PDF version of this file: New strategies needed to cut lead pollution

Simon cited a graph of hand-picked blood lead studies that the lead industry research arm - International Lead Zinc Research Organisation (ILZRO) submitted only for the final version of the OECD lead policy development process. This version was published in 1993 but two earlier drafts, in 1991 and 1992 did not include the ILZRO information so the material was never reviewed by government and environment group reviewers (such as The LEAD Group Inc) before it was published in the OECD monograph in 1993. It has since been thoroughly criticised by a Princeton University researcher, Valerie Thomas, in an extensive article available on the web at www.princeton.edu/~vmthomas/pb_elim/pb-elimination.html (where both the graphs that Simon Grose published in his article in the Canberra Times, can be found.)

The 36 studies graphed by ILZRO were dated between 1933 and 1985, showing a steady decline in the lead levels in human blood streams, at the same time as lead began to be added to petrol in 1925 peaking in 1970. Only three of the 36 carefully selected US studies were included in the 1992 WHO (World Health Organisation) review of blood lead studies which referenced over 600 mostly US studies. It is important to question how the industry determined which studies to include and which to leave out, rather than trust the industry graph without question. The lead industry graph plotted mean blood lead levels from the different studies whether there were 10 people in the study or 500, whether they were lead workers or slum children.

It makes sense that if you plot workers blood lead levels from various studies on a graph, you will see a decline over time - if you didn't, there would be an outcry about the inability of Occupational Health and Safety Regulations to protect workers. And if you plot the blood lead levels of smelter and mining town communities, you would similarly expect to find a decline over time as ways of protecting these communities from lead have been developed. Then if you take into account that legislation reducing lead in various consumer products in the US has contributed to less lead being available in newly purchased products for children eg toys, cot paint, house paint, drink and food cans, ceramics and art materials, you would also expect a decline in blood lead levels found in these studied populations. Is it so hard to comprehend a rising blood lead level in the general population up to 1970 due to lead in petrol throughout a period of declining blood lead levels for workers, smelter communities and people exposed to particular consumer products?

A more comprehensive plot of all available studies would possibly show these conflicting trends (blood lead levels from petrol lead going up to the peak of lead petrol usage in 1970 while blood lead levels from other sources were coming down). Such a graph would always have the problem though, of appearing to compare non-comparable study groups and comparing data from a period of extremely high lead contamination levels to data with relatively uncontaminated sampling and analysis techniques from the last quarter century. As Thomas (1995) says:-

Lead contamination of blood samples typically occurred through lead contamination of skin, glassware, detergents, swabs etc. Therefore, any comparison of blood lead levels over such a long period of time when sample contamination decreased drastically will show a decrease in blood lead levels.

There are many more studies than those graphed, which report on US blood lead results in the period. For instance, not one of the 65 blood lead studies comprehensively reviewed in the first draft of the OECD lead document (1991, Appendix H - "Lead Occupational Exposure Findings, by Reference") was included even though most of them were US studies. Only three of the 36 ILZRO selected studies were included in the IPCS review (1992) by the World Health Organisation which referenced over 600 studies, again, mostly US studies.

It would be interesting to know how many of the graphed studies were funded by ILZRO. According to the OECD Lead monograph (1993, p227) "Annual ILZRO expenditures for the conduct of lead health-related research will typically range between [US] $400,000 and $600,000." Industry money had been used before to criticise the US leaded petrol phase-out policy. For instance, Needleman (1992, p 278) refers to a classic example about a report written by a man called Tozzi:

The implication from Grose's use of the ILZRO graph is that the populations sampled in the blood lead studies were representative of the general population. However, until it was recognised that leaded petrol emissions were lead poisoning the whole population, only populations who were exposed to particular sources of lead were tested, as Smith (1989, p37) confirms: "Most of the early lead studies were of social disadvantaged populations, either living in close proximity to industrial premises, or in inner-city slum areas." As can be seen from the titles of the studies included by ILZRO, some of the studies were of workers, who always have higher blood lead levels than the general population, for example:

"Lead Exposure Among Decorative and House Painters",

"Relationship of Biological Indices of Lead Exposure to the Health Status of Workers in a Secondary Lead Smelter", and

"Lead Exposure in Stained Glass Workers".

The issue has been raised that the ILZRO graph indicates that blood lead levels are independent of the amount of lead used in petrol, yet half of the ILZRO graph (the post-1970 studies) confirm the downward trend in blood lead levels in the general population as lead was phased out of petrol and during the period when contamination of blood samples was being controlled. So the only part of the ILZRO "trend" graph that conflicts with the correlation found in the analysis of the National Health And Nutrition Examination Survey (NHANES) studies (Annest et al, 1983), between blood lead levels and the amount of lead used in petrol is the period from 1935 to 1970 (when the tonnage of lead used in petrol reached its peak). The pre-1970 dot points on the graph came from just 14 studies (or 13 if the Kehoe "study" is not a separate study [see below]) and involved as few as 10 people's blood lead results per study.

Look for yourself at the ILZRO graph of blood lead study results (the little circles) and of the lead used in petrol in the US over a similar period (Figure 57 & 58 at http://www.oecd.org/dataoecd/23/50/1955919.pdf). Notice how few of the graphed blood lead studies you would have to exclude (as being invalid for the above reasons) - basically the first 6 little circles from the earliest studies at the top left of the graph - to make the "downward trend" between the mid 1930s and the mid 1950s, not a trend at all.

Notice that one little circle "carries" the trend from the mid 30s to the mid 50s. The 1946 reference appears to be the same results as have already been plotted at 1935. The reference list reveals that the 1935 study and the 1946 study both had 30 participants with a mean blood lead level of 27 µg/dL. The researchers for the 1935 study were Kehoe, Cholak and Story. The 1946 study was a conference paper by Kehoe. It seems highly likely that Kehoe was presenting the 1935 results at a 1946 conference and ILZRO counted them twice. Then look at the way ILZRO has turned their series of chosen dots into an unjustified "trend line" and laid the dubious blood lead graph over the petrol lead graph. How easy this makes it to not question the accuracy of the early studies and to forget the missing hundreds of blood lead studies from the period which were excluded for no stated reason.

The ILZRO graph is often used without reference to its source and without commenting on the highly questionable uses it has been put to, for example, by an important Australian authority, in a successful attempt to convince the Australian government that international pressures were threatening Australia's lead industry. ABARE (Australian Bureau of Agricultural and Resource Economics) used the graph in a 1994 "research report" on the implications of the OECD lead strategy for Australian lead exports. It was the only blood lead graph reprinted from the OECD lead monograph, from a choice of many others provided by national governments. As the NHANES (Annest) graph from the US was not included, the ABARE authors (Cox et al) were able to cast doubt on the general trend of reduced blood lead levels as lead in petrol was reduced. Our national resources research institution only used this industry-devised graph, and did not include a reference source for it. There follows, an extract from the ABARE Acknowledgments and some pertinent quotes from the report:

Assistance from the lead mining industry to write a government report, and the use of only an industry-generated graph, has resulted in this instance in doubt being cast on the benefits of any constraint or ban on any particular use of lead.

It is interesting to compare the above evidence to a statement made by R. D. Greenway (1998): "The authorities in Australia have used every trick imaginable to support unleaded petrol, even to doctoring graphs and charts." And then to compare that statement to the more recent accusation of a trick (following).

PDF version of this file: New strategies needed to cut lead pollution