Comparing cohort and period trends of early-onset colorectal cancer: a global analysis.

Abstract

Incidence of early-onset colorectal cancer (CRC) has increased globally in recent decades. We examined early-onset CRC incidence trends worldwide for potential cohort effects, defined as changes associated with time of birth (eg, early-life exposure to carcinogens), and period effects, defined as changes associated with calendar periods (eg, screening programs). We obtained long-term incidence data for early-onset CRC diagnosed in patients aged 20 to 49 years through the year 2012 for 35 countries in the Cancer Incidence in Five Continents database. We used a smoothing method to help compare cohort and period trends of early-onset CRC and used an age-period-cohort model to estimate cohort and period effects. Cohort effects had a more dominant role than period effects in the early-onset CRC incidence in Shanghai (China), the United Kingdom, Australia, New Zealand, Canada, the United States, and Osaka (Japan). The smoothed trends show the specific birth cohorts when early-onset CRC began to increase: the 1940s-1950s birth cohorts in the United States; the 1950s-1960s birth cohorts in other Western countries; the 1960s birth cohorts in Osaka; and the 1970s-1980s birth cohorts in Shanghai. Such increases occurred earlier for early-onset cancers of the rectum than of the colon. For the other countries, the results were less clear. Recent birth cohorts may have been exposed to risk factors different from earlier cohorts, contributing to increased early-onset CRC incidence in several developed countries or regions in the West and Asia. Such increases began in earlier birth cohorts in Western countries than in developed regions of Asia.