THE DUPONT IMRON(R) STUDIES - AN EXAMPLE OF POSSIBLE MISREPRESENTATION OF DATA IN THE INDUSTRIAL-HYGIENE LITERATURE

There are a number of biases that may influence articles in the scient ific literature in spite of publishing policies and the peer-review pr ocesses used by most journals. One of these biases is caused by the in fluence of organizations with art economic interest in the results of the research. Publication of results influenced in this way can be con trary to the goals of public health practice. The purpose of this pape r is to present an example of an article that was published in the AIH A Journal after prior publication in anotherjournal, and in which the data may have been manipulated and misrepresented in order to benefit the organization conducting the research. Although isocyanates were kn own to cause respiratory sensitization, and cartridge respirators were not certified by the National Institute for Occupational Safety and H ealth for use with isocyanates because of their poor warning propertie s, DuPont and other manufacturers of polyurethane spray paints recomme nded the use of cartridge respirators on their paint labels in the 197 0s. During the 1970s and early 1980s DuPont carried out several experi ments at the Haskell Laboratories to demonstrate that organic vapor ca rtridges were effective in removing isocyanates, and that the odor of the solvent components of their paints could be used to detect cartrid ge breakthrough. Isocyanates broke through the respirator cartridges w ithin 15 minutes of exposure to DuPont's Imron(R) paint in some of the earlier studies, but only the results of the final set of experiments conducted in 1980 and 1981 were published. The one data point that in dicated isocyanate breakthrough in these experiments was omitted, and the author of the paper stated that ethyl acetate could be used as an end-of-life indicator because its odor threshold of 50% recognition is 13 ppm, and it was detected downstream of the respirator at a concent ration of 6 ppm. Specific recommendations are offered to modify the pe er-review process to minimize the possibility of scientific misconduct .