University of Sussex
Browse

EFSA on Aspartame January 2013: a lost, but not the last, opportunity

Download (478.65 kB)
report
posted on 2023-06-08, 14:30 authored by Erik Millstone
Executive summary The draft report on the safety of aspartame, issued by the European Food Safety Authority’s ANS panel on 8 January 2013, is deeply flawed. There are at least two main types of flaws: those arising from the criteria by which ‘evidence’ has been selected, and those arising from the criteria by which those studies are interpreted. The criteria of inclusion have been overly narrow, and have in particular excluded vital documents that bear directly on the scientific competence with which some pivotal studies were conducting and on the accuracy with which they were reported. Such documentary evidence is directly relevant to the reliability of the reported data, and on the truth of the claims based on those studies. The implicit criteria of interpretation of the studies that have been included are perverse and biased. The panel could only have reached its conclusion that aspartame is safe by implicitly assuming that almost all studies indicating no adverse effects are entirely reliable, even though they have numerous weaknesses and were almost all commercially funded, while all the studies indicating that aspartame may be unsafe are deemed unreliable, even though they sometimes have particular methodological strengths and even though they have all been funded independently of vested commercial interests. On each of the 80 occasions when the panel discusses a study that indicated no apparent risks from aspartame those studies are taken at face value and typically assumed to be reliable. However, on each of the 27 occasions when the panel discusses studies that indicate that aspartame may pose risks, the panel is unremittingly critical of them, implicitly assuming that they must be misleading, so dismissing them. Often those worrying studies are dismissed not because of evidence but on flimsy grounds, for example by invoking speculative hypotheses without supporting evidence. The panel reached its conclusion that aspartame is safe not by applying uniformly critical standards to all the evidence from all studies, but by routinely forgiving the shortcomings of favourable studies yet being unremitting critical of all the studies suggesting any possible risks. The panel’s overall conclusions is driven more by the panel’s biased assumptions than by the evidence adduced. One possible explanation for the asymmetric bias in the interpretation of studies might be found in the pattern of conflicts of interest that characterise the members of the panel. Of the 17 members of the EFSA panel, 7 have direct commercial conflicts of interest, and another 5 have institutional conflicts of interest – for example because their employers have already announced that aspartame is safe. Only 4 panel members are not characterised by some relevant conflicts of interest. EFSA should therefore discount the draft report, convene a new panel composed only of, and supported only by, experts who are free of any conflicts of interest. They should be asked to review all the evidence, not just some of it. They should also make explicit their criteria of interpretation, and then show that those criteria have been consistently applied, and they should be applied to prioritises the protection of consumer and public health over commercial or industrial considerations. The European Commission and the European Parliament should also take responsibility for ensuring that EFSA acts properly to protect consumers rather than assisting the food and chemical industries

History

Publication status

  • Published

File Version

  • Published version

Publisher

SPRU

Department affiliated with

  • SPRU - Science Policy Research Unit Publications

Full text available

  • Yes

Legacy Posted Date

2013-02-25

First Open Access (FOA) Date

2016-03-22

First Compliant Deposit (FCD) Date

2016-03-22

Usage metrics

    University of Sussex (Publications)

    Categories

    No categories selected

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC