Mechanisms underlying how body fatness affects postmenopausal breast cancer risk

We aim to measure sex hormones, indicators of resistance to insulin and indicators of inflammation in stored blood samples to see whether these markers predict women’s risk of breast cancer

  • Topic: Breast cancer
  • Institution: Cancer Council Victoria
  • Country: Australia
  • Status: Ongoing
Researcher: Dallas English

Grant title

Mechanisms underlying the effect of body fatness on risk of postmenopausal breast cancer: a nested case-control with causal mediation analysis

Scientific abstract

(View plain language abstract)


There is convincing evidence that adult body fatness (“adiposity”) increases the risk of postmenopausal breast cancer. The mechanisms by which adiposity causes postmenopausal breast cancer are uncertain, but may be due to metabolic dysfunction and inflammation related to increased mass and activity of adipose tissue.

Adipose tissue of obese individuals secretes higher levels of pro-inflammatory cytokines and free fatty acids, and less anti-inflammatory cytokines, compared with lean women. It also disturbs the production of sex steroid hormones and cellular sensitivity to insulin. These factors can affect cell proliferation, apoptosis and angiogenesis, which are hallmarks of cancer.

Observational epidemiological studies can assess mechanistic pathways by measuring relevant biomarkers and performing modern causal mediation analyses. Effects for each pathway can be estimated simultaneously. To explain (mediate) the effect of adiposity on breast cancer risk, markers of the pathway must be associated with breast cancer and with adiposity.

Hypothesis and objectives

Our first aim is to measure associations between markers of the adipose tissue-related pathways and risk of postmenopausal breast cancer. For the pathways that show associations, our second aim is to assess the proportion of the effect of adiposity on risk of postmenopausal breast cancer that is mediated by the pathways, singly and together.

Our objectives for this aim are to estimate the absolute risk of breast cancer associated with markers in each of the pathways, thereby also establishing the relative importance of each of the mechanistic pathways.

Settings and methods

This research is set within the Melbourne Collaborative Cohort Study (MCCS), a cohort including 24,469 women recruited from 1990–94. A second wave of data collection (Wave2) took place from 2003–07. At both waves, height, weight and waist and hip circumferences were measured and a blood sample was collected. Participants were interviewed about diet, physical activity and other lifestyle factors and for women, reproductive history.

The proposed biomarker study is a nested case-control study of postmenopausal MCCS female participants who attended Wave2. Women will be eligible if at Wave2 they were postmenopausal, not using hormone therapy, not taking antihyperglycaemic medication, and had no history of cancer. Cases will be women with incident adenocarcinoma of the breast diagnosed after Wave2; we anticipate 515 cases. Using plasma samples collected at Wave2, we will measure biomarkers of each of the three pathways.

The associations between the biomarkers and breast cancer risk will be assessed using logistic regression. The indirect and direct effects will be estimated using causal mediation methods for multiple mediators, which permit estimation of indirect effects for each of the pathways simultaneously. These methods also permit identifying indirect effects for interrelated pathways.


Few epidemiological studies have attempted to determine what biological mechanisms mediate the association between adiposity and breast cancer and no study has estimated indirect effects for multiple mediators simultaneously. Indirect effects quantify the risk mediated through each pathway, thus estimating the potential public health benefit of intervening in those pathways.

Formal mediation analyses involving multiple mediators adds substantial weight to inferences of causality. Establishing that an exposure such as adiposity asserts its carcinogenic effect through specific biological pathways is strong evidence of a causal effect.

Plain language abstract


Breast cancer in women who have passed through the menopause is one of the most common cancers in many countries. WCRF and the International Agency for Research on Cancer have both declared that body fatness causes breast cancer in postmenopausal women. How body fatness causes breast (and other) cancer is not certain.

It is known that body fatness affects sex hormones (eg oestrogen), reduces the body’s ability to react to insulin and produces inflammation in the body. Existing epidemiological studies show that postmenopausal women with higher amounts of sex hormones in their blood have increased risk of breast cancer. At present, there is not enough evidence to conclude that inflammation is associated with risk of breast cancer.

Recently developed statistical methods allow researchers to measure how much of the effect of body fatness is explained by these factors (sex hormones, inflammation and resistance to the effects of insulin).

Aims and objectives

Using the Melbourne Collaborative Cohort Study, we aim to measure sex hormones, indicators of resistance to insulin and indicators of inflammation in stored blood samples. We will use this information to see whether these markers predict women’s risk of future breast cancer. We will then perform a statistical analysis that enables us to estimate how much of the effect of body fatness on the risk of breast cancer can be explained by each of these markers.

How it will be done

The Melbourne Collaborative Cohort Study recruited 24,469 women from 1990 to 1994. These women have been followed ever since. From 2003 to 2007, 16,146 of them attended for a second face-to-face interview where they had their waist and weight measured, were asked questions about their reproductive history, diet, alcohol and tobacco consumption and physical activity. A blood sample was collected and stored in a freezer for future use.

The study will include all 515 women diagnosed with breast cancer since their second interview as well as a random sample of 1,030 women not diagnosed with cancer. A number of molecules will be measured from their blood sample, including several sex hormones, insulin and other molecules related to insulin, and several markers of inflammation. The data will then be analysed statistically to see whether the amount of these molecules in blood is different for women with breast cancer than for women without breast cancer.

Potential impact

This will be one of only a few epidemiological studies that have attempted to determine what biological mechanisms explain why body fatness causes breast cancer. It is the first study to look at several of these potential mechanisms simultaneously. Our information on biological mechanisms will also strengthen the argument that body fatness does cause breast cancer.

Unravelling the biological mechanisms offers two prospects for translation. First, it will help improve risk prediction models for breast cancer, which currently do not take mechanisms into account. Second, it offers new avenues for prevention by targeting the causal pathways in people for whom “obesity” is already a problem.