We thank Gerr and Fethke for their response to our remarks on their editorial on self-reported versus objectively recorded computer times [1]. Gerr and Fethke continue to disregard the consistent evidence that objectively recorded computer times are much more accurate and valid than self-reported computer times when compared to external “gold” standards [2,3]. They suggest that self-reported computer times may be superior to objectively recorded computer times in capturing biologically relevant risk factors for musculoskeletal disorders. Their basic argument is that musculoskeletal disorders seem to be related to self-reported but not to objectively recorded computer times. According to Gerr and Fethke this discrepancy could be due to error in self report, differences in the kind of exposure information captured, or both. They blame us that our “claims of methodological objectivity and validity … do not address this fundamental question”.

However, we did address this question [3]. We studied if the differences between self-reported and objectively recorded computer times depended on the level of computer times, variation in computer times, level of arm pain, psychosocial work characteristics, age, gender, and personality characteristics. Concerning work characteristics, self-reported computer times increased with a high degree of variation between weeks; and a one hour increase in weekly computer time was overestimated, especially at low levels of computer time. In our opinion these findings reflect self-report biases, rather than self-report “captures” of biologically relevant information. We found no musculoskeletal health effects of computer work speed, sustained activity or micropauses or their interactions with computer times [2,4]. Of course, other factors may also be considered and our findings should be replicated.

Gerr and Fethke are concerned about the generalisability of the NUDATA results with reference to our objective median exposure times [2,4]. However, these are without any relevance to generalisabilty. The NUDATA cohort was designed to include persons with low as well as high computer work hours in order to examine internal exposure-response relations. Therefore, it is the representation of variation in computer work hours that counts, not means or medians. These problems have been thoroughly discussed in our previous publications [2,4].

Sigurd Mikkelsen
Johan Hviid Andersen

References

1. Gerr F, Fethke N. Ascertaining computer use in studies of musculoskeletal outcomes among computer workers: differences between self-report and computer registration software. Occup Environ Med 2011; 68: 465-66.

2. Mikkelsen S, Lassen CF, Vilstrup I, et al. Does computer use affect the incidence of distal arm pain? A one-year prospective study using objective measures of computer use. Int Arch Occup Environ Health 2012;85:139-52.

3. Mikkelsen S, Vilstrup I, Lassen CF, et al. Validity of questionnaire self-reports on computer, mouse and keyboard usage during a four-week period. Occup Environ Med 2007; 64:541-7.

4. Andersen JH, Harhoff M, Grimstrup S, et al. Computer mouse use predicts acute pain but not prolonged or chronic pain in the neck and shoulder. Occup Environ Med 2008;65 :126-31.

Conflict of Interest: None declared

I have read with the greatest interest the convincing study on the dose-response of cadmium ions in kidneys (1). Cadmium compounds also harm the proteoglycan metabolism (2), and by using the urinary proteoglycan excretion as an indicator of cadmium effects the threshold would be at 5 microg/g creatinine (3). This agrees very well with the threshold found in
the current investigation.

Heikki Savolainen

1 Chaumont A, De Winter F, Dumont X, et al. The threshold level of urinary cadmium associated with increased urinary excretion of retinol-binding protein and beta-2-microglobulin: a reassessment in a large cohort of nickel cadmium battery workers. Occup Environ Med 2011; 68:257-264.

2 Savolainen H. Cadmium-associated renal disease. Renal Failure 1995;17:483-487.

3 Savolainen H. Studies on urinary proteoglycan excretion in occupational cadmium exposure. Pharmacol Toxicol 1994; 75:113-114.

Dr. Burstyn, in his commentary [1], underscores the critical importance of using the best exposure assessment methods possible to minimize misclassification. We agree about the value of expert formulated models for systematically and transparently documenting exposure assessment, but caution that many existing studies may not be readily adapted to such model building. For such studies, the best alternative exposure assessment methodology should be employed, such as job-exposure matrices (JEMs) or expert assessments of self-reported work histories. Even though the relationships between the true exposure and estimates by expert assessment and a JEM are unknown (which is the case for most exposure assessments) we believe that understanding the differences between the two methods is informative, especially given the considerable time and resources necessary to carry out an expert assessment.

As Dr. Burstyn indicates [1], neither assessment approach used in our study [2] allows us to claim that lead definitely causes brain tumors. However, if this is the standard for judging the success of an exposure assessment method, most methods are failures. Although only suggestive, we do see some evidence of an association and indicate that future studies would benefit from the most accurate exposure assessment method available. The intent of our analysis was to compare two widely used approaches and to encourage epidemiologists to pursue the best exposure assessment methods possible. We acknowledge limitations with the expert assessment approach and strongly support the development and use of new exposure assessment methods. However, expert assessment may be the best approach available to an existing study and could reveal important associations that future studies can explore in greater detail using more refined exposure assessment techniques.

Parveen Bhatti
Patricia Stewart
Martha S. Linet
Peter D. Inskip
Aaron Blair
Preetha Rajaraman

1. Burstyn I. The ghost of methods past: exposure assessment versus
job-exposure matrix studies. Occup Environ Med 2011;68:2-3 doi:10.1136/oem.2009.054585. (Available at http://oem.bmj.com/content/68/1/2.full?sid=f61a67b1-71a0-4c03-95b1-008ff222dad0)

2. Bhatti P, Stewart PA, Linet MS, Blair A, Inskip PD, Rajaraman P.
Comparison of occupational exposure assessment methods in a case-control study of lead, genetic susceptibility and risk of adult brain tumours. Occup Environ Med 2011;68:4-9 doi:10.1136/ oem.2009.048132. (Available at http://oem.bmj.com/content/68/1/4.full?sid=d3fd91dc-5af2-4fe6-8b38-555288c097e0)

Sastre et al described a case of occupational asthma caused by a novel low molecular weight (LMW) respiratory sensitiser, 5-aminosalicylic acid (5-ASA) [1]. They concluded that the mechanism was probably not IgE-mediated because of a negative skin prick test and a late response to bronchial challenge testing with 5-ASA. They remarked, as is the case with several other LMW asthmagens, that the mechanism by which 5-ASA caused asthma is unknown.

Such uncertainty, and likely heterogeneity in the pathophysiological mechanisms of asthma due to LMW chemicals, is one of the reasons why no single in vitro or in vivo testing method has been developed for the prediction of asthmagenic potential. One way round this problem has been to utilise an ‘in silico’ approach that makes no a priori assumptions about pathophysiological mechanism. The development of such a quantitative structure activity relationship (QSAR) model and its initial validation was published in a previous edition of Occupational and Environmental Medicine [2]. This assigns an asthma hazard index, which is a value between zero and one, and the most recent validation [3] has demonstrated that it has good global predictive value.

By entering the chemical structure of 5-ASA (fig 1) into this model, an ‘asthma hazard index’ of 0.82 was obtained indicating a high probability of asthmagenicity.





Novel LMW chemical causes of occupational asthma such as the one reported by Sastre et al appear only intermittently in the literature but respiratory physicians have used this QSAR model to provide corroborating evidence in the identification of a novel asthmagen [4]. This model may also be of use when a clinician has diagnosed occupational asthma where the cause could be one of several LMW (organic) chemicals none of which are recognized respiratory sensitisers. By ranking the possible causes according to the QSAR-generated asthma hazard index, it may be possible to prioritise the agent(s) with which to perform bronchial challenge testing. Further evaluation of the QSAR model for this purpose is planned and we would encourage respiratory and occupational physicians to utilise the model, which is freely available on the internet [5] for such purposes, and register their interest and observations by email with the undersigned.

Dr. Martin Seed, University of Manchester

Professor Raymond Agius, University of Manchester

Martin.seed@manchester.ac.uk

REFERENCES

1. Sastre J, del Potro MG, Aguado E, et al. Occupational asthma due to 5- aminosalicylic acid. Occup Environ Med 2010;67:798-9.

2. Jarvis J, Seed MJ, Elton RA, et al. Relationship between chemical structure and the occupational asthma hazard of low molecular weight organic compounds. Occup Environ Med 2005;62:243-50.

3. Seed MJ, Agius RM. Further validation of computer-based prediction of chemical asthma hazard. Occup Med 2010;60:115–20.

4. Moore VC, Manney S, Vellore AD, et al. Occupational asthma to gel flux containing dodecanedioic acid. Allergy 2009;64:1099-1107.

5. The University of Manchester. Centre for Occupational and Environmental Health. Mechanisms of Occupational Asthma; Occupational asthma hazard prediction programme. http://www.medicine.manchester.ac.uk/oeh/research/asthma/ (accessed November 2010).

17 November 2010

The statement by IARC that “Shiftwork that involves circadian disruption is probably carcinogenic to humans” [1] raises the issue of evidence-based measures for prevention, which was the focus of our letter. Assuming that nightshift work is indeed causally related to breast cancer one option for prevention would be to set limits for the individual cumulated amount of night work [2]. Such an approach would, however, only prevent cancer cases if there is a multiplicative exposure-response relation or a threshold. If this is not the case, only setting an upper limit for years of nightshifts for the individual workers would distribute the risk among more workers but not reduce the total number of cases.

Therefore, we analyzed the 9 epidemiological studies in order to examine the exposure-response relation. We agree with Hansen and Stevens that the study by Kjaer et al [3] lacks individual information on night-shift work but only proxy measures. We therefore reanalysed the data excluding this study and 3 other studies relying on proxy measures of nightshift work [4],[5],[6] and included a study recently published with relevant exposure information [7]. We obtained a meta-odds ratio of 1.004, 95% CI 1.001-1.008 by year of nightshift work. There was no indication of neither a threshold nor a meaningful exposure-response relation and thus the available limited epidemiological data do not provide evidence to warn against long-term nightshift work for the individual worker. On the other hand, there are plenty arguments for reducing the total population burden of nightshift work, among others the suspicion raised about increased breast cancer risk.

Henrik A Kolstad & Jens Peter Bonde

References

1.      International Agency for Research on Cancer. IARC Monographs on the evaluation of carcinogenic risks to humans. Painting, firefighting, and shiftwork. Lyon: International Agency for Research on Cancer; 2010. 563-766.

2.      Fritschi L. Shift work and cancer. bmj 2009;339:b2653.

3.      Kjaer TK, Hansen J. Cancer incidence among large cohort of female Danish registered nurses. Scand J Work Environ Health 2009;35(6):446-53.

4.      Tynes T, Hannevik M, Andersen A et al. Incidence of breast cancer in Norwegian female radio and telegraph operators. Cancer Causes Control 1996;7(2):197-204.

5.      Hansen J. Increased breast cancer risk among women who work predominantly at night. Epidemiol 2001;12(1):74-7.

6.      Lie JA, Roessink J, Kjaerheim K. Breast cancer and night work among Norwegian nurses. Cancer Causes Control 2006;17(1):39-44.

7.      Pronk A, Ji BT, Shu XO et al. Night-shift work and breast cancer risk in a cohort of Chinese women. Am J Epidemiol 2010;171(9):953-9.

Kolstad et al. claim in a letter: “warnings against long-term night shift work to prevent breast cancer are not evidence based” [1]. However, 24 scientists convened by the International Agency for Research on Cancer have concluded, based on sufficient evidence in experimental animals, limited evidence in humans, and strong bio-mechanistic support, that “shiftwork that involves circadian disruption is probably carcinogenic to humans”, (group 2A) [2]. This 200 pages effort by IARC was most definitely ‘evidence based’.

Kolstad et al. base their conclusion solely on a regression analysis of 9 studies. A major impact to their analysis is a large survey of nurses without any information on shiftwork [3]. Further, two of the included studies provide only information on shiftwork for the last 10 or 15 years, respectively, before diagnosis [4, 5]; it is therefore inappropriate to include these two studies in an attempt to construct a ‘dose response’ over decades of work. Despite these severe limitations, they calculate a pooled odds ratio of 1.02 per year (95% CI: 0.92 to 1.13) by increasing years of shiftwork; this would yield a relative risk of 1.49 for 20 years of shiftwork. Overall, this estimate is similar to results from a meta-analysis (RR = 1.40; 95% CI: 1.19-1.65) based on comparison of day-workers with the most extreme shiftwork exposure category in 8 of the same studies as used by Kolstad [6]. Finally, and perhaps most importantly, lack of statistical significance should not, as done by Kolstad, be interpreted as evidence for no effect, particularly when the point estimate (1.49 for 20 years) is far from 1.0.

About 15-20% of the workforce has shiftwork [2]. Therefore, clarifying of carcinogenicity is very important, and should be based on a profound synthesis of all existing documentation, which is extensive [2], but neglected by Kolstad et al. [1].

Johnni Hansen & Richard G. Stevens

References:

1 Kolstad HA, Erlandsen M, Frost P, et al. Should we warn against night shifts to prevent breast cancer? Occup Environ Med 2010;67 (11):797.

2 International Agency for Research on Cancer. IARC monographs on the evaluation of carcinogenic risks to humans. Painting, firefighting, and shiftwork. Lyon: International Agency for Research on Cancer 2010:563-766.

3 Kjaer TK, Hansen J. Cancer incidence among large cohort of female Danish registered nurses. Scand J Work Environ Health 2009;35 (6):446-53.

4 O’Leary ES, Schoenfeld ER, Stevens RG, et al. Shift work, light at night, and breast
cancer on Long Island, New York. Am J Epidemiol 2006;164 (4):358-66.

5 Davis S, Mirick DK, Stevens RG. Night shift work, light at night, and risk of breast cancer. J Natl Cancer Inst 2001;93 (20):1557-62.

6 Viswanathan AN, Schernhammer ES. Circulating melatonin and the risk of breast and endometrial cancer in women. Cancer Lett 2009;281 (1):1-7.

In his letter, Predicting chemicals causing cancer in animals as human carcinogens (Occup Environ Med 2010;67:720), Huff surprisingly finds opportunity to promote long-term carcinogenesis bioassays using animals in response to an editorial by Suarthana et al. (Occup Environ Med 2009;66:713e14), Predicting occupational diseases. In their editorial, Suarthana et al. generalize from the development of diagnostic models to predict sensitisation to occupational allergens. Suarthana et al. mention no animal studies and conclude that “[t]here is now an opportunity to develop prediction models for diverse occupational diseases, especially given the existing large epidemiological studies (emphasis added).”

In particular, Huff asserts that “findings from independently conducted bioassays on the same chemicals are consistent, albeit sometimes with additional or different target sites[.]” In fact, a comparison of 121 bioassays from the U.S. National Toxicology Program (NTP) database with those in the published scientific literature found that the studies produced consistent results only 57 percent of the time [1].

Huff also claims that “less than10% of all chemicals if evaluated in bioassays would be predicted to be carcinogenic.” In our analysis of more than 500 NTP bioassays, we show that 259 of the substances evaluated produced evidence of carcinogenicity in at least one group of animals. However, only 89 of these were subsequently classified as known or probable human carcinogens by NTP itself [2]. Even fewer, 40 and 16 respectively, were so classified by the U.S. Environmental Protection Agency and the International Agency for Research on Cancer. This high false positive rate is thought to be largely an indirect effect of increased cell proliferation in response to cell injury and death caused by the near toxic doses of test substances used in the bioassay [3]. Huff also observes that there are more similarities than differences between rodents and humans. However, species-specific modes of action operating in rats or mice but not in humans, including those mediated by α2μ-globulin, peroxisomes, and thyroid-stimulating hormone, also contribute to the high rate of false positives [4].

We must also stress the suffering that animals endure in the bioassay. Not only must they live in the barren, stressful conditions of the laboratory – often including daily forced feeding or forced inhalation – but many will also suffer from exposure to near toxic doses of test substances. This exposure is normally expected to produce lethargy, anemia, diarrhea, weight loss, and other symptoms of sickness and distress.

According to NTP’s own estimates, each bioassay requires killing 860 animals, $2-4 million and five years to plan, conduct, and evaluate. As a result, NTP has conducted an average of only 12 bioassays per year over the past several decades. Considering that humans are potentially exposed to as many as 80,000 environmental chemicals, more than 32 thousand years, 68 million animals, and $160 billion would be required to test them all at this rate.

The time has clearly come for antiquated animal tests like the bioassay to be abandoned in favor of modern, human-relevant methods such as the epidemiological studies recommended in Suarthana et al’s editorial, high-throughput in vitro methods, and computational toxicology.

Joseph Manuppello

[1] Gottmann E, Kramer S, Pfahringer B, et al. Data quality in predictive toxicology: reproducibility of rodent carcinogenicity experiments. Environ Health Perspect 2001;109:509-14.

[2] PETA. Wasted money, wasted lives: A layperson’s guide to the problems with rodent cancer studies and the National Toxicology Program. Norfolk, VA: People for the Ethical Treatment of Animals. 2006. http://www.mediapeta.com/peta/pdf/Wasted-money PDF.pdf (accessed 17 Sept 2010).

[3] Gaylor DW. Are tumor incidence rates from chronic bioassays telling us what we need to know about carcinogens? Regul Toxicol Pharmacol 2005;41:128-33.

[4] Cohen SM. Human carcinogenic risk evaluation: an alternative
approach to the two-year rodent bioassay. Toxicol Sci 2004;80:225-9.

Conflict of Interest: The author is employed by People for the Ethical Treatment of Animals.

Re: Cocco et al., Occup Environ Med 2010; 67: 341-347 (Original article)

Considering the temporal association between exposure to benzene and the later development of leukaemia it is questionable if this phenomena is also true for NHL (1). From several independent epidemiologic studies with
consistent findings it can be concluded that 10 to 15 years after exposure to benzene has been stopped, the risk of leukaemia is significantly less or even absent (2,3,4).

Assuming that the underlying mechanism for leukaemia and NHL is the same, it would be important to use the large database of the “Epilymph study” to analyze the temporal pattern between exposure and disease.

The study found increased risks for chronic lymphocytic leukaemia after exposure to toluene and xylene. A benzene exposure was excluded. These statistical associations must be discussed with respect to the genotoxicologic evidence of these aromatic hydro-carbons. Compared to benzene, toluene and xylene have not been proven as human carcinogens. The IARC Working Group concluded in 1999, that there is inadequate evidence in humans for the carcinogenicity of toluene and xylenes (5). Therefore the
associations must be interpreted with caution and do not support causality.

Gerhard Triebig

1. Triebig G. Implications of latency period between benzene exposure and development of leukemia – a synopsis of literature. Chem Biol Interact 2010; 184: 26-29.

2. Hayes RB, Song-Nian Yin, Dosemeci M. Benzene and the dose related incidence of hematologic neoplasm in China. J Natl Cancer Inst 1997; 89: 1065-1071.

3. Finkelstein MM. Leukemia after exposure to benzene: temporal trends and implications for standards. Am J Ind Med 2000; 38: 1-7.

4. Glass DC, Sim MR, Fritschi L, Gray CN, Jolley DJ, Gibbons CG. Letter to the editor. Leukemia risk and relevant benzene exposure period. Am J Ind Med 2002; 42:481-489.

5. IARC Monographs on the evaluation of carcinogenic risks to humans. Re-evaluation of Some Organic Chemicals, Hydrazine and Hydrogen Peroxide. Volume 71, Part two (Toluene) and Part three (Xylene). World Health Organization International Agency for Research on Cancer, 1999.

In our study showing the exposure to chrysotile fibres in a friction materials producing plant through urinalysis for fibres, it was shown that smaller chrysotile fibres were indeed absorbed and if shorter than 1-2 µm could also be transported in the urine (1). This property could explain effects also in other sites than respiratory tract.

Exposure can also be demonstrated by an electron microscopic analysis for fibres in the nasal lavage liquid. Thus we were able to show individual exposure to refractory ceramic fibres, the nature of which could be ascertained by X-ray diffraction technique in the same samples (2).

Thus, it seems that exposure to, and absorption of, fibres harmful to health can be determined at an individual level. Cohorts thus constituted could prove the ideas presented in this excellent paper (3).

Heikki Savolainen

1. Savolainen H, Cosca-Sliney R, Guillemin M. Detection of
occupational exposure to inorganic fibres by urinary fibre analysis. Occup Hyg 1996; 3: 351-357.

2. Linnainmaa M, Kangas J, Makinen M, et al. Exposure to refractory
cermaic fibres in the metal industry. Ann Occup Hyg 2007; 51: 509-516.

3. Loomis D, Dement J, Richardson D, et al. Asbestos fibre dimensions
and lung cancer mortality among workers exposed to chrysotile. Occup
Environ Med 2010; 67: 580-584.