Hans Eysenck book review in New Scientist, 14 April 1983:
Lead versus health, Edited by M.Rutter and R.Russell-Jones, Wiley, pp 379, £18.50
The lead scandal, By Des Wilson, Heinemann Educational, pp182, £12.50, pbk £3.95
Both these books deal with the question of whether low-level exposure to lead has important effects on behaviour and intelligence, and both come to a rather positive conclusion. Apart from these similarities they could not be more different.
Des Wilson is not a scientist but an activist; he is indeed, as is said in the book, “one of Britain’s most experienced and best-known campaigners on social issues”. The Lead Scandal is manifestly a propaganda effort, as the title suggests, and his presentation of the evidence is rather one-sided, as well as being emotional and full of accusations of scientists and politicians with whom he disagrees.
It is a very effective package, paperbacked with a threatening kind of picture on the cover: it has a subtitle (“The fight to save children from damage by lead in petrol”) which begs the question, and although it contains nothing but the truth, it certainly does not contain the whole truth.
Lead Versus Health, edited by Michael Rutter and Robin Russell-Jones, is written in a much more scientific and reasonable manner, and its 15 chapters, by various experts who are well-known as researchers in this field, give as comprehensive and impartial a survey of the evidence (including new evidence) as we are likely to get from any source.
This makes the conclusion reached all the more impressive, and certainly forces one to take seriously the possibility that even low-level exposure to lead may have gravely deleterious effects on children’s behaviour and intelligence.
The issues discussed are of considerable interest, for various reasons. The most obvious of these, of course, is the practical one. If indeed there are dangers to the health of our children in exposure to low levels of lead. and if we can reduce this exposure by banning lead additives from petrol, then something can and should be done.
For many people, this practical issue will overshadow the refinements of scientific investigation, but this would be a pity. Epidemiology presents many interesting problems to the scientist, and the controversy that has arisen around the effects of lead in petrol on the health of children highlights many of the problems and difficulties in this field. These problems figure much more prominently in the Rutter and Jones book, and quite rightly; Wilson takes the answers for granted, and hence is much less impressed with the difficulties and complexities of the scientific search for an answer.
Scientifically, there are essentially two questions. The first relates to whether there is a correlation between lead levels and behaviour and intelligence in children with blood levels below 35 pigsdl. No one doubts the toxic properties of lead above such levels, and that is not an issue in this debate.
The establishment of such a correlation is not as easy as it might seem at first, because lead is only one (and probably not one of the more important) influences on the behaviour and intelligence of children, so that the correlation would at best be low, and hence require a large number of cases to be established at an acceptable level of statistical significance. For the same reason many other factors have to be controlled in order to rule out the possibility that the observed correlations might not be due to these other causes. It is only the most recent work that has been of an acceptable standard in this respect, and the finding may be summarised as showing a modest correlation between the level of lead in the blood and hyperactivity in children, and between high blood levels and low intelligence, with the latter conclusion more firmly established. Observed correlations are usually somewhat higher than correlations corrected for various artefacts, but even the corrected correlations show a relationship of the kind posited.
I don’t think it is possible to disagree about the high probability that a relationship exists, and that indeed children with higher levels of lead in the blood tend to be less intelligent (and possibly more hyperactive) than children with lower blood lead levels, although the differences are quite small, amounting to some 3 or 4 points of IQ.
The existence of a correlation. of course, does not establish the existence of a causal relationship, and even when a causal relationship is suggested the correlation itself does not establish the direction of the causal relationship.
This problem has always bedevilled much of the research on epidemiology. The correlation between smoking the lung cancer is well known, and is usually interpreted in a causal manner: can we interpret the equally strong correlations between eating meat and cancer of the large intestine, or between sugar and breast cancer along similar causal lines? Alternative hypotheses usually exist, and have to be considered.
It is one of the weaknesses of the Wilson book, and unfortunately also of the Rutter and Jones book, that obvious alternative hypotheses are not considered—possibly because no research has been devoted to the elucidation of the causal chain.
One obvious alternative hypothesis, reversing the causal chain, would be that nonacademic, extraverted children (an alternative description of the hyperactive, low IQ type) are more likely to go out and encounter the hazards of lead in petrol. It does seem a pity that no one has bothered to look at the actual behaviour of the children in question in relation to such practices as might lead them to encounter greater dangers of lead pollution; the almost entire absence of psychologists among the medical and other experts engaged in this research may account for their overlooking a very obvious alternative hypothesis.
Oliver David and others demonstrated that penicillamine, a chelating agent whose main therapeutic action concerns its effect in increasing the loss of lead from the body, produced a curative effect on hyperactivity in children, comparable with that of methlyphenidate. This provides an important argument for the lead-causes-hyperactivity hypothesis. But there are several weaknesses in the experiment, and no measures of IQ were taken, so that the causal influence of lead on IQ is not clarified in the experiment. However, in principle this would seem to be the most promising avenue to explore to furnish us with a better understanding of causal relations.
The observed correlation of lead in the blood with IO is slight, and the few points of IQ apparently lost through lead uptake may seem negligible. But, because of the statistical properties of the normal curve of distribution, they do have a large effect at the extremes.
It seems a pity that investigators have relied almost entirely on traditional IQ tests, which can be affected by impulsivity and other personality factors related to hyperactivity; it is quite possible that the effects of lead on hyperactivity may be responsible for the apparent loss of IQ, even if we acknowledge a causal influence of lead on hyperactivity.
Perhaps the traditional type of IQ test in this context should be supplemented by the newly developed psychophysiological measures of intelligence involve ing the evoked potential of the EEG (New Scientist, vol 85, p 308) and other similar measures. Such tests, used in a design similar to that of David employing chelating agents, would perhaps get us nearer to the point where truly causal relations could be studied.
In summary it seems that a statistical relationship between low-level exposure to lead on one hand, and low intelligence and, possibly, hyperactivity on the other, has been established. but that the causal relation is still an unresolved issue. How this could or should affect the political decision about banning lead in petrol is not a question which I am competent to answer; it involves too many non-scientific aspects.
What is clear, I think, is that to refer to the whole debate as a “scandal” is grossly to exaggerate the position, and it disregards the very great scientific problems thrown up by the apparently simple question of whether low-level lead exposure does indeed produce the alleged effects.