Data are from the Surveillance, Epidemiology, and End Results program and were age-standardized to the 2000 US standard population. Solid lines represent fitted piecewise–log linear trends. Average annual percentage changes (AAPCs) with parametric confidence intervals were calculated from 1995, when incidence started to increase.
aStatistically significantly different from zero at 95% significance thresholds.
aData are from the Surveillance, Epidemiology, and End Results program. 95% Confidence intervals were derived using the Sison-Glaz method for multinomial data. Statistical tests (2-sided Pearson χ2) assessed the stage-specific changes in proportions from 1990-1994 to 2011-2015. The differential relative increase by stage in the overall stage distribution was P<.001. Error bars indicate 95% CIs.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Meester RGS, Mannalithara A, Lansdorp-Vogelaar I, Ladabaum U. Trends in Incidence and Stage at Diagnosis of Colorectal Cancer in Adults Aged 40 Through 49 Years, 1975-2015. JAMA. 2019;321(19):1933–1934. doi:10.1001/jama.2019.3076
Evidence suggests that incidence of colorectal cancer (CRC) is increasing among adults younger than 50 years in the United States.1 The underlying causes of the increased incidence are unclear. If the increase is the result of earlier detection due to increased use of colonoscopy, earlier stage at diagnosis would be expected, whereas if the increased incidence is the result of true increases in risk, relatively later stage at diagnosis would be expected. We investigated trends in CRC incidence by stage in adults younger than 50 years. We focused our study on adults aged 40 through 49 years, who account for almost 3 of 4 young-onset cases.
Data from the Surveillance, Epidemiology, and End Results (SEER) program were used to assess trends in CRC incidence by stage, with stage defined as localized, regional, distant, and unstaged according to SEER historic staging. The Stanford University institutional review board has determined that studies using deidentified data, such as SEER, do not require review. We included all cases diagnosed among those aged 40 through 49 years during years 1975 through 2015 as available from the 9 original geographic areas within the SEER program,2 which include San Francisco and Oakland, California; Connecticut; Detroit, Michigan; Hawaii; Iowa; New Mexico; Seattle and Puget Sound, Washington; Utah; and Atlanta, Georgia. JoinPoint software version 188.8.131.52 was used to age-standardize incidence rates to the 2000 US standard population, estimate piecewise–log linear trends, and derive average annual percentage changes (AAPCs) since 1995, when the overall incidence started increasing. We also assessed the CRC stage distribution of cases among patients aged 40 through 49 years for 3 different 5-year periods: 1975 through 1979, 1990 through 1994, and 2011 through 2015. R statistical software version 3.4.3 was used to derive 95% confidence intervals for the multinomial stage proportions (Sison-Glaz) and to examine whether overall stage distribution and stage-specific proportions changed at 95% statistical significance thresholds (2-sided Pearson χ2 test).
There were a total of 29 532 CRC cases diagnosed among 40- through 49-year-olds from 1975 through 2015. Trends in CRC incidence followed a similar pattern across localized, regional, and distant stages (Figure 1): a decrease from 1975 through the years 1988 through 1995, which was followed by a significant increase through 2015. In contrast to the absolute changes in incidence rates, relative increases over the last 20 years differed significantly for distant vs localized and regional stages: AAPCs were 2.9 (95% CI, 2.4-3.4) for distant disease vs 1.4 (95% CI, 1.0-1.7) and 1.3 (95% CI, 0.7-1.9) for localized and regional disease, respectively. This differential relative increase in incidence by stage represents a statistically significant shift in the overall stage distribution (Figure 2). While the proportion of localized CRCs did not change significantly since 1995 (from 33.7% to 33.8% of CRCs; P = .90) and the proportion of regional CRCs decreased (from 39.0% to 35.6%; P = .001), the proportion of distant CRCs has increased significantly (from 21.7% to 26.6%; P < .001). The increase in the proportion of distant cases (4.9% increase) cannot be explained by a decrease in unstaged cases (from 5.5% to 4.0%; P = .002).
The incidence of CRC increased among adults aged 40 through 49 years from the 1990s through 2015, and the disease was diagnosed at later stages. This suggests that there has been a real increase in risk and that the trends do not represent a shift in age at diagnosis attributable to earlier detection.
This study complements previous work documenting stage-independent increases in incidence of rectal cancer among those aged 35 through 49 years,3 and stage-progressive increases in CRC incidence among non-Hispanic whites younger than 50 years.4 The results suggest that these trends are part of an overall shift toward more advanced stage at diagnosis that may partly explain observed increases in CRC-related mortality in younger adults since 2005.5
This study is limited by reliance on historic staging data from 9 geographical areas. The results are consistent with estimated national trends for the years 1998 through 2009 based on newer database staging definitions,4 but it could not be determined to what extent the stage shift might reflect improved clinical detection of metastases.
The increasing CRC incidence and shift toward later stage at diagnosis in those aged 40 through 49 years warrant further research into the underlying causes and potential prevention methods, such as earlier screening initiation.6
Accepted for Publication: March 4, 2019.
Corresponding Author: Reinier G. S. Meester, PhD, Division of Gastroenterology and Hepatology, Stanford University School of Medicine, 430 Broadway St, Redwood City, CA 94603 (email@example.com).
Author Contributions: Dr Meester had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Meester.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Meester.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Meester, Mannalithara.
Supervision: Lansdorp-Vogelaar, Ladabaum.
Conflict of Interest Disclosures: Dr Meester reported receiving grant support from the Netherlands Organisation for Scientific Research. Dr Lansdorp-Vogelaar reported receiving grants from National Cancer Institute, European Commission, ZonMw, the Netherlands Institute of Public Health, the National Cancer Institute, and the Dutch Cancer Society. Dr Ladabaum reported receiving personal fees from Covidien, Motus GI, Clinical Genomics, Universal Dx, and Quorum and grants and personal fees from Exact Sciences. No other disclosures were reported.
Funding/Support: This work was supported by a fellowship from the Netherlands Organisation for Scientific Research (Dr Meester), and by grant U01CA199335 from the National Cancer Institute of the National Institutes of Health (Dr Lansdorp-Vogelaar).
Role of the Funder/Sponsor: The funders played no role in design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Create a personal account or sign in to: