The fractions of cancer attributable to modifiable factors: A global review

Worldwide, the burden of cancer is rising, stimulating efforts to develop strategies to control these diseases. Primary prevention, a key control strategy, aims to reduce cancer incidence through programs directed towards reducing population exposure to known causal factors. Before enacting such strategies, it is necessary to estimate the likely effect on cancer incidence if exposures to known causal factors were reduced or eliminated. The population attributable fraction (PAF) is the epidemiological measure which quantifies this potential reduction in incidence. We surveyed the literature to document and summarise the proportions of cancers across the globe attributable to modifiable causes, specifically tobacco smoke, alcohol, overweight/obesity, insufficient physical activity, solar ultraviolet (UV) radiation and dietary factors (insufficient fruit, non-starchy vegetables and fibre; red/processed meat; salt). In total, we identified 55 articles that presented PAF estimates for one or more causes. Information coverage was not uniform, with many articles reporting cancer PAFs due to overweight/obesity, alcohol and tobacco, but fewer reporting PAFs for dietary factors or solar UV radiation. At all cancer sites attributable to tobacco and alcohol, median PAFs were markedly lower for women than men. Smoking contributed to very high median PAFs (>50%) for cancers of the lung and larynx. Median PAFs for men, attributable to alcohol, were high (25-50%) for cancers of the oesophagus, oral cavity/pharynx, larynx and liver. For cancers causally associated with overweight/obesity, high median PAFs were reported for oesophageal adenocarcinoma (men 29%, women 37%), gallbladder (men 11%, women 42%) and endometrium (36%). The cancer PAF literature is growing rapidly. Repeating this survey in the future should lead to more precise estimates of the potentially preventable fractions of cancer.