In the management of occupational health hazards, medical surveillance is usually considered an established tool for secondary prevention of adverse health effects through early detection. Nevertheless, the term “medical surveillance” is used differently in various contexts and by different institutions or authors, and the borders between “surveillance”, “screening,” and “examination” are not always clear-cut. This fuzziness of terms has sometimes impacted on discussions regarding sensible and feasible medical measures targeted at managing health risks in occupationally exposed persons and very much so in the ongoing discussion about the risks for employees handling nanomaterials. To get a grip on this topic, a distinction must be made between the two types of surveillance: personal and public health surveillance. While personal surveillance focuses on the individual, public health surveillance is performed with the intention to monitor the overall health experience at population level. Furthermore, we have to differentiate between the management of known health risks associated with a given exposure and the identification of “as yet unidentified” new health risks. A useful set of definitions has been compiled by Trout and Schulte1:
* Occupational health surveillance is the ongoing systematic collection, analysis, and dissemination of exposure and health data on groups of workers for the purpose of early detection of disease and injury.
* Medical surveillance examines health status through tracking of illnesses or a change in a biological function in an exposed person or persons. It essentially involves a process of looking for health trends in a worker population.
* Medical screening is one form of medical surveillance that is designed to detect early signs of work-related illness by conducting tests in apparently healthy persons to detect those with early stages of disease or those at risk of disease.
There is thus no clear border, which strictly separates surveillance from screening, the latter being a subcategory of the former. According to these definitions, occupational health surveillance and medical surveillance may involve medical examinations, while medical screening must do so. A further distinction should be made between general and exploratory approaches on the one hand, which are more often used in broad surveillance programs, and specific or “targeted” approaches on the other hand, with a comparatively narrow focus on occupational exposures and their potential effects on health.
There are some prerequisites for targeted occupational medical screening: (a) knowledge about the existence or at least possibility of an exposure to a health hazard; (b) knowledge about specific health effects caused by such an exposure; (c) the availability of tests with a known sensitivity and specificity to detect such health effects; and (d) knowledge about the strength of an association between exposure and effect.2
At the individual level, an apparent benefit may result from any kind of medical surveillance, and specifically from screening, either if the target outcome of screening serves as an early marker of effect but is not itself a pathological condition or if the health condition expected is both diagnosable at an early stage and treatable at this point in time. A moderate suppression of cholinesterase activity after exposure to organophosphates may serve as an example for the former condition. The benefit may be less clear if no treatment option exists for the disease of interest, but it may be arguable with regard to securing a basis for compensation claims. There may, however, be no benefit at all to the individual if there is no therapeutic option and no established causal relation between exposure and health finding. In the following section, several examples for existing surveillance and screening approaches will be provided and an attempt will be made to extrapolate the usefulness and applicability of such approaches to the present situation of employees handling engineered nanomaterials.
There exists a wealth of knowledge about the health risks associated with a wide spectrum of occupations; job tasks; and chemical, biological, or physical exposures. Occupational physicians have always used their detailed knowledge about workplaces to develop and perform bespoke examination programs for exposed individuals with the aim to detect work-related health effects at the earliest possible point in time. In some instances, even individual risk factors causing enhanced individual susceptibility to work-related health hazards may be identified and appropriate preventive measures may be suggested. Existing guidelines, often issued by national competent authorities, can be hazard oriented (eg, chemicals, radiation, noise, and infectious agents), job task oriented (eg, driving, controlling, monitoring, and logging), or aimed at specific health endpoints (eg, chronic obstructive pulmonary disease, skin disease, and hearing loss).3,4
Those guidelines directed at chemical exposures often contain recommendations regarding biomonitoring, which, in the narrow sense, is the determination of chemicals or their metabolites or adducts in human tissues or body fluids. This method is particularly useful as it complements ambient air measurements by providing information about the actual uptake into the body of an exposed worker of a chemical at a workplace. It thus helps to assess the effectiveness of technical measures of exposure reduction, use of personal protective equipment and safe working practices, and to identify potential for improvement at both the individual behavioral and at a general organizational level.5 Biomonitoring is thus the part of occupational medical screening, which most obviously contributes to primary prevention, potentially triggering action to avoid or to reduce hazardous exposures even in the absence of detectable health effects.
Occupational medical screening is not only directed at current workplace hazards and ongoing exposures but, for example, in the case of carcinogens, can also be offered to formerly exposed employees even after the end of employment. This latter approach requires the establishment of registries, which exist in different countries on a variety of past exposures. Their establishment and maintenance can either be mandated by law or result from company or trade union initiatives.6 In Germany, several registries are kept under the auspices of the Employer's Liability Insurance Association (Deutsche Gesetzliche Unfallversicherung). It covers all substances classified as category one or two carcinogens under German or European Union regulations. Persons registered are offered regular screening examinations, where the screening interval and the examinations performed are chosen on the basis of knowledge about the typical target organs for the exposure in question and on experience about the natural course of such cancers.
The usefulness and feasibility of this approach has been assessed repeatedly, notably in high-risk populations with past exposure to aromatic amines for the endpoint of bladder cancer.7–10 The individual benefit of participating in bladder cancer screening is obvious for those persons, who were by means of screening diagnosed with early stage bladder cancer, because without screening this diagnosis would have been obtained later. The benefit lies in the fact that in early stage bladder cancer, therapy is more successful and the outcome more favourable than in later stages. In a cohort of more than 1700 workers from the chemical industry, we found evidence that rehabilitation costs for cohort members were 15% to 20% lower if compared with noncohort cases with bladder cancer covered by the same insurance provider.9 Another result of this study, however, was the confirmation of a deplorably poor positive predictive value of the screening protocol applied, notwithstanding its sufficient sensitivity. Consequently, an uncomfortably high number of cystoscopies were performed in the members of this cohort, where no cancer was diagnosed. This shed a somewhat different light on the perceived benefit for the latter group due to the potential for unwanted side-effects of this diagnostic procedure, such as infections and physical trauma, not to mention the fact that it is usually perceived as less than comfortable. While this situation may be deemed acceptable for a high-risk population, it certainly calls for improvement. In cooperation with urologists, epidemiologists, insurance association, and providers of diagnostic markers, we therefore started a prospective study in this cohort with the aim to evaluate alternative and innovative markers for the early detection of bladder cancer.11 We are confident that the results of this study will help to identify useful new marker panels and to redefine the cutoff values for the established ones. This example may serve as an illustration of how existing registries may yield benefits not only to the respective individuals under surveillance but also through their scientific exploitation to the general population, where they may help to optimize and improve the cost-effectiveness of screening strategies.
This optimistic view on the usefulness of targeted occupational medical screening can unfortunately not be translated to every kind of targeted screening. This is perfectly illustrated by the ongoing controversy surrounding prostate-specific antigen testing, where overdiagnosis and overtreatment of clinically insignificant prostate cancer are considered a major potential drawback by some—but not all—authors.12–16 From a patient perspective, the question is to decide, hopefully, after having received sufficient information on the pros and cons of prostate-specific antigen testing and on the preferability of either dying from prostate cancer or running the risk of impotence, incontinence, hospital-borne infection, and other possible complications after unwarranted surgery. Even after the availability of interim results from two large randomized trials, this conundrum remains unresolved.17–19
In a situation with suspected but, as yet, unproven health risks due to existing exposures, new insights may be expected from surveillance at a population level. Unrecognized health risks can be systematically researched by comparing across different subsets of employees with defined exposures medical findings obtained through or available to the occupational physician. This analysis of aggregate data can be especially useful for the identification of new sites for known health hazards.3 An unexpected exposure to isocyanates can, for instance, become apparent through measured 1-second forced expiratory volume trends in a group of workers, in which the average effect size would be insignificant for an individual but meaningful for a population if compared with an unexposed group. An apparent shortcoming of this approach is that it relies on observational data not obtained for systematic comparisons in first place. In the previous example, the analysis could be meaningless if it turned out that the group with a higher decline in 1-second forced expiratory volume consisted of heavily smoking shiftworkers while the comparison group consisted mostly of clerical workers. The feasibility of such a company-confined study is further often limited by group size, which, in most industrial settings, may be too small to enable meaningful statistical comparisons. Even more challenges result, at least from the statistician's perspective, from the fact that the frequency of and intervals between examinations may vary considerably, thus introducing potential detection bias, lead time bias, and other pitfalls to data interpretation. These limits can sometimes be overcome by purposefully designing and performing studies in exposed workers, where questions of comparability, group size, information needed on confounders, etc, can be addressed in advance. Such systematic approaches have long been used in some industries, like in the historical example, in which by data pooling across companies, it was possible to verify a preexisting suspicion regarding the manufacturing process of auramine and magenta, but not exposure to the final products, as causative in the development of bladder cancer.20
Beyond what has been said previously, basically every prospective or retrospective cohort study concerned with occupational or environmental exposures can be considered as an example for untargeted surveillance. Such research can be carried out in cohorts like the Agricultural Health Study, in which detailed information on exposure has been obtained and documented in advance of the occurrence of the outcome of interest, and socioeconomic differences are not likely to bias the comparisons to a major extent.21 In the environmental sector, the cohorts set up after the infamous Seveso accident or the studies carried out in atomic bomb survivors continue to contribute to our understanding of the effects of dioxins or radioactive radiation, respectively.22,23
It is important to keep in mind the difference between confirmatory testing of a preexisting hypothesis in an existing or—even better—newly assembled data set and an exploratory analysis, in which the possibility of chance findings is sometimes not adequately addressed by some researchers. It seems an unfortunate development that with the increasing availability and user friendliness of statistical packages, which can be used on every desktop computer, the number of studies appears to increase, where the most intricate and sophisticated statistical procedures were employed on data sets not really designed for it. This potentially leads to a plethora of “new findings” resulting from exploratory studies and “creative data modeling”, regarding old and new exposures, sometimes calling for preventive action (but at least for more funding, because further studies are needed) even before the scientific discussion about the potential significance of these findings has started. This development and a passionate “plea for epistemological modesty” has recently fuelled a lively discussion among epidemiologists.24–26 Irrespective of this controversy, it goes without saying that false alarms and thereby triggered unnecessary responses and expenses, not to mention the distress in allegedly concerned individuals, must be counted among the potential drawbacks associated with untargeted surveillance.
Having said that there shall be no doubt that each unusual pattern or frequency of health findings in a screened population warrants a closer look and thorough workup. Such observations may at first be indiscernible from chance clusters, and the significance of some of these observations has remained a subject for controversial debate for many years to come. Often, we can only in hindsight classify some of these clusters as true “sentinel health events.”
Sentinel Health Events
A special role in the detection of hitherto unknown health risks is often ascribed to “sentinel health events.” This refers to medical findings or diseases, which are unexpected either by their nature or by their frequency of occurrence, in the screened population. Admittedly, up to now, completely new insights into occupational health hazards have rarely, if ever, been obtained through purposefully designed monitoring strategies but often resulted from accumulating case series, which at length stirred suspicion in vigilant physicians or—unfortunately—pathologists. Notorious historical examples are bladder cancer resulting from aromatic amine exposure, lung cancer from hexavalent chromium, hepatic toxicity from polychlorinated naphthalenes, or even the infamous asbestos case.27–30
Sometimes, such index cases may present in a very unsuspicious manner, and it takes the specific knowledge of an experienced plant physician to find the unusual aspect in a seemingly common appearance. The following case report may serve as an example:
A technician presented himself at the site medical clinic, complaining predominantly of cough and breathing difficulties. That morning he had experienced nausea and one bout of vomiting. He had been dismantling reactors and pipes in a propionic acid plant over a period of several days during periodic maintenance activities. No specific exposure event was reported by the employee. On the basis of the knowledge of the plant operations (where nickel tetracarbonyl is used as catalyst), a urine sample was collected for the determination of the urinary nickel concentration—just in case! This examination revealed a high level of nickel in the urine and led to the diagnosis of nickel tetracarbonyl intoxication. The same diagnosis was established in retrospect for two additional employees found to have similar symptoms. They were currently being treated by their family physicians as cases of common cold and incipient pneumonia, respectively. Chest radiographs showed peribronchial infiltration in all three cases, without the signs of bronchial obstruction. Laboratory blood analyses were consistent with a nonspecific inflammatory response. The symptoms resolved, and the clinical examination findings returned to normal in all three persons within 1 week.
These were the first cases of clinically relevant nickel tetracarbonyl intoxications in BASF over a period of more than 40 years. A search in our archives identified a report on a similar incident in 1958, where a total of seven persons had been exposed, resulting in two fatalities.31 The nickel urine concentrations found in our current cases came close to the lower range observed in these historical fatal cases. Routinely performed carbon monoxide measurements at the beginning of the dismantling and maintenance work had failed to provide a clear warning sign. In retrospect, prolonged off-gassing from insoluble residue cakes formed on the reactor wall was identified as the most likely cause for this unusual exposure. As a consequence, similar tasks will be performed in the future by using self-contained breathing equipment until the absence of potentially hazardous residues on the equipment parts has been positively confirmed. The cases have been published to make responsible persons in other industries with similar processes aware of this unusual exposure scenario.32
Medical Surveillance: What Should We Not Do?
The definitions of surveillance and screening quoted at the beginning of this article are somewhat academic in that they direct a view from external on the worker involved (at least, this “unidirectional” interpretation is not explicitly ruled out in these definitions). In this context, the worker may be the object of the examination, and the information obtained on him, and from him, may primarily be used to create knowledge about the interactions between workplace exposures and individual or group health status. While such an approach can have its scientific merit and can indeed produce results that benefit working populations as a whole, it does not account for the fact that the interaction between the occupational physician at a given plant or site and the worker is “bidirectional” by default. The worker is at the same time the subject involved and may rightfully request that each and every finding obtained on him should be interpreted with regard to his current and future health and to the potential consequences for his employability. Thus, in not only occupational but also medical daily practice and outside scientific studies, the primary rule for choosing diagnostic parameters is: “Never use a method where you cannot interpret the results.” This attitude is sometimes denounced as misusing an ethical argument as a pretext for a “do-nothing” policy. Nevertheless, it has nothing to do with ethics but is simply derived from the experience of practical occupational physicians who have to answer very personal questions and concerns regarding medical findings and who often have to provide advice that may finally trigger decisions that go as far as giving up a job or leaving an employment for perceived health reasons. It is important to remember that the key question for an individual is not whether screening is effective but whether it does more good than harm.19
Medical surveillance is thus one of the cornerstones of occupational health surveillance and, as such, a vital part of the efforts to secure just and favorable working conditions in keeping with the human rights declaration. The examples discussed so far illustrate the role of occupational medical surveillance in various aspects of managing risks at workplaces and beyond. It specifically helps to
* Target known workplace-specific hazards, help to reduce exposure, detect health effects at the earliest possible point in time,
* Identify hitherto unknown health hazards or exposure possibilities,
* Enhance understanding of the significance of personal behavior for risk reduction through communication of findings to individuals under survey,
* Communicate aggregate findings to staff and management to provide the full picture of the occupational hygiene situation, to facilitate targeted intervention,
* Allow employees and management to develop informed conclusions regarding compatibility between individual health status and workplace-related health risks,
* Make employees aware of nonoccupational health risks, which are identified as a “side effect” of occupational medical screening, and
* Translate experience gained from occupational cohorts to diagnostic strategies or toxicological assessments relevant for the general population.
Nevertheless, occupational medical surveillance must be planned and performed keeping the limits, potential pitfalls, and shortcomings in mind.
Extrapolation to “Nano”: Should We Screen?
Much has been said about the potential health effects of nanoparticles, where the ongoing discussion leaves no doubt that there is no uniform common or single specific endpoint in human health. On the contrary, the health hazards associated with nanomaterials will most probably have to be assessed differently for different classes of nanomaterials, while not for any single material itself. Nevertheless, quite a bit of basic research has been carried out to date, and it has provided important clues on what may be expected. The effects may be mediated by oxidative stress, inflammation, and fibrogenesis in the widest sense, and the target organs most often mentioned are respiratory, circulatory, and central nervous system and liver.33–35 Given the lack of specificity of these endpoints and the high prevalence of respective findings in the general population, most authors agree that—while there is no evidence base for targeted “nano-specific” screening—general medical screening with methods aimed at some of the health outcomes under discussion may be performed in exposed workers.1,2,36 Such screening should be devised weighing the risk to benefit ratio for the tests in consideration, keeping in mind the risks associated with untargeted medical surveillance. The results of such screening may, after aggregate evaluation on group level, provide future insights into relevant health risks associated with the handling of nanomaterials in workplaces.
1. Trout D, Schulte PA. Medical surveillance, exposure registries, and epidemiologic research for workers exposed to nanomaterials. Toxicology. 2010;269:128–135.
2. Nasterlack M, Zober A, Oberlinner C. Considerations on occupational medical surveillance in employees handling nanoparticles. Int Arch Occup Environ Health. 2008;81:721–726.
3. Harber P, Conlon C, McCunney RJ. Occupational medical surveillance. In:McCunney RJ, ed. A Practical Approach to Occupational and Environmental Medicine. 3rd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2003:582–599.
4. Milde J, ed. Guidelines for Occupational Medical Examinations. Stuttgart, Germany: Gentner Verlag; 2007.
5. Zober A, Will W. Biological monitoring and risk assessment in occupational settings. Int Arch Occup Environ Health. 1996;68:389–393.
6. Schulte PA, Kaye WE. Exposure registries. Arch Environ Health. 1988;43:155–161.
7. Crosby JH, Allsbrook WC, Koss LG, et al. Cytologic detection of urothelial cancer and other abnormalities in a cohort of workers exposed to aromatic amines. Acta Cytol. 1991;35:263–268.
8. Mason TJ, Walsh WP, Lee K, Vogler WJ. New opportunities for screening and early detection of bladder cancer. J Cell Biochem. 1992;16:13–22.
9. Nasterlack M, Scheuermann B, Messerer P, Pallapies D, Zober A. Harnblasenkrebs in einem Risikokollektiv—klinische und epidemiologische Aspekte. Symposium Med. 2001;12:17–19.
10. Marsh GM, Cassidy LD. The Drake Health Registry Study: findings from fifteen years of continuous bladder cancer screening. Am J Ind Med. 2003;43:142–148.
11. Feil G, Horstmann M, Leng G, et al. Urine-based tumor marker tests are a helpful tool in early diagnosis of bladder cancer in high-risk populations—interim data of the prospective study UroScreen. J Urol. 2008;179:325.
12. McGregor M, Hanley JA, Boivin JF, McLean RG. Screening for prostate cancer: estimating the magnitude of overdetection. CMAJ. 1998;159:1368–1372.
13. Potosky AL, Feuer EJ, Levin DL. Impact of screening on incidence and mortality of prostate cancer in the United States. Epidemiol Rev. 2001;23:181–186.
14. Etzioni R, Penson DF, Legler JM, et al. Overdiagnosis due to prostate-specific antigen screening: lessons from U.S. prostate cancer incidence trends. J Natl Cancer Inst. 2002;94:981–990.
15. Kubota Y, Ito K, Imai K, Yamanaka H. Effectiveness of mass screening for the prognosis of prostate cancer patients in Japanese communities. Prostate. 2002;50:262–269.
16. Sirovich BE, Schwartz LM, Woloshin S. Screening men for prostate and colorectal cancer in the United States—does practice reflect the evidence? JAMA. 2003;289:1414–1420.
17. Andriole GL, Grubb III RL, Buys SS, et al. Mortality results from a randomized prostate-cancer screening trial. N Engl J Med. 2009;360:1310–1319.
18. Schröder FH, Jugosson J, Roobol MJ, et al. Screening and prostate-cancer mortality in a randomized European study. N Engl J Med. 2009;360:1320–1328.
19. Barry MJ. Screening for prostate cancer—the controversy that refuses to die. N Engl J Med. 2009;360:1351–1354.
20. Case RAM, Pearson JT. Tumours of the urinary bladder in workmen engaged in the manufacture and use of certain dyestuff intermediates in the British chemical industry. II. Further consideration of the role of aniline and of the manufacture of auramine and magenta (fuchsine) as possible causative agents. Br J Ind Med. 1954;11:213–216.
21. Alavanja MCR, Bonner MR. Pesticides and human cancer. Cancer Investigation. 2005;23:700–711.
22. Consonni D, Pesatori AC, Zocchetti C, et al. Mortality in a population exposed to dioxin after the Seveso, Italy, accident in 1976: 25 years of follow-up. Am J Epidemiol. 2008;167:847–858.
23. Dropkin G. Reanalysis of cancer mortality in Japanese A-bomb survivors exposed to low doses of radiation: bootstrap and simulation methods. Environ Health. 2009;8:56. doi:10.1186/1476-069X-8-56.
24. Boffetta P, McLaughlin JK, La Vecchia C, et al. False-positive results in cancer epidemiology: a plea for epistemological modesty. J Natl Cancer Inst. 2008;100:988–995.
25. Vineis P. The skeptical epidemiologist. Int J Epidemiol. 2009;38:675–677.
26. Boffetta P, McLaughlin JK, Vecchia CL, et al. A further plea for adherence to the principles underlying science in general and the epidemiologic enterprise in particular. Int J Epidemiol. 2009;38:678–679.
27. Rehn L. Blasengeschwülste bei Fuchsin-Arbeitern. Archiv für klinische Chirurgie. 1895;50:588–600.
28. Pfeil E. Lungentumoren als berufserkrankung in chromatbetrieben. Dtsch med Wochenschr. 1935;61:1197–1200.
29. Strauss N. Hepato-toxic effects following occupational exposure to halowax (chlorinated hydrocarbons). Rev Gastroenterol. 1944;11:381–396.
30. Wood WB, Gloyne SR. Pulmonary asbestosis. A review of one hundred cases. Lancet. 1934;2:1383–1391.
31. Ludewigs H-J, Thiess AM. Arbeitsmedizinische Erkenntnisse bei der Nickelcarbonylvergiftung. Zbl Arbeitsmed. 1970;20:329–339.
32. Pluto R-P, Trauth B, Will W, Nasterlack M, Lang S. Drei Intoxikationen mit Nickeltetrakarbonyl. Arbeitsmed Sozialmed Umweltmed. 2009;44:81–86.
33. Oberdörster G, Oberdörster E, Oberdörster J. Nanotoxicology—an emerging discipline involving studies of ultrafine particles. Environ Health Perspect. 2005;113:823–839.
34. Borm PJA, Robbins D, Haubold S, et al. The potential risks of nanomaterials: a review carried out for ECETOC. Part Fibre Toxicol. 2006;3:11. doi:10.1186/1743-8977-3-11.
35. Gwinn MR, Vallyathan V. Nanoparticles: health effects—pros and cons. Environ Health Perspect. 2007;115:1818–1825.
36. Schulte PA, Trout D, Zumwalde RD, et al. Options for occupational health surveillance of workers potentially exposed to engineered nanoparticles: state of the science. J Occup Environ Med. 2008;50:517–526.
©2011The American College of Occupational and Environmental Medicine