Attributable risk - overview of the results

In their summary chapter, Parkin et al (2011b)1 estimate that 42.7% of incident cancer cases overall predicted in the UK in 2010 were linked to sub-optimal levels of the 14 major lifestyle and environmental factors studied. This is around 134,000 cancer cases.2 Throughout the supplement, non-melanoma skin cancers were excluded. 

The estimated population attributable fraction (PAF) was higher for men (45.3% of the total number of cancers in men) than for women (40.1%). This difference is largely due to the higher number of cancers in men caused by smoking, reflecting the fact that, in the past, men were much more likely than women to smoke.2

Of the cancers linked to the lifestyle and environmental factors studied here, cervical cancer had the highest PAF (100% linked to HPV infection). Mesothelioma, the upper aero-digestive tract cancers (oral cavity and pharynx, larynx and oesophagus), lung and malignant melanoma all had PAFs higher than 85%. The PAFs for the four most common cancers in the UK were 26.8% for breast, 89.2% for lung, 54.4% for bowel and 0% for prostate cancer, which is not clearly linked to any of the factors studied.2

PAFs should not be used to indicate the percentage of cancers that can currently be prevented by practical means without reference to the individual study sections, where some of the uncertainties and assumptions are discussed in detail.

A data visualisation representing the results is available, showing the proportions of cancers in the UK linked to the 14 lifestyle and environmental factors studied.

section updated 07/12/11

Results by risk factor

The Tobacco section, published in the British Journal of Cancer on 6th December 2011, shows that the largest single cause of cancer in the UK in 2010 by a substantial margin was tobacco smoke (23% of all cancers in men and 15.6% in women, and 19.4% in persons overall). The cancer with the highest proportion of cases caused by tobacco smoke was lung, with 86% of cases in the UK in 2010 due to this factor – or around 35,000 cases (due to both active and involuntary smoking, with an estimated 2.5% of lung cancers caused by exposure of lifelong non-smokers to environmental tobacco smoke). After lung cancer, larynx was the cancer with the second largest proportion of cases caused by active smoking (79%), followed by cancers of the oral cavity and pharynx (64.5%), and oesophagus (65%).3

The Summary section, published in the British Journal of Cancer on 6th December 2011, notes the following:

• The previously rapid rate of decline in cigarette smoking in adults in Britain has slowed since 1982;
• The difference in smoking prevalence between men and women has decreased considerably since the 1970s;
• The overall reported number of cigarettes smoked per male and female smoker has changed little since the early 1980s;
• Current decreases in smoking-related cancer incidence will slow and eventually stop unless further progress can be achieved in reducing exposures.2

The study confirms that, after smoking, being overweight, drinking alcohol and having a poor diet are the most important preventable causes of cancer.2

section updated 07/12/11

Four dietary factors were studied:

• Deficient intake of fruit and vegetables;4
• Consumption of red and processed meat;5
• Deficient intake of fibre;6
• Excess consumption of salt.7

Overall, 9.2% of cancer cases in the UK in 2010, or around 29,000 cases of cancer, may have been linked to these dietary factors. The largest contribution was from deficient intake of fruit and vegetables (4.7%), followed by red and processed meat (2.7%), deficient intake of fibre (1.5%) and excess salt (0.5%).4-7 There is more uncertainty around the link between aspects of diet and cancer than for some other risk factors, such as smoking.2,4-7

section updated 07/12/11

The Overweight and Obesity section, published in the British Journal of Cancer on 6th December 2011, estimates that 5.5% of all cancers in the UK in 2010, or around 17,000 cases, were linked to excess bodyweight. The proportion of cancer cases linked to this factor was higher in women (6.9%) than men (4.1%). Excess bodyweight ranked 2nd in women and 5th in men among the factors studied. The cancer sites with the largest proportion of cases linked to overweight were uterus (33.7%), kidney (24%) and oesophagus (21.7%). It is likely that the list of cancers which can be related to bodyweight is greater than assumed in this report, and so the number of cancer cases due to excess bodyweight may be an underestimate.8

section updated 07/12/11

The Alcohol section, published in the British Journal of Cancer on 6th December 2011, estimates that 4% of cancer cases – or around 12,500 cases – in the UK in 2010 were linked to alcohol consumption. The upper aero-digestive tract cancers – oral cavity and pharynx, larynx and oesophagus – had the highest alcohol-attributable proportions (20.6%-30.4%).9

section updated 07/12/11

The Occupation section, published in the British Journal of Cancer on 6th December 2011, estimates that 3.7% of cancer cases (or around 11,500 cases) in 2010 were linked to occupation. This study is based on a detailed study in 2010 by Rushton et al.10 The estimate for men (4.9%, or around 7,800 cases) is more than twice that for women (2.4%, or around 3,700 cases).The cancer sites with the highest number of cases linked to occupation were lung (around 5,350 cases), followed by mesothelioma (around 2,400 cases) and the great majority of these cases were in men. Exposure to asbestos was the largest contributor to occupation-related lung cancers and all the occupation-attributed cases of mesothelioma were linked to this factor. Mesothelioma had the highest PAF for occupation (97% of cases in men and 82.5% cases in women linked to occupation).11

section updated 07/12/11

Overall, the Ultraviolet (UV) Radiation section, published in the British Journal of Cancer on 6th December 2011, estimates that 3.5% of all cancers (or around 11,100 cases) in the UK in 2010 were linked to this factor. All of these cases were malignant melanoma. This estimate includes exposure to UV radiation from sunbeds, but it is not possible to estimate the relative proportions due to sunbeds and sunlight, respectively. Due to the methodology used in this section, the PAF may actually be higher than estimated.12

The UV Radiation section notes that incidence rates for malignant melanoma have increased faster than those for any other common cancer over the last 30 years. Most of this is thought to reflect a real increase resulting from changes in sun exposure patterns in past decades, with an estimated 85.9% of malignant melanoma cases in the UK in 2010 linked to UV radiation.12

section updated 07/12/11

The Infection section, published in the British Journal of Cancer on 6th December 2011, estimates that 3.1% of all cancers in the UK in 2010 were linked to infections, or around 9,700 cases. The estimate was higher for females (3.7% or around 5,800 cases) than males (2.5% or around 3,900 cases). As is noted, this reflects the fact that all cervical cancers are thought to be linked to human papillomavirus (HPV) infection.13

The Infection section shows that, of the infectious agents studied, HPV was linked to the highest number of infection-associated cancers (around 5,090 cases, or 1.6% of all cancers), followed by Helicobacter pylori (around 2,560 cases), Epstein Barr virus (around 1,210 cases), hepatitis B and C viruses (around 620 cases), human immunodefiency virus and human herpes virus 8/Kaposi Sarcoma Herpes Virus (around 240 cases) and human T-cell lymphotrophic virus (around 25 cases).13

The Infection section shows that cervical cancer accounted for the largest number of infection-associated cancers (around 2,690 cases), followed by stomach (around 2,230 cases).13

section updated 07/12/11

The Ionising Radiation section, published in the British Journal of Cancer on 6th December 2011, estimates that 1.8% of all cancers in the UK in 2010, or around 5,800 cases, were linked to exposure to ionising radiation from natural and man-made sources. It was noted that radon accounts for half of the annual average radiation dose in the UK population. Despite this, diagnostic radiation (x-rays and computed tomography (CT) scans) accounted for the largest number of radiation-associated cases, estimated at 1,860 cases. Radiotherapy was estimated to be linked to around 1,380 cases of cancer in the UK in 2010, followed by radon (around 1,380 cases), background radiation (around 1,170 cases) and lastly nuclear medicine (19 cases).14

It was estimated that around 1,910 cases of lung cancer in the UK in 2010 were linked to radiation. The cancer with the second highest number of radiation-attributed cases was leukaemia (around 700 cases).14

section updated 07/12/11

The Physical Exercise section, published in the British Journal of Cancer on 6th December 2011, estimates that being physically inactive was linked to around 3,300 cases of cancer in the UK in 2010, or 1% of the total. This was based on an excess risk of breast, colon and uterine cancers. Breast cancer accounted for the largest number of linked cases (around 1,650), followed by colon (around 1,310 cases) and lastly uterus (around 310 cases).15

section updated 07/12/11

The Reproductive Factors section, published in the British Journal of Cancer on 6th December 2011, estimates that 1.7% of all cancers in women (0.9% of all cancers, or around 2,700 cases, in the UK in 2010) were linked to women breastfeeding for less than 6 months or not at all. This was made up of around 1,500 cases of breast cancer and around 1,200 cases of ovarian cancer.16

section updated 07/12/11

The Hormones section, published in the British Journal of Cancer on 6th December 2011, shows oral contraceptives have a net benefit on occurrence of cancer, due to their long-term effect on reducing risk of uterine and ovarian cancers. Therefore, only hormone replacement therapy (HRT) was included in the overall results for cancers associated with exogenous hormones. The study estimates that 1.1% of all cancers in women (0.5% of all cancers, or around 1,700 cases, in the UK in 2010), were linked to HRT. Breast cancer accounted for the majority of this total (around 1,530 cases), with uterine cancer contributing around 100 cases to the total and around 50 cases for ovarian cancer.17

section updated 07/12/11