Relationship between energy intake, anthropometrical measures and physical activity and the adenoma-carcinoma sequence in the E3N prospective study on women

  • Topic: Colorectal Cancer
  • Institution: Institut Gustave Roussy
  • Country: France
  • Status: Completed

Scientific abstract

(View plain language abstract)

Background

Most colorectal cancers develop from benign adenomatous polyps. In order to propose recommendations, it is essential to identify the factors associated with adenoma and colorectal cancer occurrence.

Aim

The aim of this project was to study the association between risk of colorectal adenomas and cancer, and anthropometry and physical activity.

Methods

We conducted this analysis in the prospective E3N cohort (Etude Epidémiologique auprès des femmes de l’Education Nationale), which included 98,995 women aged 45 to 65 years at baseline (1990), insured by the Mutuelle Générale de l'Education Nationale (M.G.E.N.), a national Health Insurance Plan, mostly covering teachers. Anthropometrical data (height, weight, Body Mass Index, silhouette, waist circumference, hip circumference, Waist-to-Hip Ratio) were collected from self-administered questionnaires.

The 1,437 subjects who developed a first adenoma between 1993 and 2005 were compared to 17,737 polyp-free subjects at colonoscopy (endoscopic examination of the large bowel). The colorectal cancer study included 414 colorectal cancer cases diagnosed between 1990 and 2005, and 69,753 subjects without colorectal cancer. Relative risks were calculated using Cox models.

The median follow-up time was 6.9 years in the adenoma study and 15 years in the cancer study.

Results

Adenoma cases were older than polyp-free subjects (54.0 ± 6.6 years for the adenoma group and 53 ± 6.5 years for the polyp-free group), less educated, had more frequently a family history of colorectal cancer in their first-degree relatives, consumed more alcohol and had a higher body mass index, weight, waist circumference, hip circumference and waist-to-hip ratio than polyp-free subjects. Colorectal cancer cases had more frequently a family history of colorectal cancer in their first-degree relatives, consumed more alcohol, had a higher body mass index, waist circumference and waist-to-hip ratio than non-cases.

In multivariate models, energy intake was not associated with risk of adenomas or cancer. Overall and recreational physical activities were not associated either with risk of colorectal adenoma or cancer.

Height was not associated with risk of colorectal tumors, whatever the stage of the adenoma carcinoma-sequence. Weight and body mass index were significantly associated with adenoma but not with cancer risk. Multivariate RRs of adenomas were 1.26; 1.09-1.46 (p for trend=0.01) and 1.17; 1.00-1.37 (p for trend=0.09) for the fourth quartile of weight and BMI respectively, as compared to the first. Corresponding figures for cancer risk were 1.18; 0.90-1.56 (p for trend=0.55) and 1.04; 0.78-1.39 (p for trend=0.91) respectively. The association of weight and BMI with adenoma risk was stronger for low-risk than high risk adenomas, and it was restricted to left colon adenomas.

Large waist and large hips, but not waist-to-hip ratio, were associated with an increased risk of colorectal adenomas, but not of cancer. The association was observed only for low-risk, left colon, tubulous, and small adenomas.

Conclusions

Our results do not support strong relationships between energy, physical activity, and anthropometry, and the colorectal adenoma-carcinoma sequence. They support a relationship between weight, body mass index and waist and hip circumferences restricted to low-risk left colon adenomas.

Plain Language Abstract

Background

Most colorectal cancers develop from a benign tumor, the adenomatous polyp. In order to propose recommendations, it is essential to identify the factors associated with adenoma and colorectal cancer occurrence.

Aim

The aim of this project was to study the association between risk of colorectal adenomas and cancer, and height, weight, abdominal corpulence, dietary energy intake, and physical activity.

We conducted this analysis in the prospective E3N cohort (Etude Epidémiologique auprès des femmes de l’Education Nationale), which included 98,995 women aged 45 to 65 years at baseline (1990), insured by the Mutuelle Générale de l'Education Nationale (M.G.E.N.), a national Health Insurance Plan, mostly covering teachers. Anthropometrical data (height, weight, Body Mass Index, silhouette, waist circumference, hip circumference, Waist-to-Hip Ratio) were collected from self-administered questionnaires.

The 1,437 subjects who developed a first adenoma between 1993 and 2005 were compared to 17,737 polyp-free subjects at colonoscopy (endoscopic examination of the large bowel). The colorectal cancer study included 414 colorectal cancer cases diagnosed between 1990 and 2005, and 69,753 subjects without colorectal cancer. Relative risks were calculated using Cox models.

The median follow-up time was 6.9 years in the adenoma study and 15 years in the cancer study.

Results

Adenoma cases were older than polyp-free subjects (54.0 ± 6.6 years for the adenoma group and 53 ± 6.5 years for the polyp-free group), less educated, had more frequently a family history of colorectal cancer in their first-degree relatives, consumed more alcohol and had a higher body mass index, weight, waist circumference, hip circumference and waist-to-hip ratio than polyp-free subjects. Colorectal cancer cases had more frequently a family history of colorectal cancer in their first-degree relatives, consumed more alcohol, had a higher body mass index, waist circumference and waist-to-hip ratio than non-cases.

In models taking into account several factors simultaneously, energy intake was not associated with risk of adenomas or cancer. Overall and recreational physical activities were not associated either with risk of colorectal adenoma or cancer.

Height was not associated with risk of colorectal adenomas or cancers. Weight and body mass index were significantly associated with adenoma but not with cancer risk. The risk of adenomas was increased by 26 % among the 25 % heaviest women as compared to the 25% lightest ones. The corresponding figure was 18 %, but not statistically significant, for cancer. The association of weight and body mass index with adenoma risk was stronger for low-risk than high-risk adenomas, and it was restricted to left colon adenomas.

Large waist and large hips, but not waist-to-hip ratio, were associated with an increased risk of colorectal adenomas, but not of cancer. The association was observed only for low-risk, left colon, tubulous, and small adenomas.

Conclusions

Our results do not support strong relationships between energy, physical activity, and anthropometry, and the colorectal adenoma-carcinoma sequence. They support a relationship between weight, body mass index and waist and hip circumferences restricted to low-risk left colon adenomas.

Grant publications