Authors: Yuqi Wu (1Department of Informatics in Health, Mayo Clinic College of Medicine, Rochester, MN), Piet C. de Groen (2Division of Gastroenterology, Hepatology and Nutrition, University of Minnesota, Minneapolis, MN), Nabil Natafgi (3Department of Health Services Policy and Management, University of South Carolina, Columbia, SC), Chao Cai (4Clinical Pharmacy and Outcomes Sciences, University of South Carolina, Columbia, SC), Dezhi Wu (5Department of Integrated Information Technology, University of South Carolina, Columbia, SC), Sudha Xirasagar (3Department of Health Services Policy and Management, University of South Carolina, Columbia, SC)
Categories: Article
Source: Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology
Authors: Yuqi Wu, Piet C. de Groen, Nabil Natafgi, Chao Cai, Dezhi Wu, Sudha Xirasagar
Black Americans experience higher colorectal cancer (CRC) incidence than Whites, despite undergoing prevention screenings similar to Whites since 2010. We compared Black and White patients’ colorectal adenoma status at screening colonoscopy, a measure of CRC risk.
Using cross-sectional, observational data, we studied colorectal adenomas at first-time screening colonoscopy of average risk patients aged 40–89 years, screened between September 2001 and July 2016 in South Carolina. We studied age-adjusted odds of Black men and women (vs White), of having adenoma, advanced adenoma, ≥3 non-advanced adenomas, and right hyperplastic polyp, and compared their total polyp burden (sum of diameters of all adenomas and right polyps detected).
Among 28,100 patients, 58.4% Black, 53.8% women, we found that Black patients had lower age- and gender-adjusted odds vs. Whites to have adenoma (OR=0.88, p<0.01), and right hyperplastic polyp (OR=0.74, p<0.01), with similar pattern within gender groups. Black and White patients were similar on advanced adenoma, and 3+ non-advanced adenoma. Among patients with lesions, mean polyp burden ranged from 8.5mm(±7.2) for Black women aged 40–49 years, to 12.3mm (±7.4) for Black men aged over 70. Age-adjusted polyp burden was 0.4mm higher for Black men than White men, and 0.3 mm lower for Black women than White women patient groups (p<0.01).
In a large, racially balanced patient sample, Black and White patients showed similar polyp profiles.
Given similar adenoma status, other evidence-supported clinical factors associated with suboptimal polyp detection should be explored to understand the continuing CRC disparities affecting Black Americans.
Colorectal cancer (CRC) is the second leading cause of cancer death in the US, with an estimated 153,000 Americans diagnosed with CRC in 2023^1^. Most CRCs develop from precancerous polyps which can be detected and removed by colonoscopy screening before they progress to cancer^1,2^. About 80% of CRCs are potentially preventable by screening colonoscopy beginning at 50 years of age^2,3^. Recent data show that Black Americans experience 15% higher CRC incidence and 33% higher CRC mortality than White Americans^1^. The recent rates represent a substantial reduction from the 2005 rates of 26% and 52%, respectively^4^. The disparities reductions have occurred after sustained increases in colonoscopy screening among the Black population^5,6,7^. However, the current CRC disparities are disproportionate to the sustained low differences in colonoscopy screening rates (by only 2–3 percentage points) observed from 2010 through 2021^5,6,7^. In 2021, colonoscopy screening rates were 60% and 63%, respectively among the US Black and White populations^6^. It has been hypothesized that the continuing CRC disparities may be due to greater CRC risk among the Black population^8^. Given that about 90% of CRCs follow the adenoma-to-carcinoma pathway, at least part of the Black-White differences in CRC risk should be reflected in the prevalence rates of adenomas and advanced adenomas at screening colonoscopy^2^.
A recent narrative review reported mixed findings on racial differences in adenoma profiles^8^. None of the studies reviewed had included large enough samples of screening colonoscopies with balanced representation of Black and White patients. We studied polyp data among a large, racially balanced series of screening colonoscopies performed in South Carolina^9^, a state from the southern US where over 50% of the US Black population lives^10,11^. The study center used a colonoscopy procedure protocol designed to maximize polyp detection and complete removal of abnormal tissue, described earlier^3,9^. All participating endoscopists were required to adhere to the colonoscopy procedure protocol and the medical records documentation protocol. The center’s completeness of polyp detection and removal are indirectly validated by prior studies showing high adenoma detection rates (24.2% to 42.3% adenoma detection rates among endoscopists with total procedure experience of 200 or more), high CRC incidence and mortality reductions relative to the South Carolina general population^9^, and high reduction in their post-colonoscopy lifetime CRC risk relative to the US SEER-18 population^3^. The observed CRC incidence reduction was similar to that of the National Polyp Study, the only documented, prospective study of colonoscopic polyp removal for CRC prevention^2^. The above findings support that the study center achieved near-complete adenoma clearance at screening colonoscopy in this cohort, in turn, supporting the validity of studying racial differences in their precancerous polyp status. We studied Black-White differences in colorectal adenoma prevalence and characteristics at screening colonoscopy among this cohort. Given the study center’s geographic location and large, racially balanced sample, polyp findings from this cohort may better generalize to the US Black population compared to previous studies that had low representation of Black patients and conducted in states with sparse Black populations.
This is a retrospective, cross-sectional, observational study of adenoma- and right hyperplastic polyp status of 28,425 adults provided screening colonoscopy at a licensed endoscopy center in South Carolina from September 4, 2001, to July 28, 2016. The center’s colonoscopy protocol and study inclusion/exclusion criteria were documented earlier^3^. By policy, the study center primarily limited its services to screening and surveillance colonoscopy of average-risk persons for CRC. Study-eligible patients were aged 40–89 years with first-time screening colonoscopy, no history of prior bowel disease or significant bowel symptoms, and no history of the known genetic pathologies/predisposition to CRC. Occasionally, patients with bowel pathology or genetic CRC predisposition were provided diagnostic colonoscopy at the center by visiting gastroenterologists; these patients were excluded by electronic text searches using suitable search terms in the “Physician notes” and “Referral notes” fields in the electronic patient record databases (Microsoft Excel prior to 2015 and G-Med software during 2015–16). Search terms for use were directed by the gastroenterologist coauthor (PCG), and records flagged to have the terms were manually reviewed to verify prior bowel disease or genetic predisposition to determine study eligibility. Search terms used were “Crohn’s disease”, “inflammatory bowel disease”, IBD, multiple polyposis, other syndromes, and bowel symptoms as well as related acronyms and synonyms.
All polyp data were documented during the procedure by a notetaker assistant (a trained endoscopy technician who also served as an additional observer of the video-screen to look for polyps). When diminutive polyps (≤5 mm) were found in multiples within a colonic segment, the notetaker documented the number in the “Polyp quantity” field associated with the respective colon segment. Per the US Multi-Society Task Force guidelines, a typical polyp or sample of polyps was retrieved for histopathology^12,13,14,15,16,17,18^, and the remaining polyps destroyed in place by hot snare technique. The detailed colonoscopy performance protocol including documentation procedures and the personnel involved are published^3,9^. Polyp quantity in each colon segment was used to calculate the patient’s total polyp burden. Total polyp burden was defined as the sum of the diameters (in mm) of all adenomas and right hyperplastic polyps, (measured along the longest dimension for oblong or irregular polyps). We excluded normal tissue polyps as well as left colon and rectal hyperplastic polyps. We also studied the odds of right (proximal) colon hyperplastic polyp separately, as this lesion is now considered a proxy for sessile serrated adenoma. The latter terminology was not consistently reported in histopathology reports until recently.
We excluded patients of other races due to small numbers and the study aim, to assess Black-White differences in adenoma profile. We used logistic regression analysis to assess the odds of Black patients vs. White having adenoma, advanced adenoma, advanced or ≥3 non-advanced adenomas, and right hyperplastic polyp. After excluding patients without adenoma or right hyperplastic polyp, among those with adenoma or right polyp, we studied Black-White differences in the total polyp burden among men and women separately, using two-part regression modeling. We used Tukey’s test to compare the mean polyp burden of Black vs. White men, and Black vs. White women within each age group. We used SAS v.9.4 and two-sided p<0.05 for statistical significance.
The study was approved by the University of South Carolina Institutional Review Board. Informed consent was waived because the data were deidentified within the endoscopy center’s data platforms before extraction into the university research team’s computers. Also, consent was deemed not practical in view of retrospective nature of the data accrued from over 25,000 patients over more than a decade. The study was conducted and reported in conformity with the Declaration of Helsinki. The manuscript complies with all the STROBE guidelines applicable to a retrospective, cross-sectional study.
Data used for this study cannot be made available due to signed undertakings with the endoscopy center management not to share data with third parties to protect patient privacy and confidentiality. Despite the data being deidentified, combinations of various variables could potentially identify individual patients. Derived data from statistical outputs reflecting the study findings are available from the corresponding author upon reasonable request.
Of 28,425 patients, 325 were excluded due to missing data (55 missing polyp histology, 270 missing gender and race). Table 1 presents the characteristics of the final study sample of 28,100 patients, 46.2% men, 54.8% Black, mean age 58.1 years, 30.8% had adenoma, 6.6% had advanced adenoma, 4% had ≥3 non-advanced adenomas, and 6.6% had right hyperplastic polyp. Overall, 34.6% of patients had adenoma or right hyperplastic polyp. The sample mean polyp burden (across all patients, with and without adenoma/right polyp) was 3.4mm (±9.9mm). Compared to White patients, Black patients had lower prevalence of adenoma, (traditional) advanced adenoma, advanced or ≥3 non-advanced adenomas, and right hyperplastic polyp, as well as 0.7mm lower total polyp burden (all p <0.05). Table 2 presents the findings by men had higher prevalence rates of all adenoma types than women, and higher mean polyp burden, mean difference 1.5mm (all p <0.05).
Table 3 presents the adjusted odds of adenomas by type. Black race was associated with lower age- and gender-adjusted odds of adenoma (OR=0.88, 95% CI 0.84–0.93, p<0.01), and right hyperplastic polyp (OR=0.74, 95% CI 0.67–0.82, p<0.01). Table 3 also shows that the age-and race-adjusted model showed lower odds of women (vs. men) to have all categories of adenoma – odds ratio for any adenoma (OR=0.64, 95% CI 0.61–0.68, p<0.01), right hyperplastic polyp (OR=0.82, 95% CI 0.74–0.90, p<0.01), advanced adenoma, and either advanced or ≥3 non-advanced adenomas, all p<0.01. Table 4 presents within-gender odds ratios by race. Among men, Black men had lower age-adjusted odds of both adenoma and right hyperplastic polyp relative to White men. Among women, Black women had lower age-adjusted odds of right hyperplastic polyp than White women. Black and White men were similar in respect of traditional advanced adenoma and advanced adenoma defined to include 3 or more non-advanced adenomas. Black and White women were also similar in respect of these adenomas.
Table 5 presents the racial differences in total polyp burden among the patient subgroup with adenoma and/or right hyperplastic polyp. Among Black men, the unadjusted mean polyp burden ranged from 8.9mm (±6.9) in the 40–49 age group, to 12.3mm (±7.4) in the 70+ age group. Among White men the corresponding range of polyp burden were 8.6mm (±4.6) and 11.8mm (±4.6). The corresponding age groups’ mean polyp burden among Black women were 8.5mm (±7.2) and 10.2mm (±5.6), and among White women, 8.8mm (±4.5) and 10.4mm (±3.8), respectively. Tukey’s tests showed that after adjusting for age, within each age group, Black men had 0.3–0.5 mm higher polyp burden than Whites (marginal, when viewed against the magnitude of mean polyp burden), all p<0.01. Among women, the polyp burden showed the opposite association with race, with small magnitudes of 0.2–0.3 mm higher polyp burden among White women compared to Black women in all age groups, all p<0.01. Statistical significance despite small magnitudes of difference is attributable to the large samples available for each comparison.
This study did not find an excess CRC risk as measured by precancerous polyp status among Black patients. On the contrary, Black men and women were slightly less likely than White men and women, respectively, to have adenoma or right hyperplastic polyp. The observed Black-White differences in polyp burden (range of differences, 0.2–0.5mm within age groups) were statistically significant owing to large samples, although negligible in terms of effect size when viewed against the magnitude of polyp burden, mean ranging from 8.6mm to 12.3mm among age-gender subgroups).
The study’s significance for the literature on precancerous polyps stems from several study strengths. First, it presents polyp data on a documented screening colonoscopy cohort that showed high levels of CRC protection (incidence and mortality) over prolonged follow-up compared to the general population^3,9^. Over 8-year follow-up, the study cohort showed 67% reduction in their lifetime CRC risk relative to the SEER-18 population, which itself is not a screening-naïve population but had 30–58% colonoscopy screening rates during the study years^3,^. Further, in our study cohort, Black and White patients showed similar CRC prevention rates. The observed cancer prevention rate supports that the cohort had received high-quality colonoscopy procedures with near-100% adenoma detection and complete removal. This in turn, presents criterion-related validity for a conclusion that precancerous polyp status is indeed similar among Blacks and Whites. By comparison, in other studies, high rates of missed adenomas and high interval cancer rates undermined the validity of polyp studies on such cohorts ^3,19^.
Second, our findings are drawn from a cohort limited to screening colonoscopy of average-risk adults, most being primary care patients referred for screening colonoscopy. The occasional diagnostic colonoscopy performed by visiting gastroenterologists using the center’s facilities were excluded by electronic text search processes of physician and referral notes followed by manual review of the flagged records. In our earlier follow-up study of the 2001–2014 segment of the cohort, 50 patients out of 26,011 patients had diagnostic colonoscopy^3^. Third, our relatively large study sample was racially balanced (55% Black), largely because of a high Black population in South Carolina (28%), and because majority of the referring primary care physicians were either Black physicians, or White physicians serving predominantly Black towns and neighborhoods. Minority physicians are known to serve predominantly Black, other minority, and indigent patients^20^. Black physicians’ patient panels are documented to be overwhelmingly Black, often over 90%^21,22^.
Fourth, the study center, located in the southern US, may showcase Black-White differences in precancerous polyp status, because the majority of the US Black population, 55%, resides in the southern states which share CRC-relevant dietary and lifestyle features^10,11^. Previous studies had low proportions of Black patients (range 4% to 12%), and were based in geographic regions with small Black populations^23,24,25^.
Other unique contributions of this study include the i) reporting on right hyperplastic polyps, a proxy for sessile serrated adenoma (little reported in the literature, and when studied, examined samples with few Black patients), ii) reporting on ≥3 non-advanced adenomas (not reported in the racial disparities literature thus far), and iii) reporting on the total polyp burden, a new measure, not reported in the literature thus far. Our study was able to calculate the total polyp burden due to the center’s policy of documenting the number of polyps in each colon segment in real-time, during the colonoscopy, a data point that was rarely documented in most endoscopy centers.
Unlike other studies, our study cohort had an unusual proportion of patients aged below 50 years despite the prevailing US Preventive Services Task Force guidelines at the time, to screen average-risk individuals from the age of 50 years^26^. The study center implemented an intentional policy of responding to the published literature on higher risk groups – persistent findings of high Black-White CRC disparities (about 6 years younger average age at CRC diagnosis among Black patients than Whites, and 26% and 52% higher CRC incidence and mortality, respectively, among the Black population during 2008–2011^4^, unchanged from 1998–2000 rates^27,28^. Other studies also showed that individuals had 2- to 6-fold higher CRC risk if they had first-degree relative(s) with CRC (increasing risk with younger age of the relative at diagnosis)^28^. The American College of Gastroenterology in its 2008 guidelines, recommended African Americans and persons having first-degree relatives with CRC to begin colonoscopy screening at age 45 (or earlier as appropriate)^29^. Responding to these findings, the study center encouraged referring primary care practices to screen early among these higher-than-average risk groups.
A recent narrative review by Rutter et al attempted to explain the excess CRC incidence among the US Black population by reviewing studies that explored whether Black people show higher CRC risk than Whites as represented by higher prevalence or progression of adenomas at colonoscopy^8^. Their review showed mixed findings. Out of eight studies reviewed, four reported higher adenoma prevalence among Black patients, two estimated similar prevalence, and two reported lower prevalence among Black patients than among Whites. The evidence on advanced adenomas and right colon (proximal) adenomas was also mixed. Our findings support studies showing no difference in adenoma rates of Black and White patients. Studies consistent with our findings include one study of 3,732 patients showing that male gender, but not race, predicted polyp prevalence,^23^ Wallace et al^24^, and Corley et al from Northern California (a large study sample with 4.2% Black patients)^25^. Based on the mixed pattern of findings, Rutter et al concluded that higher CRC incidence among the Black population may be due to differences in screening utilization^8^. However, a review of the screening colonoscopy literature suggests otherwise. The Black-White CRC screening gap has sharply narrowed since 2000 (when a wide gap existed). Since 2010, nationally representative surveys have consistently recorded 1–3 percentage point difference in Black-White colonoscopy completion rates^5,6,7^.
Our study suggests the need to look elsewhere to explain the prevailing Black-White disparity in CRC. Derivative findings from clinical studies of colonoscopy support a likely role of colonoscopy procedural challenges for adenoma detection and complete removal. Some studies suggest that the challenges may vary by race due to social determinants of health. Suboptimal bowel preparation which may impede polyp detection and removal, is encountered more often among Black patients ^30,31,32^. Poor bowel preparation may plausibly be due to poor comprehension of bowel preparation instructions, in turn, mediated by low general literacy and health literacy. In the typical pre-colonoscopy office visit, the patient is provided a printed brochure on bowel preparation, to be reviewed later ^24,25.26^. More recently, print materials were supplemented by mobile phone texts with links to online videos.^32,33.34^. The latter requires digital literacy and broadband access. General literacy, health literacy, digital literacy, and broadband access are challenges for some Black and low-SES patients, especially among the (older) CRC screening-eligible age groups^35^. In one study, low health literacy was associated with substantially poorer comprehension of a print leaflet on bowel preparation than among those with adequate health literacy^36^. In another study, 38% of Medicaid colonoscopy patients had suboptimal bowel preparation vs. 18% of private and Medicare insured patients; further, Medicaid patients were more likely to be Black and to have low SES than other insurance patients^31^. These studies suggest plausible underlying pathways by which low status on the social determinants of health may translate into poor bowel preparation, which is more frequently documented among Black patients.
In our study center, not only was suboptimal bowel preparation infrequent (6.3% of patients^3^) due to preemptive measures, but when it occurred, appropriate colonoscopy procedure adjustments were made to overcome the challenge. All pre-colonoscopy patient visits included in-person explanation of bowel preparation instructions, followed by a phone call around the time of the patient’s scheduled time to start bowel preparation. Patient navigation studies have empirically validated that low-SES patients achieve better bowel preparation with a timely phone call to reinforce in-person, verbal bowel prep instructions given at the office visit^37^. If suboptimal preparation was encountered, the study center’s colonoscopy protocol required mucosal flushing until the mucosal surface was exposed, a practice which is validated by closely similar rates of adenoma detection with optimum and suboptimal bowel preparation, 31.5% vs. 29.9% ^3^. In comparison, another study of repeat colonoscopy of patients with suboptimal bowel preparation showed a very high rate of missed adenomas, 33.8%, and high rate of missed advanced adenomas, 18%, detected at the repeat colonoscopy^38^. Assuming similar polyp rates among the Black and White population, the increased procedure time for mucosal washing (likely to be more often needed for Black patients^30,31,32^), may not be feasible in busy endoscopy practices with tight patient scheduling. This would imply greater risk of missed adenomas and consequently higher likelihood of post-colonoscopy interval cancer among Black patients. Future research should explore this aspect of screening colonoscopy quality and how it translates (or not) into CRC incidence disparities.
Another unstudied aspect of colonoscopy quality relevant to racial disparities is sedation type. Approved sedation types vary by health plan. Medicaid programs in many states including South Carolina, do not cover deep propofol sedation by continuous intravenous drip (performed by a nurse anesthetist), but Medicare and most private insurers cover this feature. Many states’ cover instead, the less costly, conscious sedation with benzodiazepine-opioid injection administered by the endoscopist prior to the procedure. Propofol sedation may facilitate longer procedure time because it maintains steady, deep sedation as long as needed for satisfactory procedure completion. Pre-procedure benzodiazepine-opioid injection could wear off, resulting in patient discomfort and pain which risks a rushed or inadequate procedure. Studies show significantly better patient comfort including less pain with propofol than pre-procedure injection^39,40^. One study, using a statewide colonoscopy registry database showed that propofol sedation was associated with 13% higher likelihood of sessile serrated polyp detection relative to conscious sedation^41^. Sedation challenges and suboptimal bowel preparation may both disproportionately affect lower SES and racial minority populations, suggested by the common factor of Medicaid associated with both challenges. There is a need for focused research studies on procedure duration, adenoma yield rate and cecum intubation rate as related to these two procedural variables.
Our study had some limitations. Although unlikely, it is possible that some colonoscopies at the study center were diagnostic procedures. Second, we found racial similarity in precancerous polyps in respect of histologic and morphologic characteristics; specifically, we did not study the molecular processes involved in the progression of abnormal tissue to invasive cancer or aggressive tumors. A third study limitation is the absence of data on individual bio-behavioral risk factors for CRC - family history of CRC, body mass index, smoking history, sedentariness, and diet. While this remains a limitation, it may be mitigated by the documented adenoma-carcinoma sequence driving more than 90% of CRCs. It is reasonable to anticipate that these underlying risk factors may be reflected in the individual’s adenoma status, as noted by earlier authors^8^. In-depth data on these factors should be studied in tandem with colonoscopy findings and molecular biology studies to evaluate the prevalence of these bio-behavioral factors by race, how these factors are associated with the patients’ adenoma status, and how these factors are associated with the type and pace of molecular processes involved in the cancerous transformation of polyps. Fourth, the study center’s databases did not have histology data on every polyp. (Diminutive polyps <5 mm were destroyed in place with a sample taken for histology when multiple polyps were found within a colon segment^12,13,14,15,16,17^). Another study limitation is single-center data from one state.
Our study does not support Black-White differences in precancerous polyp status at screening colonoscopy, thought to contribute to the prevailing CRC disparities despite racial parity in colonoscopy screening utilization for more than a decade^1,5,6,7^. Derivative findings from other studies support that research on other likely sources of CRC racial disparities may provide at least part of the explanation for CRC racial disparities, particularly, the colonoscopy procedure-related challenges to adenoma detection and complete removal which may vary systematically by race. Future research should explore racial differences in the occurrence of these challenges, the operational variables that lead to these challenges, and how these variables translate into higher CRC risk for the Black population. Research is also needed on the racial differences in the bio-behavioral risk factors for CRC and how they may translate into the occurrence of precancerous polyps and CRC risk.