Adiposity and risks of gastrointestinal cancers: A 10-year prospective study of 0.5 million Chinese adults

Abstract

Associations of adiposity with risks of oesophageal squamous cell carcinoma (ESCC) and non-cardia stomach cancer, both prevalent in China, are still inconclusive. While adiposity is an established risk factor for colorectal cancer, the relevance of fat-free mass and early-adulthood adiposity remains to be explored. The prospective China Kadoorie Biobank study included 0.5 million adults (aged 30-79 years) from 10 areas in China. Participants' body size and composition were measured at baseline and at resurveys (amongst a subset). After >10 years of follow-up, 2350, 3345 and 3059 incident cases of oesophageal (EC), stomach (SC) and colorectal (CRC) cancers were recorded, respectively. Cox regression was used to estimate hazard ratios (HRs) for these cancers in relation to different adiposity traits. General and central adiposity were inversely associated with EC (primarily ESCC) risk, with HRs of 0.81 (95% CI 0.77-0.85), 0.76 (0.72-0.81) and 0.87 (0.83-0.92) per SD increase in usual levels of BMI, body fat percentage (BF%) and waist circumference (WC), respectively. Adiposity was also inversely associated with SC risk [HR = 0.79 (0.75-0.83) and 0.88 (0.84-0.92) per SD increase in usual BF% and WC], with heterogeneity by cardia and non-cardia subsites, and positively associated with CRC [HR = 1.09 (1.03-1.15) and 1.17 (1.12-1.22) per SD higher usual BF% and WC]. Fat-free mass was inversely associated with EC [HR = 0.93 (0.89-0.98) per SD increase] but positively associated with CRC [1.09 (1.04-1.14)], while BMI at age 25 was positively associated with all three cancers. After mutual adjustment, general adiposity remained inversely associated with EC and SC, while central adiposity remained positively associated with CRC.

Keywords: adiposity; body fat; gastrointestinal cancers.

FIGURE 1

Adjusted hazard ratios (HRs) for GI‐cancers by usual levels of BMI. Analyses were stratified by age‐at‐risk (10‐year bands) and sex, and adjusted for 10 study areas, education level, household income level, family history of cancer, smoking (four categories), alcohol consumption (six categories), physical activity and dietary factors that were significantly associated with the cancer of interest (i.e. fruit, soy, preserved vegetables and spicy food intake for oesophageal cancer; and fruit intake for stomach cancer). HRs were plotted against the mean BMI values at second resurvey in baseline‐defined BMI categories. Vertical lines represent floated 95% CIs. The size of each square is inversely proportional to the variance of the log‐HR. Numbers above the squares are HRs and numbers below are number of events. Sex‐specific SDs were used. BMI, body mass index; CI, confidence interval; SD, standard deviation.

FIGURE 2

Adjusted hazard ratios (HRs) for GI‐cancers by usual levels of: (A) body fat percentage, (B) fat mass and (C) fat‐free mass. Conventions as per Figure 1.

FIGURE 3

Adjusted hazard ratios (HRs) for GI‐cancers by usual levels of: (A) waist circumference and (B) waist‐to‐hip ratio. Conventions as per Figure 1.

FIGURE 4

Associations between usual general and central adiposity levels and GI‐cancer risks, with mutual adjustments. In the main model, analyses were stratified by age‐at‐risk (10‐year bands) and sex, and adjusted for 10 study areas, education level, household income level, family history of cancer, smoking, alcohol consumption, physical activity and dietary factors (fruit, soy, preserved vegetables and spicy food intake for oesophageal cancer; and fruit intake for stomach cancer). The size of each square is inversely proportional to the variance of the log‐HR. Sex‐specific SDs were used. BMI, body mass index; CI, confidence interval; HR, hazard ratio; SD, standard deviation; WC, waist circumference; WHR, waist‐to‐hip ratio.