Increasing Colorectal Cancer Incidence Before and After Age 50: Implications for Screening Initiation and Promotion of “On-Time” Screening

Colorectal cancer (CRC) incidence rates among adults younger than 50 years are increasing, and have been attributed to a birth cohort effect after 1950 [1]. Since 2011, CRC incidence rates have also been increasing by 1% per year among ages 50–64 years [2]. These observations have led to recommendations to lower the screening initiation age to 45 years from the previously recommended 50 years, and to concerns regarding how to promote “on-time” screening at guideline-recommended ages to initiate screening [2‐8]. Differences by anatomic subsite have implications for understanding risk factors and informing cancer prevention and control strategies [1, 9]. Anatomic distribution of increases among adults aged < 50 years and changes in absolute numbers of cases have not been reported in detail by 5-year age group. To address these evidence gaps, we examined recent CRC incidence trends, stratified by anatomic subsite, among individuals age 40–59. The age range examined was just before and after age 50 years—the historically most commonly recommended initiation age for CRC screening [5‐7].

We analyzed CRC incidence data from United States Cancer Statistics (USCS), which provides 100% coverage of cancer incidence for all 50 states and the District of Columbia, utilizing the Centers for Disease Control and Prevention (CDC) National Program of Cancer Registries and the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) program [10]. Cancer subsites were defined as proximal (C18.0, C18.2–C18.5), distal (C18.6–C18.7), and rectal (C19.9, C20.9), and limited to primary colorectal adenocarcinomas [11]. Incidence rates were calculated using SEER*Stat v8.3.5, with rates expressed per 100,000 persons and age-adjusted to the 2000 U.S. standard population. The Joinpoint Regression Program (v4.8.0.1) was used to quantify Average Annual Percent Change (AAPC). AAPCs different from zero at p < 0.05 were considered to have a statistically significant “increase” or “decrease,” and otherwise were considered “stable.”

From 2003 to 2017, 470,458 cases of CRC were recorded among ages 40–59 years (Table 1). For all ages combined, rates were stable, with overall CRC incidence per 100,000 observed to be 35.4 in 2003 and 34.8 in 2017 (AAPC − 0.1; 95% CI: − 0.4–0.3). By age, overall CRC incidence per 100,000 increased between 2003 and 2017 from 13.4 to 15.8 for ages 40–44 years (AAPC 1.3; 95% CI 1.0–1.6) and from 24.8 to 29.1 for ages 45–49 years (AAPC 1.1; 95% CI 0.8–1.4). CRC incidence between 2003 and 2017 was stable for ages 50–54 years (AAPC 0.3; 95% CI: 0.0–0.7) and decreased from 72.7 to 56.7 for ages 55–59 years (AAPC − 1.8; 95% CI − 2.2–− 1.3).

Absolute number of rectal cancer cases per year increased from 1123 to 1239 for ages 40–44 years, 2011 to 2506 for ages 45–49 years, 3189 to 4173 for ages 50–54 years, and 3901 to 4566 for ages 55–59 years between 2003 and 2017 (Fig. 1). Rectal cancer incidence per 100,000 was stable for all ages combined between 2003 and 2017 (AAPC 0.5; 95% CI: − 0.2–1.3), but increased from 4.9 to 6.3 for ages 40–44 years (AAPC 2.1; 95% CI 1.5–2.7), 9.3 to 12.0 for ages 45–49 years (AAPC 1.5; 95% CI 1.1–1.4), and 16.7 to 19.5 for ages 50–54 years (AAPC 1.0; 95% CI 0.7–1.3). Incidence per 100,000 decreased from 24.7 to 20.8 for ages 55–59 years (AAPC − 1.2; 95% CI − 2.2– − 0.1; Fig. 2 and Table 1).

Absolute number of distal colon cancer cases per year increased from 936 to 1039 for age 40–44, 1646 to 1938 for ages 45–49 years, 2900 to 3327 for ages 50–54 years, and 3336 to 3682 for ages 55–59 years between 2003 and 2017 (Fig. 1). From 2003 to 2017, distal colon cancer incidence per 100,000 was stable for all ages combined between 2003 and 2017 (AAPC 0.2; 95% CI: − 0.7–1.1), but increased from 4.1 to 5.3 for ages 40–44 years (AAPC 2.0; 95% CI 1.5–2.4), and from 7.6 to 9.3 for ages 45–49 years (AAPC 1.5; 95% CI 1.1–1.8). Distal colon cancer incidence per 100,000 between 2003 and 2017 was stable for ages 50–54 years (15.2 in 2003 vs 15.6 in 2017, AAPC 0.4, 95% CI − 0.2–1.1) and decreased from 21.1 to 16.8 for ages 55–59 years (AAPC − 1.7; 95% CI − 2.6– − 0.7; Fig. 2 and Table 1).

Absolute number of proximal colon cancer cases per year between 2003 and 2017 decreased from 1010 to 821 for ages 40–44 years, 1726 to 1643 for ages 45–49 years, 3015 to 3012 for ages 50–54 years, and 4254 to 4209 for ages 55–59 years (Fig. 1). From 2003 to 2017, proximal cancer incidence per 100,000 decreased for all ages combined (AAPC − 1.2; 95% CI: − 2.1– − 0.3), but was stable for ages 40–44 years (4.4 vs 4.2 in 2017; AAPC − 0.1; 95% CI − 0.6–0.4) and for ages 45–49 years (7.9 vs 7.9 in 2017; AAPC 0.0; 95% CI − 0.3–0.4; Fig. 2). Incidence of proximal cancer per 100,000 decreased from 15.8 to 14.1 for ages 50–54 years (AAPC − 0.6; 95% CI − 1.0– − 0.3) and from 26.9 to 19.2 for ages 55–59 years (AAPC − 2.3; 95% CI − 2.8– − 1.8) between 2003 and 2017.

Utilizing a national sample of over 470,000 CRC cases observed between 2003 and 2017 within the US Cancer Statistics dataset, we found that increased risk trends for CRC under age 50 are also present after age 50. Specifically, we observed that rectal cancer incidence rates increased for ages 40–54 between 2003 and 2017. By 2017, counts and rates of rectal and distal cancers for ages 50–54 approached those reported for ages 55–59. We postulate that these trends are a consequence of increasing rates of rectal and distal cancer for ages 50–54 years along with decreasing rates of distal cancer among ages 55–59 years. Our data suggest that the observed increased risk for CRC under age 50 for birth cohorts born after 1950 is also present after age 50, despite the prior availability of screening for this group. As such, availability of screening for persons age 50 and beyond does not appear to “reset” the birth cohort effect for increased CRC risk observed under age 50. Rectal cancers often require multiple modalities for treatment, including radiation, chemotherapy, and surgery, and survivors often experience significant post-treatment morbidity [12]. The observed increase in absolute number of distal and rectal cancers, and in the incidence rate of rectal cancers age 50–54, highlights importance of screening uptake to reduce morbidity and mortality. CRCs, particularly rectal cancers, are detectable and preventable with currently available screening modalities such as the fecal immunochemical test and colonoscopy. However, in 2018, the proportion up-to-date with screening ranged from 61 to 63% for adults ages 50–64 years compared to 71–79% for adults ages ≥ 65 years [2, 13]. Furthermore, the proportion of individuals ages 50–54 years who were up-to-date with screening was just 48–50% [2, 12].

Our sample was large enough to allow for anatomic subsite-specific analysis of CRC trends, which enables observation of the drivers of increasing trends in CRC. We found changes in distal and rectal cancer incidence for individuals ages 40–44 and 45–49 years, and changes in rectal cancer incidence for individuals ages 50–54 years. We observed decreasing proximal cancer incidence for ages 50–54 and 55–59 years, and stable proximal incidence for ages 40–44 and 45–49 years. Data reported by Siegel et al., which has significant overlap with the US Cancer Statistics dataset data we used, have shown trends toward increasing CRC incidence from birth to age 49 for proximal, distal, and rectal cancer for time periods 2016 and earlier [2]. For the age group 50 to 64 years of age, Siegel et al. reported increasing incidence trends for cancers of the distal colon and rectum from 2012 to 2016, but not the proximal colon [2]. Our data confirm and extend Siegel et al.’s report by providing finer detail on incidence trends among age groups just before and after age 50. Additionally, our data emphasize that the increasing trend for the 50–64-year-old group reported by Siegel is currently driven by individuals age 50–54, underscoring the importance of promoting early initiation of screening at age 45, and the need to increase screening uptake for 50–54-year-olds in order to address current trends, and possibly head off future, even larger increases in age-specific incidence as birth cohorts from 1950 and younger age and grow older. Further, as the embryologic origins, molecular tumor characteristics, and associated microbial milieu may differ by anatomic site, our observation of differences in trend by site highlights the importance of considering site in studies examining risk factors for early-onset CRC, as well as for CRC early in middle age just before and after age 50 [9, 14‐17].

A strength of this study is the use of USCS, which provides a national assessment of case counts between 2003 and 2017. A limitation of our analysis is that we do not have data on whether increases in CRC incidence, particularly for those age 50 and older, were occurring among people exposed versus unexposed to colorectal cancer screening tests, and cannot therefore determine whether the increased risks observed were among people who were unscreened.

Taken together, increasing CRC incidence for both ages 45–49 and 50–54 years and the low prevalence of “on-time” screening at age 50 years underscore the importance of implementing interventions to increase screening for individuals age 50–54 years. A variety of interventions may be useful in increasing screening uptake in this patient population, including offering choice of tests, and organized mailed outreach invitations to complete screening [18]. Increasing CRC incidence among adults aged 45–49 and 50–54 years also supports recent American Cancer Society [4], American College of Gastroenterology [8], and United States Preventive Services Task Force [19] recommendations to initiate average risk screening at age 45, instead of age 50 [6, 7].