Colonoscopy and Sigmoidoscopy Use among the Average-Risk Population for Colorectal Cancer: A Systematic Review and Trend Analysis

Colorectal cancer was estimated to be among the three most commonly diagnosed cancers and the second leading cause of cancer mortality worldwide in 2018, accounting for over 1.8 million new cases and nearly 900,000 deaths (1). Despite its high burden, colorectal cancer is characterized by a relatively long preclinical phase, thus allowing for prevention and early detection by screening (2–5).

Recognizing its capacity to reduce colorectal cancer cases and deaths, several expert groups and societies have been recommending screening for persons at average risk for colorectal cancer, starting at ages 45, 50, or older (6–11). Consequently, an extensive roll-out of screening programs has been taking place worldwide over the past two decades (12). Among the various screening offers, lower gastrointestinal endoscopy has been proved to be particularly effective in reducing colorectal cancer incidence and mortality through observational studies (colonoscopy and sigmoidoscopy) and randomized controlled trials (sigmoidoscopy) (3, 13–15).

Two reviews have been previously conducted to summarize data on lower gastrointestinal endoscopy use (16, 17). They have reported substantial increases in utilization among the average-risk population in the US up to 2012 and a lower and wide variation in uptake in other countries. Given the recent implementation of colorectal cancer screening in many countries and the relevance of monitoring endoscopy use to estimate colorectal cancer burden in the future, we conducted a systematic review to update knowledge on lifetime and recent use of colonoscopy and sigmoidoscopy in the average-risk population for colorectal cancer. Furthermore, we analyzed recent trends in colonoscopy utilization in Germany and the US, where colonoscopy has been used as a primary screening modality.

Two previous systematic reviews had summarized evidence reported on colonoscopy and sigmoidoscopy utilization in the general population at average risk for colorectal cancer by original articles published up to October 31, 2016 (16, 17). Using the electronic databases PubMed and Web of Science and the search terms previously applied by Chen and colleagues (17) [(colonoscopy OR sigmoidoscopy OR ((colon OR colorectal) AND (endoscop* OR examinat*))) AND ((surveillance AND system) OR (health AND survey) OR (population AND based)], we updated the previous reviews by including the search period November 1, 2016–December 31, 2018. In addition to the search in the databases, the references of the eligible reports were screened and relevant websites were checked for further studies. The PRISMA recommendations for reporting systematic reviews were followed (18).

Studies were considered eligible for inclusion if they reported nation- or region-wide estimates of lifetime or recent use of colonoscopy and/or sigmoidoscopy in the general population (or in specific ethnic groups) ages 50 and older. Studies with sample sizes smaller than 150 were not included. Moreover, studies which only included individuals at higher risk for colorectal cancer (e.g., family history of colorectal cancer, ulcerative colitis, Crohn's disease and other inflammatory bowel diseases) were excluded, as were those which merely focused on colonoscopy use as follow-up of positive stool-based tests. Finally, studies with less recent estimates which have already been summarized in the two previous reviews or in other included reports with more detailed data were not considered. No language restrictions were applied.

Two researchers (R. Cardoso and T. Niedermaier) independently extracted relevant data from the included studies into a standardized form. The information extracted from the studies were: first author, year of publication, country, data source, year(s) of data collection, target population, sample size, age range, geographical representativeness of the study and data regarding lower gastrointestinal endoscopy use (type of test, time frame, and overall, sex- and age-specific estimates of utilization). Disagreements in data extracted were resolved through additional review and discussion.

Estimates of lower gastrointestinal endoscopy use were summarized and analyzed by type of test, that is, (i) colonoscopy; (ii) sigmoidoscopy; or (iii) colonoscopy or sigmoidoscopy. Given the large number of reports from the US, we separated our analysis into US and all other countries. Weighted estimates taking the sampling scheme into account were selected rather than unweighted (crude) estimates when both were provided. For some studies, estimates were calculated in case they were not directly available but could be determined on the basis of additional data provided. For example, if estimates were available by race/ethnicity or body mass index, we calculated overall utilization by combining all subgroup estimates. To explore time trends in lower gastrointestinal endoscopy use, data from the two previous reviews were extracted and plotted by year of data collection and country. Because of limited available data on sigmoidoscopy use for several time points, such analysis was restricted to colonoscopy and colonoscopy/sigmoidoscopy use combined. In addition, age- and sex-specific estimates were analyzed when available.

To analyze recent trends in colonoscopy use within 10 years in Germany, estimates reported in Starker and colleagues (19) based on data from the first wave of the German Health Interview Survey (DEGS1), were used alongside estimates determined from the second wave of the European Health Interview Survey (EHIS) data (20). Both the DEGS1 and the EHIS data allow for calculation of nationally representative estimates from self-reports of adult residents in Germany. The DEGS1 was carried out in 2008–2011 and reached approximately 8,000 participants (21). In 2014–2015, the EHIS was conducted which involved nearly 25,000 participants (22). Individual weights accounting for the sampling scheme and demography, structure and population size of the federal states were applied to obtain nationally representative estimates from the EHIS data.

For the US, we analyzed data from the National Health Interview Survey (NHIS), a cross-sectional in-person interview survey carried out by the Centers for Disease Control and Prevention (23). The NHIS is directed at the non-institutionalized US population and aims to produce nationally representative estimates of sociodemographic and health topics based on self-reports. For this analysis, publicly available data for the years 2010 (n = 27,157) and 2015 (n = 33,672) were used and weighted for the complex study design to obtain estimates of colonoscopy use within 10 years (24).

The statistical software SAS 9.4 (SAS Institute Inc.) was used to calculate the estimates from the EHIS and the NHIS data.

The electronic search yielded 983 reports, which were screened for the title and abstract. Among them, 72 qualified for full-text screening and 23 were considered eligible for the review. Fourteen reports provided estimates for the US (25–38). The other studies were conducted in Australia (one study; ref. 39), Canada (two studies; refs. 40, 41), England (one study; ref. 42), Germany (one study; ref. 43), Saudi Arabia (one study; ref. 44), and South Korea (three studies; refs. 45–47). The study selection procedure is summarized in Fig. 1.

Several studies from the US reported estimates from two nationwide surveys [NHIS (refs. 25, 32, 33) and the Behavioral Risk Factor Surveillance System (refs. 34, 37)]. From the remaining US reports, seven provided prevalence estimates from epidemiological studies [three from baseline measures of randomized controlled trials (refs. 27, 30, 35), two from cohort studies (refs. 28, 29) and another two from a cross-sectional study (refs. 31, 38)], one from medical records (36) and one from Medicaid and commercial claims (26). Three studies provided nationally representative estimates (based on NHIS; refs. 25, 32, 33) and another three only included specific ethnic groups [Chinese or Chinese Americans (27), African Americans (30) and Koreans (35)].

As for non-US countries, most of the estimates were determined from cross-sectional survey data (seven out of nine studies; refs. 39, 42–47). The other two reports provided estimates from a cohort study (41) and registry data (40). In total, six studies conducted in England (42), Germany (43), Saudi Arabia (44), and South Korea (45–47) provided nationally representative estimates.

For the US (Table 1), lifetime utilization of colonoscopy was reported in two studies, both on specific racial groups [Chinese or Chinese Americans living in San Francisco: 31% (27); African Americans from two metropolitan areas: 78% (30)]. Recent use of colonoscopy, mostly defined as colonoscopy use within 10 years, was reported in seven studies (25, 26, 28, 29, 31, 36, 38) and varied from 9% (members of a large community-based healthcare delivery system in Northern California in 2000; ref. 29) to 80% (White or Black adults living in the 34 Indiana counties with above-average cancer mortality rates in 2018; ref. 38). A study based on NHIS data from 2013 was the only one that provided nationally representative estimates (25). Overall, the extracted data point toward an ongoing increase in colonoscopy use over the last two decades [observed in Smith and colleagues (36) and Levin and colleagues (29)]. This trend can be clearly identified in Fig. 2, which shows estimates of endoscopy use reported in the literature up to now.

As far as sigmoidoscopy utilization is concerned, six studies were included [four on sigmoidoscopy use within five years (26, 29, 31, 38), one on lifetime sigmoidoscopy use (27) and one on sigmoidoscopy use assessed over a four-year period (28)]. Overall, the estimates for sigmoidoscopy use were substantially lower when compared with colonoscopy use (0.4%–27%) and seem to have strongly decreased over time. For instance, in Northern California sigmoidoscopy uptake fell from 27% in 2000 to 2% in 2015 (29).

Of the five articles reporting on colonoscopy or sigmoidoscopy use (combined; refs. 32–35, 37), nationally representative estimates were provided in two studies, both on colonoscopy use within 10 years or sigmoidoscopy use within 5 years (32, 33). Estimates ranged from 51% to 60% (data from NHIS 2010 and 2015, respectively). No differences were observed between men and women. Age-specific estimates were provided in one study and revealed higher uptake among older individuals (≥65 years) than among those ages 50 to 64 years (33). Sex- and age-specific estimates obtained from the NHIS data are further addressed in this review under Trend analysis of colonoscopy utilization in Germany and the US.

In addition, three studies reported race-specific estimates (31, 33, 38). Two studies (including one nationally representative study; ref. 33) described Whites as the group with the highest uptake (33, 38). The other study, conducted in the context of an organized screening program in Southern California, identified a higher proportion of Asians than non-Hispanic Whites being up-to-date with screening colonoscopy (31).

With respect to non-US countries (Table 2), six studies on colonoscopy utilization were included [one study each from Australia (39), Canada (40), Germany (43), and South Arabia (44) and two studies from South Korea (45, 46)]. Estimates of recent use of colonoscopy varied considerably among the different countries. Estimates for Germany (data from 2014–2015: 59%) were comparable to the estimates for the US in the same time period and were substantially higher than for other non-US countries. The lowest prevalence estimate of colonoscopy utilization among the included reports was seen in Saudi Arabia (data from 2006–2007: 1%; ref. 39).

As far as sigmoidoscopy use is concerned, one study carried out in England was identified (42). Overall prevalence was very low among the general adult population (8%); however, among those who had ever been invited for sigmoidoscopy, the uptake was substantially higher (69%).

Concerning colonoscopy and sigmoidoscopy utilization combined, two studies were identified [one conducted in the Canadian province Alberta (41) and the other one in South Korea (47)]. Lower gastrointestinal endoscopy uptake was rather low in Alberta (12%); however, the estimates dated back to 2000–2006 when utilization rates were rather low in any country other than the US. As for South Korea, estimates (2010–2012) were close to 60%, and hence in line with those reported for the US.

When exploring time-trends in lower gastrointestinal endoscopy use, and particularly colonoscopy use, substantial increases were seen in Germany and South Korea where uptake levels close to the ones reported in the US have been reached in recent years (Fig. 2). There was also a considerable increase in colonoscopy/sigmoidoscopy use in Canada from 2008 (24%) to 2012 (37%) and the most recent region-wide study also points in that direction (colonoscopy use in Toronto, 2014: 45%). For other countries, no evaluation of trends was possible due to limited data at different points in time.

Analyzing sex-specific estimates, unlike the pattern observed for colonoscopy/sigmoidoscopy use in Alberta, colonoscopy use was higher in men than in women in both Germany and South Korea. Age-specific estimates were provided in three studies (43, 45, 46). In South Korea, colonoscopy use was higher in age group 60–69 than in age groups 50–59 and 70–74. In Germany, estimates were highest in age group ≥70 among men and in age group 65–69 among women. Complementary analyses for Germany are presented under Trend analysis of colonoscopy utilization in Germany and the US and displayed in Fig. 3A.

The estimates of colonoscopy use within 10 years in Germany for 2008–2011 and 2014–2015 are displayed in Fig. 3A (see Supplementary Table S1 for table containing the corresponding estimates). There was an increase in colonoscopy uptake over the years among both men and women (from 54.8% to 58.5% for both sexes combined), but the increase was much larger among men. Although in 2008–2011 the utilization of colonoscopy was slightly higher in women (men: 54.1%; women: 55.4%), in 2014–2015 men reported substantially higher uptake (men: 61.0%; women 56.3%). The estimates were lowest for the youngest age group for which screening colonoscopy was offered in Germany (55–59 years), and even decreased over time among women by approximately four percentage points (from 47.2% to 43.1%).

In the US, the estimates of colonoscopy use within 10 years also increased from 2010 (54.5%) to 2015 (59.0%) and the same age distribution pattern identified for Germany was observed. However, contrary to the trend described for Germany, uptake increased across all age groups for both men and women, including among women ages 50 to 59 years. Colonoscopy use among women was higher than among men in the younger age group (50–59 years) and lower than among men in the age groups ≥60 (Fig. 3B; Supplementary Table S2).

This review presents an updated summary on lifetime and recent use of colonoscopy and sigmoidoscopy in various countries. In comparison with the previous reviews, further studies with more recent data on colonoscopy and sigmoidoscopy use among the average-risk population for colorectal cancer became available. Specifically, new national estimates were identified for the US, Germany, and South Korea, where endoscopy use has been monitored through nationwide cross-sectional surveys carried out on a regular basis, and for England and Saudi Arabia for the first time. Yet, data keep lacking for many regions and countries, notably all countries in Africa and Central and South America.

In the US, colonoscopy use within 10 years has steadily increased over the past two decades and reached about 60% of those at average risk of developing the disease. At the same time, sigmoidoscopy utilization has sharply declined and seems to be rarely used any more as suggested by the regional studies we identified. The establishment of colonoscopy as the primary mode of colorectal cancer screening and the decline in sigmoidoscopy use can be potentially related to a variety of factors, such as the physicians' overall conviction that colonoscopy constitutes the best available test (48) and recommendations from medical societies. Specifically, the American College of Gastroenterology recommends colonoscopy as the preferred colorectal cancer prevention test and sigmoidoscopy as an alternative option in certain settings only (6). Besides, robust evidence concerning the public health benefits of colonoscopy use in the US has been generated (4, 49–51). For instance, Chen and colleagues (4) estimated that the proportion of colorectal cancer–related deaths prevented by colonoscopy was 29% in 2008–2009 and 34% in 2010–2011.

In spite of overall increases in colonoscopy use in the US, our results raise concerns with regards to the younger age groups (50–60 years) even though Medicare and most private health insurances cover screening colonoscopy from age 50 on. This trend, which is not justified by an increasing use of stool-based tests (NHIS 2015: 6% among those ages 50–64), becomes particularly disquieting when looking at the colorectal cancer burden (notably, the increasing incidence rates) in younger birth cohorts in the US (33), and led the American Cancer Society to recommend starting screening at age 45 instead of 50 (11).

Regarding lower gastrointestinal endoscopy use in European countries, limited data have become available for recent years (only for Germany and England). As of 2002, screening colonoscopy has been offered to the average-risk population ages 55 or older in Germany, and prevalence estimates of colonoscopy utilization close to those described for the US have been reached in recent years. Already in 2008 to 2011, 31% of colorectal cancer–related deaths were estimated to have been prevented due to colonoscopy use (4). Nonetheless, in the same vein as the US, the utilization among the younger age groups continues to be rather low, with no further increase observed in recent years. Furthermore, such estimates have been accompanied by decreasing uptake levels of fecal occult blood test (52), suggesting that a large proportion of this younger population is not up-to-date with colorectal cancer screening.

As for England, the low uptake observed for sigmoidoscopy on the national level has to be interpreted in the context of the gradual introduction of sigmoidoscopy in the English Bowel Screening Program. Roll-out of an organized program with invitations to screening sigmoidoscopy for people at age 55 started in 2013 and full coverage is expected to be achieved by 2021 only (12, 13, 53). Geurts and colleagues (54) have estimated that if uptake is 50%, about 10,000 colorectal cancer cases and 2,000 colorectal cancer–related deaths will be prevented by 2030. The high uptake rate of 69% among participants invited to sigmoidoscopy suggests that even more cases and deaths may be prevented in the years to come.

Given the lack of nation- and region-wide reports on lower gastrointestinal endoscopy use from European countries, surveys such as the EHIS, which are able to generate detailed and comparable data for all European Union member states, are of great relevance for monitoring purposes and constitute an extremely useful tool to drive policy action. The second wave of the survey, which took place between 2013 and 2015, included questions concerning colonoscopy and fecal occult blood test utilization (20). When disclosed for all countries, its results will strongly enhance the empirical basis for monitoring and further improvement of screening strategies.

Among Asian and Middle Eastern countries, to our knowledge only Palestine and Singapore (for which estimates have been summarized in a previous review; ref. 17), and Saudi Arabia and South Korea have reported prevalence estimates of endoscopy use among the average-risk population for colorectal cancer. Except for South Korea, utilization rates are extremely low and reflect the lack of colorectal cancer screening implementation, public health education as well as financial, cultural, religious, and psychological barriers to screening (12, 44, 55, 56). South Korea, although with relatively lower utilization rates than the US and Germany, has seen the steepest increase in colonoscopy use in recent years. If the trend remains, the high colorectal cancer burden in the country (age-standardized incidence rate, 2018: 31.3/100,000; ref. 57) is expected to be considerably reduced in the next years.

Like South Korea, Australia is among the countries with the highest colorectal cancer incidence rates in the world (36.9/100,000; ref. 57). Although no nationwide report was identified from Australia, a study carried out in outpatient clinics of a major regional hospital in New South Wales revealed rather low utilization of colonoscopy. The low use of colonoscopy (as a first-line screening exam) among the average-risk population would be expected given the ongoing roll-out of an organized screening program with fecal immunochemical test (12, 58). Nevertheless, the low uptake of stool tests reported in the included study (47%) and nationally (2014–2015: 41%; ref. 59) calls for further action to reduce the colorectal cancer burden in the country.

As far as African and Central and South American countries are concerned, no studies were identified that met our eligibility criteria. The absence of studies from these regions has also been reported in the two previous reviews (16, 17), suggesting limited endoscopy utilization. The need for, and feasibility of, population-based colorectal cancer screening in these regions has been questioned due to the relatively low colorectal cancer incidence and mortality, high burden of communicable diseases and lack of resources (60, 61). Nevertheless, the absence of quality registry data (ref. 62; which might largely underestimate colorectal cancer cases and deaths) and the ongoing transition toward the adoption of Western/unhealthy lifestyles (which will most likely contribute to a substantial increase in colorectal cancer burden in the near future; ref. 63) call for developments in terms of secondary prevention of the disease. A modeling study conducted by Ginsberg and colleagues (64) estimated that screening colonoscopy at age 50 together with treatment can be cost-effective even in sub-Saharan Africa.

When looking at the large variations in lower gastrointestinal endoscopy use across the different countries, the distinct characteristics of the health care systems in relation to colorectal cancer screening offer seem to play an important role. For instance, in Germany (65), South Korea (45), and the US (66), colorectal cancer screening is fully, or almost fully, covered by health or social insurances, whereas no such offers are in place in most other countries. Especially in the US, the Affordable Care Act has been considered key to achieve national screening goals and reach the most deprived stratum of the population (67, 68).

Although our study focused on populations ages 50 years and older, the search terms used for literature search were not limited to any specific age group. Nevertheless, we did not identify any report which provided estimates for the age group <50 years separately. A study conducted in South Korea across all age groups was identified; however, no age-specific estimates were revealed (69). Given the recent increase in colorectal cancer incidence among those younger than 50 years in many regions and countries (33, 70–72), and potential reconsideration of current screening guidelines, assessing the uptake levels among these younger age groups may be of great importance in the near future.

This study has strengths and limitations that should be considered when interpreting its findings. The main strengths of this study include the use of two main databases for literature search and adherence to the recommendations for reporting systematic reviews by following PRISMA guidelines. In addition, time trends and subgroup analyses of colonoscopy use in Germany and the US could be explored by using data files of nationwide surveys. These trends and recent estimates might be used in future studies as model input to accurately project colorectal cancer incidence and mortality and to eventually adapt and optimize screening offers. Its limitations pertain to the studies and data included. First, a great proportion of the included articles reported estimates of lower gastrointestinal endoscopy use based on self-reports, which may be subject to over- or underreporting of actual use (73). Second, many studies focused on specific regions and their estimates may differ from the national estimates, which may also limit comparability across countries.

In summary, the majority of studies on lower gastrointestinal endoscopy use were identified from the US, with studies from other countries still being very limited. Although colonoscopy use has continuously increased over the last years, sigmoidoscopy use seems to have almost disappeared in the US. Moreover, this study emphasizes the need to develop strategies to reach the younger age groups at average risk for colorectal cancer. Besides the US, Germany and South Korea also show increasing colonoscopy utilization over the past few years. In Germany, recent estimates of colonoscopy utilization are very similar to the US estimates, and the same age distribution pattern was observed. For other countries, notably African and Central and South American countries, data on lower gastrointestinal endoscopy use were not identified, most likely reflecting low activity in screening colonoscopy and sigmoidoscopy.

Although colorectal cancer is one of the most preventable cancers and diseases, with enormous opportunities for primary, secondary, and tertiary prevention (74), additional efforts are needed if we are to reduce its burden. Specifically, increasing the utilization of colonoscopy or sigmoidoscopy as primary colorectal cancer screening methods or colonoscopy to follow-up positive stool-based tests would very likely lead to a substantial reduction in colorectal cancer burden in the future, as has already been shown in Germany and the US (4, 49–51, 75).

No potential conflicts of interest were disclosed.

The results and conclusions are authors' responsibility and do not represent the views of Eurostat and the European Commission.

Conception and design: H. Brenner

Development of methodology: R. Cardoso, H. Brenner

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C. Chen

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R. Cardoso, T. Niedermaier, C. Chen, H. Brenner

Writing, review, and/or revision of the manuscript: R. Cardoso, T. Niedermaier, C. Chen, M. Hoffmeister, H. Brenner

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R. Cardoso

Study supervision: M. Hoffmeister, H. Brenner

This study used data from second wave of the EHIS that we gratefully acknowledge. This work was supported in part by a grant from the German Federal Ministry of Education and Research (No. 01GL1712).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.