While observational epidemiological studies have consistently shown that beta-carotene is associated with decreased cancer risk, particularly of lung cancer, findings of seven randomised trials testing the effect of beta-carotene supplementation on cancer incidence and mortality generally have not been supportive. Two of these trials even suggested the possibility of harmful effects.

Two large trials of beta-carotene conducted among persons at average risk of cancer found no statistically significant benefit or harm associated with beta-carotene supplementation [16-17].
Two other large trials tested beta-carotene among persons at high risk of cancer [18, 19]. One (The Alpha-Tocopherol, Beta carotene Cancer Prevention Study, 1996) reported a statistically significant 18% increase in lung cancer incidence after 5–8 years of treatment with beta-carotene among male Finnish smokers 
[18]. Another that used a combination of beta- carotene and retinol, reported a statistically significant 28% increase in lung cancer incidence among United States Smokers, former smokers and asbestos workers[19]. Only one large trial, which tested a combination of beta-carotene, vitamin E and selenium in a poorly nourished Chinese population, found that after 5 years, the treated group experienced a statistically significant 9% reduction in total mortality, primarily as a result of a statistically significant 21%lower stomach cancer mortality rate [20].

The indirect evidence that beta-carotene may protect from stomach cancer comes from the randomised, controlled double-blinded chemoprevention trial in subjects with gastric dysplasia in an area with a very high gastric cancer risk in Columbia. Gastric biopsy taken at baseline was compared with those taken at 72 months. Treatment with 30 mg beta-carotene resulted in a statistically significant increase in the frequency of regression of preneoplastic lesions of the stomach [relative risk (RR) = 5.1, 95% CI 1.6–14.2][21]. One small trial of 1805 people with previous non-melanoma skin cancer that tested treatment with beta-carotene (50 mg per day) to reduce the occurrence of new skin cancer did not find any effect of this intervention[22]

It can be concluded that there is evidence at present that beta-carotene supplements have no value as cancer chemoprevention agents and cannot be recommended for use in the general population in this context.