Appraisal of the RAND Report entitled "A Review of the Scientific Literature as it Pertains to Gulf War Illnesses, Volume II: Pyridostigmine Bromide" by Beatrice Alexandra Golomb

General

1. The RAND Report on Pyridostigmine Bromide (PB) is one of a series of reports commissioned by the Office of the Secretary of Defense, US Department of Defense (DoD) from the RAND corporation. The report was released, by the US DoD, on 14 October 1999. It can be found on the internet at http://www.rand.org/publications/CT/CT164/

2. It was carried out jointly by the RAND Health Center for Military Health Policy Research and the Forces and Resources Center of the National Defense Research Institute. The other reports in this series deal with other factors which have been considered to be possible causes of Gulf Veterans Illnesses (GVI) and include chemical and biological warfare agents, depleted uranium (DU), pesticides, oil well fires, immunisations, infectious diseases and stress. In this appraisal reference to the RAND report refers specifically to the RAND report on PB.

3. The Report on PB was prepared by Beatrice A Golomb, a RAND consultant. Dr Golomb is a staff physician at the San Diego VA Medical Centre who also has a Ph.D. in biology specialising in neurobiology. She is an assistant professor of medicine at the University of California, San Diego and a research associate professor in the USC psychology department.

4. The RAND report, which is over 300 pages long, is a review of the scientific literature and does not report any new scientific research or studies. The scope of the review is comprehensive and has addressed most, but not all, of the known scientific information drawn from over 900 literature citations. It does not propose new hypotheses but brings together much of the speculation and debate that already exists in the literature.

Aims

5. The aim of the report, stated in general terms in its conclusions as " to identify theories which may relate use of PB to development of chronic illness in PGW (Persian Gulf War) veterans", shows that the author has not set out to assess critically the possible causes of GVI, but simply to identify those that might in theory exist. The report’s summary states that: "If sufficient evidence cannot be marshalled to rule out a hypothesis, this does not imply that it is necessarily a causal factor, only that the possibility cannot be dismissed". By taking this approach, the report fails to adopt a critical approach to the hypotheses proposed, and rules out only one, i.e. bromism. All others are given equal weight.

Hypotheses

6. The hypotheses considered within the report are listed below. It is important to note that some of the hypotheses are not mutually exclusive. A range of other hypotheses are put forward by the report which are sub-sets of the hypotheses listed below.

Hypotheses relating to mechanisms of possible increased susceptibility to pyridostigmine bromide.

1. The possible entry of PB into the brain as a consequence of increased permeability of the blood-brain barrier produced by stress.

2. Physiological differences between individuals that might lead to different degrees of susceptibility to PB and its known effects.

3. Interactions between PB and other environmental chemicals to which individuals might have been exposed.

Hypotheses relating to mechanisms that might produce possible chronic effects of PB.

4. The ingestion of PB might lead to neuropsychiatric symptoms produced by the bromide in PB (bromism).

5. That veterans have developed "multiple chemical sensitivity " (MCS), (a medical condition that is not universally recognised and has a symptomatology similar to that of PGW veterans).

6. The known pharmacological effects of PB disrupt the neuromuscular junction.

7. The known biochemical effects of PB (AChE inhibition) lead to changes in the regulation of neurotransmitter systems in the body.

Methodology, approach and assessment

7. The approach and methodology adopted for the report lack the academic rigour that would normally be expected. The author has identified most of the available scientific literature but has also included a number of hypotheses seeking to link GVI to pyridostigmine, for which no peer-reviewed evidence is presented. This is very unusual in a scientific paper.

8. The scientific literature cited in the report has not been critically assessed in a consistent and rigorous manner. Most papers are given similar weight resulting in poor and less robust studies being given too much credence, whilst some very robust studies are given limited attention. There is little attempt to critically appraise the scientific value or validity of the literature reviewed with regard to experimental design and statistical validity. Many of the design faults in studies are reported, but the author does not use this information to judge the value and weight of the study, eg., in one instance the report clearly states that a cited reference is severely limited by a lack of control subjects but nonetheless, goes on to suggest the development of MCS in Gulf veterans (page 150, lines 1-8). Hence, in general, the conclusions are uncritical, unfocused and often unsupported by the evidence.

9. There are many examples of re-interpretation or misquoting of the original papers in the cited literature. For example, the author has reworked the statistical analysis of a paper from CBD Porton Down (Gleadle, Kemp and Wetherell, 1983) using inappropriate methodology to prove something that the original authors did not claim (page 44, lines 8-21); there is re-interpretation of some research findings from a paper from Guy’s, King’s and St Thomas’s Medical School (Unwin et al, 1999) in which the RAND report has not used the conclusions provided by the authors following detailed statistical analysis, but has selected certain other features which support a different angle (page 65, lines 12-36); and occasional misquoting (Gordon et al, 1983) where the term neuropathic has been replaced in the RAND report with neuropsychiatric, which has a completely different meaning and is crucially important and critical to subsequent arguments (page 69, Chronic Effects). This weakens the value of what is undoubtedly a very significant compilation of existing scientific literature.

10. The report also introduces considerable speculation based upon poorly founded, and in some cases, scientifically inappropriate extrapolation. For example, by the selective citation of research literature and the misrepresentation of studies in animal species, a false case has been developed. Species differ in their response to drugs and to allow for this, doses are adjusted to achieve pharmacological equivalence in response. Provided that one understands the basis for this, extrapolation between species is acceptable and provides the basis for much drug development and safety testing in the pharmaceutical industry. In the RAND report, the author uses the literature inappropriately by using examples, where doses are not always equivalent, to imply that extrapolation between guinea pigs, rodents and man is not sound, eg., the report states that the extrapolation of oxime effects from guinea pigs to primates is problematic (footnote, page xxi, summary). The report does not reference the source for this statement. There is little evidence to support this speculation and it is the MOD view, based on research in many allied laboratories, that there is no substance in the author's argument. Guinea pigs were chosen, by all allied nations, because they were indeed a good model for primates. It is of concern that the report does not mention at all that mice and rats are an unacceptable species for nerve agent studies, as they are much more resistant to nerve agents than other species. This is a consequence of them having very high levels of carboxyesterase in their blood stream.

11. The report is written from the point of view that PB has indeed been shown to be related to Gulf War illness. Hence, even when evidence which does not support this argument is presented, it is frequently followed by a statement that the evidence does not diminish the possibility that PB has an effect. This approach is compounded by omission of some important studies and evidence from Canada, Australia and the USA which discuss the benefits and effectiveness of PB (Goss Gilroy Inc: Study of Canadian Gulf War Veterans; Dawson (1996); Madsen (1998)).

Role of Pyridostigmine Bromide

12. The report proposes that there is no need to use PB unless there is a threat of attack from the nerve agent soman. There is much research within NATO nations, including the USA, which does not support this statement. For example, the US Army Medical Research Institute for Chemical Defense (Madsen 1998) advised US army clinicians on the effectiveness of PB against a range of agents. Therapeutic oximes alone do not provide universal, broad spectrum, protection against all nerve agents. The current UK and US oximes provide excellent protection against sarin and VX for example, but less so against other agents. In consequence, there is a clear need for pre-treatment to provide maximum protection against a broad range of nerve agents and PB is the only method currently available.

13. The report also suggests that PB reduces the protective effect of oximes against sarin. In some studies with PB the level of protection, with oxime administration, was