Food chemical safety

Scientific data on the artificial sweetener aspartame

The following are supporting documents for an article by Erik Millstone, published in Archives of Public Health (July 2019) they were originally compiled by Prof Erik Millstone in response to a European Food Safety Authority (EFSA) call for scientific data on the artificial sweetener aspartame issued on 1 June 2011 as per his letter to the EFSA (republished below). In regards to the journal article:

Millstone, E; Dawson, E (2019) EFSA’s toxicological assessment of aspartame: was it even-handedly trying to identify possible unreliable positives and unreliable negatives? Archives of Public Health 77:34

See the press release: New research casts doubts on safety of world’s most popular artificial sweetener

Read Daily Mail article: Experts call for fresh ban on 'seriously flawed' artificial sweetener used in thousands of products including Coca-Cola and Pepsi over safety fears

To whom it may concern

Re EFSA’s 2013 assessment of the safety and/or toxicity of aspartame (E 951)

In response to a call for scientific data on aspartame, issued by EFSA on 1 June 2011, I provided the Secretariat of the ANS Unit a list of the contents of a dossier of 30 documents that I had then assembled, ie in July 2011, that could be of direct and important interest to EFSA’s re-assessment of aspartame’s safety and acceptability.

The following is the text of my letter to EFSA dated 11 July 2011. The table below provides a narrative commentary to the documents that comprise the dossier, with hotlinks directly to the relevant document.

Dear Sir/Madam

Re EFSA Call for scientific data on Aspartame (E 951)

In response to the call for scientific data on aspartame, issued by EFSA on 1 June 2011, I provide below a list of the contents of a dossier of 30 documents that I have so far assembled that could be of direct and important interest to EFSA’s re-assessment of aspartame’s safety and acceptability.

I have been researching into the safety and approval of aspartame since 1984, and the documents listed below is a small fraction of the set of documents I have accumulated. I estimate that I have something like 400 documents on aspartame, organised and indexed chronologically. I believe that I have the most comprehensive set of documents on the history of the testing and approval of aspartame available anywhere in Europe. If you seek documents that are not listed below, please do not hesitate to ask me, in case I have copies of them.

My view is that the documents I have assembled collectively indicate that when aspartame was initially tested by and for G D Searle in the 1970s, several of the pivotal tests were incompetently conducted and then misleadingly reported. Furthermore, when a senior US FDA toxicologist uncovered the problems, and they were then investigated by FDA task forces, evidence emerged indicating that no reliance could be placed on the supposed results of those test. The documents listed below also reveal than in response Searle, and then Nutrasweet, went to considerable lengths to orchestrate a cover-up, and those efforts have in most respects so far been successful. The information contained in these documents cast doubt on the adequacy of previous evaluations of aspartame that have been conducted by the US FDA, by the Joint Expert Committee on Food Additives of the United Nations Food and Agriculture Organisation and the World Health Organization (or JECFA), by the Scientific Committee for Food of the EEC (or SCF), by the Committee on the Toxicity of Chemicals in Food, Consumer Products and the Environment (or CoT) at the UK’s Department of Health and many others.

The information contained in the documents listed below does not prove that aspartame is unsafe, but it does show that it should never have been approved in the first place, and no one can yet confidently conclude that it is acceptably safe. It follows therefore than the EFSA should recommend to the European Commission that the use of aspartame in the EU should be suspended.

The following table identifies each of the 30 documents in the dossier, and outlines what they show. In addition, for many of these documents, a set of my observations and their significance is included.

Appendices
Additives: A Guide for Everyone (Penguin Books): Erik Millstone: 1988 Excerpt
Sweet and Sour. The Unanswered Questions about Aspartame: Erik Millstone: 1994
ASPARTAME Chronology: Senator Howard Metzenbaum’s staff: 1986 with subsequent elaborations
Who’s Who: Senator Howard Metzenbaum’s staff: 1986
Document List: Senator Howard Metzenbaum’s staff: 1986

This Document List about the G D Searle testing and approval process indicates the thoroughness with which Metzenbaum’s team detailed the shortcomings in the testing and approval processes of aspartame in the USA. Since subsequently the UK, EC/EU and the JECFA all relied on US data, and presumed the adequacy of those data and the review of them by the US authorities, they too suffer from identical problems.

Observations

The scandal about the short-coming in the conduct and reporting of test on aspartame was first uncovered by scientists from the FDA’s drug control division. The FDA decided to investigate Aspartame after one of its leading pathologists, Dr Adrian Gross, noticed what he thought was a worrying anomaly in a Searle report on a pharmaceutical product, Flagyl; the summary at the start of Searle’s submission did not accurately reflect the detailed data presented in subsequent chapters. The report was returned by Gross to Searle in the expectation that it would be re-submitted with the summary revised to fit the data. Gross was surprised when a fresh submission arrived with data altered to fit the summary!

In response, officials from the FDA’s Drugs Division made an unannounced investigative visit to Searle’s offices and laboratories. In the course of the FDA’s investigation questions were raised about the conduct and reporting of tests on the safety of Aspartame: as with Flagyl, the documents submitted by Searle did not accurately represent the conduct of the experiments which they were supposed to be reporting, and consequently may have underestimated the toxicity of the sweetener. In response the FDA established two Special Task Forces; one under the auspices of the Bureau of Drugs reviewed Searle’s safety evaluations of their pharmaceutical products, while the second under the Bureau of Foods, examined Aspartame.
Index of Studies Submitted to the FDA in Support of Aspartame: NutraSweet company: undated, obtained from G D Searle’s London representatives in 1987
Establishment Investigation Report (EIR): FDA’s initial Searle Task Force: 18 February 1978
The Establishment Investigation Report covers several studies conducted with aspartame under contract to Searle by Hazleton Laboratories Inc. based in Virginia.

Observations
- Study E77/78 - 30 serious errors noted in the conduct and reporting of that study.
- Study E32 21 - serious errors noted
- Study E28 14 - serious errors noted
- Study E27/35 - 16 serious errors noted including (p10): Observation for drug effect records are inconsistent. Records for hamster No. N9LM indicate this animal was not alive on October 23 1970, was alive on November 20, was not alive on December 18 and January 15 1971, was alive February 12, and was found dead on February 25
It is difficult to see how any reliance can be placed on studies that have been so poorly conducted and reported.
Final Report of Investigation of G D Searle Company: FDA’s Searle Investigation Task Force: 24 March 1976

The Final Report of Investigation of G.D.Searle Company reviews serious inadequacies in the general conduct of toxicity tests, conducted both by Searle and by a sub-contractor Hazleton Laboratories.

Observations

“At the heart of FDA’s regulatory process is its ability to rely upon the integrity of the basic safety data submitted by sponsors of the regulated products. Our investigation clearly demonstrates that, in the case of the G.D. Searle Company, we have no basis for such reliance now…Through our efforts, we have uncovered serious deficiencies in Searle’s operations and practices which undermine the basis for reliance on Searle’s integrity in conducting high quality animal research to accurately determine or characterise the toxic potential of its products.” (page 1)

Under Recommendation #1 the report states: “…the Bureau of Foods should make a determination on the disposition of the Aspartame studies currently under official FDA seal at Searle and Hazleton Laboratories.” (page 8)
Memo on Draft Agreement for Validation of Searle Asparame Studies: Dr Adrian Gross, US FDA: 4 November 1976

A Memo on Draft Agreement for Validation of Searle Asparame Studies from Dr Adrian Gross of the US FDA to one of his superiors, Carlton Sharp, concerning the FDA’s intention to contract out to the Universities Associated for Research and Education in Pathology Inc (or UAREP) the responsibility for reviewing 12 of the 15 safety studies on aspartame that the initial FDA investigation had indicated were problematic. The memo indicates that Gross doubted that the UAREP had the relevant expertise, and that he was doubted the wisdom of having G D Searle pay for the UAREP’s work. He was concerned that G D Searle might influence the conduct and reporting of the UAREP’s work. In the event, Gross’s concerns turned out to be well-grounded. The first few pages, including the Table of Contents, of the UAREP report appear as item 15 in the current dossier.

Observations

The memo shows that not only were the tests on aspartame unreliable, but that serious questions arise in connection with the steps taken by the FDA ostensibly to ‘validate’ the Searle studies.
Letter from Richard Merrill, then FDA’s Chief Counsel in Washington, to Samuel Skinner, then US Federal Attorney in Chicago: 10 January 1977 (the discussion of aspartame starts on page 17)
Internal memo from Samuel Skinner to his subordinates: 8 March 1977

In the internal memo from Samuel Skinner to his subordinates Skinner indicates that he has decided to recuse (i.e. excuse) himself from the Searle investigation because he had been invited to join the law firm Sidley and Austin, which was representing G D Searle.

Observations

Under those circumstances Skinner had no choice but to withdraw from the Searle investigation, but the development contributed to delays that eventually resulted in the time limit set under the Statute of Limitations expiring before the next stage of due process could be completed.
Bressler Report – first part: 18 July 1977

An Establishment Investigation Report (EIR) covering the period from 2 March 1977 to 8 July 1977 that documents in great detail serious inadequacies in two of the studies (E5 and E89) conducted on aspartame, both intended to evaluate the embryotoxicity and teratogenic potential of aspartame when administered orally to albino rats (E5) and to mice (E89).

Observations

Establishment Investigation Reports were a tool developed by the FDA, before the rules on Good Laboratory Practice were established, if serious concerns were raised about the ways in which experiments had been conducted and/or reported.
Bressler Report – second part: 15 August 1977

An Establishment Investigation Report (EIR) that documents in great detail the serious inadequacies in a pivotal test on DKP (diketopiperazine), which is one of aspartame’s breakdown products.

Establishment Investigation Reports were then a tool developed by the FDA, before the rules on Good Laboratory Practice were established, if serious concerns were raised about the ways in which experiments had been conducted and/or reported.

Observations

The study in question (PD 988S73, SC-19192) was a 115 week rodent oral carcinogenicity study on diketopiperazine, a decomposition product of aspartame. The main findings of the 77 page EIR are summarised in Appendix 1 below.
Internal FDA memo from the Bureau of Foods Task Force to the Acting Director of the FDA Bureau of Foods, Howard R. Roberts: 28 September 1977

Along with its appendices, this Internal FDA memo provides an analysis that was supposed to summarise the two Bressler Reports. This memo is, however, a curious document because, despite the numerous problems highlighted by Bressler and his colleagues, it states on page 3 that: “The differences observed and documented...are not of such a magnitude that they would significantly alter the conclusions of the studies” and “...the three studies appear to be authentic.” These conclusions are puzzling because they gloss over many of the issues raised in, and to be inconsistent with many of the facts recorded in, the two Bressler Reports.

The reason for this apparent anomaly is explained in the quotation from Dr Jacqueline Verrett that follows.

Observations

In May 1987 Dr Jacqueline Verrett, one of the signatories to the 28 September 1977 memo, provided the following explanation. This is from her statement on the record at a hearing of the US Senate Committee on Labor and Human Resources (3rd November 1987, item 24, page 383 et seq.)

Her comments explain the divergence between the summary and conclusions of the Bureau of Foods Task Force Report and the evidence on which it was supposed to have been based.

“We 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 considerations. We were supposed to figure out what the conclusions would have been if the studies had been fully and correctly reported. We were obliged to ignore the protocols and the non-homogeneity of the DKP. [i.e. the Bureau of Food Task Force was instructed to ignore the difference between the way in which the study was supposed to have been conducted and the way in which it had actually been conducted and also to ignore the failure to mix the test compound properly into the feed of the laboratory animals.] The Bressler Report did show that non-homogeneity. Some animals 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. Our remit was limited to a comparison of the Bressler data against the original data. 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.”
Authentication Review of Selected Materials Submitted to the Food and Drug Administration Relative to Application of Searle Laboratories to Market Aspartame – table of contents: Universities Associated for Research and Education in Pathology (UAREP): 18 November 1977

These are just the first few pages of the 1,062 pages UAREP document. An entire copy could be provided if the FSA finds it difficult otherwise to obtain a copy. The Table of Contents lists those studies that the UAREP reviewed.

Observations

A critique of the UAREP Report is provided by Dr Adrian Gross, the FDA pathologist, in his letters to Senator Metzenbaum which are included in the record of the US Senate Committee on Labor and Human Resources, 3 November 1987. (See item 25 below, p.430 et seq, especially pages 435 – 6)
Two letters from Dr Adrian Gross to Senator Howard Metzenbaum: 3 November 1987
Aspartame. Decision of the Public Board of Inquiry: Department of Health and Human Services. Food and Drug Administration: 30 September 1980

In 1979, in an attempt to resolve the controversy once and for all, the FDA set up a Public Board of Inquiry (PBOI) that published its conclusions in October 1980.

Observations

The PBOI confined itself to examining two questions, both relating to Aspartame’s possible effects on the brain. On brain damage resulting in mental retardation, it took the view that Aspartame consumption would not pose an increased risk of mental retardation.

On brain tumours, it concluded (by reference to data from two of the studies examined by UAREP) that it was unable to rule out 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 tests.

The conclusion of the document states that “The Board has not been presented with proof of a reasonable certainty that aspartame is safe for use as a food additive under its intended conditions of use.” (p. 49)
Aspartame – Dissenting Opinion on the Brain Tumor Issue: internal FDA memo from Robert Condon, FDA toxicologist: 19 May 1981

In the internal FDA memo Aspartame – Dissenting Opinion on the Brain Tumor Issue Robert Condon explains to his superior why he is not convinced that aspartame had been shown to be safe in respect of the risk of brain tumours.

Observations

Condon highlighted the positive results in a study (also known as E33/34) in female rats, inadequacies in the conduct of studies E70 and E78, and the limited power of those studies. This document was written after the PBOI had reported but before the then Commissioner at the FDA, A H Hayes, approved aspartame for use in the USA. It indicates that Hayes did not have the endorsement of some of his staff for his decision, and indicates why not.
Aspartame: Review of Rat Tumor Issue: Memo from Dr Douglas Park, a senior FDA toxicologist, to his superior: 19 May 1981

In the memo Aspartame: Review of Rat Tumor Issue Dr Park argues that the evidence of the brain tumour risk from aspartame was equivocal, but he explains that the information provided to the PBOI was, in some key respect, inaccurate, and that the correct information would have reinforced their concerns about the risks that aspartame consumption could pose. He concluded that aspartame had not been shown to be safe, and recommended against approval.

Observations

On 15 July 1981, Arthur Hull Hayes, US FDA commissioner, overruled the Public Board of Inquiry and approved aspartame, initially for dry products, asserting that aspartame has been shown to be safe.

In September 1983, ex-Commissioner Hayes joined Burston-Marsteller, Searle’s public relation firm.

On 8 July 1983, Acting FDA commissioner, Mark Novitch, approved aspartame for use in carbonated beverages.
Letter from Professor Richard Wurtman at the Massachusetts Institute of Technology to the New England Journal of Medicine: 18th August 1983 pp. 429 – 430)
Congressional Record: 1st August 1985
Senator Metzenbaum’s letter to the Chair of the Senate Judiciary Committee: 3 February 1986
Letter from Samuel Molinary, then the Vice President, Regulatory Affairs of the Nutrasweet Company, to Erik Millstone: 20 April 1987
Text of a Washington news article by Gregory Gordon: 12 October 1987
Examining the Health and Safety Concerns of Nutrasweet (Aspartame): Hearing Before the Committee on Labor and Human Resources United States Senate: 3rd November 1987 Cover page and table of contents
Increasing Brain Tumour Rates: Is There a Link to Aspartame?: John W. Olney, MD, Nuri B. Farber, Edward Spitznagel, and Lee N. Robins: Journal of Neuropathology and Experimental Neurology: November 1996

In Increasing Brain Tumour Rates: Is There a Link to Aspartame? The authors analysed cancer statistics from the US National Cancer Institute covering a sample of approximately 10% of the US population for the period from 1975 to 1995.

Observations

The authors found that the introduction of aspartame into the USA, into dry goods in 1981 and soft drinks in 1983, was followed by an abrupt increase in the reported incidence of brain tumours. The change was most noticeable between 1984 and 1985, and it corresponded to approximately 1,500 extra cases of brain cancer per year in the USA.

Their second main finding was that there had also been a marked change in the incidence of particular types of brain tumours, with a reduction in the proportion of a relatively less aggressive (and often preliminary) type of tumour (astrocytomas) and a sharp increase in the incidence of a far more aggressive (and all too often terminal) type of tumour (glioblastomas). The investigators argued that the reported changes in tumour incidence were unlikely to have been artefacts of improvements in diagnostic technologies. The introduction and rapid diffusion of computerised tomography in the early to mid- 1970s, and of magnetic resonance imaging technology in the early to mid-1980s, certainly improved diagnostic precision. But they contend that the impact of those innovations upon the reported incidence of these central nervous system (CNS) tumours had fully worked their way through before aspartame was introduced. Before those imaging technologies were introduced, it was far harder to diagnose brain cancer. Consequently, it was often not until tumours developed into glioblastomas that they were diagnosed, and a relatively high portion of tumours at the earlier astrocytoma stage went undetected. When the imaging technologies were introduced, brain tumours tended to be detected at the earlier stage, and consequently in the late 1970s the number of reported astrocytomas went up, while the number of glioblastomas exhibited a corresponding decline. After aspartame was introduced, however, the opposite pattern can be observed. The incidence of glioblastomas rose sharply, and starting in the late 1980s the number of astrocytomas declined even more sharply. Since those latter changes run counter to the direction which could be attributed to the introduction of better diagnostic technologies, it is hard to see how the reported changing tumour incidence could be ascribed to innovations in diagnosis. If the apparent increase in overall incidence had been due to improved diagnostics, then we should expect a marked change in post-diagnostic survival rates, but no such change was evident.

Olney and his colleagues suspect aspartame to be implicated in the aetiology of the extra cases of brain cancer for three main reasons. Firstly, the type of CNS 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 aspartame in the 1970s. Secondly Olney and colleagues also drew attention to the results of a study by Shephard et al published in 1993. (Shephard S. E. et al, ‘Mutagenic activity of peptides and the artificial sweetener aspartame after nitrosation’, Food and Chemical Toxicology, 1993, Vol. 31, pp. 323-329) thereby indicating a biochemical mechanism that could account for a tumorigenic effect, and thirdly epidemiological evidence apparently confirming the predictions previously derived from animal and in vitro studies.
Shephard S. E. et al, ‘Mutagenic activity of peptides and the artificial sweetener aspartame after nitrosation’, Food and Chemical Toxicology, 1993, Vol. 31, pp. 323-329

In the paper ‘Mutagenic activity of peptides and the artificial sweetener aspartame after nitrosation’ (1993) Shephard and her colleagues attempted to simulate in vitro the conditions that can occur in the human digestive tract, and in particular the conditions that result in the nitrosation of dietary ingredients. They reported that the nitrosated aspartame had significant mutagenic action.

Observations

This evidence may be important because it suggests not only a mechanism through which aspartame could exert a carcinogenic action, but also why the interval between the compound's introduction and the elevation of brain cancer rates appears to have been so brief.

Olney et al also suggested that aspartame may reasonably be suspected of responsibility because the other main possible candidates for responsibility, such as ionising radiation, smoke inhalation, pesticides, electromagnetic fields and various other chemicals were gradually introduced over recent decades rather than all at once in the early 1980s. Exposures to those potential hazards are, furthermore, occupationally linked and it is hard to see how they could explain why males and females seem to be equally affected.

If Olney’s hypothesis is to be substantiated, it may be helpful to analyse several long-term brain cancer time-series data sets for other countries covering the period both before and since aspartame was introduced. That has proved difficult because while aggregate brain cancer statistics are readily available, information on tumours types is hard to obtain. If aspartame were to act by modifying an already present or nascent brain cancer, we should expect its impact to vary in different countries in ways which depend on the age structure of the consumers of this sweetener and their patterns of consumption. Anecdotal evidence suggests that a far larger proportion of 50 to 70 year old Americans consume aspartame-sweetened products than is the case in the UK or in other European countries. According to Sebastian Bizzari et al of the American Dietetic Association, US residents account for about 75% of all the aspartame consumed in the world, and therefore the evidence from the USA is likely to be clearer than from other countries.
(Bizzari et al, ‘Position of the American Dietetic Association: Use of Nutritive and Nonnutritive Sweeteners’, Journal of the American Dietetic Association, Vol 104 No 2, Feb. 2004, p. 263)
Soffritti et al, 'Aspartame induces lymphomas and leukaemias in rats', European Journal of Oncology: July 2005

In the paper 'Aspartame induces lymphomas and leukaemias in rats' Soffritti et al reported evidence that aspartame induced lymphomas and leukaemias in rats, in a consistent dose-related manner.

Observations

This study is crucial because, for many years, critics of the original Searle studies have argued that the original studies that were flawed had never been repeated. The study conducted by Soffritti and his colleagues at the Fondazione Ramazzini, was initiated precisely because the pivotal studies on aspartame toxicity had been so poorly conducted in the first place, and incompetently evaluated. It therefore serves to fill a crucial gap in our scientific knowledge about the toxicity of aspartame. The Ramazzini study used a larger sample of laboratory rodents than the sum of the sizes of the samples of rodents used in all the previous chronic toxicity long-term feeding studies. In these new circumstances, it is now appropriate to conclude that sufficient information has now emerged to indicate that we can now be confident that aspartame is not acceptably safe.

While there were quite possibly some shortcomings in the studies conducted by the Fondazione Ramazzini they were far less serious than those that characterized the initial studies conducted by Searle and Hazleton laboratories in the 1970s. To continue to treat the Searle as Hazleton data as reliable but the Fondazione Ramazzini data as unreliable is profoundly unscientific.
New research data on the sweetener aspartame to be considered by EFSA's scientific experts: Press Release from the European Food Safety Authority: 14 July 2005

The EFSA press release 'New research data on the sweetener aspartame to be considered by EFSA's scientific experts' is in response to the publication of the study by Soffritti et al (item 27). The document includes the statement: “EFSA has already held initial discussions with the scientists concerned and will ask its Scientific Panel on Food Additives, Flavourings, Processing Aids and Materials in Contact with Food (AFC), as a matter of high priority, to review these results, in the context of the previous extensive safety data available on aspartame.” (emphasis added).

Observations

The phrase emboldened for emphasis in this quotation suggests that the approach being taken by EFSA and by the Scientific Panel on Food Additives, Flavourings, Processing Aids and Materials in Contact with Food is predicated on the assumption that the results of the Soffritti study can and should be juxtaposed to ‘previous safety data’. The analysis implied by this commentary and the dossier of documents that it accompanies is that the previous data do not reliably indicate safety, and therefore that those previous data do not provide an appropriate counter-weight to the findings of the study by Soffritti et al.
Press Release from the Food Standards Agency: 14 July 2005