Vol. 7 No. 1 March 1987



I. EPA review of 2,4-D article by Hoar et al.
II. Antibiotic Resistance and Human Infections
III. Occupational Asthma from Inhaled Egg Protein -- Iowa
IV. 1986 Surgeon General's Report: The Health Consequences of Involuntary Smoking
V. Trends in Mortality from Cirrhosis and Alcoholism - United States, 1945-1983
VI. Tularemia -- New Jersey
VII. Communicating Concepts in Toxicology


This newsletter contains a variety of short articles concerning a wide range of toxicology issues. I have also included an article on Tularemia, because it points out a common perception that people often have about the role of toxic substances, and that is that they are the first, and most likely suspect cause of inexplicable death, disease, or environmental problems. Too often, we forget about other causes of illness, and this may be due in part to the great attention toxic substances receive from the communications media. Two of the articles will be of particular interest and these are the review of the Kansas 2,4-D epidemiological study, and the review of a very recent report published about transfer of antibiotic resistant bacteria from animals to humans. Copies of the antibiotic resistance articles may be obtained from my office if you would like to see the original. As always, feedback from readers is always welcomed as are articles which might be included for distribution.

I will close this introduction with a plea to all Cooperative Extension personnel and other readers, to please refrain from using the term "toxics". The term is a media invention, and is really just the corruption of an adjective into a noun. The word "toxic" is an adjective. The appropriate word to use is "toxicants", or "toxic substances". Many toxicologists find the term "toxics" irritating. Help stamp out irritation among toxicologists (and English teachers too). Avoid the term "toxics"!

I. EPA review of 2,4-D article by Hoar et al.

NOTE: This article is reproduced in its entirety from a copy of the review which was conducted by Dr. Brian MacMahon, who was contracted by EPA to review the study. It was obtained under the Freedom of Information Act and is the best overall review of the meaning of this recent, widely publicized article about the possible link between herbicide exposure and cancer in farmers. Portions printed in bold-face or italics were done for emphasis, and were not printed in that way in the copy of the original review.

Review of Hoar et al.1 and related literature

This review is prepared in response to EPA Purchase Order 6W-3948-NASA dated September 10, 1986. According to Jerome Blondell's letter accompanying the purchase order: "The key question is: What does the 'weight of evidence' say about the risk of lymphoma for agricultural workers exposed to 2,4-D? Is 2,4-D a likely cause of lymphoma?" This question was prompted by the referenced publication of Hoar et al in the September 5, 1986 issue of JAMA.

Hoar et al

This is a population-based case-control study of all male cases of soft-tissue sarcoma (STS), Hodgkin's disease (HD) and non-Hodgkin's lymphoma (NHL) identified in the State of Kansas over a 7-year period. 3 controls, matched for age and living-or- dead status, were selected - either by random digit telephone dialing (for living cases under 65 years of age), from Medicare files (for living cases 65 or older) or from Kansas state mortality files (for dead cases). Information on occupation and exposure to herbicides was obtained by telephone interview - with the case or control for half of the subjects with STS or NHL (and corresponding controls) and one-third of the HD cases and controls, and with the next of kin for the remaining, deceased subjects. This study shows every indication of having been carefully and competently carried out. I see no methodologic problems that are likely to have produced the reported positive association between use of herbicides (predominantly uracil and phenoxyacetic acids) and NHL. The strong and statistically significant increasing risk of NHL with increasing frequency of herbicide use (days per year) supports the idea that the association is real, but the weak, and barely significant association with years of use argues somewhat against it. There are some points of detail which should be noted, although none jeopardize the principal findings, so far as I can judge:

- presumably to have series of the three tumors of approximately equal size (200, 173 and 200), the investigators selected a sample of 200 cases of NHL from the 293 available. The relevance of this sampling is that, if the investigators had had any inkling of what their results would be, they would probably not have discarded 93 cases of NHL, and it must be presumed that it was not an a priori hypothesis that an association would be found only for NHL.

- there is an unexplained, but statistically highly significant, difference between the three groups of cases in the proportion of identified cases which were interviewed. This is primarily due to the low proportion of NHL cases which were excluded, either because they were not confirmed histologically (i.e. were not eligible) or because, if eligible, they were not interviewed. The differential loss occurs at several levels. Thus, the percentages of STS, HD and NHL cases not histologically confirmed were 19, 15 and 10 percent, respectively. Of the eligible cases, the percentages not interviewed were 4, 8 and 1, respectively. It is difficult to see any relevance of these differences to the study conclusions, but it is curious that determination of eligibility and success in interviewing were both more complete in the group of cases for which an association is found.

- for a high proportion of subjects (50% of cases of STS and NHL and their controls), the exposure information was obtained from surrogates since the subjects themselves were dead. One would suspect that surrogate-supplied information on occupation would be reasonably accurate, but one must question surrogates' knowledge of what specific herbicides were used and on how many days of the year. Since any inaccuracy involved would presumably apply to both cases and controls this cannot be regarded as a possible explanation of the association noted. In fact, it would tend to reduce any true association that exists. One might even wonder, in fact, whether it could explain the lack of association found for STS and HD - tumors for which others have reported associations with phenoxyacetic acid exposures.

- although both years of herbicide use and days of use per year show statistically significant trends for NHL risk, it is useful to note the small numbers on which these trends rest. Only two individual categories show significant differences between NHL cases and controls - use for 16 years or more (based on 16 cases, RR 2.0 and of marginal statistical significance), and use for 21 or more days per year (based on 7 cases, RR 6.0 and more clearly significant). It is not stated to what extent these categories overlap - i.e. contain the same individuals - but it is noteworthy that, in the most persuasive category (use for 21 or more days per year), where there are 7 cases, the expected number based on the controls would be about 2.3. It would take only 2 or 3 cases misclassified to this category (or controls misclassified out of it) to render the difference not statistically (or biologically) significant.

- the authors' Table I shows that farmers for whom no use of herbicides were reported had RR higher than non-farmers (RR = 1.3) and while this is not formally significant (at p less than 0.05) it is approaching significance (the lower 95% confidence limit is 0.8) and is not much different from that for all farmers (1.4). The latter similarity results, of course, from the small size of the group of herbicide users that does show substantially increased risk.

- the paper's Table 2 shows risk ratios associated with every-use of specific herbicides. Among 8 groups of herbicides (including a category "nonspecified"), the RR associated with phenoxyacetic acid is lower than that for any other group except the uracils. The RRs range from 1.3 to 12.5, that for the phenoxy acetic acids being 2.2. Besides the phenoxyacetic acids, RRs significantly above 1.0 are seen for triazines (2.5), amides (2.9), trifluralin (12.5) and nonspecified (5.8). The focus on the phenoxyacetic acids seems to stem from their frequency of use (second only to the uracils), rather than from the level of risk associated with their use.

In summary there are some questions and uncertainty in the data from this study - as there are in all epidemiologic studies - but, if there were no other evidence available, this study would stand as a good basis for the hypothesis that the risk of non-Hodgkin's lymphoma is increased by agricultural exposure to the phenoxyacetic acids - principally 2,4-D - and perhaps other herbicides. I would not accept this study as grounds for concluding that such associations do exist - only as a basis for hypotheses which must be tested in other data.

Other Studies

Paradoxically, it is unfortunate that this study is not the only one to provide evidence on this topic. In fact this study was prompted by previous studies suggesting that STS, HD and NHL were all increased in persons exposed to phenoxyacetic acid and other herbicides. It is when one tries to fit the results of the Kansas study into the context of previous work that matters become difficult.

I do not believe that the authors' conclusion that "the study confirms the reports from Sweden and several US states that NHL is associated with farm pesticide use, especially phenoxyacetic acids" is justified. The Swedish studies of Hardell et al2 showed elevated risks of 5- 6-fold for all three cancers investigated by Hoar et al. Exposure in the Swedish study was defined as "ever exposed" - principally on the basis of occupational history - and it is not possible to compare levels of exposure in the two studies to determine whether lower exposures could account for the lower RRs found in the US study (among all exposed). However, the important discrepancy is that the Swedish study found significant associations for all three tumors and the US study only for one. Before concluding that the US study is confirmatory of the Swedish one with respect to NHL one must understand the reason for the discrepancy with respect to STS and HD. The reasons for these discrepancies - whether in the exposures studied, the method of study, or simply chance - are as cogent as is the agreement with respect to NHL. Until there is an adequate explanation for the discrepancies one can have little confidence that the agreement represents reality.

It is beyond the scope of this contract to review all the published relevant literature but perusal of the articles accompanying the review request does not lead to any clear impression of support for or evidence against the conclusion of Hoar et al. Pearce et al3 report a case-control study of 83 cases of NHL and conclude that there was no significant difference between cases and controls with regard to potential exposure to phenoxy herbicides. However, in this relatively small study, the results (RR 1.4, 90% CL 0.7-2.5) are not statistically incompatible with the RR (2.2) reported by Hoar et al for ever- use of phenoxyacetic acid.

Lynge4 reports a cohort study of persons exposed in the manufacture of various pesticides. The numbers are very small but are more suggestive of an association for STS (obs. 5, exp. 1.84) than for lymphoma (HD and NHL not distinguished) (obs. 7, exp. 5.37) among all employees, and there was no case of NHL among the 41 cancer deaths among persons employed in the manufacture and packing of phenoxy herbicides specifically. The total number of cancer deaths expected in this group was 41.46. Lung cancer showed a significant excess (obs. 12, exp. 6.11).

Other studies, because of small numbers, lack of specificity of exposure and/or other reasons, carry little evidential weight.

The key question

The key question in Mr. Blondell's letter quoted earlier is in fact two questions - what does the 'weight of evidence' say about the risk of lymphoma for agricultural workers exposed to 2,4-D, and is 2,4-D a likely cause of lymphoma? The second question cannot be answered (except perhaps by animal experiment) until the first is answered, since without an association there is no causation.

In my opinion the weight of evidence does not support the conclusion that there is an association between exposure to 2,4-D and NHL. It is axiomatic that, except when relative risks are very high - and sometimes even then - no single study will establish an association between an exposure and an outcome. The acceptance of an association depends on a number of studies showing consistent results across populations and across different epidemiologic methods. The study of Hoar et al is a strong study - strong enough on its own to establish a hypothesis of relationship of exposure to 2,4-D with some small proportion of cases of NHL - a hypothesis that clearly deserves attempts at refutation or support in other populations. When one attempts to place the results of this study among the results of those published previously, the picture becomes very confusing - much more so than if Hoar et al had been the only study published. Taken as a whole, I believe that the weight of evidence indicates that an association between 2,4-D and NHL remains a hypothesis that is still to be tested. I am unwilling to speculate as to whether 2,4-D causes NHL (or some cases of NHL) until the evidence is clear that there is an association between them.

Brian MacMahon, M.D., Ph.D.
September 29, 1986

1 Hoar SK, Blair A, Holmes FF et al. Agricultural herbicide use and risk of lymphoma and soft-tissue sarcoma. JAMA 1986; 256:1141-7.

2 Harden L, Eriksson M, Lenner P, et al. Malignant lymphoma and exposure to chemicals, especially organic solvents, chlorophenols and phenoxy acids: a case-control study. Br J Cancer 1981; 43:169-76.

3 Pearce WH, Smith AH, Howard JK, Sheppard, RA, Giles HJ, Teague CA. Non-Hodgkin's lymphoma and exposure to phenoxyherbicides, chlorophenols, fencing work, and meat works employment: a case- control study. Br J Ind Med 1986; 43:75-83.

4 Lynge E. A follow-up study of cancer incidence among workers in manufacture of phenoxy herbicides in Denmark. Br J Cancer 1985; 52:259-70.

II. Antibiotic Resistance and Human Infections

A recent article published in the New England Journal of Medicine reports evidence that drug resistant Salmonella newport from dairy farms was responsible for an epidemic of human Salmonellosis cases in California during 1985. The authors (all fourteen of them), traced the particular strain of Salmonella to hamburger which contained the bacteria, and then back to the slaughtering plant, and ultimately to a small number of dairy farms. The presence of a unique plasmid (small portion of DNA which can confer antibiotic resistance to bacteria) made this tracing possible. The authors also reported that dairies which had used chloramphenicol to treat calves, had the highest incidence of this particular resistant bacteria. The study appears to have been well done, and the route of transmission to humans well established. The authors note that thorough cooking of contaminated meat will prevent any possibility of bacterial transmission, and also mention that other procedures to identify infected animals before slaughter, would also be beneficial.

I found this study particularly interesting for two reasons. The first was that this epidemic was clearly traced, and was presumably due to the use of chloramphenicol by dairies. Chloramphenicol is not registered for use in food animals in this country, but is extensively used in food animals by other countries. The FDA did not allow the use of chloramphenicol in food animals because of potential residue problems in foods, and the potential of these residues to cause human illness. FDA did not focus on possible emergence of resistant bacteria, and other countries will now have to consider this. Secondly, this was not caused by subtherapeutic use of antibiotics (low level feeding), but apparently by the use of therapeutic doses, a situation much more likely to lead to the emergence of resistant strains.

Additionally, in late 1985, the FDA came out with a very strong program to discourage the use of any Chloramphenicol in any food animal, and over the last year, the word has been spread very effectively. I don't think that this antibiotic will be used in food animals in this country any more, although it is still available for use in companion animals (dogs, cats, etc.). While the possibility of other drugs which are registered for use in food animals and in humans might indeed follow similar scenarios, it is possible to prevent this by proper preparation and processing of animal derived foods.


Chloramphenicol-Resistant Salmonella newport Traced Through Hamburger to Dairy Farms: A major persisting source of human Salmonellosis in California. New England Journal of Medicine, 316(10):565-570, 1987.

III. Occupational Asthma from Inhaled Egg Protein - Iowa

In January 1984, workers at an Iowa egg processing plant requested an investigation by the National Institute for Occupational Safety and Health (NIOSH) of the causes of "asthma- like" symptoms (wheezing, shortness of breath, tightness in chest) believed to be work-related. This plant daily processes up to a million and a half raw eggs into powdered whole egg or powdered egg yolk and liquid egg white.

After an initial site visit in March 1984, NIOSH investigators returned in August 1984 for an environmental and questionnaire survey. They sampled for total and respirable dust and for several chemicals because the original request had listed cleaners, sanitizers, and germicides as possible irritants.

Results showed employees' levels of exposures to dust near the American Conference of Governmental Industrial Hygienists' guideline of 10 mg/m3 for total dust. Dust samples had a 50% protein content and an amino acid composition resembling egg yolk protein.

Based on medical examinations and clinical histories showing temporal association with workplace exposures, the physician diagnosed five employees as having occupational asthma.

NIOSH made specific recommendations for local exhaust ventilation to control egg dust during plant operations and recommended appropriate medical therapy for selected individuals. They also reported the problem to all other plants producing dried egg products in the United States, to the trade association representing the companies, and to major trade unions representing the workers.

Editorial Note: Chicken egg white is a common allergen; ingestion may provoke pruritus (intense itching) in atopic (allergic) individuals and exacerbation (worsening) of atopic (allergic) dermatitis (inflammation of the skin), rhinitis (inflammation of the nasal passages), urticaria (hives), angioedema (swelling of the blood vessels), and bronchial asthma. Egg yolk also contains proteins antigenically related to proteins in egg white. A previous report of allergy to inhaled egg protein involved eight of 13 bakery workers who developed respiratory symptoms from spraying meat rolls with a 25% mixture of egg white and yolk in water. The current investigation, by contrast, identified five asthmatic individuals at the egg processing plant: four were nonatopic; all had evidence of lgE-mediated allergic reactions to egg protein; and only one had an elevated total serum lgE level.

Liquid egg products are dried at 23 plants in the United States. An estimated 1,600 workers may be exposed to powdered egg dust in this industry (U.S. Department of Agriculture, personal communication). Currently, no standard exists for occupational exposure to egg protein, and no generic standard has been established for occupational exposure to dust of organic origin. The only enforceable standard applicable to this situation is the Occupational Safety and Health Administration's nuisance dust standard of 15 mg/m3. By definition, nuisance dusts are presumed to be biologically inert. Consideration must be given to developing a strategy for controlling adverse health effects from exposure to powdered egg dust in this industry.

MMWR, Vol. 36/No. 2, January 23, 1987.

IV. 1986 Surgeon General's Report: The Health Consequences of Involuntary Smoking

Inhalation of tobacco smoke during active cigarette smoking remains the largest single preventable cause of death and disability in the United States.

Data in the 1986 report present evidence that the chemical composition of sidestream smoke (smoke emitted into the environment by a smoker between puffs) is qualitatively similar to the mainstream smoke inhaled by the smoker and that both mainstream and sidestream smoke act as carcinogens in bioassay systems. Data on the environmental levels of the components of tobacco smoke and on nicotine absorption in nonsmokers suggest that nonsmokers are exposed to levels of environmental tobacco smoke (ETS) that would be expected to generate a lung cancer risk. In addition, epidemiological studies of populations exposed to ETS have documented an increased risk for lung cancer in those nonsmokers with increased exposure. Of the 13 epidemiological studies that were available for review in the scientific literature, 11 reported a positive relationship and six of these observed statistically significant results. It Is rare to have such detailed exposure data or human epidemiologic studies on disease occurrence when attempting to evaluate the risk of low-dose exposure to an agent with established toxicity at higher levels of exposure. The relative abundance of data reviewed in the report, their cohesiveness, and their biologic plausibility allow a judgment that involuntary smoking can cause lung cancer in nonsmokers.

The 1986 Surgeon General's Report comes to three major conclusions:

Involuntary smoking is a cause of disease, including lung cancer, in healthy nonsmokers.

Compared with children of nonsmoking parents, children whose parents smoke have an increased frequency of respiratory symptoms and infections. They also have slightly smaller rates of increase in lung function as the lung matures.

Simple separation of smokers and nonsmokers within the same air space may reduce, but does not eliminate, ETS exposure.

The Surgeon General, in his preface to the report, states, "Cigarette smoking is an addictive behavior, and the individual smoker must decide whether or not to continue that behavior; however, it is evident from the data presented in this volume that the choice to smoke cannot interfere with the nonsmokers' right to breathe air free of tobacco smoke."

Editorial Note: A review recently published by the National Academy of Sciences states that approximately 20% of the estimated 12,200 lung cancer deaths occurring annually in nonsmokers are attributable to environmental tobacco smoke.

MMWR, Vol. 35/No. 50, December 19, 1986

V. Trends In Mortality from Cirrhosis and Alcoholism - United States, 1945-1983

Approximately 10.6 million adults in the United States can be classified as alcoholics, and an additional 7.3 million either are alcohol abusers or have experienced negative consequences of alcohol use such as arrest or involvement in an accident. In addition, an estimated 4.6 million young people aged 14 to 17 are problem drinkers. The public health consequences of problem drinking include injuries and deaths from fires, falls, drowning, homicide, suicide, family abuse and other violence as well as industrial and motor vehicle accidents. An estimated one-third to one-half of all unintentionally and intentionally injured adult Americans involved in accidents, crimes, and suicides had been drinking alcohol. Problem drinking also causes medical damage including pancreatitis, nutritional deficiencies, malignancies, fetal alcohol syndrome, and cirrhosis (the ninth leading cause of death among adults in the United States).

Per capita rates of alcohol consumption rose approximately 21% during the 1960s and 10.3% during the 1970s. Data from 1977 through 1984 (the latest year for which complete data are available) show that overall per capita consumption reached a plateau in 1980 and 1981 and then declined until 1984. The 1984 consumption rate, which approximated that of 1977, was estimated at 2.65 gallons of absolute ethanol per year for U.S. residents aged 14 or older.

ALC NOTE: Some quick calculations on this consumption data reveal the following dose information for the populace. The calculations assume as the report does, that everyone over the age of 14 is included, and that the standard weight of a person is 70 kg. 2.65 gallons of ethanol converts to 237.4 oz/person/yr, which is equivalent to 6,737.4 grams of ethanol/person/yr. This converts to a dose of 96 g/kg/year, or 0.264 g/kg/day (264 mg/kg/day). I cannot think of another depressant drug, teratogen, hepatotoxicant, or carcinogen that rivals this dosage rate in humans, and the intake is voluntary.

MMWR, Vol. 35/No. 45, November 14, 1986.

VI. Tularemia - New Jersey

On December 3, 1985, a 67-year-old woman died from tularemia in a New Jersey hospital. On admission, she had an "unhealing sore" on the first finger of her right hand. Initial treatment included gentamicin and cefazolin, as well as insulin for uncontrolled, late-onset diabetes. After 3 days, the treatment was changed to streptomycin. Despite these measures, disseminated intravascular coagulation, respiratory failure, and hypotension developed, and the woman died.

The case history showed that on November 9, 1985, an 18- year-old neighbor had shot two rabbits behind his home in Gloucester County, New Jersey. After eviscerating the animals, he gave them to the patient and her 64-year-old husband, who skinned and froze the rabbits. During the summer, the young man had noticed several dead rabbits around his house and had attributed their deaths to insecticide that had been sprayed on local fields. One of the two rabbits he shot was noted to be losing its fur.

Two days after dressing out the rabbits, the young man became ill with an ulcerated hand lesion, axillary lymphadenopathy, and a fever. He was examined at the local hospital; no diagnosis was made, but he was treated with antipyretics. On November 23, his two neighbors--the recipients of the rabbits--were admitted to the local hospital. They both had sepsis and hand lesions. On November 26, following instructions from the hospital, the young hunter was started on streptomycin, and he recovered rapidly.

MMWR, Vol. 35/No. 48, December 5, 1986.

VII. Communicating Concepts in Toxicology

A special poster demonstration session was held at the 26th Annual Meeting of the Society of Toxicology in Washington, D.C. from February 23-26, 1987. The focus of this session was to present methods currently being used to instruct nonscientists about toxicology issues. The materials presented ranged from commercial programs developed for training employees who handle hazardous chemicals, to free brochures and pamphlets written by people from the Environmental Defense Fund. I presented the slide tape shows that were developed several years ago called "Toxicology: The Science of Poisons", and "Toxicity Testing". Because there may be interest about these materials among newsletter readers, I will give a brief overview of those that impressed me the most.


Dr. June Fessenden-Raden and her colleagues at Cornell are developing some very nice software that present a wide range of toxicological topics in interactive computer assisted instruction programs. Their programs use pictures and graphics to help emphasize points and I was impressed with the speed of the programs. They currently run on IBM-PC's with EGA cards, but will soon be available for IBM-PC's with other color graphics adapters. At this time, this software seems to be most useful for individual, self-paced training, or for use with small groups of people.


Dr. Mike Kamrin and colleagues at Michigan State University have developed an LD50 demonstration/teaching program for the IBM-PC. The program was designed for use with undergraduate courses in toxicology, and allows the user to do acute toxicity "testing" of 5 different chemicals. The program is fast, and uses graphics effectively for demonstrating dose response relationships. This program has great potential for integration with other teaching methods.


Dr. Herb Thier from the Lawrence Hall of Science in Berkeley, presented materials developed by the Chemical Education for Public Understanding Program (CEPUP). These are teaching materials developed specifically for use in public schools (jr. high and high school). The CEPUP modules focus on issues like solutions and pollution, risk perception, and waste management. All of the modules include hands-on experiments using simple chemicals to demonstrate specific chemical processes. The chemistry (science) and policy issues are discussed in relation to the outcome of the experiments. These modules have also been tested on adult groups such as the Women's League of Voters, with good success. Links of CEPUP and 4-H are underway, and CEPUP offers great potential in the adult education arena in the future.


Drs. Jim Witt and Frank Dost from Oregon State University presented a syllabus they developed for PCA and PCO training in Oregon. The booklet syllabus, titled "The Chemistry, Biochemistry and Toxicology of Pesticides" is an excellent syllabus for teaching PCA's and PCO's principles of chemistry and toxicology, using pesticides as examples. They use the syllabus in a 2-day course held in different parts of Oregon, once a year. I recommend this highly to anyone wishing to explore the topic in depth, in an easily understandable way.

If anyone wishes more information, contact these individuals directly at the addresses and phone numbers which follow:

Dr. Jim Witt and Dr. Frank Dost
Department of Agricultural Chemistry
Oregon State University
Corvallis, OR 97331
(Witt: 503-754-2906; Dost: 503-754-3791)

Dr. Mike Kamrin
Center for Environmental Toxicology C-231 Holden Hall
Michigan State University
East Lansing, MI 48824

Dr. Herb Thier
Lawrence Hall of Science CEPUP
University of California Berkeley, CA

Dr. June Fessenden-Raden
Program of Science, Technology, Society 630 Clark Hall
Cornell University
Ithaca, New York 14853

Arthur L. Craigmill, Ph.D.
Extension Toxicology Specialist
U.C. Davis