The prevention of occupational asthma

Pre-employment selection

If the development of OA involves an important element of individual risk, then identification of susceptibility markers could conceivably be used in the selection of innately low-risk employees. Leaving aside its doubtful morality (or even legality), this approach appears to be highly inefficient. More broadly, a focus on the “susceptible host” may discourage efforts to reduce risks and prevent disease in populations 2. Asthma, occupational or otherwise, is a complex disease, likely to have multiple genetically determined influences and currently it appears improbable that any single marker of predisposition, or even manageably small group of markers, will ever be sufficiently discriminatory to be useful. Age and sex appear to be unimportant, independent risk factors, although prognosis may be poorer in older persons 10. The absence of “atopy” (variously defined) is used in several industries during the selection of new employees, but this is almost certainly a very inefficient approach. Some 30–40% of young adults in industrialised countries are now “atopic” in one way or another and their general exclusion dramatically reduces the pool of potential new employees. There is an increasing volume of published work that points towards human leukocyte antigen (HLA)‐restricted susceptibility to several occupational agents 11, although this body of evidence is not internally consistent. No wholly sensitive or specific markers of individual (genetic) susceptibility have been identified.

As with most “competing causes” of disease, it may be that susceptibility has a more powerful effect at lower levels of environmental exposure 12. Thus, their consideration may become more important as environmental controls improve. It is interesting to consider how a (genetic) screening marker might be used if a reasonably efficient one was available. The experience of berylliosis may prove instructive. Screening for Glu69 in HLA‐DPB1, a marker of susceptibility with high sensitivity but low specificity, has recently been offered to some employees of the beryllium industry in the USA. The test is voluntary and the results are made known only to the employee, who is therefore free to act on the knowledge or otherwise; exclusion from employment is not the only available course of action. It remains to be seen whether this will result in a reduced incidence of beryllium sensitisation and to what extent the practice is acceptable.

It seems reasonable to apply some rules of exclusion at the stage of first employment. Persons with established OA from a particular agent ought not to be employed in a new job where there is further exposure to the same agent. More contentiously, it is not uncommon for prospective employers to decline offers to those with a history of OA to an agent which is entirely unrelated to any sensitiser they may encounter in the new job. There is no reliable evidence that such persons are, on account of their original OA, at increased risk of acquiring a second variant of the disease. Nonetheless this practice is a common cause of employment handicap to those who have acquired asthma at work.

More common still is the screening-out of potential employees with “community” asthma from jobs where there is a risk of exposure to sensitising agents, as advocated by some 13. For those with severe or moderately severe asthma, this may reasonably be justified on the grounds that it would be unwise to put them at risk of developing an additional respiratory impairment. Such practice may fall foul of anti-discriminatory legislation, although in most countries this is yet to be tested in law. For those with mild asthma, or indeed those with a past history of the disease that is now quiescent, the value of screening is far less obvious; and will become increasingly so as the number of children who have at some stage acquired an asthmatic label rises. Nonetheless, there may be a role here for better education of those (with or without asthma) attending vocational schools and even those of school age considering a suitable career.

Preproduct screening

For many years, there have been attempts to identify, prior to their widespread use, which newly introduced agents are likely to act as human respiratory sensitisers. This toxicological approach generally relies on animal models 14. Using a mouse intranasal test, for example, the potencies of several protease and nonprotease enzymes in causing specific immunoglobulin (Ig)G1 production varied 60‐fold 15. More broadly, the hazard evaluation of low molecular weight agents always includes an assessment of their dermal sensitisation potential. In most cases this is now done using an animal-based local lymph node assay, which tests the ability of topically applied chemical agents to induce proliferative lymphocyte responses in draining nodes 16. Cytokine fingerprinting may offer a more specific index of respiratory sensitising potential 17–19. To the extent that the potential of chemicals to cause immunological sensitisation depends on their mode of interaction with antigen-presenting cells and the subject's subsequent lymphocyte responses, and since antigen-presenting cells and lymphocytes are not confined to the skin, there is no a priori reason to assume that the dermal route of administration is critical in eliciting a sensitisation response. Arguably, chemicals that are positive in tests involving dermal application should be considered as sensitisers regardless of their mode of contact with the body, including through inhalation 20. This is contentious, and further research into the relationship between dermal and respiratory sensitisation is required. Not only would this improve the identification of respiratory sensitisers, it might also assist in the development of more appropriate labelling and (primary) preventive measures at work. However, it is not yet established whether chemicals that are negative in skin sensitisation tests are thereby unlikely to cause respiratory sensitisation. Other means of premarketing product testing include clinical trials in human volunteers, such as that used to test an enzyme in a personal cleansing product 21.

Controlling exposure in the workplace

In the case of high-molecular weight agents, there is sufficient evidence, gathered from analytical epidemiology, of an exposure-response gradient for occupational allergens. There are, however, important limitations as follows. 1) Much, although not all, of the evidence relates more strongly to occupational sensitisation (the development of specific IgE antibodies) than to asthma. The relationships between these two outcomes are poorly understood, although most would argue that the acquisition of IgE sensitisation to some agents is a strong predictor (if not precursor) of asthma. 2) The “shape” of the exposure-response relationship is unknown and there is, as yet, no human evidence which reliably allows the setting of indisputable no-effect thresholds 22. 3) There is some evidence that high or continued exposures lead to immunological “tolerance” 23, although it is unclear whether these observations reflect survival pressures. 4) On the whole, it seems probable that the timing of high‐intensity exposure is important. For instance, exposure/response relationships are more easily demonstrated when they consider the intensity of exposure at the time of onset of disease rather than at the time of study 23.

Given the evidence for a broad exposure/response relationship, primary preventive efforts should be concentrated in the control of workplace exposures. There are, however, few examples (see below) of situations where a reduction in exposure alone has been demonstrated to result in a reduced disease incidence. This is probably due more to a lack of research interest in this area than to a lack of efficacy of the approach. Wegman 24 suggests, in addition, an increasing but potentially damaging demand for “perfect” understanding of disease aetiology. Whilst laudable in principle, this may simply act (passively or otherwise) to defer timely preventive action and, moreover, is probably an unrealistic appeal. Much responsible public health action in other areas has been taken in the face of scientific uncertainty.

Methods of control generally follow standard occupational hygiene tenets and range from elimination or substitution, through enclosure and ventilation to work practices, personal protective equipment and appropriate monitoring, and administrative processes 25. Elimination of an asthmagen from the workplace is unusual but occasionally practised. Enforcing suitable exposure levels to occupational asthmagens is difficult, because there are very few health-based legal standards. Those that exist are generally set on the basis of a substance's other hazardous properties (such as the irritant effects of isocyanates) or are of doubtful value (for example the current standard of 60 ng·m⁻³ for detergent subtilisins). The reasons for such omissions include: the lack of detailed evidence for most occupational sensitising agents concerning exposure-response relationships; the lack of consensus over whether there are or can be thresholds at which exposure(s) to sensitising agents induces no adverse health effects 26; and the technical difficulties in measuring airborne allergens at the low intensities at which they probably exert their effects.

It is probable that these obstacles will persist for the near future. Leaving aside the theoretical debates, there are far too few resources for detailed examination of the large number ofoccupational asthmagens. Legislative controls generally require, therefore, exposures to be kept to the minimum that is technically possible without undue financial, social or economic disruption. Where levels are set in such cases they are therefore “pragmatic” rather than “health based”.

In so far as any occupational exposure standards are valuable, these approaches seem reasonable. Indeed there appears little alternative. Difficulties arise, however, over issues of compliance. In almost all cases compliance is easier for large firms than small 27 and especially when the former is engaged in large-scale manufacture of only a few products. Compliance may be perceived as inimical to productivity, at least in the short term. There is also serious concern over howmeaningful many of these concepts are to those working in industry. Surveys of managers in UK chemical firms, mostemploying <10 workers, revealed an alarming level of misunderstanding or ignorance over legislated requirements concerning employees' exposure 28, 29. Although most were taking steps to control workplace exposures, it was apparent that set exposure limits were of limited influence in their decision making. These findings prompted the UK Health and Safety Executive to their current efforts to devise simplified methods of setting and conveying occupational exposure limits.

Methods to enforce or encourage compliance can, broadly, be divided into the legal and the economic. Since legal approaches are backed by economic penalties and economic approaches must be sanctioned in law, the distinction is perhaps less than real. In the case of OA, a legal approach is generally taken whereby workplace control measures are legally enforceable and infringements penalised by fine or occasionally factory closure. These approaches are fairly slow, require extensive regulatory surveillance and have not been proven to be effective alone in reducing the incidence of OA. Sanctions directed against the primary manufacturers of sensitisers, and not against the users, may be more rapidly effective. At present, the latter are not liable if they have made full disclosure of their product's hazardous nature. Many would argue that extension of this responsibility would be both unfair and unworkable; it again appears to be untested. Outright bans on substance manufacture and use have not, asfar as we are aware, been instituted, unlike for other hazardous substances, such as 2‐naphthylamine in the UK rubber and dyestuff industries.

Economic measures are either punitive, the taxing of hazardous substances, or exhortatory, whereby financial incentives are provided by government to support the development of safer alternatives. Both may be used in (phased) conjunction. In general, such approaches are more acceptable if there is limited use of the substance in question and if the costs ofreplacement are likely to be low. This is seldom the case foroccupational sensitising agents. Insurance premiums are generally set in the knowledge of estimated or measured risk at a particular workplace, although a flat rate may be applied to smaller enterprises; this may be unwise if the risks are generically higher in such settings. In the environmental field, it is recognised that tough regulation encourages technological development and, in the long run, often leads to cost savings (Porter's paradox). Interestingly, compliance with general environmental regulations appears to be much higher than that for occupational health risks; this may reflect differences in public perceptions of the relative importance ofthe two types of hazards.

Enzymes in the detergent powder industry

Following very high incidence rates of OA and reports of a small number of cases in detergent users, granulated proteases (accompanied by more stringent engineering controls) were introduced in the 1970s. Two publications 45, 46, each sponsored by major manufacturing companies, describe dramatic reductions in the prevalence of OA following these interventions (fig. 1⇓). Unfortunately, neither report incidence rates and it is not entirely clear how much of the falling prevalence might be due to factors such as the “survival” of unaffected employees. Nonetheless, the implication from each is that concerted (and successful) attempts to reduce workplace exposures can be highly effective in the primary prevention of occupational asthma.

Some 20 yrs later (fig. 1⇓) other enzymes were introduced in a similar, encapsulated format and most biological washing powders now contain protease in combination with an amylase, cellulase or lipolase. Recently, a very large outbreak of OA in a single factory, with most cases sensitised to more than one enzyme type, was described 47. The outbreak has not been fully explained but the rapid introduction of new enzyme types together with management failures to put industry guidelines into practice may have been important. An additional factor may have been the factory's obligation, through its position in the marketplace, to manufacture a large number of different detergent products in rapid succession. The lessons, perhaps, are that disease prevention may be contingent on forces beyond the immediate workplace environment; and that vigilance must be eternal, especially where new (potential) allergens are being introduced.

Diisocyanates

An example of more structured evaluation is provided by the experience of OA due to diisocyanates. Legislation, introduced in l983 by the Ontario Ministry of Labour, required monitoring of diisocyanate levels to maintain 8‐h concentrations <5 parts per billion (ppb) and short-term exposure levels <20 ppb. In addition, mandatory medical surveillance measures were introduced: a questionnaire and spirometry pre-employment, with repeated respiratory questionnaires every 6 months and spirometry at least on an annual basis. Workers with lower respiratory symptoms on questionnaire, or changes in spirometry, were required to have a medical assessment. There was no equivalent legislation to provide surveillance for other respiratory occupational sensitisers, although some nondiisocyanate-using firms (for example, those using enzymes) had their own programmes.

Throughout the evaluation period, diisocyanates remained the most commonly recognised cause of compensated occupational asthma (fig. 2⇓). Indeed, there was an initial increase in annual compensation claims after introduction of the surveillance programme, consistent with increased case-finding 48. This, however, was followed by reductions in the proportionate and actual numbers of accepted diisocyanate-induced asthma claims for the last few years in which data were reviewed. In addition, among workers in companies known to be in compliance with the programme, there was an earlier diagnosis of OA: a mean of 1.7 yrs after onset of symptoms compared to 2.7 yrs for workers without documented compliance. Similarly, there was a reduction in the mean time to diagnosis of diisocyanate asthma from 3 yrs to 2.1 yrs (p=0.014). Perhaps as a consequence, indices of asthma severity at the time of diagnosis suggested milder disease in those diagnosed in the second period of the study, both for the group accepted for diisocyanate asthma and the group accepted for other causes of OA.

When measured levels of diisocyanates were compared among diisocyanate-using companies who had compensated claims for OA with companies without accepted claims, the former were significantly more likely to have had a measured level of diisocyanates ≥0.005 parts per million 49. This is consistent with a previous report suggesting a low rate of sensitisation in a new plant engineered to minimise exposure levels 44.

These figures represent one of the first, large-scale attempts to measure the effectiveness of a prevention programme in OA. Although directed against a single agent, the intervention was aimed at a large number of firms of different sizes. The findings suggest that a combination of regulated exposure and medical surveillance can be successful in the (secondary) prevention of asthma at work; and even that the beneficial effects may spill over to other industries.

Laboratory animal allergy

Botham et al. 50, working in the occupational health unit of a large research establishment in the UK, studied a retrospectively assembled cohort of new employees working with laboratory animals. Those who entered employment between 1979 and 1982 were followed for 3 yrs; the cohorts starting in the subsequent 2 yrs were studied for 1–2 yrs. Figure 3⇓ depicts the incidence rates of respiratory symptoms attributed to laboratory animal allergy in each of the annual cohorts. In 1981 a new code of practice for working with laboratory animals was introduced at the site, accompanied by a series of educational lectures designed to increase awareness of the laboratory animal allergies.

Although the design of this evaluation is fairly crude, “one-group pretest-posttest” in Cook and CAmpbell's nomenclature 51, it seems reasonable to attribute the improvements, in part at least, to the new preventive programme. The study is unusual in this field for its relatively careful design and analysis; it remains a useful example of how primary preventive measures may be evaluated with relative ease at a single site, albeit one with an apparently very high incidence of disease.

A similar study has been reported from a pharmaceutical research institution in the USA 52. As above, there is some lack of clarity over the calculations of incidence used, but the findings suggest that laboratory animal allergy can be effectively prevented by a targeted programme; in this case including education and training, modification of work practices, engineering controls, the use of personal protective equipment and a standardised system of surveillance. The latter included yearly measurement of specific IgE antibodies to a variety of animal allergens encountered either at home or at work. A positive result to any of these had a low predictive value (30%) for self-reported symptoms; the comparative figure for a positive result to a workplace allergen was not reported. Botham et al. 53 suggest that the presence of specific IgE to laboratory species is highly predictive of subsequent clinical allergy.

Natural rubber latex

The annual number of allowed claims for latex-induced OA declined after the Ontario Workplace Safety and Insurance Board encouraged hospitals to use powder-free, low-protein or nonlatex gloves in 1996 54. In Germany, a widespread information campaign for hospital administrators was accompanied by a revision in the (compulsory) technical regulations for dangerous substances, which stated that only low-allergen, powder-free latex gloves should be used in healthcare settings 55. These changes took place between 1997 and 1998. Data on the number of suspected cases of skin or respiratory latex allergy reported to a large insurance company, covering around half of the nation's hospitals, were collected between 1996 and 2001 (fig. 4⇓). The temporal relationship, with a 2‐yr lag, between the decline in purchases of powdered gloves by acute care hospitals and the fall in the number of cases of OA was interpreted as evidence of success for the preventive programme. A decline in latex allergy has also been documented, although on a far smaller scale, in hospitals and dental schools where powdered latex gloves were substituted by low-protein or powder-free latex gloves 56, 57.

The story of latex allergy is instructive. There is increasing evidence, albeit largely anecdotal, that after only some 15 yrs, the epidemic of occupational allergy among healthcare workers is coming to an end. This has probably been achieved through the rapid understanding that glove dusting powder (and protein content) was directly related to the risk of sensitisation and, importantly, that each of these could be rectified with relative ease. It is tempting to suggest too that the very high profile of the disease, which was undoubtedly helpful in understanding its aetiology, was a result of its heavy impact on healthcare professionals.