ABSTRACT
Beta-carotene is a provitamin, i.e. a precursor of vitamin A (retinol), which is classified as an essential nutrient for humans. Beta-carotene is one of many carotenoids found in plants, fungi and bacteria. Carotenoids are therefore predominantly obtained through foods of plant origin or food supplements. Carrots contribute approximately half of the total beta-carotene intake in the Norwegian diet, followed by mixed frozen vegetables, tomatoes, fruits and berries. VKM emphasises that this opinion on upper level (UL) for beta-carotene addresses beta-carotene in food supplements only. Beta-carotene from regular foods such as vegetables and fruits is not considered to be a health concern. In 2002, the Scientific Committee on Food (SCF) established a tolerable upper intake level (UL) for vitamin A (SCF, 2002). However, the SCF opinion covers only retinol compounds (various forms of vitamin A). The bioconversion of carotenoids to vitamin A in the body is well regulated and therefore only intake of vitamin A has been considered relevant for vitamin A toxicity (Blomhoff et al., 2003; EFSA, 2008). The Norwegian Food Safety Authority is considering whether betacarotene should be regulated separately from retinol compounds. Beta-carotene seems to have a carcinogenic effect in smokers. A number of studies have been published where possible mechanisms of this negative health effect are discussed. The suggested mechanisms are either related to effects on cytochrome P450-related activities, altered retinoid signalling or to a pro-oxidant activity of beta-carotene. No UL has been established for beta-carotene. Several risk-assessment bodies have, however, previously attempted to establish safe levels or temporary guidelines, summarised in the following table: Previous reports Conclusion SCF, 2000, EU No dose-response relationship could be derived. Supplementation of 20 mg beta-carotene per day or more is contraindicated for use in current heavy smokers. There is insufficient evidence to set an UL for beta-carotene. IOM, 2000, USA No UL was established for beta-carotene or carotenoids. Beta-carotene supplementation is not recommended in the general population. EVM, 2003, UK The LOAEL was set to 20 mg/day. An uncertainty factor of 3 was applied to extrapolate from LOAEL to a NOAEL. A Safe Upper Level for beta-carotene supplements was set at 7 mg/day (equivalent to 0.11 mg/kg body weight/day for a 60 kg adult). NNR, 2012, Nordic countries No specific beta-carotene recommendation or UL. Rasmussen, 2006, Denmark A Temporary Guidance Level for beta-carotene equal to the average dietary level of 5 mg/day for all age groups was suggested. Seven randomised controlled trials (RCTs) have been included in this VKM opinion, conducted either in Europe or the USA, with almost 47 100 participants in the beta-carotene groups. In six of these RCTs there were no observed increased risk of cancer, but the large Finnish ATBC study found an increased risk of lung cancer in the beta-carotene group. Two prospective studies were included, one Danish and one from the USA, with 125 000 participants all together. The Danish study found that risk of lung cancer increased in smokers with increasing doses of beta-carotene supplements. In addition, eleven meta-analyses were included; one with age-related macular degeneration as endpoint, one with a mixture of cardiovascular disease (CVD) and cancer as endpoints, four on cancer as only endpoint, one on a mixture of CVD and all-cause mortality as endpoints and four on all-cause mortality alone. One of the meta-analyses on all-cause mortality was later excluded. There were no significant findings in the meta-analysis on macula degeneration. One of the two meta-analyses on CVD found a small increased risk in the beta-carotene arm (Vivekananthan et al., 2003). The combined CVD and cancer meta-analysis did not have sufficient statistical power to get significant results, but found a probable increase in lung cancer incidence in high-risk subgroups (smokers and asbestos workers (Fortmann et al., 2013). In the five meta-analyses studying cancer, there were no effect on other cancer forms than lung cancer. The four meta-analyses on all-cause mortality used information from the same RCTs as included in this VKM-opinion. They extracted information on numbers of death in each study and used these numbers to analyse risk of death in the beta-carotene versus placebo groups. Alarmingly, they all found an increased risk of all-cause mortality. These meta-analyses have been discussed thoroughly. To complete the risk characterisation of beta-carotene, VKM has followed the steps 1 – 4 as suggested by SCF in their Guidelines for the development of tolerable upper intake levels for vitamins and minerals (SCF, 2000a). Step 1 and 2. Hazard identification and characterisation Up until two decades ago, beta-carotene was thought to be harmless even in large doses. In the wake of the Finnish ATBC study which found an increased risk of lung cancer and death in male smokers, animal studies have indicated three possible mechanisms for such a detrimental effect. Although conclusive mechanistic explanations for the negative effects have not yet been agreed upon, there is a scientific rationale for the argument that population groups with vulnerable lungs may also have increased risk from beta-carotene supplements. The dose used in the Finnish ATBC study was 20 mg beta-carotene/day. The effect was only observed during the intervention period; in follow-up studies conducted after the active period was finished, the risk declined and was no longer significant. 20 mg beta-carotene may thus be considered as a LOAEL. The Danish prospective study found a dose-dependent increase in lung cancer risk with increased intake of supplemental beta-carotene. Unfortunately, the paper does not allow for setting a NOAEL or LOAEL based on the published data. In the four meta-analyses on all-cause mortality, all found a 6-7% increased risk of death. One of the meta-analyses also found an increased risk of CVD in smokers. However, all results were driven, statistically, by the ATBC study. Studies with a more mixed population (both men and women) and with a more typical prevalence of smokers (10–20%), found no such increased risk. Step 2, continued: Derive at a UL, taking into account the scientific uncertainties in the data. ULs may be derived for various life-stage groups within the population VKM found it extraordinary challenging to decide which uncertainty factor to use for beta-carotene. The present SCF guidelines for establishment of tolerable upper intake levels do not give clear guidance/advice in deciding the numeric level of the uncertainty factor. This seems to leave the decision to scientific judgement. If the NOAEL is based on human data, an uncertainty factor of 10 is recommended as a starting point to encompass inter-individual variation and sensitivity. The SCF guidelines state that a small uncertainty factor is to be used if the judgement is that little population variability is expected for the adverse effects, and a larger uncertainty factor (close to 10) may be used if variability is expected to be large. For beta-carotene, a NOAEL is not available, and an uncertainty factor may be applied to account for the uncertainty in deriving a UL from the LOAEL. The size of the uncertainty factor involves a judgement based on the severity and incidence of the observed effect at the LOAEL and the steepness (slope) of the dose response, if this is possible to estimate. For beta-carotene, we have not found the data necessary to make a dose-response curve. In addition, the following considerations were discussed before deciding on an uncertainty factor: · The study which found a negative effect of beta-carotene supplementation (the Finnish ATBC study) was very large (n=29 133) which indicates that it encompasses inter-individual variation and sensitivity. Additionally, the most vulnerable groups, in this case smokers, was an inclusion criteria. Both these factors indicate that the uncertainty factor can be in the lower end. · The meta-analyses for the endpoint “increased risk of all-cause mortality” found an increased risk of death in the beta-carotene groups. This is severe, and indicates that a maximum uncertainty factor should be applied. However, as all results in the all-cause mortality metaanalyses were driven statistically, by the ATBC study on smokers, we choose to use a lower factor. Based on the above considerations, VKM has chosen to use 5 as an uncertainty factor for betacarotene. An UL for beta-carotene cannot be derived, but a tentative upper level (TUL) is set at 4 mg/day, based on a LOAEL of 20 mg and the uncertainty factor of 5. Smokers and anyone else in the population with vulnerable lungs (e.g. asthmatics, COPD patients) should be discouraged from taking beta-carotene containing supplements all together. Step 3. Exposure assessment – evaluates the distribution of usual total daily nutrient intakes among members of the general population In the food survey Småbarnskost 2007, the mean intake of beta-carotene in 2-year-olds was 1.5 mg/day. In Norkost 3, the estimated mean intake in adults was 2.4 mg/day and 6.9 mg/day in the 95th percentile. About 3% of the adults reported use of beta-carotene supplements. The use of tanning pills containing beta-carotene may have been underreported. Beta-carotene from regular foods such as vegetables and fruits is, however, not considered to be of any health concern. Negative health effects from beta-carotene in natural foods have never been reported. On the contrary, the consumption of vegetables and fruits should be increased, and the recommendation of “5 a day” should be achieved in all age groups of the population. Step 4. Risk characterisation – analyses of the conclusions from steps 1 through 3 and characterises the risk. The risk will depend on the fraction of the population