Coffee drinking and colorectal cancer and its subsites: A pooled analysis of 8 cohort studies in Japan

Abstract

Coffee is a rich source of bioactive compounds that have potential anticarcinogenic effects. However, it remains unclear whether coffee drinking is associated with colorectal cancer. Also, despite different etiological factors involved in gut physiology, few studies have investigated this association by anatomical site of the lesion. To address these issues, this study examined the association between coffee drinking and colorectal cancer in a pooled analysis from 8 cohort studies conducted in Japan. Among 320,322 participants followed up for 4,503,274 person-years, 6,711 incident colorectal cancer cases were identified. Study-specific hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using Cox proportional hazards models and then pooled using the random effects model. Coffee drinking was not materially associated with colorectal cancer risk in men or women (pooled HR 0.92, 95% CI 0.82-1.03 in men and pooled HR 0.90, 95% CI 0.76-1.07 in women). Analysis by subsite showed a lower risk of colon cancer among female drinkers of ≥3 cups coffee/day (pooled HR 0.80, 95% CI 0.64-0.99). There was no such association in men. Coffee drinking was not associated with risk of rectal cancer in men or women. Results were virtually the same among never smokers except for an increased risk of rectal cancer associated with frequent coffee consumption. Coffee drinking may be associated with lower risk of colon cancer in Japanese women.

Coffee Consumption and Colon Cancer Risk: A Meta- Epidemiological Study of Asian Cohort Studies

“…In other words, the selection criteria were defined that a prospective cohort follow-up study conducted to evaluate coffee consumption and risk of colon cancer in Asian and showed adjusted relative risk (RR) and its 95% confidence interval (95%CI). In response, five cohorts selected by Sartini et al, (2019) were subject to priority consideration (Oba et al, 2006;Lee et al, 2007;Naganuma et al, 2007;Yamada et al, 2014;Kashino et al, 2018). Meanwhile, the RR and its 95%CI was extracted using the 'highest versus lowest method' (HLM) in the selected papers as performed by Sartini et al (2019).…”

“…Statistical significance level was set as 0.05 Kashino et al, (2018) in order to include eight Japanese cohort data, three papers (Lee et al, 2007;Naganuma et al, 2007;Yamada et al, 2014) were excluded to avoid duplication. As a result, 9 cohort data from two papers (Oba et al, 2006;Kashino et al, 2018) were selected for meta-analysis, finally. All were cohorts of Japanese population.…”

“…Jong-Myon Bae* particular, in the meta-analysis of Asian women showing statistical significance, authors included Kashino et al, (2018) having the results calculated from 8 Japanese cohort data. By including Kashino et al, (2018), three cohort databases (Lee et al, 2007;Naganuma et al, 2007;Yamada et al, 2014) were duplicated. In other words, there is a need to exclude duplication from the Sartini et al, (2019) and then conduct a new meta-analysis in Asians.…”

“…In the two final papers selected for meta-analysis (Oba et al, 2006;Kashino et al, 2018), the RR values extracted by applying the HLM method are presented in Figure 1. Both papers showed statistically significant protective effects in women.…”

“…Thus, it might be regarded as a marginally statistical significance and it should be interpreted carefully. Second, the articles selected for meta-analysis included not only original articles but also two special articles using pooled analysis that combined and analyzed databases from several cohorts (Zhang et al, 2010;Kashino et al, 2018). In…”