The impact of obesity and insulin on cancer risk

Paul Brennan’s research will assess the role of obesity and insulin across multiple cancers using Mendelian randomisation

  • Topic: Combination of cancers
  • Institution: International Agency for Research on Cancer (IARC)
  • Country: France
  • Status: Ongoing

"We are delighted with this support from WCRF that will allow us to further elucidate the role of obesity in multiple cancers"

– Paul Brennan


It is estimated that about four per cent of cancer cases in high-income countries are due to obesity. This estimate is based on numerous studies that typically measure a person’s weight at one point in time. Because of this, the effect of obesity on cancer risk may be underestimated. An alternative approach is to use genetic measures of obesity and assess their risk with cancer. Genetic measures of obesity may be a more accurate reflection of a person’s body weight across their entire life span, and may provide a better measure of the lifelong risk associated with obesity. Preliminary data from our team indicate that this approach results in obesity having more of an impact on risk of both kidney and pancreatic cancer than with the risks observed using traditional methods. They also report an increased risk with insulin levels that may mediate the obesity-cancer relation.

Aims and objectives

The primary aim is to use our understanding of the genetics of overweight and obesity to evaluate the effect of body weight on cancer risk. This will combine very large population-based studies with extensive genetic data (such as UK Biobank) with large genome-wide studies of different cancers thought to be associated with obesity. We will also use recently collated data within the EPIC cohort of over 300,000 individuals to determine if changes in weight over adulthood are associated with cancer risk, in particular for postmenopausal breast cancer. Finally, we will follow up on our preliminary genetic findings for an association between insulin levels and three cancers (kidney, pancreas and lung) to evaluate if there is an association between these cancers and insulin levels when one measures these levels in fasting blood samples among people who develop these cancers and people who do not.

How it will be done

Much of these analyses will be conducted using genetic data that are available from very large population-based studies as well as cancer-specific studies. This will include developing genetic instruments within UK Biobank, a study of 500,000 individuals in the UK with extensive genomic characterisation. It will also include collating and analysing genome-wide data for specific cancers, typically involving between 10,000 and 100,000 case-control pairs. Research on adult weight gain will involve an analysis within the EPIC cohort of over 300,000 individuals who had BMI and other lifestyle factors assessed over an average of nine years after baseline measurements. Finally, analysis of fasting insulin in three specific cancers will involve measuring insulin levels from 800 cancer cases (500 lung, 150 pancreas and 150 kidney) and 800 matched controls. All blood samples were fasting blood samples and collected as part of the Northern Sweden Health and Disease Study, a population-based cohort study of over 100,000 individuals.

Potential impact

Body weight is thought to explain about four per cent of cancers in high-income countries. This study may result in a re-evaluation of this estimate. It will also investigate weight gain during adulthood, and investigate a wide range of mechanisms that may mediate the excess risk due to increased body weight.