Following a request from the European Commission, the European Food Safety Authority was asked to deliver a scientific statement on a carcinogenicity study in mice with transplacental exposure to aspartame, as reported by Soffritti et al. (2010), and on a prospective cohort study on the association between intakes of sugar-sweetened and artificially sweetened soft drinks and preterm delivery in pregnant women, as reported by Halldorsson et al. (2010).
Aspartame has been authorised for use in foods and as a table-top sweetener by several Member States since the 1980s. The European legislation harmonised its use in foodstuffs in 1994 following thorough safety evaluations by the Scientific Committee on Food (SCF) in 1984 and 1988. Further reviews of the aspartame data were carried out by the SCF in 1997 and 2002. No concerns regarding possible reproductive and developmental toxicity, genotoxicity or carcinogenicity were identified. In 2006 and 2009, the Scientific Panels on Food Additives, Flavourings, Processing Aids and Materials in Contact with Food (AFC) and the Food Additives and Nutrient Sources added to food (ANS) assessed two long-term carcinogenicity studies in rats orally exposed to aspartame performed by the European Ramazzini Foundation (ERF). In both studies the authors reported a significant dose-related increase of malignant tumours in male and female rats. The AFC and ANS Panels concluded in their opinions that on the basis of all the evidence available, there was no indication of any genotoxic or carcinogenic potential of aspartame and that there was no reason to revise the previously established Acceptable Daily Intake (ADI) for aspartame of 40 mg/kg body weight. In a report published in 2010, National Experts nominated by the EU Member States reviewed the scientific literature on aspartame that had become available since 2002 and concluded that there was no requirement to reconsider the previous opinions on aspartame published by the EFSA Panels and the SCF.
In the Soffritti et al. (2010) study on aspartame in mice, dietary concentrations of 2 000, 8 000, 16 000 and 32 000 mg aspartame/kg diet equivalent to doses of 242, 987, 1 919 and 3 909 mg aspartame/kg bw/day were used. The details provided by the authors include the following. The maternal animals were orally exposed to aspartame from the 12th day of gestation throughout lactation and their offsprings were subsequently exposed to aspartame until natural death or for 130 weeks. Food intake, weight gain and clinical signs were monitored throughout the study. Unhealthy or moribund animals were sacrificed and necropsied as were all the mice that were alive at week 130. All tissues and organs were microscopically examined.
The authors reported a statistically significant dose-related increase in hepatocellular carcinomas in the 16 000 and 32 000 mg aspartame/kg diet dose groups of male mice compared to controls. In addition, a statistically significant increase in the incidence of alveolar/bronchiolar carcinomas in males was reported in the highest dose group compared to controls. No compound-attributed carcinogenic effects were reported by the authors in female mice at any of the doses tested. There was no change in the number of hepatocellular or alveolar/bronchiolar adenomas or the total incidence of carcinomas in male mice compared to controls and female mice compared to controls; however, a dose-dependent yet marginal and not statistically significant increase in the total number of adenomas in female mice was noted. Based on these observations, the authors conclude that aspartame induces cancer in the livers and lungs of male Swiss mice, thereby reiterating their previous conclusions that aspartame is a carcinogenic agent in rodents.
For the Soffritti et al. (2010) publication, EFSA noted that the descriptions of the statistical analyses and of the study design (e.g., outcome variable, covariates and assumptions, power calculation, blinding) are not sufficiently detailed, and thus, the relevance of the statistical analyses presented by the authors cannot be assessed at present.
In relation to the design of the study, EFSA also noted that it is generally accepted that life time studies until or close to natural death can lead to erroneous conclusions because of the following limitations. Older animals are more susceptible to illness and have increased background pathology, which includes spontaneous tumours and have a higher probability of autolysis than younger animals. These attributes can differ between treated and control animals; thus, it is very difficult to causally link tumours in treated animals to treatment or some intervening factor.
EFSA observed that the hepatic and pulmonary tumour incidences reported by Soffritti et al. (2010) all fall within their own historical control ranges for spontaneous tumours. EFSA also noted that Swiss mice are known to have a high background incidence of spontaneous hepatic and pulmonary tumours and that hepatic tumours in mice are generally considered as irrelevant for human risk assessment.
EFSA noted that Soffritti et al. (2010) suggested that the metabolism of aspartame leading to the formation of methanol might have played a role in the development of hepatocellular tumours. However, no evidence to support this hypothesis was provided.
Halldorsson et al. (2010) examined the association between intakes of sugar-sweetened and artificially sweetened soft drinks and preterm delivery in a cohort of 59 334 pregnant women. The authors used data from the “The Danish National Birth Cohort” and the objective of their study was to investigate whether there was a link between sugar-sweetened and artificially sweetened soft drinks and an increased incidence of preterm delivery.
The authors identified seven non-dietary factors that are well-recognized determinants of preterm delivery, namely, maternal age, height, pre-pregnancy Body Mass Index (BMI), cohabitant status, parity, smoking during pregnancy and familial socio-occupational status. To separate out the effects of food and energy intakes the authors also included total energy intake as a covariate. To establish a dose response relationship soft drink intake was also modelled (ranging from never to ≥4 servings per day). Secondary analyses were conducted by three categories of pre-pregnancy BMI.
The authors report that there was no association for the sugar sweetened carbonated and noncarbonated soft drinks. A monotonically increasing association was observed for the carbonated artificially sweetened soft drinks. A similar but more modest association was observed for noncarbonated artificially sweetened soft drinks. Similar results were observed from the secondary analysis, for carbonated artificially sweetened soft drinks, for late preterm and moderately preterm although for the early preterm no association was observed. The results were consistent for the three categories of BMI (i.e., an increase in risk for the intake of artificially sweetened carbonated with a further increase in risk with an increase in servings). The authors also presented some results for medically induced and spontaneous deliveries and found that their results were driven by the medically induced deliveries.
Halldorsson et al. (2010) conclude from their prospective cohort study that the daily intake of artificially sweetened soft drinks may increase the risk of preterm delivery in pregnant women.
EFSA considered that the study design, set up, conduct and quality of the Halldorsson et al. (2010) study were adequate to meet the defined objectives. However, as stated by the authors, due to the very nature of this type of studies, only an association between intake of artificially sweetened carbonated and noncarbonated soft drinks and an increased risk of preterm delivery can be identified. This is the first study on this subject, and EFSA noted that the study cannot be interpreted to support a causal relationship between the consumption of artificially sweetened soft drinks and preterm delivery in pregnant women. The authors put forward the hypothesis that the aspartame metabolite methanol might be associated with the observed preterm delivery but no direct supporting evidence was provided. In addition, EFSA observed that several sweeteners are used alone or in combination in soft drinks and that there is no justification to focus on any specific sweetener.
EFSA shared the view of the authors that further studies would be needed to reject or confirm their findings. In particular, EFSA wished to highlight that future studies should take into account potentially important confounders that could have affected the outcome of the study. For example, since the association identified by the authors was mainly driven by medically induced delivery, medical history and criteria on which the medical decisions to induce anticipated delivery are factors that should be investigated further. Also, differences in levels of the different artificial sweeteners between carbonated and non-carbonated drinks, and potential exposure to artificial sweeteners from sources other than soft drinks should be taken into consideration in the analysis. Other potentially important confounding factors such as exposures from soft drinks or other dietary sources to substances that might have an effect on pregnancy (e.g. caffeine) should be investigated. Furthermore, in view of the fact that studies can contribute directly to the risk assessment of a specific food additive only when they can ascribe the effects to this additive, EFSA advised that studies aiming at confirming an association should focus on a specific food additive.
In conclusion, EFSA has assessed the carcinogenicity study with transplacental exposure of mice to aspartame as reported by Soffritti et al. (2010). EFSA concluded that, on the basis of the information available in the publication, the validity of the study and its statistical approach cannot be assessed and its results cannot be interpreted. Furthermore, in view of the generally recognised lack of relevance for human risk assessment of the tumours observed in Swiss micewhen they are induced by non-genotoxic compounds , EFSA concluded that the results presented in the publication by Soffritti et al. (2010) do not provide sufficient scientific evidence to reconsider the previous evaluations by EFSA on aspartame that concluded on the lack of genotoxicity and carcinogenicity of the sweetener.
EFSA also assessed the prospective cohort study on the association between intakes of sugar-sweetened and artificially sweetened soft drinks and preterm delivery by Halldorsson et al. (2010). Since the intention of this study has been to generate hypothesis, that such studies are exploratory in nature and given the fact that there is no evidence available to support a causal relationship between the consumption of artificially sweetened soft drinks and preterm delivery, additional studies are required to confirm the association. For such studies, potential confounding factors should be investigated. Therefore, EFSA concluded that there is no evidence available to support a causal relationship between the consumption of artificially sweetened soft drinks and preterm delivery and that additional studies are required to reject or confirm an association.
Overall, EFSA concludes that the information available from the Soffritti et al. (2010) and Halldorsson et al. (2010) publications do not give reason to reconsider the previous evaluations of aspartame or of the other food additive sweeteners authorised in the European Union.
EFSA will continue monitoring the scientific literature in order to identify new scientific evidence on sweeteners that may indicate a possible risk for human health or which might otherwise affect the safety assessment of these food additives.
This EFSA statement follows the scientific statement adopted by the Food Additives and Nutrient Sources added to food Panel (ANS) on 3 February 2011 (EFSA, 2011). In its scientific statement, the ANS Panel has indicated that it will undertake a detailed analysis of the study results and conclusions reported by Soffritti et al. (2010), including the suggested implication of methanol.