONSHIP BETWEEN DISTAL FINDINGS AND PROXIMAL NEOPLASIA IN COLORECTAL CANCER SCREENING: A SYSTEMATIC REVIEW AND META-ANALYSIS Jason L. Huang, Yanhong Wang, Johnny Y Jiang, Chun Pong Yu, Yun Lin Wu, Ping Chen, Xiao Qin Yuan, Miaoyin Liang, Harry Hao-Xiang Wang, Martin C Wong Importance: Flexible sigmoidoscopy (FS) and colonoscopy are recommended as screening tests of choice for colorectal cancer (CRC). FS is now gaining popularity in some European countries and Asia Pacific regions since it can be performed by primary care professionals. Whether screening participants with distal hyperplastic polyps (HPs) detected by FS should be followed by subsequent colonoscopy, however, is still controversial as previous studies presented mixed findings. Objective: to evaluate the association between distal HPs and proximal neoplasia (PN)/advanced proximal neoplasia (APN) in asymptomatic, average-risk patients undergoing endoscopic CRC screening. Data Sources: Ovid Medline, EMBASE and the Cochrane Library from inception to 31 July 2016. Data Extraction and Synthesis: We included all screening studies that examined the relationship between different distal findings and PN/APN for average-risk, asymptomatic subjects. Data were independently extracted by two reviewers and in cases of disagreement, consensus was made via referral to a third reviewer. Main Outcomes and Measures: We pooled absolute risks and odds ratios with a random effects meta-analysis. Four subgroup analyses were performed according to study characteristics. Heterogeneity was characterized with the I2 statistics. Results: Twenty eight eligible studies were finally identified, including a total of 104,961 subjects (Figure 1). Overall, the prevalence of colorectal neoplasia was 25.0%; advanced colorectal neoplasia 4.9%; PN 13.2% and APN 2.2%. When compared with normal distal findings, distal HP was not associated with PN (OR=1.16, 95% CI 0.89-1.51, p=0.14, I2=40%) or APN (OR= 1.09, 95% CI 0.87-1.36, p=0.39, I2=5%) (Figure 2), whilst subjects with distal non-advanced or advanced adenoma had higher odds of PN and APN. Higher odds of PN/APN were observed for more severe distal lesions. Weaker association between distal and proximal findings was noticed in studies with higher quality, larger sample size, population-based design and better endoscopy quality control measures. There was no significant heterogeneity when the associations between distal HP and PN (I2=40%, p=0.28) /APN (I2=5%, p=0.39) were examined. The Egger's regression tests showed that publication bias did not exist (all
S-21
AGA Abstracts
AGA Abstracts
p > 0.50). Conclusions and Relevance: The presence of distal HP is not associated with PN/ APN in asymptomatic screening population. Our findings do not support routine colonoscopy workup for subjects with distal HPs detected by sigmoidoscopy, and provided concrete evidence in support of the US Preventive Services Task Force Recommendation Statements published in 2016. We anticipate that this clinical implication has a substantial potential to reduce unnecessary colonoscopy procedures, complications and healthcare costs, particularly in regions where colonoscopic capacity may be limited.
(53.9%), unknown in 335 (4.6%), and synchronous in both sides in 70 (1.0%). Cases were significantly less likely than controls to have undergone any colonoscopy (adjusted OR= 0.47; 95% CI, 0.44-0.50). Colonoscopy was associated with reduced risk of death from leftsided cancer (aOR=0.34; 95% CI, 0.31-0.38) and right-sided cancer (aOR=0.66; 95% CI, 0.59-0.73). In the screening colonoscopy subgroup, the adjusted ORs were 0.37 (95% CI, 0.31-0.43) overall, 0.27 (95% CI, 0.21-0.34) for left-sided cancer, and 0.49 (95% CI, 0.320.63) for right-sided cancer. Sensitivity analysis varying the time window before CRC diagnosis and colonoscopy exposure did not affect the primary findings (Table). Conclusions: In this national VA-Medicare study, colonoscopy was associated with significant reductions in CRC mortality among Veterans. Colonoscopy was associated with significant reductions in left-sided and right-sided CRC mortality, although the magnitude of the benefit was less for right-sided cancer.
85 RACIAL AND ETHNIC DIFFERENCES IN HEALTH BEHAVIORS AND COLORECTAL CANCER SCREENING AMONG UNITED STATES ADULTS: A 2015 BEHAVIORAL RISK FACTOR SURVEILLANCE SYSTEM ANALYSIS Nicolette Rodriguez, Thomas R. McCarty Introduction and Aim: Previous studies have demonstrated significant healthcare disparities associated with preventative healthcare screenings in the United States (U.S.). However, with the Affordable Care Act and expanded access to care, little is presently known about the association between these sociodemographic disparities and colorectal cancer screening (CRCS). The primary aim of this study was to evaluate the racial and ethnic prevalence of health-related characteristics with CRCS among U.S. adults. Methods: Using the Behavioral Risk Factor Surveillance System (BRFSS), the largest continuously conducted health survey system in the world, we conducted a secondary analysis based upon 2015 abstracted data. The population examined consisted of U.S. adults aged 50-75 years old who had valid responses for CRCS and race/ethnicity (n=42693). CRCS was defined as self-reporting a history of colonoscopy, sigmoidoscopy, or fecal occult blood testing. Survey adjusted and population weighted-multivariable logistic regression analyses were conducted to analyze the association between CRCS and race/ethnicity while adjusting for sex, marital status, employment, education and income level, health insurance, routine access to a healthcare provider, smoking, and binge alcohol use. Results: Approximately 79% (n=33569) of respondents reported CRCS. Self-reported screening was highest in the White (n=27526; 77%) racial/ethnic group; followed by Black and Other (n=2850; 72% and n=1285; 72%, respectively); Asian (n=1087; 70%); and Hispanic (n=821; 61%). On multivariate logistic regression, self-identification as Asian or Hispanic was associated with reduced odds of CRCS - Table. History of smoking, lack of an established primary care provider, and lack of health insurance were also associated with reduced odds of CRCS. Additionally, lower annual salary demonstrated less odds of obtaining CRCS - a step-wise decline when stratified by income level. Male/female sex and a history of binge alcohol use were not associated with changes in screening patterns. Discussion: Despite improved access to care, large racial/ethnic and socioeconomic disparities in CRCS persist. These differences in health behaviors underscore the need to develop culturally sensitive, evidence-based interventions to improve CRCS. Hispanics were the least likely racial and ethnic group to receive recommended CRCS. Future studies are needed to better elucidate this low rate of screening and provide annual trends to assess overall impact of these interventions.
84 COLONOSCOPY IS ASSOCIATED WITH DECREASED COLORECTAL CANCER MORTALITY IN THE VETERANS AFFAIRS SYSTEM Charles J. Kahi, Heiko Pohl, Laura Myers, Dalia A. Mobarek, Douglas J. Robertson, Thomas F. Imperiale Background and objectives: Colonoscopy is widely used in the Veterans Affairs (VA) system for colorectal cancer (CRC) prevention. However, the effect of colonoscopy on CRC mortality in the VA system is unclear. We aimed to determine whether exposure to colonoscopy is associated with decreased CRC mortality in Veterans, and whether the effect of colonoscopy on CRC mortality differs by anatomic location of the cancer. Methods: We designed a casecontrol study using national VA-Medicare administrative data. Cases were Veterans aged 52 or older diagnosed with CRC between 1/1/2003 and 12/31/2008, and died of CRC by 12/ 31/2010. Cancer diagnoses and death from CRC were ascertained from the VA Central Cancer Registry and National Death Index, respectively. Cases were age-, sex-, and facilitymatched to 4 controls that had not died of CRC. Patients with a prior history of CRC, colon surgery, inflammatory bowel disease, or familial polyposis were excluded. We determined the exposure of cases and controls to colonoscopy from 1997 to a date 6 months before CRC diagnosis in cases, and to a corresponding date in controls. Cancers were categorized as right-sided (cecum, ascending colon, hepatic flexure, transverse colon) and left-sided (splenic flexure, descending colon, sigmoid colon, rectum). Odds ratios for exposure to colonoscopy were calculated, adjusted for Charlson comorbidity score, aspirin use, and family history of CRC. Colonoscopy indication was classified based on a validated algorithm, and subgroup analysis was performed for patients who had undergone a colonoscopy for screening. Results: 7363 cases and 29,452 controls were identified; mean age was 73.3 years, and 98.8% were male. CRC was right-sided in 2993 (40.7%), left-sided in 3965
AGA Abstracts
S-22
factors, including age, sex, cancer stage, grade of differentiation, subsite, alcohol use, smoking status, physical activity, body mass index, regular use of aspirin, and intake of total fat, folate, calcium, and vitamin D. Results: We observed 731 deaths, of which 145 were classified as colorectal cancer-specific deaths over a median of 10 years of follow-up. High fiber intake after diagnosis was associated with lower risk of colorectal cancer-specific mortality (P for trend=0.04) and all-cause mortality (P for trend=0.01). Compared to patients in the lowest quartile, those in the highest quartile had 54% lower risk of colorectal cancerspecific mortality (HR, 0.46; 95% CI, 0.24-0.87) and 28% lower risk of all-cause mortality (HR, 0.72; 95% CI, 0.54-0.96). We did not detect any statistically significant nonlinear relationship by restricted cubic spline analysis. The results remained essentially unchanged after further adjusting for dietary glycemic or insulin load. For different fiber sources, vegetable fiber was associated with lower colorectal cancer-specific mortality (comparing extreme quartiles: HR, 0.54, 95% CI, 0.30-0.97), whereas no association was found for fiber from cereals or fruits. Conclusion: Among patients with colorectal cancer, increasing intake of fiber may reduce the risk of all-cause and colorectal cancer-specific mortality. Table 1. Post-diagnostic total fiber intake and mortality among colorectal cancer patients (n=1,487)*
Abbreviation: CI, confidence interval; HR, hazard ratio. *Post-diagnostic intake was assessed at least one year after diagnosis to minimize the influence of active treatment. †Cox proportional hazards regression model stratified by age groups at diagnosis (<60, 60-64, 65-69, 70-74, and ≥75 years), sex, and cancer stage (I, II, III, and unspecified), with additional adjustment for age at diagnosis (continuous). ‡Further adjusted for tumor grade of differentiation (1-3 and unspecified), subsite (proximal colon, distal colon, rectum and unspecified), pre-diagnostic fiber intake (in quartiles), alcohol consumption (<0.15, 0.15-1.9, 2.0-7.4, ≥7.5 g/d), pack-years of smoking (0, 1-15, 16-25, 26-45, >45), BMI (<23, 23-24.9, 2527.4, 27.5-29.9, ≥30 kg/m2), physical activity (women: <5, 5-11.4, 11.5-21.9, ≥22 METhours/week; men: <7, 7-14.9, 15-24.9, ≥25 MET-hours/week), regular use of aspirin (yes or no). §Further adjusted dietary intake of total fat, folate, calcium, and vitamin D.
Figure 1. Dose-response relationship between post-diagnostic fiber intake and colorectal cancer-specific mortality (A) and all-cause mortality (B) among colorectal cancer patients. Dashed lines represent the 95% confidence intervals of the hazard ratio (HR). Multivariable model was adjusted for the same set of covariates as in Table 1. No spline variable was selected. P for linearity = 0.04 for colorectal cancer-specific mortality, and 0.005 for allcause mortality.
83 THE RELATIONSHIP BETWEEN DISTAL FINDINGS AND PROXIMAL NEOPLASIA IN COLORECTAL CANCER SCREENING: A SYSTEMATIC REVIEW AND META-ANALYSIS Jason L. Huang, Yanhong Wang, Johnny Y Jiang, Chun Pong Yu, Yun Lin Wu, Ping Chen, Xiao Qin Yuan, Miaoyin Liang, Harry Hao-Xiang Wang, Martin C Wong Importance: Flexible sigmoidoscopy (FS) and colonoscopy are recommended as screening tests of choice for colorectal cancer (CRC). FS is now gaining popularity in some European countries and Asia Pacific regions since it can be performed by primary care professionals. Whether screening participants with distal hyperplastic polyps (HPs) detected by FS should be followed by subsequent colonoscopy, however, is still controversial as previous studies presented mixed findings. Objective: to evaluate the association between distal HPs and proximal neoplasia (PN)/advanced proximal neoplasia (APN) in asymptomatic, average-risk patients undergoing endoscopic CRC screening. Data Sources: Ovid Medline, EMBASE and the Cochrane Library from inception to 31 July 2016. Data Extraction and Synthesis: We included all screening studies that examined the relationship between different distal findings and PN/APN for average-risk, asymptomatic subjects. Data were independently extracted by two reviewers and in cases of disagreement, consensus was made via referral to a third reviewer. Main Outcomes and Measures: We pooled absolute risks and odds ratios with a random effects meta-analysis. Four subgroup analyses were performed according to study characteristics. Heterogeneity was characterized with the I2 statistics. Results: Twenty eight eligible studies were finally identified, including a total of 104,961 subjects (Figure 1). Overall, the prevalence of colorectal neoplasia was 25.0%; advanced colorectal neoplasia 4.9%; PN 13.2% and APN 2.2%. When compared with normal distal findings, distal HP was not associated with PN (OR=1.16, 95% CI 0.89-1.51, p=0.14, I2=40%) or APN (OR= 1.09, 95% CI 0.87-1.36, p=0.39, I2=5%) (Figure 2), whilst subjects with distal non-advanced or advanced adenoma had higher odds of PN and APN. Higher odds of PN/APN were observed for more severe distal lesions. Weaker association between distal and proximal findings was noticed in studies with higher quality, larger sample size, population-based design and better endoscopy quality control measures. There was no significant heterogeneity when the associations between distal HP and PN (I2=40%, p=0.28) /APN (I2=5%, p=0.39) were examined. The Egger's regression tests showed that publication bias did not exist (all
S-21
AGA Abstracts
AGA Abstracts
p > 0.50). Conclusions and Relevance: The presence of distal HP is not associated with PN/ APN in asymptomatic screening population. Our findings do not support routine colonoscopy workup for subjects with distal HPs detected by sigmoidoscopy, and provided concrete evidence in support of the US Preventive Services Task Force Recommendation Statements published in 2016. We anticipate that this clinical implication has a substantial potential to reduce unnecessary colonoscopy procedures, complications and healthcare costs, particularly in regions where colonoscopic capacity may be limited.
(53.9%), unknown in 335 (4.6%), and synchronous in both sides in 70 (1.0%). Cases were significantly less likely than controls to have undergone any colonoscopy (adjusted OR= 0.47; 95% CI, 0.44-0.50). Colonoscopy was associated with reduced risk of death from leftsided cancer (aOR=0.34; 95% CI, 0.31-0.38) and right-sided cancer (aOR=0.66; 95% CI, 0.59-0.73). In the screening colonoscopy subgroup, the adjusted ORs were 0.37 (95% CI, 0.31-0.43) overall, 0.27 (95% CI, 0.21-0.34) for left-sided cancer, and 0.49 (95% CI, 0.320.63) for right-sided cancer. Sensitivity analysis varying the time window before CRC diagnosis and colonoscopy exposure did not affect the primary findings (Table). Conclusions: In this national VA-Medicare study, colonoscopy was associated with significant reductions in CRC mortality among Veterans. Colonoscopy was associated with significant reductions in left-sided and right-sided CRC mortality, although the magnitude of the benefit was less for right-sided cancer.
85 RACIAL AND ETHNIC DIFFERENCES IN HEALTH BEHAVIORS AND COLORECTAL CANCER SCREENING AMONG UNITED STATES ADULTS: A 2015 BEHAVIORAL RISK FACTOR SURVEILLANCE SYSTEM ANALYSIS Nicolette Rodriguez, Thomas R. McCarty Introduction and Aim: Previous studies have demonstrated significant healthcare disparities associated with preventative healthcare screenings in the United States (U.S.). However, with the Affordable Care Act and expanded access to care, little is presently known about the association between these sociodemographic disparities and colorectal cancer screening (CRCS). The primary aim of this study was to evaluate the racial and ethnic prevalence of health-related characteristics with CRCS among U.S. adults. Methods: Using the Behavioral Risk Factor Surveillance System (BRFSS), the largest continuously conducted health survey system in the world, we conducted a secondary analysis based upon 2015 abstracted data. The population examined consisted of U.S. adults aged 50-75 years old who had valid responses for CRCS and race/ethnicity (n=42693). CRCS was defined as self-reporting a history of colonoscopy, sigmoidoscopy, or fecal occult blood testing. Survey adjusted and population weighted-multivariable logistic regression analyses were conducted to analyze the association between CRCS and race/ethnicity while adjusting for sex, marital status, employment, education and income level, health insurance, routine access to a healthcare provider, smoking, and binge alcohol use. Results: Approximately 79% (n=33569) of respondents reported CRCS. Self-reported screening was highest in the White (n=27526; 77%) racial/ethnic group; followed by Black and Other (n=2850; 72% and n=1285; 72%, respectively); Asian (n=1087; 70%); and Hispanic (n=821; 61%). On multivariate logistic regression, self-identification as Asian or Hispanic was associated with reduced odds of CRCS - Table. History of smoking, lack of an established primary care provider, and lack of health insurance were also associated with reduced odds of CRCS. Additionally, lower annual salary demonstrated less odds of obtaining CRCS - a step-wise decline when stratified by income level. Male/female sex and a history of binge alcohol use were not associated with changes in screening patterns. Discussion: Despite improved access to care, large racial/ethnic and socioeconomic disparities in CRCS persist. These differences in health behaviors underscore the need to develop culturally sensitive, evidence-based interventions to improve CRCS. Hispanics were the least likely racial and ethnic group to receive recommended CRCS. Future studies are needed to better elucidate this low rate of screening and provide annual trends to assess overall impact of these interventions.
84 COLONOSCOPY IS ASSOCIATED WITH DECREASED COLORECTAL CANCER MORTALITY IN THE VETERANS AFFAIRS SYSTEM Charles J. Kahi, Heiko Pohl, Laura Myers, Dalia A. Mobarek, Douglas J. Robertson, Thomas F. Imperiale Background and objectives: Colonoscopy is widely used in the Veterans Affairs (VA) system for colorectal cancer (CRC) prevention. However, the effect of colonoscopy on CRC mortality in the VA system is unclear. We aimed to determine whether exposure to colonoscopy is associated with decreased CRC mortality in Veterans, and whether the effect of colonoscopy on CRC mortality differs by anatomic location of the cancer. Methods: We designed a casecontrol study using national VA-Medicare administrative data. Cases were Veterans aged 52 or older diagnosed with CRC between 1/1/2003 and 12/31/2008, and died of CRC by 12/ 31/2010. Cancer diagnoses and death from CRC were ascertained from the VA Central Cancer Registry and National Death Index, respectively. Cases were age-, sex-, and facilitymatched to 4 controls that had not died of CRC. Patients with a prior history of CRC, colon surgery, inflammatory bowel disease, or familial polyposis were excluded. We determined the exposure of cases and controls to colonoscopy from 1997 to a date 6 months before CRC diagnosis in cases, and to a corresponding date in controls. Cancers were categorized as right-sided (cecum, ascending colon, hepatic flexure, transverse colon) and left-sided (splenic flexure, descending colon, sigmoid colon, rectum). Odds ratios for exposure to colonoscopy were calculated, adjusted for Charlson comorbidity score, aspirin use, and family history of CRC. Colonoscopy indication was classified based on a validated algorithm, and subgroup analysis was performed for patients who had undergone a colonoscopy for screening. Results: 7363 cases and 29,452 controls were identified; mean age was 73.3 years, and 98.8% were male. CRC was right-sided in 2993 (40.7%), left-sided in 3965
AGA Abstracts
S-22