Long-term status of predicted body fat percentage, body mass index and other anthropometric factors with risk of colorectal carcinoma: Two large prospective cohort studies in the US

Abstract

Anthropometric measurements, such as body mass index (BMI), waist circumference, and body fat percentage, have been used as indicators of obesity. Despite evidence that excess body fat is a risk factor for colorectal carcinoma (CRC), the magnitude of the association of BMI and other obesity indicators with the long-term risk of CRC remains unclear. Utilizing a Cox proportional hazards regression model, we examined differential associations between predicted body fat percentage and BMI with the risk of CRC (n = 2,017). The associations between CRC incidence and different adiposity measurements were also assessed. Predicted body fat percentage had a similar increased risk of CRC risk as BMI. In multivariable-adjusted analyses, the hazard ratio for CRC in the second to fifth quintiles (compared to the first quintile) of predicted body fat percentage were 1.32, 1.31, 1.53 and 2.09 for men (p_trend < 0.001) and 0.91, 0.90, 0.98 and 1.15 for women (p_trend = 0.03). Among various anthropometric measurements, predicted fat mass and waist circumference were slightly more strongly associated with CRC risk than BMI. In conclusion, the novel anthropometric prediction equations provided further evidence that a greater amount of body fat might contribute to CRC risk in both sexes. An innovative approach enabled us to estimate the susceptibilities of specific body composition with CRC risk, in an inexpensive and minimally invasive manner. Furthermore, the typically used measures of BMI and waist circumference are robust measures of adiposity to predict cancer risk in a relatively healthy population.

Keywords: body composition; body fat percentage; colorectal cancer; obesity; visceral adiposity.

Figure 1.

Risk of colorectal carcinoma according to deciles of predicted body fat percentage (FP) and body mass index (BMI). All models were adjusted for the same set of covariates in Table 2. Abbreviations: BMI, body mass index; FP, predicted body fat percentage.

Figure 2.

Joint association of BMI and predicted body fat percentage (a for women and b for men)/waist circumference (c for women and d for men) with risk of colorectal carcinoma in both sexes. Participants were first categorized according to BMI (<25, 25–27.5, ≥27.5 kg/m²) and then within each BMI category they were further grouped by tertiles of predicted body fat percentage/waist circumference. The multivariable Cox proportion hazards regression models were used, adjusted the same sets of covariates in Table 2, with the individuals low in both BMI and predicted body fat percentage/waist circumference as the reference group. Abbreviations: BMI, body mass index.

Figure 2.

Joint association of BMI and predicted body fat percentage (a for women and b for men)/waist circumference (c for women and d for men) with risk of colorectal carcinoma in both sexes. Participants were first categorized according to BMI (<25, 25–27.5, ≥27.5 kg/m²) and then within each BMI category they were further grouped by tertiles of predicted body fat percentage/waist circumference. The multivariable Cox proportion hazards regression models were used, adjusted the same sets of covariates in Table 2, with the individuals low in both BMI and predicted body fat percentage/waist circumference as the reference group. Abbreviations: BMI, body mass index.

Figure 2.

Joint association of BMI and predicted body fat percentage (a for women and b for men)/waist circumference (c for women and d for men) with risk of colorectal carcinoma in both sexes. Participants were first categorized according to BMI (<25, 25–27.5, ≥27.5 kg/m²) and then within each BMI category they were further grouped by tertiles of predicted body fat percentage/waist circumference. The multivariable Cox proportion hazards regression models were used, adjusted the same sets of covariates in Table 2, with the individuals low in both BMI and predicted body fat percentage/waist circumference as the reference group. Abbreviations: BMI, body mass index.

Figure 2.

Joint association of BMI and predicted body fat percentage (a for women and b for men)/waist circumference (c for women and d for men) with risk of colorectal carcinoma in both sexes. Participants were first categorized according to BMI (<25, 25–27.5, ≥27.5 kg/m²) and then within each BMI category they were further grouped by tertiles of predicted body fat percentage/waist circumference. The multivariable Cox proportion hazards regression models were used, adjusted the same sets of covariates in Table 2, with the individuals low in both BMI and predicted body fat percentage/waist circumference as the reference group. Abbreviations: BMI, body mass index.