ArticlePDF Available

EFSA’s toxicological assessment of aspartame: was it even-handedly trying to identify possible unreliable positives and unreliable negatives?

Authors:
  • Science Policy Research Unit University of Sussex

Abstract and Figures

Background: A detailed appraisal is provided of the most recent (December 2013) assessment of the safety and/or toxicity of the artificial sweetener aspartame by the European Food Safety Authority's Panel on Food Additives and Nutrient Sources Added to Food. That appraisal is prefaced with a contextualising chronological account drawn from a documentary archive of the key highlights of the antecedent scientific and policy debates concerning this sweetener from the early 1970s onwards. The appraisal focuses specifically on Section 3.2 of the panel's review, which is headed 'Toxicological data of aspartame'. Methods: The methodology of the appraisal focusses on the extent to which the panel was symmetrically alert to possible false positives and false negatives, which in toxicological terms denote misleading indications of possible toxicity or misleading indications of safety. The methodology involved identifying and tabulating the prima facie indications of each of 154 empirical studies, and then comparing them with the way in which the panel chose to interpret the studies' findings, by focussing primarily on whether the panel deemed those studies to be reliable or unreliable. If the panel had been even-handed, the criteria for assessing reliability should have been the same for both putative positive and negative studies. Results: Eighty-one studies were identified that prima facie did not indicate any possible harm, and of those the panel deemed 62 to be reliable and 19 as unreliable. Seventy-three studies were identified that prima facie did indicate possible harm; of those the panel deemed all 73 to be unreliable; none were deemed reliable. A qualitative comparative review of the criteria of appraisal invoked by the panel to judge the reliability of putative negative and positive studies is also provided. Conclusion: The quantitative result indicate that the panel's appraisal of the available studies was asymmetrically more alert to putative false positives than to possible false negatives. The qualitative analysis shows that very demanding criteria were used to judge putative positive studies, while far more lax and forgiving criteria were applied to putative negative studies. Discussion: That quantitative and qualitative patterns are very problematic for a body supposed to prioritise the protection of public health. Given the shortcomings of EFSA's risk assessment of aspartame, and the shortcomings of all previous official toxicological risk assessments of aspartame, it would be premature to conclude that it is acceptably safe. They also imply that the manner in which EFSA panels operate needs to be scrutinised and reformed.
Content may be subject to copyright.
R E S E A R C H Open Access
EFSAs toxicological assessment of
aspartame: was it even-handedly trying to
identify possible unreliable positives and
unreliable negatives?
Erik Paul Millstone
*
and Elisabeth Dawson
Abstract
Background: A detailed appraisal is provided of the most recent (December 2013) assessment of the safety and/or
toxicity of the artificial sweetener aspartame by the European Food Safety Authoritys Panel on Food Additives and
Nutrient Sources Added to Food. That appraisal is prefaced with a contextualising chronological account drawn
from a documentary archive of the key highlights of the antecedent scientific and policy debates concerning this
sweetener from the early 1970s onwards. The appraisal focuses specifically on Section 3.2 of the panels review,
which is headed Toxicological data of aspartame.
Methods: The methodology of the appraisal focusses on the extent to which the panel was symmetrically alert to
possible false positives and false negatives, which in toxicological terms denote misleading indications of possible
toxicity or misleading indications of safety. The methodology involved identifying and tabulating the prima facie
indications of each of 154 empirical studies, and then comparing them with the way in which the panel chose to
interpret the studiesfindings, by focussing primarily on whether the panel deemed those studies to be reliable or
unreliable. If the panel had been even-handed, the criteria for assessing reliability should have been the same for
both putative positive and negative studies.
Results: Eighty-one studies were identified that prima facie did not indicate any possible harm, and of those the
panel deemed 62 to be reliable and 19 as unreliable. Seventy-three studies were identified that prima facie did
indicate possible harm; of those the panel deemed all 73 to be unreliable; none were deemed reliable. A qualitative
comparative review of the criteria of appraisal invoked by the panel to judge the reliability of putative negative and
positive studies is also provided.
Conclusion: The quantitative result indicate that the panels appraisal of the available studies was asymmetrically
more alert to putative false positives than to possible false negatives. The qualitative analysis shows that very
demanding criteria were used to judge putative positive studies, while far more lax and forgiving criteria were
applied to putative negative studies.
Discussion: That quantitative and qualitative patterns are very problematic for a body supposed to prioritise the
protection of public health. Given the shortcomings of EFSAs risk assessment of aspartame, and the shortcomings
of all previous official toxicological risk assessments of aspartame, it would be premature to conclude that it is
acceptably safe. They also imply that the manner in which EFSA panels operate needs to be scrutinised and
reformed.
Keywords: Aspartame, EFSA, Appraisal, Asymmetry, Public health
© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
* Correspondence: e.p.millstone@sussex.ac.uk
Science Policy Research Unit, University of Sussex, Brighton BN1 9SL, England
Millstone and Dawson Archives of Public Health (2019) 77:34
https://doi.org/10.1186/s13690-019-0355-z
Background
Amongst food additives, aspartame is one of the most
controversial, especially in the USA, but also in the UK
and the EU. The most recent official attempt to settle
the controversy was provided by the European Food
Safety Authoritys (or EFSA) Panel on Food Additives
and Nutrient Sources added to Food (or ANS) in De-
cember 2013 [1]. The ANS Panel: “… concluded that as-
partame was not of safety concern at the current
aspartame exposure estimates or at the ADI [acceptable
daily intake] of 40 mg/kg bw/day[2]. An ADI is a level
of consumption officially deemed to be acceptably safe.
In the context of a set of proposals to transform the
European Food Safety Authority into an Open EFSA,a
2014 discussion paper from the EFSA Board highlighted
some of the intended benefits from the openness and trans-
parency to which it claimed to aspire. The text stated:
The proactive and committed adherence to openness
and transparency values by the Authority facilitates an
informed debate, both among experts and the public,
on scientific issues within EFSAs remit. Thereby this
represents a prerequisite for constructive and
informed dialogue between the agency and any
interested organisation or individual.[3].
This paper is in part intended to contribute to a construct-
ive and informed dialogue with EFSA and with other
stakeholders. This paper will highlight the limited extent
to which the ANS panels risk assessment of the artificial
sweetener aspartame was in practice transparent. EFSA
also stipulated that its risk assessments should be repro-
ducible[4], correspondingly this paper will also specify
the conditions under which it might be reproducible, in
terms of the assumptions that would need to be made.
This paper has two main sections. The first provides a
chronological account drawn from a documentary arch-
ive of the key highlights of the antecedent scientific and
policy debates concerning the safety and/or toxicity of
aspartame from the early 1970s onwards, while the sec-
ond provides a critical review of the December 2013
ANS report and explains why it did not settle the con-
troversy but rather contributed to it. The central ques-
tion addressed in the second section asks whether the
ANS panels review of toxicological evidence was sym-
metrically sceptical? In other words, did it even-
handedly try to identify possible unreliable positives (ie
studies indicating adverse effects, but unreliably) and un-
reliable negatives (ie studies not indicating adverse ef-
fects, though unreliably), or was it asymmetrically
focused more on one than the other?
The answer that emerges from a detailed quantitative
and qualitative scrutiny of the studies, and the ANS
panels representations and interpretations of those
studies, is that the panel frequently treated studies that
provided no prima facie evidence of harm as if they were
unproblematically reliable, even when they were very
weak studies, yet discounted the results of every single
one of 73 studies that indicated that aspartame could be
harmful, deeming all those studies to be unreliable and/
or insufficient, even though some were more powerful
and sensitive than some of the seemingly negative stud-
ies that the panel deemed reliable.
Furthermore, the evidence shows that, if the bench-
marks that the ANS panel used to evaluate the results of
negativestudies had been consistently used to evaluate
the results of positivestudies then the panel would have
been obliged to conclude that there was sufficient evi-
dence to indicate that aspartame is not acceptably safe.
Correspondingly, if the benchmarks that the ANS panel
used to evaluate the results of positive studies had been
consistently used to evaluate the results of negative stud-
ies then the panel would have been obliged to conclude
that there was insufficient evidence to conclude that as-
partame is acceptably safe. Instead it said that aspartame
is safe when consumed at currently estimated rates of
consumption. This paper therefore questions the pro-
cedure followed by the ANS panel and its conclusion.
Given that animal experiments conducted to test, for
example, the safety and/or efficacy of chemicals often
generate findings that are equivocal and/or conflicting,
considerable efforts have, in recent years, been devoted
to developing suitable methods by which to provide sys-
tematic reviews of the diverse results of such studies to
indicate the overall implications of multiple datasets.
One promising example of such novel methods is known
as SYRCLE, which is an adapted version of the Cochrane
Risk of Bias tool [5]. An alternative approach is known
as CAMARADES, which is an acronym for Collabora-
tive Approach to Meta-Analysis and Review of Animal
Data from Experimental Studies[6]. Both SYRCLE and
CAMARADES focus on trying to identify risk of biasin
animal studies, though their applicability does not ex-
tend to epidemiological or clinical studies, though
Cochrane Collaboration .guidelines can be applied to
those studies. Given that this paper is a review of a cen-
tral part of one single EFSA document, those approaches
are not applicable in this context. However, as and when
future official public policy reviews of the safety and/or
toxicity of aspartame and other chemical additives, con-
taminants and nutrients will be conducted, such system-
atic approaches should be adopted, and their selection
and application explained and justified.
Section 1
Regulatory, scientific and corporate context
A good case can be made for the claim that no industrial
food additive has been more controversial than
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 2 of 22
aspartame [7]. G D Searle, a US pharmaceutical com-
pany, first petitioned the US Food & Drug Administra-
tion (FDA) for permission to market aspartame in 1973;
in 1974 the FDA announced its intention to grant per-
mission for its use in dry goods, such as table-top sweet-
eners [8]. Before that decision could be implemented,
objections were raised by independent scientists alleging
that aspartame could cause mental retardation, brain le-
sions and neuroendocrine disorders [9]. Before those is-
sues were resolved, further objections were raised
focussing on documentary and interview evidence indicat-
ing that Searle, and its sub-contractor Hazleton, had failed
to conduct properly at least 15 toxicology tests, and that
their subsequent reports had been misleading [10].
The key allegation was that the conduct of those stud-
ies was seriously incompetent, and that when Searle
managers recognised the failures that had taken place
they mis-reported the studies to conceal those failures,
portraying the studies as if they had been conducted
competently and reported accurately. The FDA could
not simply discard or discount the evidence it had col-
lected, which revealed that it had initially been misled by
unreliable studies and reports, so instead it was dis-
carded and discounted through several separate
processes.
The FDA arranged for the 15 problematic studies to
be reviewed, but in two separate sets and by different in-
stitutions. The FDAs Bureau of Foods convened 5 of its
staff into a Task Force and assigned it to review just 3 of
those 15 studies [11]. The job of reviewing the other 12
studies was assigned to an organisation called the US
Universities Association for Research and Evaluation in
Pathology (UAREP). The FDA negotiated an agreement
with G D Searle under which Searle would pay the costs
of the UAREP work, in exchange for which the FDA
agreed that Searle would contribute to setting the
UAREP reviews terms of reference [12]. Dr. Adrian
Gross, who was then an FDA pathologist, and who first
uncovered the problems with Searles laboratory work
and reporting, argued in 1976 that the members of the
UAREP team were not appropriately qualified to con-
duct the kind of investigation that was required, and
consequently that their eventual conclusions could not
be considered to be reliable or definitive [13]. Gross had
been instrumental in uncovering the shortcomings in
Searletests on aspartame as well as on two pharmaceut-
ical products (flagyl - an antibiotic, and aldactone - a di-
uretic) [14]. Gross persistently criticised all 15 of the
studies, ie both those reviewed by the Bureau of Foods
Taskforce and those reviewed by the UAREP.
Both the Bureau of Foods Task Force and the UAREP
subsequently issued reports containing reassuring con-
clusions, but in both cases they only contrived to reach
those conclusions because their terms of reference had
been set particularly narrowly and in ways that failed to
address the critical shortcomings of the studies. Both
teams focussed their attention on characterising and
comparing the remaining parts of the documentary re-
cords of those studies alongside the detectable features
of laboratory samples, including samples of animalstis-
sues on glass slides, but without examining the prior
processes that had resulted in those documents being
written and the samples remaining available [15]. The
report of the UAREP never explicitly said that the re-
viewers took Searles documentary and laboratory evi-
dence at face value, but that in practice was what
happened. On a few occasions, the UAREP highlighted
omissions from documents, and inconsistencies between
them, but otherwise treated them as if they were un-
problematically reliable. The UAREP was not provided
with evidence of the incompetence of some of the la-
boratory staff or the fictional aspects of some of the
documentation. They took the documents and the path-
ology slides at face value, and checked the arithmetic.
Although UAREP noted “… a substantial number of
minor and inconsequential discrepancies …” during its
review, it found “… few, if any, discrepancies which
would produce a change of greater than five percent in
the final numerical data being compared[16].
The FDA bureau of food task force report
The conclusion of the Bureau of Foods Task Force stated
that while the three tests had not been properly con-
ducted, and although there were marked differences be-
tween raw data and the summaries submitted in the
petition to the FDA, those differences: “…were not of
such a magnitude that they would significantly alter the
conclusions of the studies[17]. The three studies
reviewed by the Bureau of Food Task Force were listed
using Searles numbering system as: E5, E-89 and E-77/
78. The last of those three studies has been omitted
from our quantitative analysis below because it was a
study of the toxicity of diketopiperazine (or DKP) which
is one of aspartames breakdown products, rather than a
study of aspartame itself. It was however included in
Section 7.2.4.1 of the ANS panels report, where it was
interpreted by the panel as a reliable negative.
The Task Force had considerable difficulty in evaluat-
ing the studies, in part because in some cases there were
no raw data with which to compare the reported results.
In other cases, it was impossible to determine which the
real raw results were and which were subsequent revi-
sions or summaries. In some contexts, the Task Force
had to rely on information and assumptions provided by
Searle employees who had not been involved in the ori-
ginal work. At worst, it was impossible to identify the
occasion on which a particular animal had died. For ex-
ample, the report said in relation to E78/79, a study of
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 3 of 22
the toxicity of DKP: Observation records indicated that
animal A23LM was alive at week 88, dead from week 92
through week 104, alive at week 108, and dead at week
112[18]. Most scientists do not believe in reincarna-
tion, and we should not expect that the FDA or the ANS
panel to do so either.
When reviewing the test (E78/79) on DKP, the report
listed no fewer than 52 major discrepancies in the Searle
submission [19]. One of the central problems concerned
the quantities of DKP supposedly consumed by the rats.
The FDA investigators found no fewer than 3 separate
documents with different specifications for the content
and the purity of the test substance, and they were un-
able to establish precisely which specification, if any, was
correct. It was impossible to reconcile the quantity of
the chemical requisitioned from stores with the quan-
tities supposedly fed to the animals. There were ques-
tions raised as to the extent to which the DKP was
uniformly incorporated into the animalsfood. There is
clear evidence to show that the test substance was not
properly ground, and inadequately mixed, so that it
might have been possible for the animals to avoid the
DKP while eating their food [20].
Ten years later, at a November 1987 hearing of a US
Senate Committee hearing Dr. Jacqueline Verrett, an
FDA toxicologist who had been a member of the Bureau
of Foods Task Force, explained that the three studies it
had examined were “… woefully inadequate …” (p 387)
and that: Almost any single one of these aberrations
would suffice to negate a study designed to assess the
safety of a food additive, and most certainly, a combin-
ation of many such improper practices would, since the
results are bound to be compromised. It is unthinkable
that any reputable toxicologist giving a completely ob-
jective evaluation of this data resulting from such a
study could conclude anything other than that the study
was uninterpretable and worthless and should be re-
peated.[21]. Nonetheless, the ANS panel included E5
in Section 3.2 of its December 2013 review, but deemed
it an unreliable positive; though it deemed E89 to be a
reliable negative.
When asked to explain the contrast between the 1977
report of the Bureau of Foods Task Force and her subse-
quent statements to the 1987 Senate committee, Jacque-
line Verrett explained that the Task Force members
were: “… limited in what we could actually conclude
about the studies. We were not allowed to comment on
the validity of any study. It was an explicit instruction
based on administrative rather than scientific consider-
ations. We were supposed to figure out what the conclu-
sions would have been if the studies had been fully and
correctly reported. We were obliged to ignore the proto-
cols and the non-homogeneity of the DKP Some ani-
mals did reject the DKP. Searle initially said that it may
not have been fully mixed but that that did not matter,
they later said that it had been fully mixed. We were not
allowed to consider those issues by the Bureau of Foods
administrator We were ham-strung in being able to
comment. The fact is that the studies should not have
been considered at all, and that was the position from
the beginning.[22].
The report of the UAREP
In 1978, the UAREP delivered its 1062-page report,
which concluded that the 12 studies they had audited
were authentic[23]. Gross subsequently told the No-
vember 1987 US Senate committee hearing that:
“… no amount of additional examinations of pathology
material such as undertaken by the UAREP [or]
new additional statistical analyses and no judgmental
evaluations or interpretations of any data arising from
those studies can in any way rectify the basic problem
: in the absence of reasonable expectation that the ex-
perimental animals were administered the correct dos-
ages of the test agent, any observational data carried out
on those animals must be regarded as questionable or
flawed. This is to say nothing of all the myriad of other
problems involving the competence of those conducting
such studies, and the [lack of] care they exercised in
their execution. Once a study is carried out and the test
animals are disposed of, all that remains are the number
of tiny bits of tissue preserved from their organs for
microscopic examination and the written records of ob-
servations made by those who actually carried out that
study. While the tissues themselves can be examined by
others long after the remains of those animals no longer
exist, the reliability of the written records has already
been found to be unacceptable in a great variety of ways.
Once a study is compromised in its executions, it is
beyond salvation by anyone. Even with respect to those
small portions of tissue preserved for microscopic exam-
ination for an indefinite period of time after any study is
completed there are serious problems there is little if
any assurance that such samples of tissues as were pre-
served actually originate from the specific animals said
to have been their source Furthermore, due to the
unacceptably high rate of post-mortem autolysis, a great
many such tissues were not collected at all from the ex-
perimental animals.[11].
The 12 studies for which the UAREP was responsible
for reviewing were listed as: E-28, E-33 & 34, E-70, E-75,
E-76, E-86, E-87, E-9, E-11, E-19, E-88 and E-90. Of
those, E-76 was not included in the ANS panels discus-
sion of the toxicity of aspartame, but it was included in
relation to DKP. It is consequently omitted from our
quantitative analysis below. When referring to the stud-
ies reviewed by the UAREP, Verrett said: “… the safety of
aspartame and its breakdown products has still not been
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 4 of 22
satisfactorily determined, since many of the flaws cited
in these three studies [those reviewed by the FDA Task
Force] were also present in all the other studies submit-
ted by Searle [including those reviewed by UAREP]
[24].
Despite the fact that those two reviews had provided
reassurances, several objectors remained dissatisfied, and
furthermore a new complex set of objections to the
safety of aspartame were introduced [25]. In an attempt
to resolve the controversy once and for all, the FDA pro-
posed the establishment of a so-called Public Board of
Inquiry (or PBoI). This was a unique institution; the pro-
cedure had never previously been used, and in all prob-
ability will not be used again [26]. The PBoI first met in
early 1980, and published its conclusions in October
1980 [27]. Two sets of issues were on its agenda. On one
of the crucial questions its view was that aspartame con-
sumption would not pose an increased risk of brain
damage resulting in mental retardation, but on the other
issue it concluded that the evidence available did not
preclude the possibility that aspartame could induce
brain tumours. Consequently, the Board recommended
that aspartame should not be permitted for use, pending
the results of further studies.
The FDAs attempted follow-up
In 1976, the evidence of misconduct by Searle, in rela-
tion to two pharmaceutical products and aspartame, was
sufficient to convince the FDAs Chief Counsel (Richard
Merrill) to instruct the Federal Attorney in Chicago to
convene a Grand Jury to investigate: “… apparent viola-
tions of the Federal Food, Drug, and Cosmetic Act
and the False Reports to the Government Act by G.D.
Searle and Company and three of its responsible officers
for their willful and knowing failure to make reports to
the Food and Drug Administration required by the Act
and for concealing material facts and making false
statements in reports of animal studies conducted to es-
tablish the safety of Aspartame.[28].
A team of investigators working for US Senator Met-
zenbaum subsequently gathered and then released a set
of documents showing that soon after the Chicago Fed-
eral Attorney (Samuel Skinner) received Richard Mer-
rills April 1976 letter, he was invited to join the board of
the firms of lawyers (Sidley & Austin) then representing
G D Searle; he accepted the invitation. The Searle dos-
sier remained suspended until the next Federal Attorney
(Thomas Sullivan) was appointed; he too was then in-
vited to join the board of Sidley & Austin, and accepted
the invitation [29]. Those processes served to delay legal
proceedings until the interval, specified by the Statute of
Limitations, had expired. Searle was therefore not prose-
cuted, but that did not establish the corporations
innocence.
Informing EFSA
Documents providing detailed empirical support for the
above account of the toxicological and regulatory history
of aspartame were included in a dossier provided by Erik
Millstone to the secretariat of the ANS panel in October
2011. The Secretariat had issued a public request for “…
all necessary data (published, unpublished or newly gen-
erated) …” on 1st June 2011 [30]. Millstone responded
by sending EFSA an annotated list of 30 relevant docu-
ments. The Head of the ANS Unit replied requesting by
4 November 2011 digitised copies of 27 of the docu-
ments [31]. 26 of the 27 documents were then fully digi-
tised and dispatched to EFSA on a CD-ROM. The
exception was the UAREP report, because at 1062 pages,
only the Table of Contents was provided. In the event,
the ANS panels reports of both January and December
2013 failed to mention that dossier, or most of the docu-
ments of which it was comprised, which was unjustified
when judged by reference to both scientific and policy
criteria. Copies of those documents are available at
http://www.sussex.ac.uk/spru/research/projects/fcs.
Aspartames eventual approval
It was not until 1981, and the arrival of the Reagan ad-
ministration, that the FDA permitted aspartames com-
mercial use, restricting it initially to dry goods [32]. In
1983 the FDA approved its use in beverages, which be-
came its major market for which it was marketed under
the name Nutrasweet[33]. The eventual approval
process was seriously problematic. The FDA Commis-
sioners decision was taken against the advice of FDA
toxicologists and the Public Board of Inquiry [34]. Com-
missioner Hayes approved aspartame as one of his first
decisions in that post, and was subsequently employed
by the US National Soft Drinks Association [35].
When the Reagan administration assumed office in
early 1981, Searles former CEO accepted the job as the
new US ambassador to Beirut, in exchange for a promise
that the administration would get aspartame onto the
market [36]. The Reagan administration took office a
short while after a terrorists group truck-bombed the US
embassy in Beirut. Reagan appointed as his new ambas-
sador to Beirut the former chief executive of GD Searle,
Donald Rumsfeld.
Aspartame was deemed acceptably safe by the World
Health Organisation and UN Food and Agriculture
Organisation's Joint Expert Committee on Food Addi-
tives (JECFA) in 1980 and by the European Commis-
sions Scientific Committee for Food in 1985 [37].
Aspartame was also approved in the United Kingdom
(UK), on advice from the Committee on Toxicology
(CoT), the chair of which had his research indirectly
funded by Searle [38]. All of those committees based
their judgements on sets of studies that included those
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 5 of 22
that Verrett had accurately characterised as the woefully
inadequate. They were included and evaluated as if they
were no more problematic than any other studies; their
severe shortcoming were ignored or discounted, or
maybe they were not drawn to the attention of the
members of those committees. Several of the members
of those committees were, moreover, acting as paid con-
sultants to relevant food, beverage and chemical com-
panies, in circumstances when declarations of conflicts
of interest were then not required [39]. It would be naive
to presume that undeclared conflicts of interest could
not have influenced the judgements of those committee
members.
Subsequent twentieth century developments
In October 1985 it emerged that Searle had been ac-
quired by the large chemical company Monsanto, which
had long been one of the major manufacturers of Sac-
charin [40]. Monsanto subsequently detached the aspar-
tame business from the remainder of Searles operations
and established the NutraSweet Company [41].
After aspartame had been approved the controversy
shifted to discussions of reports from consumers of
acute adverse effects [42]. The most common symptoms
were (and are) neurological problems including severe
headaches and blurred vision; thankfully reports of
epileptic-type seizures, though serious, are rare. Such
evidence has repeatedly been officially dismissed as an-
ecdotal, though the sufferers often report that when con-
sumption ceases so too do the symptoms. Moreover,
when symptoms recur, the sufferers of those symptoms
often discover that they had inadvertently consumed as-
partame [43].
In the 1990s one key development was a paper by
Olney et al. in the Journal of Neuropathology and Ex-
perimental Neurology suggesting that the introduction of
aspartame into the USA may have resulted in a rapid in-
crease in the incidence of a particularly aggressive type
of brain tumour [44]. Their evidence was however offi-
cially discounted, despite providing convergent indica-
tions from animal studies, in vitro mutagenicity tests
and human epidemiological data.
Twenty-first century debates
At the end of the twentieth century there were, conse-
quently, good grounds for concluding that no one could
be confident that aspartame was acceptably safe. In this
century, the main toxicological contributions were pro-
vided by the Ramazzini Foundations rodent carcinogen-
icity studies [45].
The conventional protocols for long-term rodent
feeding studies typically involve feeding a test com-
pound to four dose-groups of animal, with 50 males
and 50 females in each of low-dose, mid-dose and high-
dose groups as well as a corresponding control group,
making a total of 400 animals. The typical duration for
carcinogenicity studies in rodents is 104 consecutive
weeks [46], although the OCED has indicated that,
while the duration will normally be 24 months for ro-
dents, for specific strains of mice, 18 months may be
more appropriate [47].
The first Ramazzini study on aspartame, published in
2005, used 1800 rats. Instead of testing the compound at
three dose levels (plus a control group) they tested it at
six dose levels plus controls. Instead of killing the rats at
2 years of age, the rats were allowed to live longer so
that long-term effects could be studied. Subsequently,
the Ramazzini Institute published data from a mouse
study of aspartame, and a further rat study that included
in utero exposure [48]. Gift et al. have argued that: The
protocols characteristic of RI [Ramazzini Institute] stud-
ies can cause interpretive challenges, but aspects of the
RI design, including gestational exposure, life span ob-
servation, and larger numbers of animals and dose
groups, may impart advantages that provide chemical
risk assessors with valuable insights for the identification
of chemical-related neoplasia not obtained from other
bioassays[49].
In these and many other ways, the Ramazzini study was
more thorough, sensitive, reliable and relevant to human
exposure than those conducted in accordance with con-
ventional protocols. The authors reported in 2005 that
their study: “… demonstrated for the first time that APM
[aspartame] is a multipotent [ ] carcinogenic agent …”
with dose-related tumour increases in both males and fe-
males [50]. In 2010 the Ramazzini team published the re-
sults of a study showing that aspartame induced tumours
in the livers and lungs of male mice [51].
Several official bodies, including the US FDA, JECFA,
the Scientific Committee on Food (SCF) of the European
Commission and the UKs CoT discounted those find-
ings, complaining that the Ramazzini studies had not
followed standard protocols. While the Ramazzini proto-
cols were non-standard, their deviations from the stand-
ard, by using more animals in more dose groups and not
sacrificingthem prematurely entailed that the Ramaz-
zini studies provided greater sensitivity than could be
obtained from a standard study. Keeping the animals
until they die may not be common practice, but since
European Union (EU) food safety policy legislation stip-
ulates that Assuring that the EU has the highest stan-
dards of food safety is a key policy priority …” [52]we
might have expected that EFSAs benchmark would be
the protection of all consumers throughout their entire
lives, rather than, for example, only until they reach re-
tirement age. Those considerations imply that the
Ramazzini protocol can be expected to provide a better
model of the risks to the population of Europe than
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 6 of 22
studies that sacrificethe animals prematurely, however
orthodox they might be. Premature sacrifice might well
result in unreliable negatives, and unacknowledged ones
at that.
One reason why the findings of the Ramazzini rat
study had been officially discounted was because of rela-
tively high rates of respiratory infections in the elderly
rats [53]. However, as the rates of infection in the test
groups were not significantly different from that in the
control group, those infections could not explain the
dose-related increase in tumours. Caldwell et al. convin-
cingly rebutted the hypothesis that lymphomas and leu-
kaemias were induced by infection. They noted, for
example, that while respiratory infections frequently
occur in old rats, and in most Ramazzini Institute rat
bioassays, leukaemia and lymphoma were only reported
in a few animals, namely 8 out of 112, implying that a
link between the respiratory infections and those path-
ologies was improbable [54].
Another reason why the findings of the Ramazzini In-
stitutes studies have been officially deemed not to be re-
liable has been because the tumour rates in treated
animals were within the ranges reported for historical
controls, even though they showed significantly higher
rates that those in the concurrent controls. One reason
why that criterion of interpretation is problematic was
articulated by the WHOs International Agency for Re-
search on Cancer, which had emphasised in January
2006 that: It is generally not appropriate to discount a
tumour response that is significantly increased compared
with concurrent controls by arguing that it falls within
the range of historical controls .[55].
In 2013 the EFSA ANS panel, in line with other statu-
tory risk assessors, discounted the Ramazzini findings as
a set of unreliable positives, while accepting as reliable
negatives, the evidence of studies that had many more,
and far more serious, imperfections including those Ver-
rett previously characterised as woefully inadequate. In-
evitably, there were imperfections in the Ramazzini
studies, but all studies are characterised by some imper-
fections. The ANS panel chose to treat many of the 15
earlier studies [56] (discussed above) as reliable, despite
the fact that their imperfections were very substantially
greater than those that characterised Ramazzinis work.
Despite the efforts of EFSA, and a coalition of indus-
trial and commercial stakeholders, to provide reassur-
ances about the safety of aspartame, the accumulation of
fresh evidence and public concerns provoked the Euro-
pean Parliaments Public Health and Consumer Protec-
tion Committee to call on the Commission formally to
instruct EFSA urgently to initiate a review of aspartames
toxicity and safety, rather than keeping to a previously-
set target-date of 31 Dec 2020 before doing so. In May
2011 the European Commission asked EFSA to re-
evaluate the safety on aspartame (E951) as a food addi-
tive, and to do so by 31 July 2012 [57].
The ANS panel issued a 245-page draft reportin Janu-
ary 2013, and requested comments by 15 February 2013
[58]. The abstract of that document stated that the Panel
had: “… concluded that there were no safety concerns at
the current ADI of 40 mg/kg bw/day. Therefore, there
was no reason to revise the ADI for aspartame.[59]The
panels draft was problematic in numerous respects; it
failed to address the key issues concerning the unreliability
of the 15 studies, namely those that had previously been
reviewed by the FDA Bureau of Foods Task Force (ie E5,
E-89 and E-77/78) and those previously reviewed by the
UAREP (ie E-28, E-33 & 34, E-70, E-75, E-76, E-86, E-87,
E-9, E-11, E-19, E-88 and E-90), on which Erik Millstone
had provided the ANS panels secretariat with detailed
documentary evidence. All of those 15 studies were cited
in the ANS panels report, and the only relevant comment
on them cited the UAREPs document, but failed to refer
to Grosss devastating critique of the relevance and reli-
ability of the UAREP review [60].
The way in which the studies, which had been in-
cluded, had been interpreted appeared consistently in-
consistent [61]. The ANS panel had portrayed most of
the studies that did not indicate any possible harm, ex-
cept at dose levels above the nominal no-observed-ad-
verse effect-level(or NOAEL) of 4000 mg of aspartame
per kilogramme of the body weight of the test-species
per day (ie mg/kg bw/day) as unproblematically reliable,
while portraying each and every one of the studies indi-
cating possible harm as unreliable and/or inconclusive,
even though many of the studies providing positive evi-
dence of toxicity were far more sensitive and rigorously
conducted and reported than some of the apparently
negative studies [62].
EFSA issued a final reporton 10 Dec 2013. It culminated
with: Overall, the Panel concluded from the present assess-
ment of aspartame that there were no safety concerns at
the current ADI of 40 mg/kg bw/day. Therefore, there was
no reason to revise the ADI for aspartame.[63].
Millstone responded on 14 December with a 3-page
critique arguing that the panel had consistently treated a
very large majority of negative studies as reliable, while
discounting each and every one of the studies providing
evidence of harm as unreliable [64]. In response, EFSAs
Head of Regulated Products and senior colleagues
held a video-conference on 14 April 2014 with Mill-
stone. On 14 November 2014 EFSAs Head of Regu-
lated Products wrote to Millstone, referring to a full
list of studies. The letter reported that an internal re-
view had concluded that:
“…we found that the number of studies not
indicating harm considered in the [ANS] opinion on
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 7 of 22
aspartame as unreliable was substantially higher (35%)
than the number claimed in your letter (20%).
Likewise our analysis showed that the number of
studies indicating harm by aspartame and treated by
the ANS Panel as reliable was not zero as stated in
your letter. Instead, we found a similar proportion of
studies in which an adverse effect by aspartame was
described (typically at a dose above the NOAEL) and
that were considered by the ANS Panel as unreliable
as compared to reliable (43% vs 57%). Moreover, the
relative proportion of studies produced or funded by
industry and found reliable to be used in the safety
assessment of aspartame to be very similar to the
proportion of studies carried out using no-commercial
funds, irrespective of the outcome of the study (54%
vs 46%). Those findings do not support your claim
that EFSA took a pro-industry views. I remain
confident after reviewing our internal analysis that the
analysis of the studies and the literature reported in
the aspartame opinion was conducted in a scientific-
ally rigorous manner and that there was no evidence
of bias.[65].
In other words, EFSA argued that it had been symmet-
rically sceptical, with respect to both putative false posi-
tives and putative false negatives. This paper is in part a
response to that claim, but it is also intended to deepen
understandings how some regulatory scientific panels
discharge their duties.
Section 2
Methods and approach
It is against the background of this contested saga that
we report the results of our characterisation of the ways
in which the EFSA ANS panel interpreted the individual
studies that were included in Section 3.2 of the ANS
Panels December 2013 review; the section is headed
Toxicological data of aspartame; it extended from pages
56 to 102. All the studies and documents are listed in
full in EFSA Panel report (available at https://efsa.online-
library.wiley.com/doi/epdf/10.2903/j.efsa.2013.3496, see
pages 151170).
The goal of this investigation was to establish if the
ANS panel even-handedly tried to identify possible unre-
liable positives and unreliable negatives, or whether it
asymmetrically focused more on one than the other?
That issue is important because an asymmetry would
constitute evidence of bias. If greater effort had been de-
voted to identifying and discounting false negatives that
would indicate bias against aspartame, and in favour of
consumer protection, whereas if greater effort had been
devoted to identifying and discounting false positives
then the bias would have favoured commercial interests
to the detriment of consumer protection. Symmetry
would imply that the ANS panel was neutral as between
those two competing interests. It is important however to
acknowledge that a symmetrical perspective need not entail
equal numbers of true and/or false positives and true and/or
false negatives. That would depend, amongst other things,
on the truth regarding the safety or otherwise of aspartame.
A toxicological risk assessment requires more than the
collection of data. The data generated by empirical studies
need to be evaluated and interpreted. All studies have
shortcomings, but some are significantly more robust and
reliable than others. That is especially the case when the
data derive from studies of model systems, such as labora-
tory rodents or microbes in glass dishes rather than from
clinical or epidemiological studies of people. The rele-
vance of, for example, the results of a rodent feeding study
to the probable effects on people can never be taken for
granted. Not all studies are equally relevant, and judge-
ments of extrapolative relevance are firstly unavoidable
and secondly never determined solely by reference to the
results of the study under consideration.
Judging extrapolative relevance is closely related to
judgements about the existence, extent and implications
of uncertainties, and about how the benefit of the conse-
quent doubts should be allocated. This is the basis for
the contrast between so-called positiveand negative
regulatory list systems. With a negative list systems che-
micals are assumed to be safe until shown to be harmful,
while a positive list system assumes that they may be
risky unless and until sufficient evidence of safety is pro-
vided. The EU is supposed to have a positive list system
for many categories of food additives, including intense
sweeteners. With respect to Section 3.2 of the ANS
panelsDecember 2013 review of the safety of aspar-
tame, three main types of answers might be forthcoming
to the question of whether the ANS panel's treatment
was biased.
1. If the ANS panels treatment was symmetrically
concerned with, and sensitive to, identifying
apparently positivefindings of toxicity and
negativefindings then its perspective could be
characterised as neutral as between commercial and
consumersinterests. For the purposes of this
appraisal, symmetry was the null hypothesis.
2. If the panel was more concerned to detect and
discount putative unreliable negatives than
unreliable positives then it could be characterised as
favouring the interests of consumers over
commercial ones.
3. If, however, the panel was more concerned to
detect and discount putative unreliable positives
than unreliable negatives then it should be
characterised as favouring commercial interests
over those of consumers.
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 8 of 22
We have used, as the indicators of symmetry or asym-
metry, firstly the quantitative frequency with which stud-
ies were variously deemed by the ANS panel to be
reliable or unreliable, and secondly the qualitative strin-
gency of the hurdles that had to be satisfied for the panel
to deem a putative positive as a reliable positive or an
apparently negative study as a reliable negative. The fea-
tures and severity of those hurdles were sometimes ex-
plicit in the panels text, but often they had to be
inferred from the comments in the text, some of which
were rather enigmatic, and from the subsequent discus-
sion leading to the Panels conclusions.
The data for this analysis were obtained solely by focus-
sing on the studies cited and discussed in Section 3.2 of
the December 2013 report (pages 56 to 101), which was
entitled Toxicological data of aspartame. This analysis
therefore does not cover parts of the document, such as
Section 2.8, that reviewed evidence for an exposure as-
sessmentor Section 3.1 on Absorption, distribution, me-
tabolism and excretion of aspartame. The corresponding
discussions of the toxicity of aspartame-derived methanol
or diketopiperazine (DKP), in Section 7 pp. 127133, are
also outside the scope of this analysis. While our analysis is
not exhaustive, it focusses on a pivotal section and is suffi-
cient to sustain robust and relevant conclusions. There is
also no reason to think that Section 3.2 might be unrepre-
sentative of the document as a whole; though if it were, that
would be no less problematic. A detailed tabulation of all of
the studies cited by the ANS panel in Section 3.2 of its De-
cember 2013 report is provided in Additional file 2.
Those studies are allocated to the 7 categories, which
are listed below in Table 1, in Additional file 1.The
resulting statistics are provided in Tables 2and 3below.
Criteria of categorisation
The following analysis has been constructed by develop-
ing two distinct binary characterisations, firstly the tenor
of the authorsconclusion of their reports of their stud-
ies and secondly whether the ANS panel had subse-
quently deemed those studies to be reliable or as
unreliable. The first categorisation differentiates the
studies into two main groups: those for which the
authors provided some evidence of adverse effects on
the one hand and those providing no evidence of adverse
effects on the other.
As no animal toxicology studies, of which we are
aware, has shown adverse effects in all types of cells, tis-
sues or systems, our default assumption has been to cat-
egorise studies as positive, ie indicating possible harm, if
the evidence concerned one or more parameters. On the
other hand, studies have been categorised as negative if
no indications of harm were reported by the authors.
The panels approach was to allow one important excep-
tion to that interpretative criterion. The panel would
sometimes report evidence of adverse effects in particu-
lar studies that only occurred in animals that had re-
ceived high doses of the test material [66]. The panel
chose to interpret those studies as providing quantitative
indicators of lower levels of exposure at which such ad-
verse effects were not apparent; it referred to them in
toxicologically-conventional terms as indicating no ob-
served adverse effect levels(or NOAELs), which are re-
ported in terms of the dose measured in milligrams per
kilogram of the body weight of the recipients per day (or
mg/kg bw/day), The panel followed orthodox official
practices by interpreting such levels as if they were
thresholds of exposure below which adverse effects did
not occur, in that variety of that species, or (at any rate)
a level at which adverse effects were not observed.
Food additive policies typically invoke what are called
ADIs (or acceptable daily intakes), where ADIs are de-
fined as the lowest observed NOAEL, in the most sensi-
tive laboratory species, divided by a safety factor(or SF)
. The most commonly quoted safety factor is 100, osten-
sibly on the assumption that a factor of 10 accommo-
dates the difference between average humans and
average laboratory animals, and another factor of 10 ac-
commodates the variations amongst humans [67]. The
prevailing ADI for aspartame in the EU is 40 mg/kg bw/
day, as the ANS panel applied a safety factor of 100. We
therefore took particular note of which studies provided
evidence of adverse effects below or above the desig-
nated NOAEL of 4000 mg/kg bw/day. It is important to
appreciate, however, that most professional toxicologists
Table 1 analytical categories used in this report
Category
1 rP Study result(s) deemed reliable by the panel as indicating adverse effects on humans, ie reliably positive
2 uP Study result(s) deemed unreliable by the panel as indicating adverse effects on humans, ie unreliable positive
3 rN Study result(s) deemed reliable by the panel as indicating no adverse effects on humans, ie reliable negative
4 uN Study result(s) deemed unreliable by the panel as indicating no adverse effects on humans ie unreliable negative
5 Cont Contradictory (when comparing the wording in the main report with text in the Appendices of the ANS panel's report)
6 ELlow Study indicating NOAEL at/or below 4000 mg/kg bw/day
7 E Lhigh Study indicating adverse effects, but only at doses above 4000 mg/kg bw/day
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 9 of 22
think that it is inappropriate to set an ADI for com-
pounds deemed to be carcinogenic by reference to an
estimated NOAEL from an animals study. This is be-
cause one mechanism by which compounds can act as
carcinogens is by damaging the DNA in the nuclei of
cells, that is by being genotoxic, and for genotoxic car-
cinogens one molecule could be sufficient to initiate or
promote tumours, and therefore no level of exposure
should be deemed acceptably safe. Given that three ro-
dent studies conducted by the Ramazzini institute had
provided evidence indicating that aspartame is a rodent
carcinogen, the decision of the ANS panel to allocate an
NOAEL to aspartame and then to ascribe an ADI to it
was problematic.
We gave serious consideration to the possibility of
categorising studies only showing evidence of harm at
doses above 4000 mg/kg bw/day as providing positive
evidence of harm, if only at the higher doses, but for the
purposes of this analysis we chose to categorise such
studies in line with the ANS panels portrayal of them,
as not showing adverse effects at levels below the
NOAEL. For completeness, our analysis nonetheless in-
cludes an estimate of the number of studies in which ad-
verse effects were only evident at levels of exposure
above 4000 mg/kg bw/day.
In the following discussion, individual studies and/or
papers are referred to either by reference to the family
name of the first author, along with the year of publica-
tion, or in terms of the number assigned to the study by
G D Searle in its submission to the US FDA in the
1970s. All of the early Searle studies were assigned num-
bers prefaced by the letter E, and that is how the ANS
panel referred to them. This paper follows that practice.
To differentiate studies in terms of whether or not
they indicated possible harm, we relied in the first place
on the ANS panels account. For example, when the
panel said of studies E97 and E101 that: Aspartame was
tested for mutagenicity in [5]Salmonella typhimurium
strains Aspartame was not mutagenic in this test
system, either in the absence or in the presence of the
metabolic activation system(p 59) it was straightfor-
ward to categorise those studies as not indicating harm.
On the other hand, when the panel said of Halldorsen et
al. 2010 (A large prospective cohort study based on
data from the Danish National Birth Cohort [that] inves-
tigated associations between consumption of artificially
sweetened and sugar- sweetened soft drinks during preg-
nancy and subsequent pre-term delivery(p 86)) Statis-
tically significant trends were found in the risk of pre-
term delivery with increasing consumption of artificially
sweetened drinks (both carbonated and non-
carbonated), but not for sugar-sweetened drinks. In the
highest exposure groups (4 servings/day) the odds ra-
tios relative to non-consumption were 1.78 and 1.29
respectively for carbonated and non-carbonated artifi-
cially sweetened drinks.(p 87) it was straightforward to
categorise that study as providing some indication of
possible harm from artificially sweetened beverages, a
market that aspartame currently dominates.
Categorisation was not, however, always straightfor-
ward because in some cases the panels text was unclear,
evasive or even self-contradictory. In all of those cases
we checked the text of the original publication, to iden-
tify any toxicological effects that were reported by the
researchers. In a few cases we made our own interpret-
ative judgements about whether or not the findings re-
ported by the author(s) should be characterised as
adverse. We did not check all the original documents
because, for the purposes of this analysis, whenever the
panel reported prima facie evidence of adverse effects
that was sufficient reason to categorise the studies
accordingly.
To differentiate studies, in terms of whether or not the
panel deemed them reliable or unreliable, we drew in
the first place on the panels text. For example, when the
panel said of study E81 that it: “… noted some discrep-
ancies in description of doses and that the test system
employed has not received further validation and is
Table 2 ANS panels interpretation of the reliability of studies for those that had, and had not, indicated possible harm, by number
of studies
Number of studies
reviewed
Number treated as
reliable
Number treated as
unreliable
Number of studies on which the panel was
self-contradictory
Studies not indicating
possible harm
81 62 19 2
Studies indicating possible
harm
73 0 73 8
Table 3 Numbers in Categories 6 and 7
Number of studies
ELlow = Studies implying a NOAEL less than 4000 mg/kg bw 16
ELhigh = Studies indicating adverse effects but only at doses greater than 4000 mg/kg bw/day 5
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 10 of 22
presently considered obsolete and therefore, the results
of the study were not included in the assessment…” it
was straightforward to categorise the panels judgement;
the panel deemed it unreliable.
Similarly, when the panel commented of E43 that:
The methods implemented were thought to be suffi-
ciently robust to support the results reported(p 209) it
was straightforward to categorise the study as one that
the panel had deemed reliable. On other occasions, how-
ever, the panel failed to provide any explicit indication as
to whether or not particular studies were deemed reli-
able. In such cases (eg E55, 1973, 3.2.5.1.2 page 75) cate-
gorisations were based on an exercise of informed
judgement. For example, where the panel suggested in
passing that some symptoms emerged that might pos-
sibly have been related to exposure to the test com-
pound, but subsequently never again mentioned that
evidence, particularly not in the summary and conclu-
sions of the relevant section, we judged that the puta-
tively positive evidence had been discounted, and
deemed to be unreliable.
The seven categories into which we differentiated the
studies, and the panels interpretations of their results,
are set out in Table 1.
The reasons for selecting the first four of those cat-
egories should be self-evident; the others deserve brief
explanations. Categories 5, 6 and 7 were not ones that
we had expected would be required, but the need for
them emerged as the panels account of each individual
study was examined. Category 7, Contwas necessary
because in several cases the main texts in Section 3.2
were subsequently contradicted by comments included
in one of the appendices.
Our quantitative analysis, set out in Table 2, did not
include every single study or paper referred to in Section
3.2 of the panels December 2013 report. For example, to
minimise avoidable double counting we omitted Iwata
2006 because it was not a separate study, but a re-
evaluation of the previous Ishii et al. 1981 study. It is
also important to note that our categorisations were not
always exclusive. For example, we assigned E81 and E44
to both uN and Cont, because the text in the appendix
contradicted the wording of the main text. In adopting
that categorisation, the wording in Section 3.2 is treated
as representing the panelsdefinitive judgement while
the comment in the appendix is treated as a secondary
qualifier. E11, E9, E39, E54, E55, E56, E90 and McAnulty
et al. 1989 belong in both uP and ELlow. E47, E63, E51,
E52 and E79 belong in uP, Cont and ELlow. E48 belongs
in uP and Cont. Reynolds et al. 1976 belongs in uP and
uN. Bunin et al. 2005 belongs in uP and ELhigh. Conse-
quently the aggregate count of the number of exemplifi-
cations of the 7 categories in Table 2is greater than the
total number of individual studies.
Results
Using those criteria, in relation to Section 3.2 of the De-
cember 2013 report of the ANS panel, we generated the
descriptive statistics shown in Tables 2and 3. Table 2
specifies the numbers of studies that provided either no
evidence that aspartame is harmful (in the first row, for
which we identified 81 studies) or evidence of possible
harm (in the second row, for which we identified 73).
Those totals are partitioned into two further columns,
namely those that the Panel deemed to be reliable and
those it deemed unreliable; thereby providing figures for
Categories 15.
The figures show that, while the ANS panel deemed
~ 84% of negativestudies to be reliable, it deemed
everysingleoneoftheputativepositivestudies as
unreliable, ie the panel discounted 100% of the evi-
dence of harm. Those figures constitute prima facie
evidence of asymmetry. On the basis of its interpret-
ation of the studies, the panel concluded that the
ADI for aspartame of 40 mg/kg bw need not be chan-
ged. A tabulation of the individual studies from which
those figures were derived isavailableasAdditional
file 1.Table3provides corresponding numbers for
categories 6 and 7.
While those figures may be starkly revealing, they
do not provide a comprehensive picture of the prob-
lematic features of the panels treatment of the puta-
tive toxicity of aspartame. Other considerations
emerge from a careful scrutiny of the panelstext,
and it is to those problems that this discussion now
turns. The discussion contextualises them by refer-
ence to guidance provided to all its risk assessment
panels by the Board of EFSA.
EFSAs 2009 benchmarks of performance for all scientific
risk assessment panels
The Commissions request of EFSA to review all the evi-
dence relating to the safety or toxicity of aspartame was
issued in May 2011, and consequently the panel deliber-
ations and conclusions ought to have conformed to
some procedural guidance that had been issued by
EFSAs Scientific Committee in 2009, which referred to:
Transparency in the Scientific Aspects of Risk Assess-
ments. That guidance stipulated that:
"scientific outputs must be transparent with regard
to the data, methods of analysis and assumptions that
are used in the risk assessment process
Transparency is needed in all parts of the risk
assessment
The sources of all data used for the assessment,
including unpublished data and personal
communications, must be referenced and
evaluated to determine their quality and relevance to
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 11 of 22
the assessment.These should be reflected in the
relative weight given to them in the assessment and
taken into account in the overall evaluation of
uncertainty
The inclusion/exclusion criteria applied to the
data should be explained and described within the
risk assessment. If data are excluded, this should
be stated along with the rationale for their
exclusion."
All assumptions should be documented and
explained. Where alternative assumptions could
reasonably be made, the related uncertainties can be
evaluated together with other uncertainties "[68].
(Emphases added)
In the panels December 2013 treatment of aspartame,
all of those guidelines were breached. The following dis-
cussion substantiates those assertions, and compares the
criteria of interpretation applied by the ANS panel to
particular studies.
Inclusions, exclusions and asymmetries
Although the panel did provide an account of the cri-
teria of inclusion and exclusion that it purported to have
used, the meaning of that text was vague [69], leaving
the panel with considerable discretion over what to treat
as evidence and what to exclude and those discretion-
ary exercises remained both unacknowledged and asym-
metric. Evidence was often selectively excluded without
due acknowledgements or explanations, thereby failing
to comply with the 2009 guidance to EFSAs scientific
committees on Transparency.
One important issue concerns which reports and doc-
uments were deemed suitable for inclusion in the panels
review and which were omitted. The discussion above
highlighted the fact that detailed documentary evidence
is in the public domain indicating that at least 15 of the
early studies on aspartame, and of its breakdown prod-
uct DKP, conducted in the laboratories of Searle, and its
sub-contractor Hazleton Laboratories, were incompe-
tently conducted and misleadingly reported, thereby
concealing the incompetence. Several commentators
have long argued that the defects in those studies and
their reports were so serious that they should never have
been accepted, and needed to be repeated and con-
ducted properly [70].
The dossier that Millstone provided to the ANS
Unit in October 2011 included the relevant documen-
tary evidence in relation to the account provide in
Section 1 above. An anomalous and problematic fea-
ture of the panels December 2013 report is that,
while it cited passages from documents suggesting
that any shortcoming in those studies were inconse-
quential, it failed to acknowledge, let alone engage
with, the contrary evidence, such as the reasons pro-
vided by Verrett and Gross to the 1987 Congressional
hearing, explaining why reassuring narratives concern-
ing those 15 studies were profoundly flawed. The
panel mentioned the report of the UAREP, referring
to it “… the authentication review of selected mate-
rials submitted to the Food and Drug Administration
…” [71],butitfailedtociteorengagewithGrosss
explanation of why the UAREP document was not a
genuine authenticationor with Verretts contention
that none of those 15 studies could be deemed reli-
able. Nonetheless, all 15 of those studies were in-
cluded in the ANSs analysis, and treated as if they
were unproblematically reliable, although as Verrett
had explained, they were woefully inadequateand
worthless.Thepanels failure to cite or discuss the
evidence from eg Gross and Verrett entails that the
panels assessment was neither thorough nor reliable.
In 1996 Olney et al. published a paper highlighting the
increase in the incidence of an aggressive type of brain
tumour in US adults following the introduction of aspar-
tame [72]. There are several reasons why that report de-
serves to be deemed credible. Firstly, the type of central
nervous system tumour found to be increasing most
rapidly in the USA is the same kind of lesion as was
found in one of the animal studies conducted on aspar-
tame in the 1970s [73]. Secondly, Olney et al. drew at-
tention to a study by Shephard et al., in which they
simulated in vitro the conditions that can occur in the
human digestive tract, especially the conditions that re-
sult in the nitrosation of dietary constituents. They re-
ported that nitrosated aspartame had significant
mutagenic action [74]. That is important because it sug-
gests not only a mechanism by which aspartame could
act carcinogenically, but also why the interval between
the compounds introduction and the elevation of US
brain cancer rates was so brief. The paper by Olney et
al. was omitted from the panels discussion of human
epidemiology (Section 3.2.7.1.3) though it was men-
tioned in Section 2.7.1, where the panel cited and en-
dorsed a 2002 report from the French food agency
(AFSSA) that said: “… Olney et al. (1996) did not pro-
vide any scientific evidence of a relationship between ex-
posure to aspartame and the appearance of brain
tumours.[75] But that remark was misleading because
while it did not provide proof of such a relationship, it
certainly provided evidence; which illustrates the panels
practice of discounting any evidence of possible harm
falling short of causal proof.
Puzzling anomalies inconsistent and unacknowledged
assumptions
Table 3above indicates that in relation to 10 studies the
panel contradicted itself, when comparing the main body
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 12 of 22
of the text with the appendices. The following para-
graphs highlight several examples.
Study E44 [76] was a so-called host-mediated mutage-
nicity testin which rats were injected with a micro-
organism that could reveal induced gene mutations, and
then tested to see if aspartame is a mutagen. In relation
to that study, which in our analysis has been categorised
as not indicating harm, the panel criticised the test be-
cause: “… the test system employed has not received fur-
ther validation and it is presently considered obsolete,
and therefore the results of the study were not included
in the assessment.Yet in Appendix H on page 210, the
text states that: The methods implemented were
thought to be sufficiently robust to support the results
reported.A similar problem arises in relation to E81 for
which the text in Sec 3.2.3.1 on page 59 suggests that
the results of study were not included in the assessment,
while Appendix H page 210 also indicates that the
methods were thought sufficiently robust. In those two
cases, the panels comments were contradictory.
In relation to E54, which was a developmental toxicity
study on rabbits, the panel reported: “… a significant de-
crease in fetal body weight and skeletal anomalies
[yet] the Panel concluded from these observations
that the developmental effects on body weight and skel-
etal development reported in the aspartame feeding
studies may be caused by the significant depression of
feed consumption in the high dose group.[77] They
chose to discount the indication of adverse effects in the
high dosegroup even though: The actual aspartame
doses were reported to be 1880 and 1870 mg/kg bw/day
for the intended 2000 and 4000 mg/kg bw/day groups,
respectively.But as the intakes of the two groups were
so similar, the panels attribution of decreased fetal
body-weight to a difference in consumption was flawed.
In our analysis this study was categorised as indicating
possible harm, though it had in effect been deemed un-
reliable by the panel.
In relation to E55, which was another developmen-
tal toxicity study on rabbits, the panel failed to com-
ment on its findings or reliability. In Appendix I page
229,however,welearnthat:Feed consumption de-
creased by up to 29% in the high dose group[78].
That change may well have been a consequence of a
reduction in the palatability of the diet to the rabbits
in the high dose group. Nonetheless the panel prob-
lematically portrayed the highest dose of 4000 mg/kg/
bw as a NOAEL. Our analysis categorised this study
as not indicating possible harm, as it did so only at a
high dose, but the panel failed to provide an explan-
ation of why the reduction in food consumption was
not counted as an adverse effect and therefore why
the NOAEL in this study was not specified at a dose
lower than 4000 mg/kg/bw.
In relation to E62, which was another developmental
toxicity study on rabbits, the panel “… noted that the ac-
tual dose to which the rabbits were exposed did not ex-
ceed 1880 mg/kg bw/day (range 1160-1880 mg/kg bw/
day) when the intended dose was 4000 mg/kg bw/day
(administered as 4.8 % (w/w [weight for weight]) aspar-
tame in the diet). This was a result of the considerable
decrease in feed intake observed in the pregnant rabbits
in this dose group. as such the Panel noted that
the actual aspartame doses in the low and high aspar-
tame dose groups were often comparable[77]. In Ap-
pendix I, page 229, however the text reports that at the
high dose, 3 fetuses exhibited minor malformations, yet
the panel deemed 2000 mg/kg bw/day as a NOAEL, but
if the doses were the same in the highand lowdose
groups, then 2000 mg/kg bw/day was a level at which
adverse effects were observed, and so portraying 2000
mg/kg bw/day as a NOAEL was a misrepresentation. In
our analysis this study was categorised as indicating pos-
sible harm, although it was deemed by the panel to be
unreliable.
In relation to E90, a developmental toxicity study in
rabbits, the panel explained that in this study: A num-
ber of animals died spontaneously during the study
mainly due to misdosing No abortions were detected
in the control, mid dose and L-aspartic acid groups, two
abortions in the low dose aspartame group and 24 abor-
tions were observed in the high dose aspartame group (a
significant increase compared to controls) and four in
the L-phenylalanine groupSince the decrease in body
weight started several days before (13 and 18 days) the
abortions (28 days), the authors concluded that abortion
was a consequence of significant and rapid body weight
loss caused by decreased feed consumption[79]. Firstly,
the panel accepted the authorsreasons for attributing
the abortions to maternal body-weight loss, even though
both could equally reasonably have been interpreted as
adverse toxicological effects. Secondly, Appendix I (page
231) portrayed the low-dose level of 1000 mg/kg bw/day
as a NOAEL, but it was problematic to portray two
abortions as no adverse effects, particularly without ref-
erence to any estimate of possible statistical significance.
Consequently in our analysis this study was categorised
as indicating possible harm, though deemed unreliable
by the panel.
In relation to Brunner et al., which was a reproductive
study on rats, the panel said: Increased offspring mor-
tality was observed in rats fed the highest aspartame
dose and in the phenylalanine-exposed animals[80]. So
while the death of infant rats deserved to be deemed as
adverse, the panel claimed that it: “… agreed with Brun-
ner et al …” implying that the adverse effects should be
discounted as only occurring at highest doses. The prob-
lem is that that is not what the original paper asserted.
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 13 of 22
It said: “… the addition of aspartame (6% by weight)
to the standard diet of rats has a profound effect on
early postnatal development. That the majority of ad-
verse effects of Asp 6%are transmitted during the
period of lactation was clearly shown in a cross-fostering
experiment Conversely as 6% administered during
lactation, but not earlier, resulted in high mortality and
somewhat retarded development similar to that observed
in the offspring of dams receiving Asp 6% continuously
from prior to breeding .[81]. Moreover as Magnuson
et al. observed: “… this difference was statistically signifi-
cant.[82]. Consequently, in our analysis this study was
categorised as indicating possible harm, though it was
deemed unreliable by the panel.
Low-hurdle for the acceptability of negative studies
In section 3.2.3.1 on genotoxicity, Searle studies E97 and
E101 using salmonella bacteria, were reported by the panel
as having used methods that failed to comply with the con-
ventional requirements for good laboratory practice, yet
the December 2013 report of the ANS panel did not dis-
count the findings from them. We, by contrast, would
count them as unreliable for precisely those reasons.
In relation to E4, a sub-chronic rat feeding study, with
only 5 animals per dose group, lasting only 9weeks, no
adverse effects were reported by the panel except for “…
reduced body weight gain …” [83]. The panel reported:
No significant treatment-related changes in haematology,
clinical chemistry, urinalysis were recorded and no
treatment-related changes in relative organ weight or
pathology were observed. Reduced body weight gain was
noted for both treatments and both sexes (though less
marked in females) …” [83]. (emphasis added) The panels
use of the word significant, when combined with its fail-
ure to comment on the very small size (ie 5) of the test
groups, reveals a willingness to accept a very weak nega-
tive study as if reliable, despite its obvious weaknesses.
Other examples of where the panel treated negative stud-
ies as if they were reliable, despite the small size of the ex-
perimental samples, are provided for example by E86 (Sec
3.2.4.1 p 69, which examined brain tissue cells) which was
treated as a reliable negative [84], despite having only 5
dogs per dose group, and Sasaki et al. 2002, which was
treated as a reliable negative despite using only 4 mice per
group, and which received only a single dose [85].
In relation to E51, the panel suggests that no embryo-
toxic effects were observed in a reproductive study on rab-
bits. The text also explained that: The study was
confounded by poor health of the animals and the gavage
technique issues. As a result, maternal mortality was high.
The administration of aspartame was associated with de-
pression of feed consumption by up to 40%.[86]. In Ap-
pendix I, however, the text reveals that the number of
abortions and maternal deaths in the high dose group
were double those in the control group [87]. That infor-
mation contradicts the claim that no embryotoxic effects
were observed. Furthermore the dose at which such ef-
fects emerged was 2000 mg/kg bw/day, so the suggestion
that that figure could be deemed a NOAEL is problematic.
Yet again negative indications were treated as reliable,
while positive findings were discounted as unreliable.
In relation to E43, which was a study in rats investigat-
ing chromosome aberrations in bone marrow cells, the
panel complained that no mitotic indexhad been deter-
mined, and therefore deemed the study limited[88]. A
mitotic index is an estimate or measure of the number of
cells per unit (usually 1000 cells) undergoing cell division
in a specific time period; it provides an estimate of the rate
of tissue growth. In Appendix H the panel commented
that: The mitotic index should be determined as a meas-
ure of cytotoxicity in at least 1000 cells per animal for all
treated animals (including positive controls) and untreated
negative control animals. Overall judgement:The
methods implemented were thought to be sufficiently ro-
bust to support the results reported[89]. (emphasis in
the original) In other words, even though the mitotic
index had not been estimated, the study had been deemed
sufficiently robust. By contrast, Mukhopadhyay et als
findings, in a study of chromosomal aberrations in mice,
using a blend of aspartame and acesulfame K, and which
indicated adverse effects, was discounted because no mi-
totic index had been determined and because a mixture of
sweeteners was used [90]. But if the lack of a mitotic index
did not count against E43 why did it count in this case?
The panels failure to acknowledge that blends of aspar-
tame and acesulfame K are often used in food and bever-
age products (eg Coke Zero) was also conspicuously
absent [91]. The panel therefore invoked a mitotic index
criterion inconsistently. Representatives of the sweetener
industry have also claimed that while a blend of aspartame
and acesulfame K is significantly sweetener than the sum
of their separate sweetness, ie they act synergistically on
our taste buds, they are entirely confident that toxicologic-
ally they cannot act synergistically [92].
When responding to seemingly positive evidence of
toxicity the ANS panel never set a requirement, or is-
sued requests, for follow-up studies, or further informa-
tion. It used the lack of follow-up confirmation as
grounds for discounting evidence of possible harm, but
not as grounds for requiring further tests; a precaution
that had at least sometimes been taken previously by the
UK, European and WHO advisory panels.
(Unreachably) high hurdles for positivestudies
indicating adverse effects assumptions not
acknowledged but revealed
Official toxicologists routinely use so-called no observed
adverse effects levels(or NOAELs) in animal studies as a
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 14 of 22
basis for setting an acceptable daily intake(or ADI) for
regulated compounds, simply by applying a safety factor
which is often 100. If the protection of public health is the
primary goal of policy, the selected NOAEL should be no
higher than the dose below which no adverse effects were
observed in the most sensitive parameters of the most sen-
sitive variety of the most sensitive laboratory species tested.
For reasons that have not been explained, the ANS
panel discounted several pieces of evidence indicating
that the prevailing NOAEL for aspartame of 4000 mg/kg
bw/day was too high. Section 3.2 included at least 12 ex-
amples of the panel noting, but then discounting, evi-
dence of adverse effects, and therefore NOAELs, below
4000 mg/kg bw/day. For example, E39 [93], which stud-
ied the effect of aspartame on peri and postnatal devel-
opment in female rats, was reported as indicating a
NOAEL of 2000 mg/kg bw (as did E9, E47, E48, E53,
E54 and E62) while the lowest reported figure was 750
mg/kg bw in study E79 [94]. For E62 the panel even gen-
erously accepted 2000 mg/kg bw as a NOAEL, despite
reporting minor malformationsin infant rabbits at that
dose [78]. Those examples illustrate the forgiving gener-
osity the panel showed towards studies and dose levels
at which ostensibly adverse effects were observed and
reported. Those examples thereby reveal the demanding
heights of the hurdles that studies providing evidence of
possible harm would have had to have reached before
the panel might have deemed them relevant and reliable.
Another example of the panel discounting evidence of
possible harm by implicitly setting a very high hurdle is
provided by two studies of the consequences of nitrosat-
ing aspartame, namely Meier et al. 1990 and Shephard
et al. 1993. The panel discounted the evidence from
those studies by suggesting that the conditions for nitro-
sation were harsh, without acknowledging that nitrosa-
tion is precisely the kind of process that happens during
human digestion, and that conditions in the human
stomach, especially after consuming highly acidic cola
beverages, are similarly harsh[95].
In relation to Searle study E20, a 2-month oral admin-
istration study in 10 male and 10 female rats, the panel
discounted “… a significantly higher liver to body weight
ratio for the high dose group males compared to con-
trols[96]. But the high doselevel was only 125 mg/kg
bw/day [97]. Changes were discounted as being neither
systemic nor pharmacological, which was implicitly in-
voking a very high hurdle.
In relation to Searle study E87, which examined rat
brain tissue from E3334 and E70 looking for intracra-
nial neoplasms, the panel reported that tumours were
found, which were random with respect to dose and
gender; the panel discounted the tumours as unrelated
to aspartame [98]. What the panel failed to acknowledge
or explain was why it (implicitly) assumed that cancers
in rats only count if they affect males and females to a
similar extent, and only if the dose-response relationship
is monotonic. Both those assumptions implicitly
favoured commercial over public health interests, and
are scientifically contested by reference to both empirical
evidence and theoretical understandings. The panels
treatment of E87 therefore failed to conform to the in-
junction to EFSA panels that: All assumptions should
be documented and explained …” [99]. Moreover it re-
veals the asymmetry of the height of the implicit hurdles,
and therefore a failure to be even-handed.
In relation to Searle study E20, an 8-week rat feeding
study using 50 animals of which 10 served as controls,
the panel reports that notwithstanding “… organ weights
were unaltered except in the high dose males where a
significantly higher liver to body weight ratio was ob-
served for the high dose group males compared to con-
trols[100]. In Appendix G the text indicated: “…
increased terminal blood sugar at 125 mg/kg bw/day but
no other significant changes in clinical chemistry …” as
well as “… no unequivocal effect on bodyweight or food
consumption.[101]. In other words, the liver-to-body-
weight ratio was significantly disturbed in the high dose
males, which in such a small group (5 animals) suggests
a systemic problem, which nonetheless the panel dis-
counted. Characterising and discounting effects on body
weight and food consumption as not unequivocal, and
implicitly demanding unequivocal findings from a study
with a group size of 5, provides further examples of the
panel raising the bar for positive evidence, despite keep-
ing it very low for reassuring evidence and studies.
When the panel discussed Ishii et als 1981 2-year
chronic toxicity and carcinogenicity rat study [102]. it
reported dose related changesincluding focal mineral-
isation of the renal pelvis, but those findings were dis-
counted as of “… minimal toxicological significance …” .
But several other small studies that failed to find tu-
mours were portrayed as if toxicologically significant
and reliable, revealing a further example of asymmetric
benchmarks of appraisal and criteria of interpretation.
Indeed the panel: “… noted that the study provided in-
formation on the lack of toxicity of aspartame when ad-
ministered in conjunction with DKP[102]. This is an
example of the panels willingness to portray the absence
of evidence as if it constituted evidence of absence,
which is always an invalid inference. The panel also
failed to acknowledge any of its underlying assumptions,
thereby failing to follow the procedural stipulations that
the EFSA Board had specified.
In relation to Halldorsson et al. 2010, a prospective co-
hort study on the association between intakes of artifi-
cially sweetened soft drinks and pre-term delivery, the
panel reported that Statistically significant trends were
found in the risk of pre-term delivery with increasing
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 15 of 22
consumption of artificially sweetened drinks (both car-
bonated and non-carbonated), but not for sugar-
sweetened drinks.[103]. Nonetheless, the panel
declined to treat those findings as reliable, noting that:
“… the prospective design and large size of the study
sample were major strengths, and there were no import-
ant flaws in the methods used. However, risk estimates
may have been inflated by residual confounding (includ-
ing by year of delivery). No account was taken of other
dietary sources of methanol, and use of aspartame spe-
cifically was not distinguished from that of other artifi-
cial sweeteners). Therefore, given these limitations, the
Panel agreed with the authors who concluded that repli-
cation of their findings in another setting was war-
ranted.[104]. The panels remark that the results: “…
may have been inflated by residual confounding …”
[104] was shallow and opportunistic. The results might
have been affected by unidentified confounding factors,
but they may not have been, and the fact that residual
confounding was a logical possibility cannot constitute
sufficient grounds for disregarding the indications of
harm, especially as the panel never raised the possibility
of over-adjusting for confounding variables in the nu-
merous negative studies that were treated as if reliable.
The next study the panel discussed was Englund-Ögge
et al. 2012, which was portrayed as if it had been an at-
tempt to replicate the study by Halldorson et al. [105].
The panel said that: No significant trends were found in
the risk of pre-term delivery with increasing consump-
tion either of artificially sweetened drinks or of sugar-
sweetened drinks. [but] Small elevations of risk were ob-
served with higher consumption of artificially sweetened
soft drinks, but after adjustment for covariates, these
reached statistical significance only when categories of
consumption were aggregated to four levels, and then
the odds ratio for the highest category (1 serving/day)
was only 1.11 (95 % CI 1.00-1.24) in comparison with
non-consumption.[106]. In other words after adjust-
ment for covariatesthe Endlund-Ögge et al. study re-
ported similar findings, but only more weakly than
Halldorson et al.
In relation to those two studies the panel remarked
that: Both Halldorsson et al and Englund-Ögge et al
appear to have been well designed and conducted.
Noting this, the Panel concluded that even at high levels
of exposure to artificially sweetened soft drinks the risk
of pre-term delivery is likely to be small, if any. The ob-
served associations could be a consequence of uncon-
trolled residual confounding, and the inconsistencies in
the patterns of association reinforce this uncertainty.
[106]. The panel interpreted the results of Englund-
Ögge et al. as if they refuted those of Halldorsson et
al., when it would have been no less reasonable to in-
terpret Englund-Ögge as supporting Halldorson,
though with a weaker signal, which constitutes an-
other example the panels underlying unacknowledged
asymmetric assumptions.
In relation to Camfield et al. 1992, which was a
double-blind study with eight girls and two boys to as-
certain aspartames relevance to their neurological sei-
zures, the panel: “… noted that the combination of the
two parameters (number and length of spike-wave
bursts) into a single measure was not adequately ex-
plained, and lack of control of food and drink intake be-
fore and after dosing may have affected the results. The
Panel further noted that aspartame was given in a single
dose at the ADI.[107]. But since both the frequency
and severity of seizures are toxicologically significant ad-
verse effects, integrating them into a single indicator can
be revealing rather than misleading. It is also noteworthy
that the panel never questioned the choice of indicators
used in negative studies. Once again the panel picked on
tiny imperfections in a putative positive, while in relation
to negative studies appearing indifferent, not just to
similarly tiny imperfections, but to substantially more
serious ones.
When discussing Kulczyckis 1986 report of a case of
aspartame-induced urticaria, the panel said that it dis-
counted all self-reported allergic-like reactions on the
grounds that the European Commissions Scientific
Committee for Food had previously discounted them in
2002 as they had not been confirmed double blind [108].
Nonetheless, lower down the same page the panel ac-
knowledged that: Kulczycki reported a case of aspar-
tame induced urticaria confirmed by double blind
challenge …” [109]. (Emphasis added) One page later the
panel said: However, the Panel cannot exclude the pos-
sibility that in rare instances individuals could be suscep-
tible to allergic reactions following aspartame ingestion
[110]. Yet the panel treated that evidence and the possi-
bility as if it had no implications whatsoever for the ac-
ceptable daily intake of aspartame; which leaves
unanswered the question acceptable to whom:to
NutraSweet, to average healthy adults, or to all likely
consumers? Unfortunately no answer is provided to that
implied question; instead ADIs are treated as if they
were natural constants rather than value-based social
judgements about how much risk should be tolerated,
and by whom.
In relation to Veien et al. 2012, which was entitled
Systemic allergic dermatitis presumably caused by for-
maldehyde derived from aspartame, the panel acknowl-
edged that a: “… few cases of presumed systemic allergic
dermatitis in patients with contact sensitivity to formal-
dehyde, apparently caused by the intake of aspartame in
artificial sweeteners, have been described. The four pa-
tients described in the literature all had eyelid derma-
titis[111]. Those findings were then discounted by the
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 16 of 22
panel on the grounds that: “… the studies available were
performed on a limited number of participants.A study
with a small sample size is unavoidably insensitive and
yet effects were still noted and should not have been dis-
counted because the sample size was small.
In relation to Roberts 2001, the panel mentioned: “…
case reports published in peer reviewed journals and
reports compiled by Dr H.J. Roberts and published
under the title Aspartame Disease An ignored Epi-
demic The total number of symptoms reported from
all sources was 4281, as most cases reported more than
one symptom. Headache was the most frequently re-
ported adverse effect (28.5 %), followed by dizziness and
giddiness (19.2 %). Although the results of a
questionnaire-based study (Lipton et al., 1989) and two
double-blind out-patient investigations (Koehler and
Glaros, 1988; Van den Eeden et al., 1994) employing
daily doses of up to 30 mg/kg bw/day indicated a poten-
tial association between aspartame intakes and headache,
it is still not possible to deduce causality, as the effect of
diet has not been adequately controlled for and the in-
terpretation of the data was complicated by a high drop-
out rate and a limited experimental design. The Panel
noted that the number of cases is low when compared
with the widespread use and that the effects were mild
to moderate[112]. (emphases added) Using the phrase
it is still not possible to deduce causalityreveals that
when faced with apparently positive evidence of harm
from two double-blind investigations the panel invoked
the astonishingly demanding hurdle of deductive proof
of a causal relationship; a criterion that was radically
more demanding than any or all of those utilized to in-
terpret negative studies, for which mere plausibility was
often treated as sufficient.
Selecting 4000 mg/kg bw/day as a NOAEL for aspartame
Our comprehensive tabulation (see Additional file 2
and summarised below in Additional file 1, which in
turn is summarised in Tables 2and 3above) identifies
16 studies that showed adverse effects at dose levels of
less than, or equal to, 4000 mg/kg bw/day. This is im-
portant because, supposedly, the dose level at which a
NOAEL should be specified ought to be the highest level
producing no adverse effects in the most sensitive variety
of the most sensitive species. Since the set of studies the
panel reviewed in Section 3.2 included 12 that did show
evidence of adverse effects at or below 4000, it is clear that
the panels designation of that dose level as a NOAEL was
flawed and unjustifiable. The studies that provided evi-
dence of adverse effects at or below 4000mg/kg bw/day
are: E20, Ishii et al. 1981, E11, E9, E39, E47, E53, E54, E55,
E63, E51, E52, E79, E90, McAnulty et al. 1989 and NTP-
CERHR Report 2003.
Summary of results of qualitative analysis
If all of the implicit benchmarks, which the ANS panel
invoked, as grounds for discounting putative positive
evidence of adverse effects from aspartame, are aggre-
gated together they collectively imply that, for the ANS
panel, nothing could count as a reliable positive unless:
1) the magnitude of the evidential difference between
test and control groups satisfied the customary
criterion of statistical significance, namely that there
was less than one chance in 20 of it having been a
random fluctuation (or a p-value < 0.05) (cf E2
1972; E3 1972; E44 1972; E75 1974);
2) the results were entirely unequivocal (cf NTP 2005;
E9, 1972; E11, 1971; Collison et al. 2012b);
3) the findings were consistent eg consistently and
monotonically dose-related (cf E21);
4) were obtained from a long-term study conducted
with a large sample of people or large groups of la-
boratory animals (cf Abhilash et al. 2011); and.
5) entirely free of any imperfections (cf Kamath et al.
2010; Sasaki et al. 2002).
Those are some of the assumptions that would need to
be adopted in any attempt to reproduce the ANS panels
conclusions. Without those particular assumptions, the
panels conclusions could not have been reached.
On the other hand, studies were treated by the panel
as if they were reliable negatives despite numerous im-
perfections (eg E43) of the sort that the panel highlights
when discounting putative positives. For example: “…the
findings reported might be ascribed specifically to the
conditions of the study…” [113].
Discussion
This paper has provided numerous pieces of evidence
showing that the implicit criteria by reference to which
the ANS panel evaluated toxicological studies on aspar-
tame were not even-handed. On the contrary, they were
markedly asymmetric as between apparently positive and
negative studies. In quantitative terms, the panel treated
~ 76% of apparently negative studies as if they were reli-
able, despite their numerous shortcomings, yet it
deemed 100% of positive studies as unreliable, despite
the fact that their shortcomings were often fewer and
less serious than those that characterised the negative
studies. This asymmetry favoured commercial and in-
dustrial interests over the protection of consumers.
The qualitative analysis showed that, while very de-
manding criteria were used to judge the reliability of pu-
tative positive studies, far more lax and forgiving criteria
were applied to judge the reliability of apparently nega-
tive studies. If the benchmarks that the ANS panel used
to evaluate the results of negativestudies had been
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 17 of 22
consistently used to evaluate the results of positive
studies then the panel would have been obliged to con-
clude that there was sufficient evidence to indicate that
aspartame is not acceptably safe. Correspondingly, if the
benchmarks that the ANS panel used to evaluate the re-
sults of positive studies had been consistently used to
evaluate the results of negative studies then the panel
would have been obliged to conclude that there was in-
sufficient evidence to show that aspartame is acceptably
safe. In the event, it did the opposite.
The ANSs December 2013 treatment of evidence con-
cerning aspartame toxicity failed to conform to basic stan-
dards of scientific rigour and more specifically failed to
comply with EFSAsGuidance of the Scientific Committee
on Transparency in the Scientific Aspects of Risk Assess-
ments carried out by EFSA, in all of the following respects.
In the following list, key words and phrases used in the
EFSA Guidance are italicised.
The output of the ANS panels review of aspartame
was not transparent with regard to the data,
methods of analysis and assumptions that were used
in the risk assessment process.
The sources of all data used for the assessment,
including unpublished data and personal
communications, were not properly evaluated to
determine their quality and relevance. The relative
weight given to the available evidence in the overall
evaluation was not even-handed but asymmetric as
between positiveand negativestudies. The uncer-
tainties that were acknowledged by the panel related
far more frequently to positivestudies than to
negativeones.
The criteria of inclusion and exclusion of evidence
were applied inconsistently and were neither
properly described nor explained.
When studies and data were excluded from the risk
assessment the reasons were often not given, and
when some were stated, they were not applied
consistently.
Numerous assumptions were neither documented
nor explained.
Where alternative assumptions could have reasonably
been made, the panel failed to identify them, or to
appraise uncertainties in an even-handed manner.
The assumptions that the panel implicitly made were
applied in a consistently inconsistent manner, in ways
that favoured commercial interests over consumer
protection.
There is little point requiring animal tests and, as the
euphemism has it, sacrificingthe animals unless those
studies are conducted and interpreted so as to protect
public health.
Conclusion
This paper provides compelling quantitative and qualita-
tive evidence of asymmetric criteria of interpretation
and appraisal on the part of the ANS panel, but it does
not provide an explanation of why such an asymmetrical
approach was taken. A hypothesis that some members
of the ANS panel may have had commercial conflicts of
interest has been advanced [114], but as the panels
meetings all took place behind closed doors, such hy-
potheses are difficult to test. An alternative hypothesis
that regulatory institutions exhibit forms of institutional
inertia, meaning that they are singularly unwilling to
criticise their previous judgements or the judgements of
other official bodies including their predecessors, is
plausible but it will also remain hard to test until risk as-
sessment processes become substantially more
transparent.
On the issue of the transparency of EFSA risk assess-
ment the chief executive of EFSA, Bernhard Url, was re-
ported, by the commercial food industry news service
NutraIngredients.com, as having said in October 2015
that:
In a scientific process which is what we call
organised scepticism the scientist must have the
freedom to ask stupid questions that are out of the
box, to challenge their peers, trial and error. And
there Im not convinced we help the process of
finding the nearest approximation of truth by putting
every single question for the whole life of a scientist
on Youtube .I think that science needs parts of the
process in a closed room and then many steps of the
process in an open atmosphere.But I think it also
needs a protected room where they can speak
completely freely, openly and challenge each other.
Thats also science." [115].
Those remarks were difficult to reconcile with EFSAs
commitment to transparency and reproducibility. No-
one is proposing to put the entire lives of EFSAs scien-
tific advisors on YouTube, but in this context the issue
relates to the transparency of specific professional activ-
ities of expert advisors when providing judgements and
advice to EFSA. A good case could be made for the ar-
gument that anyone who is not willing to be fully ac-
countable for their judgements in respect of EFSAs food
safety risk assessments should be disqualified from serv-
ing on EFSA panels. Since March 2002, expert commit-
tees that advise the UKs Food Standards Agency have
been required to hold their meeting in publicly open ses-
sions, and there is no obvious reason why the same re-
quirement could not be applied by EFSA [116]. It would
contribute substantially to enhancing EFSAs credibility
with citizens of the EU.
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 18 of 22
Url argued (see above) that EFSAs scientific advisors
must be free to ask stupid questions. What counts as a
stupidquestion may often be an issue about which
people can legitimately disagree, but it is vitally import-
ant to ensure that members of EFSA panels are consist-
ently asking wise and thoughtful questions that
challenge the claims put before them. The discussion
above suggests that the ANS panel questions were not
always consistently wise or sceptical. Moreover the lack
of transparency risks the continuation of bad practices.
Given the shortcomings of EFSAs risk assessment of
aspartame, and the shortcomings of previous official
toxicological risk assessments of aspartame, it would be
premature to conclude that it is acceptably safe. They
also imply that the manner in which EFSA panels oper-
ate needs to be scrutinised and reformed.
EFSA, the European Commission, the European Par-
liament as well as EU consumers, have been poorly
served by the ANS panel. The ANS panel may have
wished to portray the December 2013 document as its
final report, but that may well have been over-
optimistic.
A fresh review of all the data from all the studies, as
well as all other relevant scientific and policy docu-
ments, should be conducted. None of the members of
that review panel should have any relevant conflicts of
interest. Its conduct should be fully transparent and
demonstrably compliant with all EFSA Guidance docu-
ments, and it should also be guided by bias-detection
tools such as those provided by SYRCLE [117]. and
CAMARADES [6], the selection and application of
which should be explained and justified.
Additional file
Additional file 1: List of studies allocated to categories 1 to 7. (PDF 462 kb)
Additional file 2: Comprehensive tabulation of all studied cited by EFSAs
ANS panel in Section 3.2 Toxicological data of aspartame. (PDF 655 kb)
Abbreviations
ADI: Acceptable daily intake; ANS Panel: EFSAs Panel on Food Additives and
Nutrient Sources added to Food; APM: Aspartame; Cont: Contradictory
(when comparing the wording in the main report with text in the
Appendices of the ANS report); CoT: Committee on Toxicology, of the UK;
DKP: Diketopiperazine, a breakdown product of aspartame; EFSA: European
Food Safety Authority; ELhigh: Study indicating NOAEL at/or below 4000 mg/
kg bw/day; ELlow: Study indicating adverse effects, but only at doses above
4000 mg/kg bw/day; EU: European Union; FAO: UN Food and Agriculture
Organisation; FDA: US Food & Drug Administration; JECFA: Joint Expert
Committee on Food Additives of the WHO and FAO; mgs/kg bw/day
: milligrams of aspartame per kilogramme of the body weight; NOAEL: No-
observed-adverse effect-level; PBOI: Public Board of Inquiry; rN: Study result(s)
deemed reliable by the panel as indicating no adverse effects on humans, ie
reliably negative; rP: Study result(s) deemed reliable by the panel as
indicating adverse effects on humans, ie reliably positive; SCF: Scientific
Committee on Food, of the European Commission; SF: Safety factor;
UAREP: US Universities Association for Research and Evaluation in Pathology;
UK: United Kingdom; uN: Study result(s) deemed unreliable by the panel as
indicating no adverse effects on humans, ie unreliable negative; uP: Study
result(s) deemed unreliable by the panel as indicating adverse effects on
humans, ie unreliable positive; w/w: weight for weight; WHO: World Health
Organisation
Acknowledgements
Not applicable.
Authorscontributions
ED prepared a first draft of the Tabulation in Additional file 2. EM prepared a
first draft of the text of this paper, drawing on the tabulation. Both authors
reviewed the text of the paper, and the detailed tabulation, which is
available as Additional File 2. Both authors read and approved the final
manuscript.
Funding
The only contribution that funding made to the preparation of this paper
drew from Professor Millstones research residue account that is held by the
University of Sussex, and it paid for some of Elisabeth Dawsons time.
Availability of data and materials
All relevant data and material will be available on publication. Additional file
2and Millstones dossier is available at http://www.sussex.ac.uk/spru/
research/projects/fcs.
Ethics approval and consent to participate
As a document-based study, ethical approval was not required, and issues of
consent to participate did not arise.
Consent for publication
No consent is required for publication, other than from the authors.
Competing interests
The authors declare that they have no competing interests.
Received: 6 January 2019 Accepted: 21 May 2019
References
1. EFSA Journal 2013; 11(12):3496. In January of that year the ANS panel had
issued what it termed a Draft Opinion, which is available as EFSA Panels,
DRAFT Scientific Opinion on the re-evaluation of aspartame (E 951) as a food
additive, 8th January 2013, it is available at https://efsa.onlinelibrary.wiley.
com/doi/epdf/10.2903/j.efsa.2013.3496
2. EFSA Journal 2013; 11(12)p 1 in The Abstract and p 152 in The Conclusion.
3. EFSA 2014 Discussion Paper Transformation to an Open EFSAPublic
Consultation, 17 July - 15 September 2014, http://www.efsa.europa.eu/sites/
default/files/corporate_publications/files/openefsadiscussionpaper14.pdf
Accessed September 2014; cited in EFSA Preliminary Implementation Plan
Transformation to an Open EFSA, page 3 footnote 1, available at http://
www.efsa.europa.eu/sites/default/files/corporate_publications/files/
openefsapreliminaryimplementationplan150327.pdf Accessed 3 Apr 2019.
4. EFSA 2014 Discussion Paper Transformation to an Open EFSAPublic
Consultation, 17 July - 15 September 2014p 9.
5. Hooijmans C R, Rovers M M, de Vries R BM, Leenaars M, Ritskes-Hoitinga M,
Langendam M W. SYRCLEs risk of bias tool for animal studies. BMC Med
Res Methodol. 2014;14(43). https://doi.org/10.1186/1471-2288-14-43.
6. See http://www.dcn.ed.ac.uk/camarades/. Accessed 12 June 2019.
7. Bressler J et al (1977), Establishment Investigation Endorsements, of Searle
Laboratories Division of G.D. Searle, Chicago, for the FDA Bureau of foods,
18th July 1977 and 7th august 1977; Bureau of Foods Task Force, (1977)
Authentication Review of Data in Reports Submitted to the FDA
Concerning Aspartame, FDA memo from the Bureau of Foods Task Force to
Howard R. Roberts, acting director of foods (HFF-1), 28th September 1977;
UAREP 1978, Authentication Review of Selected Materials Submitted to the
Food and Drug Administration Relative to Application of Searle Laboratories
to Market Aspartame, universities associated for research and education in
pathology Inc, 18th November 1978; FDA (1980), Aspartame: Decision of
the Public Board of Inquiry, Department of Health and Human Services,
Food and Drug Administration, docket no. 75F-0355, 30th September 1980;
Olney J, (1984), 'Excitotoxic Food additives - relevance of animal studies to
human Safety', Neurobehavioral Toxicology and Teratology, Vol. 6, pp. 455-
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 19 of 22
462; Stegink L D & Filer L J (eds), (1984) Aspartame: Physiology and
Biochemistry, Marcel Dekker, Basel; Wurtman R J, (1985), aspartame: possible
effect on seizure susceptibility, The Lancet, 9th November 1985, p 1060;
SCF (1985) Sweeteners. Reports of the scientific Committee for Food
(sixteenth series), pp 8-9, EUR 10210 EN, Commission of the European
Communities, Luxembourg; Millstone E (1986) Food Additives, penguin
books, 1986; GAO (1986), Food and Drug Administration: Six Former HSS
Employees' Involvement in Aspartames Approval, US general accounting
office, briefing report to Senator Howard Metzenbaum, reference number
GAO/HRD-86-109BR, July 1986; US senate (1987), Nutrasweet- Health and
Safety Concerns, hearing before the committee on labor and human
resources of the US senate, 3rd November 1987; Millstone E (1988)
Additives: A Guide for Everyone, penguin books; Wurtman R J & Ritter-
Walker E (eds), Dietary Phenyalanine and Brain Function, Birkhaeuser,
Boston, 1988; SCF (1989), commission of the EC Scientific's Committee for
Food, 21st series reports, pp. 22-23; McCann J E, (1990) Sweet Success: How
Nutrasweet Created a Billion Dollar Business, business one Irwin,
Homewood, Illinois; Millstone E (1994) sweet and sour: the unanswered
questions about aspartame, The Ecologist, Vol. 24, no. 2, pp. 71-74; Starr
Hull J, Sweet Poison: How The Worlds Most Popular Artificial Sweetener Is
Killing Us, new horizon press, 1998; Millstone E (2006) aspartame: the litmus
test for the FSA and EFSA, The Food Magazine, Issue 73, pp. 810.
8. Verrett in S US Senate. Nutrasweet- health and safety concerns, hearing
before the committee on labor and human resources of the US senate, 3rd
November 1987; 1987. p. 387.
9. McCann JE. Sweet Success: How nutrasweet created a billion dollar
business. Homewood: Business One Irwin; 1990. Ch. 2
10. US FDA Searle investigation task force report, March 24 1976; US FDA
Bureau of Foods Task Force, September 28, 1977, Memo to Howard R
Robert, Bureau of Foods re authentication Review of data in reports
submitted to the Food and Drug Administration concerning aspartame;
Bureau of Foods Task Force; UAREP 1978, Authentication Review of Selected
Materials Submitted to the Food and Drug Administration Relative to
Application of Searle Laboratories to Market Aspartame. Using the
identification numbers of the studies, which were assigned to them by G D
Searle, the 3 studies reviewed by the FDA Bureau of Food Task Force were:
E5, E-89 and E-77/78, and the 12 studies reviewed by the UAREP were: E-28,
E-33 & 34, E-70, E-75, E-76, E-86, E-87, E-9, E-11, E-19, E-88 and E-90.
11. Gross in US Senate (1987), Nutrasweet- Health and Safety Concerns,
hearing before the committee on labor and human resources of the US
senate, 3rd November 1987.
12. Letter from Dr Adrian gross to Mr Carl sharp at the Food and Drug
Administration, 4th November 1976, reproduced in Nutrasweet-Health
and Safety Concerns, hearing before the committee on labor and human
resources of the US senate, 3rd November 1987, pp. 440-442.
13. GLetterfromDrAdrianGrosstoMrCarlSharpattheFoodandDrug
Administration, 4th November 1976, reproduced in Nutrasweet-Healthand
Safety Concerns, Hearing before the Committee on Labor and Human Resources
of the US Senate, 3rd November 1987, pp. 440-442; Gross & Verrett in US Senate
(1987), Nutrasweet- Health and Safety Concerns, Hearing before the Committee
on Labor and Human Resources of the US Senate, 3rd November 1987.
14. Searle Investigation Task Force Report, March 24 1976; and Grosss
testimony in US Senate, Nutrasweet- Health and Safety Concerns, Hearing
before the Committee on Labor and Human Resources of the US Senate,
3rd November 1987.
15. Adrian Grosss testimony in US Senate, Nutrasweet-Health and Safety
Concerns, hearing before the committee on labor and human resources of
the US senate, 3rd November 1987.
16. UAREP 1978, Authentication Review of selected materials submitted to the
Food and Drug Administration relative to application of Searle laboratories
to market aspartame, Universities Associated for Research and Education in
Pathology Inc, 18th November 1978.
17. Bureau of Foods Task Force, (1977) Authentication Review of Data in Reports
Submitted to the FDA Concerning Aspartame, FDA Memo from the Bureau of
Foods Task Force to Howard R. Roberts, Acting Director of Foods (HFF-1),
28th September 1977 p. 3.
18. Bressler J, et al. Establishment Investigation Endorsements, of Searle Laboratories Division
of G.D. Searle, Chicago, for the FDA Bureau of foods, 18th July 1977; 1977. p. 2.
19. Bressler J, et al. Establishment Investigation Endorsements, of Searle
Laboratories Division of G.D. Searle, Chicago, for the FDA Bureau of foods,
18th July 1977; 1977. p. 28.
20. Bressler J, et al. Establishment Investigation Endorsements, of Searle
Laboratories Division of G.D. Searle, Chicago, for the FDA Bureau of foods,
18th July 1977; 1977. p. 34.
21. Verrett in US Senate. Nutrasweet- Health and Safety Concerns, hearing
before the committee on labor and human resources of the US senate, 3rd
November 1987; 1987. p. 385.
22. J Verrett, Pers. Comm. Washington DC May 1987.
23. UAREP 1978, Authentication Review of selected materials submitted to the
Food and Drug Administration relative to application of Searle laboratories
to market aspartame, Universities Associated for Research and Education in
Pathologsy Inc, 18th November 1978.
24. Verrett in US Senate. Nutrasweet- Health and Safety Concerns, hearing
before the committee on labor and human resources of the US senate, 3rd
November 1987; 1987. p. 386.
25. Graves F, (1984), How safe is your diet soft drink?,Common Cause, July/august
1984, pp. 25-43; McCann J E, (1990) Sweet Success: How Nutrasweet Created a
Billion Dollar Business, business one Irwin, Homewood, Illinois, pp. 37-48.
26. Brannigan VJ, (1983), The first FDA public Board of Inquiry: the aspartame case
chapter 9 in Law and Science in Collaboration, Nyhart JD & Carrow MM (eds.),
Lexington bo oks, Mass.; Smyth T R, (1983), the FDA's public Board of Inquiry
and the aspartame decision,Indiana Law Journal, Vol. 58 pp. 627-649.
27. FDA (1980), Aspartame: Decision of the Public Board of Inquiry, Department
of Health and Human Services, Food and Drug Administration, docket no.
75F-0355, 30th September 1980.
28. Merrill R A, Letter from Richard A Merrill, Chief Counsel, Food and Drug
Administration to Mr Samuel K Skinner, United States Attorney regarding
clinical (animal) testing of drugs and food additives by G D Searle Co,
Washington DC; 1976.
29. Dossier of documents released from the office of US Senator Howard
Metzenbaum, 6.30 pm 6 February 1986.
30. EFSA. Call for Scientific Data on Aspartame (E 951). Parma; 2011. see http://
www.efsa.europa.eu/sites/default/files/assets/110601.pdf. Accessed 4 A pr 2019.
31. Millstone E, Dossier of 28 documents, plus covering letter; available at
http://www.sussex.ac.uk/spru/research/projects/fcs. Accessed 12 June 2019.
32. Federal Register (1981) Friday July 24, 1981, part IV, Department of Health
and Human Services, Food and Drug Administration, aspartame;
CommissionersFinal Decision, Docket No 75F-0355, pp. 38284-38308.
33. Federal Register (1983) Friday July 8, 1983, Vol. 48, No. 132, Food Additives
Permitted for Direct Addition to Food for Human Consumption: Aspartame
Docket No. 82F-0305, pp. 31376-31382.
34. FDA (1980), Aspartame: decision of the public Board of Inquiry, Department
of Health and Human Services, Food and Drug Administration, docket no.
75F-0355, 30th September 1980; Smyth T R, (1983), the FDA's public Board
of Inquiry and the aspartame decision, Indiana law Journal, Vol. 58 pp. 627
649; Brannigan VJ, (1983), the first FDA public Board of Inquiry: the
aspartame casechapter 9 in law and science in collaboration, Nyhart JD &
Carrow MM (eds.), Lexington Books, Mass.
35. GAO (1986), Food and Drug Administration: Six Former HSS Employees'
Involvement in Aspartames Approval, US general accounting office, briefing report
to Senator Howard Metzenbaum, Reference number GAO/HRD-86-109BR, July 1986.
36. Cockburn A, Rumsfeld: an American Disaster, Verso, New York & London,
2007, Chapter 4: How $weet Is It?
37. JECFA (1980), WHO/FAO Joint Expert Committee on Food Additives,
Toxicological Evaluation of Certain Food Additives, Food Additives Series
No. 15, Geneva, pp. 18-86; SCF (1985) Sweeteners.Reportsofthe
Scientific Committee for Food (Sixteenth Series), pp 8-9, EUR 10210 EN,
Commission of the European Communities, Luxembourg.
38. Veitch A (1984) Charity linked to makers of suspect drug, The
Guardian, 13 February 1984, p 4.
39. JECFA (1980), WHO/FAO Joint Expert Committee on Food Additives,
Toxicological Evaluation of Certain Food Additives, Food Additives Series
No. 15, Geneva, pp. 18-86; SCF (1985) Sweeteners. Reports of the Scientific
Committee for Food (Sixteenth Series), pp 8-9, EUR 10210 EN, Commission
of the European Communities, Luxembourg; Millstone E (1994) Sweet and
sour: the unanswered questions about aspartame, Ecologist, Vol. 24, No. 2,
pp. 71-74.
40. McCann JE. Sweet Success: how nutrasweet created a billion dollar
business. Homewood: Business One Irwin; 1990. p. 98.
41. McCann 1990 Ch. 4.
42. CDC, Evaluation of consumer complaints related to aspartame use, November
1, 1984; Novick N L, Aspartame-Induced Granulomatous Panniculitis,Annals
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 20 of 22
of Internal Medicine, Vol. 102, No. 2, February 1985; Wurtman R J, (1985),
Aspartame: Possible Effect on Seizure Susceptibility,The Lancet, 9th
November 1985, p 1060; Kiritsy PJ & Maher TJ, Acute Effects of Aspartame
on Systolic Blood Pressure in Spontaneously Hypertensive Rats,Journal of
Neural Transmission, 1986; Walton R G, Seizure and mania after high intake
of aspartame,Psychosomatics, Vol. 27, Part 3, pp. 218-220, 1986.
43. Pers. Comms. 6 Oct 1986; 5 March 1987; 18 July 1987; 14 Dec 1987.
44. Olney JW, Farber NB, Spitznagel E, Robins LN. Increasing brain tumor rates:
is there a link to aspartame? J Neuropathol Exp Neurol. 1996;55:111523.
45. Soffritti M, Belpoggi F, Esposti DD and Lambertini L, Aspartame induces
lymphomas and leukaemias in ratsEuropean Journal of Oncology, 2005,
10,107116; see https://www.researchgate.net/publication/225029050_
Aspartame_induces_lymphomas_and_leukaemias_in_rats; Soffritti M,
Belpoggi F, Degli Esposti D, Lambertini L, Tibaldi E and Rigano A, First
experimental demonstration of the multipotential carcinogenic effects of
aspartame administered in the feed to Sprague-Dawley ratsEnvironmental
Health Perspectives, 2006, 114, 379-385; Soffritti M, Belpoggi F, Tibaldi E,
Degli Esposti D and Lauriola M, Life-span exposure to low doses of
aspartame beginning during pre-natal life increases cancer effects in rats,
Environmental Health Perspectives, 2007, 115, 1293-1297.
46. See eg https://www.fda.gov/Food/GuidanceRegulation/GuidanceDocu
mentsRegulatoryInformation/IngredientsAdditivesGRASPackaging/ucm07
8388.htm#ftnIV and https://ntp.niehs.nih.gov/ntp/test_info/finalntp_toxcars
pecsjan2011.pdf. Accessed 12 June 2019.
47. See OECD, (http://www.oecd.org/env/test-no-451-carcinogenicity-studies-
9789264071186-en.htm. Accessed 12 June 2019.
48. Soffritti M, et al. Aspartame administered in feed, beginning prenatally
through life span, induces cancers of the liver and lung in male Swiss mice,
American Journal of Industrial Medicine, 2010, 53:11971206; Soffritti M et al,
first experimental demonstration of the multipotential carcinogenic effects
of aspartame administered in the feed to Sprague-Dawley rats. Environ
Health Perspect. 2006;114:37985. https://doi.org/10.1289/ehp.8711.
49. Gift JS, Caldwell JC, Jinot J, Evans MV, Cote I, Vandenberg JJ. Scientific
considerations for evaluating cancer bioassays conducted by the Ramazzini
institute. Environ Health Perspect. 2013;121:125363. https://doi.org/10.1289/
ehp.1306661 Available at https://www.ncbi.nlm.nih.gov/pubmed/24045135,
Accessed Feb 2019.
50. Soffritti M, Belpoggi F, Esposti DD, Lambertini L. Aspartame induces
lymphomas and leukaemias in rats. Eur J Oncol. 2005;10:10716 see https://
www.ramazzini.org/wp-content/uploads/2008/03/Aspartame-induces-
lymphomas-and-leukaemias-in-rats_2005.pdf.
51. Soffritti M, Belpoggi F, Manservigin M, Tibaldi E, Lauriola M, Falcioni L, Bua L.
Aspartame administered in feed, beginning prenatally through life span,
induces cancers of the liver and lung in male Swiss mice. Am J Ind Med.
2010;53:1197206.
52. Commission of The European Communities, White Paper On Food Safety,
Brussels, 12 January 2000, COM (1999) 719 final.
53. EFSA Scientific Opinion of the Panel on Food Additives, Flavourings,
Processing Aids and Materials in Contact with Food (AFC) related to a new
long-term carcinogenicity study on aspartame,The EFSA Journal, 356, 1-44; UK
Committee on Carcinogenicity of Chemicals in Food, Consumer Products and
the Environment, Statement on a Carcinogenicity Study of Aspartame by the
European Ramazzini Foundation. Statement COC/06/S2 December 2006.
54. JCaldwell JC, Jinot J, DeVoney D & Gift JS, Evaluation of evidence for infection
as a mode of action for induction of rat lymphoma,Environmentaland
Molecular Mutagenesis, 19 December 2007, https://doi.org/10.1002/em.20356,
see https://www.ncbi.nlm.nih.gov/pubmed/18095346.
55. Preamble to the IARC Monographs, B. SCIENTIFIC REVIEW AND
EVALUATION, https://monographs.iarc.fr/preamble-to-the-iarc-monographs/
preamble-to-the-iarc-monographs-10/], although that is exactly what the
ANS panel did. [EFSA Dec. 2013, p. 71, Sect. 3.2.4.2.
56. The 3 studies reviewed by the FDA Bureau of Food task force were: E5, E-89
and E-77/78. The 12 studies that the UAREP reviewed were: E-28, E-33 & 34,
E-70, E-75, E-76, E-86, E-87, E-9, E-11, E-19, E-88 and E-90.
57. Request from the European Commission to EFSA, Question No EFSA-Q-
2011-000406, 4 May 2011.
58. EFSA Panels, DRAFT Scientific Opinion on the re-evaluation of aspartame (E
951) as a food additive, 8th January 2013, available at www.efsa.europa.eu/
sites/default/files/consultation/130108.pdf.
59. EFSA Panels, DRAFT Scientific Opinion on the re-evaluation of aspartame (E
951) as a food additive, 8th January 2013page 1 lines 25-6.
60. Gross A (1987),Letter to Senator Howard Metzenbaum from Dr Adrian
Gross, October 30th 1987, reproduced in Nutrasweet" - Health and Safety
Concerns, Hearing before the Committee on Labor and Human Resources
of the US Senate, 3rd November 1987, pp. 430-439.
61. Millstone E, 2013, EFSA on Aspartame January 2013 a lost, but not the last,
opportunity, letter to the European Food safety authority, 20 February 2013,
available at http://sro.sussex.ac.uk/43821/
62. Millstone E, 2013, EFSA on Aspartame January 2013: a lost, but not the last,
opportunity, 22 February 2013, available at https://www.sussex.ac.uk/
webteam/gateway/file.php?name=em-letter-to-efsa-on-aspartame-
20feb2013.pdf&site=25
63. EFSA Journal, 2013, 11(12):3496, p 7.
64. Millstone E, EFSA on Aspartame December 2013, available at: https://www.
sussex.ac.uk/webteam/gateway/file.php?name=millstone-on-efsa-on-
aspartame-16dec2013.pdf&site=25.
65. Letter from Per Bergman, Head of regulated products, EFSA to Erik
Millstone, 14 November 2014.
66. Eg G D Searle studies E33-34 and E39.
67. Wallace Hayes A, editor. Principles and methods of toxicology. New York:
Raven Press; 1982.
68. Guidance of the EFSA Scientific Committee on transparency in the scientific
aspects of risk assessment carried out by EFSA. Part 2: general principles. in
response to Question No EFSA-Q-2005-050Ba, and adopted on 7 April 2009,
The EFSA Journal (2009) 1051, 1-22.
69. EFSA Journal 2013; 11(12):3496 Appendix A pp 182-185.
70. See eg Gross A (1987),Letter to Senator Howard Metzenbaum from Dr
Adrian Gross, October 30th 1987, reproduced in Nutrasweet" - Health and
Safety Concerns, Hearing before the Committee on Labor and Human
Resources of the US Senate, 3rd November 1987, pp. 430-439; Verrett J in
US Senate (1987), Nutrasweet-Health and Safety Concerns, Hearing before
the Committee on Labor and Human Resources of the US Senate, 3rd
November 1987; Turner J; http://www.swankin-turner.com/lawof.html;
Millstone E (1994) Sweet and Sour: The Unanswered Questions about
Aspartame,The Ecologist, Vol. 24, No. 2, pp. 71-74; Millstone E (2006)
Aspartame: the litmus test for the FSA and EFSA,The Food Magazine, Issue
73, pp. 8-10.
71. EFSA Journal 2013;11(12):3496, footnote page 67 in respect of E102a, E102b
and E102c, and in Appendix L pp 259-260.
72. Olney JW, et al. Increasing brain tumor rates: is there a link to aspartame? J
Neuropathol Exp Neurol. 1996;55(11):1115-23.
73. See Two year toxicity study in the Rat: Final Report and Appendix, Hazelton
Laboratories study number P-T 838H71, Submitted to the FDA 25 January
1973, Master File numbers E-33 and E-34.
74. Shephard SE, et al. Mutagenic activity of peptides and the artificial
sweetener aspartame after nitrosation. Food Chem Toxicol. 1993;31:3239.
75. EFSA Journal 2013; 11(12):3496, p 33.
76. EFSA Journal 2013; 11(12):3496, Sec 3.2.3.1 p 60.
77. EFSA Journal 2013; 11(12):3496, Sec 3.2.5.1.2 page 75.
78. EFSA Journal 2013; 11(12):3496, Appendix I page 229.
79. EFSA Journal 2013; 11(12):3496,Sec 3.2.5.1.2 pp 76-7.
80. EFSA Journal 2013; 11(12):3496, Sec 3.2.5.2.1 page 78.
81. Brunner RL, Vorhees CV, Kinney L, Butcher RE. Aspartame: assessment of
developmental psychotoxicity of a new artificial sweetener. Neurobehav
Toxicol. 1979;1:7986.
82. Magnuson BA, et al. Aspartame: a safety evaluation based on current use
levels, regulations, and toxicological and epidemiological studies. Crit Rev
Toxicol. 2007;37:629727. https://doi.org/10.1080/10408440701516184 p 680.
83. EFSA Journal 2013; 11(12):3496, Sec 3.2.2.1 p 58.
84. EFSA Journal 2013; 11(12):3496, Sec 3.2.4.1 p 69.
85. EFSA Journal 2013; 11(12):3496, Sec 3.2.3.2 page 62.
86. EFSA Journal 2013; 11(12):3496, Sec 3.2.5.1.2 page 76.
87. EFSA Journal 2013; 11(12):3496, pages 229- 230.
88. EFSA Journal 2013; 11(12):3496, Sec 3.2.3.1 pages 59-60.
89. EFSA Journal 2013; 11(12):3496, page 209.
90. EFSA Journal 2013; 11(12):3496, Sec 3.2.3.2 page 61.
91. See https://www.coca-cola.co.uk/drinks/coca-cola-zero/coke-zero. Accessed
12 June 2019.
92. See eg Von Rymon Lipinski G-W, Sweeteners and sugar substitutes,
Lebensmittelchemie, 1994, 48-2, pp 34-9; Von Rymon Lipinski G-W, the
blending of sweeteners applications and safety issuesCh 14 of Grenby
TH ed. Advances in Sweeteners, Chapman Hall, London 1996.
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 21 of 22
93. EFSA Journal 2013; 11(12):3496, Sec 3.2.5.1.1 page 73.
94. EFSA Journal 2013; 11(12):3496 Appendix I page 230, though that was not
acknowledged in Sec 3.2.5.1.2 on page 76.
95. EFSA Journal 2013; 11(12):3496 page 63.
96. EFSA Journal 2013; 11(12):3496 pages 58 and Appendix G204.
97. EFSA Journal 2013; 11(12):3496 page 57.
98. EFSA Journal 2013; 11(12):3496 pages 66-67 including Tables 19 and 20.
99. Guidance of the EFSA Scientific Committee on transparency in the scientific
aspects of risk assessment carried out by EFSA. Part 2: general principles. in
response to Question No EFSA-Q-2005-050Ba, and adopted on 7 April 2009,
The EFSA Journal (2009) 1051, page 3.
100. EFSA Journal 2013; 11(12):3496, Sec 3.2.5.1.1 page 57.
101. EFSA Journal 2013; 11(12):3496 page 204.
102. EFSA Journal 2013; 11(12):3496 Sec 3.2.4.1 page 68.
103. EFSA Journal 2013; 11(12):3496 Section 3.2.7.1 page 86.
104. EFSA Journal 2013; 11(12):3496 page 87.
105. EFSA Journal 2013; 11(12):3496 Sec 3.2.7.1 page 87.
106. EFSA Journal 2013; 11(12):3496 page 88.
107. EFSA Journal 2013; 11(12):3496 Sec 3.2.7.5 page 98.
108. EFSA Journal 2013; 11(12):3496 Sec 3.2.7.8 page 100.
109. EFSA Journal 2013; 11(12):3496 page 100.
110. EFSA Journal 2013; 11(12):3496 page 101.
111. EFSA Journal 2013; 11(12):3496Sec 3.2.7.8 page 101.
112. EFSA Journal 2013; 11(12):3496 Sec 3.2.7.9 page 101-2.
113. EFSA Journal 2013; 11(12):3496 page 84.
114. Corporate Observatory Europe, Conflicts of interest scandals at EFSA: A non-
exhaustive chronology of recent events, available at https://corporateeurope.
org/food-and-agriculture/efsa/chronology.Accessed12June2019.
115. Harrison-Dunn A, EFSAs Bernhard Url: science should reserve one private
room for open debate, 19 October 2015, https://www.nutraingredients.
com/Article/2015/10/20/EFSA-s-Bernhard-Url-Science-should-reserve-one-
private-room-for-open-debate, Accessed Oct 2015 and Mar 2019.
116. UK Food Standards Agency (2002) Report on the Review of scientific
committees, 15 April 2002.
117. See https://doi.org/10.1186/1471-2288-14-43. Accessed 12 June 2019.
PublishersNote
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
Millstone and Dawson Archives of Public Health (2019) 77:34 Page 22 of 22
... This letter is in response to a recent paper by Millstone and Dawson (2019) in which the authors criticise the reevaluation of the high intensity sweetener aspartame in 2013 by the former EFSA's Panel on Food Additives and Nutrient Sources added to Food, on the grounds that EFSA did not follow its own procedures for its risk assessment. Moreover, the authors claim that the appraisal of the available studies was asymmetrically more alert to putative false positives than to possible false negatives. ...
... This included an uncertainty analysis. Overall, and in contrast with the claims in [1], these elements demonstrate that the approach followed by the ANS Panel was a rigorous one, ensuring an independent and in-depth analysis of all existing data [6]. ...
Article
Full-text available
Abstract This letter is in response to a recent paper by Millstone and Dawson (2019) in which the authors criticise the re-evaluation of the high intensity sweetener aspartame in 2013 by the former EFSA’s Panel on Food Additives and Nutrient Sources added to Food, on the grounds that EFSA did not follow its own procedures for its risk assessment. Moreover, the authors claim that the appraisal of the available studies was asymmetrically more alert to putative false positives than to possible false negatives. In this letter it is shown that the methodology for collection and selection of the scientific information used as a basis for the aspartame risk assessment, and the inclusion/exclusion criteria applied were defined a priori and documented in the published opinion. Furthermore, the Panel applied a Weight-of-Evidence approach combined with an analysis of the biological relevance of the appraised and validated evidence for its analysis, integration and interpretation, followed by an uncertainty analysis. Finally, an analysis of the distribution of negative versus positive outcome of the studies in the context of reliability showed that the claim of bias in the scientific risk assessment of aspartame is not substantiated.
... Publication of the RI findings on the carcinogenicity of aspartame generated intense controversy [15]. At the heart of this debate were doubts raised about the accuracy of the RI's histopathological diagnoses -in particular the RI's diagnoses of pulmonary lymphomas and leukemiasin animals exposed to aspartame [14,16]. ...
Article
Full-text available
Background Aspartame is one of the world’s most widely used artificial sweeteners and is an ingredient in more than 5000 food products globally. A particularly important use is in low-calorie beverages consumed by children and pregnant women. The Ramazzini Institute (RI) reported in 2006 and 2007 that aspartame causes dose-related increases in malignant tumors in multiple organs in rats and mice. Increased cancer risk was seen even at low exposure levels approaching the Acceptable Daily Intake (ADI). Prenatal exposures caused increased malignancies in rodent offspring at lower doses than in adults. These findings generated intense controversy focused on the accuracy of RI’s diagnoses of hematopoietic and lymphoid tissue tumors (HLTs). Critics made the claim that pulmonary lesions observed in aspartame-exposed animals were inflammatory lesions caused by Mycoplasma infection rather than malignant neoplasms. Methods To address this question, RI subjected all HLTs from aspartame-exposed animals to immunohistochemical analysis using a battery of markers and to morphological reassessment using the most recent Internationally Harmonized Nomenclature and Diagnostic (INHAND) criteria. Findings This immunohistochemical and morphological re-evaluation confirmed the original diagnoses of malignancy in 92.3% of cases. Six lesions originally diagnosed as lymphoma (8% of all HLTs) were reclassified: 3 to lymphoid hyperplasia, and 3 to chronic inflammation with fibrosis. There was no evidence of Mycoplasma infection. Interpretation These new findings confirm that aspartame is a chemical carcinogen in rodents. They confirm the very worrisome finding that prenatal exposure to aspartame increases cancer risk in rodent offspring. They validate the conclusions of the original RI studies. These findings are of great importance for public health. In light of them, we encourage all national and international public health agencies to urgently reexamine their assessments of aspartame’s health risks - especially the risks of prenatal and early postnatal exposures. We call upon food agencies to reassess Acceptable Daily Intake (ADI) levels for aspartame. We note that an Advisory Group to the International Agency for Research on Cancer has recommended high-priority reevaluation of aspartame’s carcinogenicity to humans.
... The final report of EFSA (released December 2013), which comprised the analysis of the in vitro, in vivo and clinical studies concerning ASP conducted up to 2011, concluded that "there were no safety concerns at the current ADI of 40 mg/kg bw/day. Therefore, there was no reason to revise the ADI for aspartame" [38,39]. Another aspect mentioned in the report was the estimated ASP daily intake values (data acquisition from 26 studies conducted in 17 European countries), presented as the mean (1.2 to 5.3 mg/kg/day) and highest values (1.9 to 15.6 mg/kg/day), indicating that these values represent less than 50% of the ADI and the risk of toxicity is low [40,41]. ...
Article
Full-text available
Aspartame (ASP), an artificial sweetener abundantly consumed in recent years in an array of dietary products, has raised some concerns in terms of toxicity, and it was even suggested a link with the risk of carcinogenesis (colorectal cancer), though the present scientific data are rather inconclusive. This study aims at investigating the potential role of aspartame in colorectal cancer by suggesting two experimental approaches: (i) an in vitro cytotoxicity screening in HT-29 human colorectal carcinoma cells based on cell viability (Alamar blue assay), cell morphology and cell migration (scratch assay) assessment and (ii) an in ovo evaluation in terms of angiogenic and irritant potential by means of the chorioallantoic membrane method (CAM). The in vitro results showed a dose-dependent cytotoxic effect, with a significant decrease of viable cells at the highest concentrations tested (15, 30 and 50 mM) and morphological cellular changes. In ovo, aspartame (15 and 30 mM) proved to have a pro-angiogenic effect and a weak irritant potential at the vascular level. These data suggest new directions of research regarding aspartame’s role in colorectal cancer.
... We are grateful to George Kass and Frederica Lodi for their letter, recently published in Archives of Public Health [1] in response to the Archives of Public Health paper we co-authored entitled 'EFSA's toxicological assessment of aspartame: was it even-handedly trying to identify possible unreliable positives and unreliable negatives?', published in Archives of Public Health in July 2019 [2]. ...
Article
Full-text available
On behalf of the European Food Safety Authority (EFSA), Kass and Lodi recently published a letter purporting to ‘refute’ our July 2019 analysis of EFSA’s December 2013 assessment of the risks of aspartame. We had previously claimed inter alia that the EFSA panel had evaluated studies that had indicated that aspartame might be harmful far more sceptically than those that had not indicated harm. We reported that EFSA had deemed every one of 73 studies suggesting harm to have been unreliable. Kass and Lodi provided a tabulation with figures that differed from ours in every detail. This commentary shows that, while Kass and Lodi provided a response to our analysis, they have not come close to refuting it. Our analysis provided detailed characterisations of each of the studies and how the panel interpreted them, but Kass and Lodi provide no corresponding information at all. Kass and Lodi claim that EFSA deemed 21 of 35 studies that had indicated possible harm to have been reliable. But if that is so, we now ask: why did the EFSA panel not recommend that aspartame should be banned, or at least tightly restricted?
... EFSA concluded that aspartame is safe at a dose of 40 mg/kg body weight/day (117). However, aspartame safety on human health is still under debate; in fact, a recent study highlighted several important shortcomings in the EFSA document (118). ...
Article
Full-text available
Since excessive sugar consumption has been related to the development of chronic metabolic diseases prevalent in the western world, the use of sweeteners has gradually increased worldwide over the last few years. Although low-and non-calorie sweeteners may represent a valuable tool to reduce calorie intake and prevent weight gain, studies investigating the safety and efficacy of these compounds in the short-and long-term period are scarce and controversial. Therefore, future studies will need to elucidate the potential beneficial and/or detrimental effects of different types of sweeteners on metabolic health (energy balance, appetite, body weight, cardiometabolic risk factors) in healthy subjects and patients with diabetes, obesity and metabolic syndrome. In this regard, the impact of different sweeteners on central nervous system, gut hormones and gut microbiota is important, given the strong implications that changes in such systems may have for human health. The aim of this narrative review is to summarize the current evidence for the neuroendocrine and metabolic effects of sweeteners, as well as their impact on gut microbiota. Finally, we briefly discuss the advantages of the use of sweeteners in the context of very-low calorie ketogenic diets.
... The Chief Executive of EFSA has acknowledged that EFSA is supposed to be a purely scientific body, giving advice but with no statutory responsibility for policy decisions [39]. In practice, however, numerous policy choices have been autonomously made by EFSA risk assessors concerning the problem-framing they adopted, without the involvement of, or reference to, risk managers [40,41]. ...
Article
Full-text available
We identify the origins, and key characteristics, of the current regulatory framework for genetically modified organisms (GMOs) in the European Union (EU). We focus on the approach of the European Food Safety Authority (EFSA) to assessing GMO risks to public and environmental health. An historical account informs our critical appraisal of the current practices of EFSA’s GMO panel, and helps to explain how and why it fails to satisfy the objectives of the EU’s GMO legislation. While those legislative texts set appropriate objectives, their concrete implementation has fallen far short of the legislative goals. EFSA’s prevailing approach to GM crop and food risk assessment starts from what it terms a ‘comparative safety assessment’. Those comparisons require the scrutiny of sets of molecular, chemical and phenotypic data from GM plants and non-GM varieties (many of which may be only remotely related to the GM variety). Those data are, however, inadequate for predicting adverse biological, toxicological and ecological effects. EFSA’s ‘comparative safety assessments’ draw over-optimistic conclusions from too little data from too few studies. When GM products are deemed to have passed a ‘comparative safety assessment’, EFSA has interpreted that as grounds for conducting only very narrowly circumscribed risk assessments, which have not required meaningful data from studies of ecological or eco-toxicological impacts. This is a reductionist approach to risk assessment, when a more inclusive and comprehensive approach, which we outline, is scientifically available, and also more likely to meet the specified legislative aims. Instead, however, the reductionist choice is systematically applied, but never justified nor acknowledged as such. Indeed, it is concealed, by EFSA and by its main policy client the European Commission, as if it were only for specialist expert scientific deliberation. Thus, key questions that sound scientific assessments should ask about potential harm are not even posed, let alone answered—or at least, they are ‘answered’ only by default, given that the implicit burden of proof requires harm to be demonstrated. Furthermore, and a key point of this paper, we show how the problematic features of EFSA’s approach have been premised on a set of evaluative policy judgements, rather than purely scientific considerations. Responsibility for selecting how EFSA frames its scientific approach should however lie with European Commission risk managers, and not with EFSA. These problems might have been and could be avoided if explicit commitments entered into by the EU at the Codex Alimentarius Commission were implemented by the European Commission and EFSA were instructed accordingly.
Article
This article discusses Christian theological-ethical questions related to the processing offood and especially to the widespread use offood additives. First, the aims, parameters and methodology of the article are discussed. Secondly, the theme of food in the Bible is briefly explored, along with its theological implications. Thirdly, the wider cultural context of food production and processing is noted, along with the commercialisation of food. Next, the nature of food additives of various kinds, the reasons for their use and their effects on human beings are analysed to determine the extent to which their use is injurious or harmless. Throughout, and especially in the final section, a theological-ethical analysis of the use of food additives, especially in a South African context, is provided.
Article
Full-text available
Background: This study was performed to see the effect of chronic administration by gavage of aspartame on kidneys and behavior in Wistar rats. Methods: In this study, 28 Wistar rats were used, divided into 04 groups including a control group treated with distilled water and three other groups treated with aspartame at different doses (40 mg / kg, 60 mg / kg, 1000 mg / kg body weight) for 90 days orally (gastric gavage). Results: A highly significant increase (p <0,001) of blood urea level ASP 3 group compared to the ASP1 group (ADI Acceptable Daily Intake) , however we observe a highly significant reduction in blood creatinine level (p <0, 01) as compared to the ASP1 tolerated daily intake group. In the open field test, we notice highly significant decrease (p <0,01) in the ASP3 group treated with the 1000 mg per kg body weight dose ASP 3 compared to the group treated with the 40 mg / kg bw dose or the ADI ( Acceptable Daily Intake) ASP1. The histopathological study of kidneys showed multiple vascular congestion with inflammatory infiltration in Group ASP3 compared to the control. Conclusion: The chronic administarion of aspartame by gavage on Wistar rats causes kidney damage and neurobehavioral troubles.
Article
Full-text available
Systematic Reviews (SRs) of experimental animal studies are not yet common practice, but awareness of the merits of conducting such SRs is steadily increasing. As animal intervention studies differ from randomized clinical trials (RCT) in many aspects, the methodology for SRs of clinical trials needs to be adapted and optimized for animal intervention studies. The Cochrane Collaboration developed a Risk of Bias (RoB) tool to establish consistency and avoid discrepancies in assessing the methodological quality of RCTs. A similar initiative is warranted in the field of animal experimentation. We provide an RoB tool for animal intervention studies (SYRCLE's RoB tool). This tool is based on the Cochrane RoB tool and has been adjusted for aspects of bias that play a specific role in animal intervention studies. To enhance transparency and applicability, we formulated signalling questions to facilitate judgment. The resulting RoB tool for animal studies contains 10 entries. These entries are related to selection bias, performance bias, detection bias, attrition bias, reporting bias and other biases. Half these items are in agreement with the items in the Cochrane RoB tool. Most of the variations between the two tools are due to differences in design between RCTs and animal studies. Shortcomings in, or unfamiliarity with, specific aspects of experimental design of animal studies compared to clinical studies also play a role. SYRCLE's RoB tool is an adapted version of the Cochrane RoB tool. Widespread adoption and implementation of this tool will facilitate and improve critical appraisal of evidence from animal studies. This may subsequently enhance the efficiency of translating animal research into clinical practice and increase awareness of the necessity of improving the methodological quality of animal studies.
Article
Full-text available
The Ramazzini Institute (RI) has completed nearly 400 cancer bioassays on over 200 compounds. Study design and protocol differences between RI and other laboratories have been suggested by the European Food Safety Authority (EFSA) and others to contribute to controversy regarding cancer hazard findings, principally lymphoma/leukemia diagnoses. Evaluate RI study design, protocol differences, and accuracy of tumor diagnoses for their impact on carcinogenic hazard characterization. We analyzed the findings from a recent Pathology Working Group (PWG) review of RI procedures and tumor diagnoses, evaluated consistency of RI and other laboratory findings for chemicals identified by RI as positive for lymphoma/leukemia, and examined evidence for a number of other issues raised regarding RI bioassays. The RI cancer bioassay design and protocols were evaluated in the context of relevant risk assessment guidance from International authorities. Although the PWG identified close agreement with RI diagnoses for most tumor types, it did not for lymphoma/leukemia of the respiratory tract and neoplasms of the inner ear and cranium. We discuss (1) the implications of the PWG findings, particularly lymphoma diagnostic issues, (2) differences between RI and other laboratory studies that are relevant to evaluating RI cancer bioassays and (3) future work that may help resolve some concerns. We conclude that (1) issues related to respiratory tract infections have complicated diagnoses at that site (i.e., lymphoma/leukemia) and for neoplasms of the inner ear and cranium, and (2) there is consistency and value in RI studies for identification of other chemical-related neoplasia.
Chapter
Intense sweeteners have become more popular and their use has been expanding for many years. Whereas in the past they were chiefly used in diabetic products, with saccharin regarded as a cheap sugar substitute, intense sweeteners are now common ingredients in a range of low-calorie and calorie-reduced foods and beverages. In the course of the last 15 years the new sweeteners acesulfame-K (Sunett®, Hoechst AG) and aspartame (e.g. NutraSweet®, NutraSweet Co., Sanecta®, Holland Sweetener Co. or Aspartil®, Enzymologa S.A.) have been approved in many countries and are being increasingly used in a variety of applications. With the availability of improved sweeteners with a good taste profile, perceived by consumers as safe food ingredients, new and extended fields of application are being developed or will be possible in countries progressing towards approval for additional applications e.g. by implementing the EUC Sweetener Directive.
Article
The EFSA ANS Panel provides a scientific opinion on the safety of aspartame (E 951). Aspartame is a sweetener authorised as a food additive in the EU. In previous evaluations by JECFA and the SCF, an ADI of 40 mg/kg bw/day was established based on chronic toxicity in animals. Original reports, previous evaluations, additional literature and data made available following a public call were evaluated. Aspartame is rapidly and completely hydrolysed in the gastrointestinal tract to phenylalanine, aspartic acid and methanol. Chronic and developmental toxicities were relevant endpoints in the animal database. From chronic toxicity studies in animals, a NOAEL of 4000 mg/kg bw/day was identified. The possibility of developmental toxicity occurring at lower doses than 4000 mg/kg in animals could not be excluded. Based on MoA and weight-of-evidence analysis, the Panel concluded that developmental toxicity in animals was attributable to phenylalanine. Phenylalanine at high plasma levels is known to cause developmental toxicity in humans. The Panel concluded that human data on developmental toxicity were more appropriate for the risk assessment. Concentration-response modelling was used to determine the effects of aspartame administration on plasma phenylalanine using human data after phenylalanine administration to normal, PKU heterozygote or PKU homozygote individuals. In normal and PKU heterozygotes, aspartame intakes up to the ADI of 40 mg/kg bw/day, in addition to dietary phenylalanine, would not lead to peak plasma phenylalanine concentrations above the current clinical guideline for the prevention of adverse effects in fetuses. The Panel concluded that aspartame was not of safety concern at the current aspartame exposure estimates or at the ADI of 40 mg/kg bw/day. Therefore, there was no reason to revise the ADI of aspartame. Current exposures to aspartame - and its degradation product DKP - were below their respective ADIs. The ADI is not applicable to PKU patients.
Article
Aspartame, a widely used artificial sweetener, was administered with feed to male and female Sprague-Dawley rats (100-150/sex/group), 8 weeks-old at the start of the experiment, at concentrations of 100,000; 50,000; 10,000; 2,000; 400; 80 and 0 ppm. Treatment lasted until spontaneous death of the animals. In this report we present the first results showing that aspartame, in our experimental conditions, causes a statistically significant, dose-related increase in lymphomas and leukaemias in females. No statistically significant increase in malignant brain tumours was observed among animals from the treated groups as compared to controls.
Article
Aspartame (APM) is a well-known intense artificial sweetener used in more than 6,000 products. Among the major users of aspartame are children and women of childbearing age. In previous lifespan experiments conducted on Sprague-Dawley rats we have shown that APM is a carcinogenic agent in multiple sites and that its effects are increased when exposure starts from prenatal life. The aim of this study is to evaluate the potential of APM to induce carcinogenic effects in mice. Six groups of 62-122 male and female Swiss mice were treated with APM in feed at doses of 32,000, 16,000, 8,000, 2,000, or 0  ppm from prenatal life (12 days of gestation) until death. At death each animal underwent complete necropsy and all tissues and organs of all animals in the experiment were microscopically examined. APM in our experimental conditions induces in males a significant dose-related increased incidence of hepatocellular carcinomas (P < 0.01), and a significant increase at the dose levels of 32,000  ppm (P < 0.01) and 16,000  ppm (P < 0.05). Moreover, the results show a significant dose-related increased incidence of alveolar/bronchiolar carcinomas in males (P < 0.05), and a significant increase at 32,000  ppm (P < 0.05). The results of the present study confirm that APM is a carcinogenic agent in multiple sites in rodents, and that this effect is induced in two species, rats (males and females) and mice (males). No carcinogenic effects were observed in female mice. Am. J. Ind. Med. 53:1197-1206, 2010. © 2010 Wiley-Liss, Inc.
Article
The effects of continuous exposure (before conception through 90 days of postnatal life) to aspartame (Asp) were evaluated in rats through observations of reproduction, physical development and behavior. Three dose levels (2%, 4% and 6% of diet) were fed in standard laboratory chow, and data from these subjects were compared to normally fed negative controls and a positive control group exposed to 550 mg/kg of hydroxyurea on the 12th day of gestation and a phenylalanine (3%) fed group. Aspartame (6%) and phenylalanine (3%) both significantly increased offspring mortality. Aspartame (6% and 4%) and phenylalanine significantly reduced dam weights during lactation. Offspring weights were reduced in Asp 6%, phenylalanine and positive control groups. It was determined from a cross-fostering experiment that offspring weight reduction and mortality occurred in rats reared by aspartame fed dams but aspartame exposure during gestation alone had no untoward effects. Behaviorally, Asp 6% or phenylalanine delayed eye opening, righting, swimming and resulted in hypoactivity. The Asp 6% group showed delayed appearance of the startle response and forward locomotion. More minor, but statistically reliable delays in swimming development appeared in the Asp 4% and Asp 2% groups. Among the adult tests, in open field, the Asp 6% and positive control females were slower to leave the center starting section. At 90 days of age, brain weights of the positive control group were significantly reduced. Regional brain cell counts made at 21 days of age were lower in the cerebella of Asp 6% rats and in the olfactory bulbs of Asp 6% and Asp 2% rats. In general, the adverse effects of 6% aspartame appear indistinguishable from those attributable to its phenylalanine content.
Article
Exogenous tyrosine lowers blood pressure in spontaneously hypertensive rats (SHR). The artificial sweetener aspartame also elevates blood and brain tyrosine levels in rats by being hydrolyzed to phenylalanine, which is then rapidly hydroxylated to tyrosine in the liver. Hence we tested the ability of aspartame; its hydrolytic products phenylalanine, aspartic acid and methanol; and of tyrosine itself to lower blood pressure in SHR. For one week prior to experimentation rats were acclimated to the indirect blood pressure measurement technique; on the day of an experiment they received I.P. injections (mg/kg) of aspartame (12.5-200), tyrosine (25-200) or phenylalanine (100-200), or of aspartic acid or methanol in the doses theoretically contained within 200 mg/kg aspartame. Animals receiving 50, 100 or 200 mg/kg of aspartame exhibited maximum falls in blood pressure of 17.3, 24.2 and 19.3 mmHg, respectively. All changes were significant, as determined by ANOVA and the Newman-Keuls test (p less than 0.05). Tyrosine or phenylalanine also lowered blood pressure, but aspartic acid or methanol produced no significant effects. Co-administration of aspartame with valine, a large neutral amino acid that competes with phenylalanine or tyrosine for brain uptake, attenuated aspartame's hypotensive effect. These observations suggest that the neurochemical changes produced by aspartame lead to predicted tyrosine-induced changes in blood pressure.
Article
Evidence is reviewed supporting the view that excitotoxic food additives pose a significant hazard to the developing nervous system of young children. The following points are stressed: (1) although blood-brain barriers protect most central neurons from excitotoxins, certain brain regions lack such protection (a characteristic common to all vertebrate species); (2) regardless of species, it requires only a transient increase in blood excitotoxin levels for neurons in unprotected brain regions to be "silently" destroyed; (3) humans may be at particularly high risk for this kind of brain damage, since ingestion of a given amount of excitotoxin causes much higher blood excitotoxin levels in humans than in other species; (4) in addition to the heightened risk on a species basis, risk may be further increased for certain consumer sub-populations due to youth, disease or genetic factors; (5) despite these reasons for maintaining a wide margin of safety in the use of excitotoxins in foods, no safety margin is currently being observed, i.e., a comparative evaluation of animal (extensive) and human (limited) data supports the conclusion that excitotoxins, as used in foods today, may produce blood elevations high enough to cause damage to the nervous system of young children, damage which is not detectable at the time of occurrence but which may give rise to subtle disturbances in neuroendocrine function in adolescence and/or adulthood.