INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Adult-onset asthma is common among people of working age and its incidence appears to be increasing (1). More than 200 specific agents encountered at work can cause asthma (2). The role of nonspecific respiratory irritants at workplaces has been less studied but may be more important than usually thought (3, 4, 5). The median of the published estimates indicates that 9% of adult asthma might be attributable to work-related factors (6). The work-related portion of adult-onset asthma may, however, be underestimated as not all occupational etiological factors have been identified and very large studies would be needed to reliably discern the etiological share even for occupational causes so far recognized.

The above-mentioned median estimate of work-related attributable fraction in adult-onset asthma is mainly based on case-control studies, cross-sectional population surveys, or incidence data from compensation or surveillance schemes of occupational asthma (6). Only one population incidence study has so far been published, but did not cover an entire national population and included only a 3-mo follow-up (7). Large population incidence studies could, however, detect the variability of asthma incidence between occupational groups. This would allow for the estimation of the attributable fraction of work even when not all specific causative factors are known. Finland provides a unique opportunity for such studies. In Finland, individuals with clinically well-established persistent asthma are eligible for special reimbursement of medication from the national health insurance scheme and are therefore registered. By combining, at an individual level, these data with the population census data we estimated the attributable fraction and risk of adult-onset asthma that was associated with occupation in a follow-up study of the entire employed population of Finland in 1986-1998.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All employed Finns were followed during 1986-1998. Technically, we included three consecutive census cohorts, each followed for 5 (or 3) yr, that is, all employed Finns without preexisting asthma and aged 25-59 yr in December 31, 1985, 1990, or 1995 were classified according to their occupation and were followed for asthma during 1986-1990, 1991- 1995, or 1996-1998, respectively.

The follow-up was done through two national registries: the Medication Reimbursement Register of the Social Insurance Institution (SII) and the Finnish Register of Occupational Diseases. Duplicate notifications between the registries were identified and the registries were matched to the censuses by the unique personal identification code. The matching succeeded in 99.9% of cases. The cases and person-years were calculated according to the occupation the individual held at the time of the census. The study protocol was approved by the ethical committees of the participating institutions.

In Finland, the medication for asthma can be reimbursed at a higher than ordinary level. The patient has to apply for the entitlement from SII and provide a medical certificate written by a chest physician. The criteria of persistent asthma must be fulfilled (Table 1). The national health insurance system and the registration cover the entire population. As the only exception, patients with recognized occupational asthma are compensated for medication from the Statutory Accident Insurance. They are registered by the Finnish Register of Occupational Diseases (FROD) and a diagnosis of asthma made by a chest physician and evidence of causality between a specific workplace exposure and the disease are required.

An individual was defined as an incident case of asthma when he or she received rights for reimbursement for asthma medication from SII (95% of cases) or was notified for recognized occupational asthma to FROD (5% of cases).

The incidence rate for each occupation was calculated by summing up the case and person-year data from the three census cohorts. Incidence rate ratios (hereafter relative risks) were estimated by a log-linear model adjusted for age. Administrative workers were used as a reference group.

In addition, a dichotomized occupation was formed to compare the risk of asthma between those employed in administrative work ("unexposed") with those employed in the remaining occupations ("exposed"). The population attributable fraction was estimated according to the formulas given for adjusted analysis (8).

The use of a reference group including only the lower social class administrative employees did not change the results. Individual smoking habits were not known, but information on smoking habits by occupation was available from pooled samples (n = 34058) of national annual lifestyle surveys in 1978-1991 (9).

Granted reimbursement of medication has been previously used in studies on asthma epidemiology but has not been validated (10, 11). We assessed the reliability of the asthma diagnosis in a representative sample of the Finnish population (n = 12936) examined by SII (12). Altogether 205 individuals had received reimbursement rights for asthma medication in 1986-1998 and were aged 25-59 yr at that time. We analyzed the SII medical certificates and, if the medical certificate was incomplete, data from hospital records (with the patient's permission).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

There were 49,575 incident cases of asthma during the follow-up, of which 2,464 were cases of recognized occupational asthma. The incidence rate of asthma was 1.65/1,000/yr for men and 2.47/1,000/yr for women. The age-adjusted risk was increased especially in occupations of agricultural, mining, manufacturing, and service work (Table 2). When nonadministrative work (exposed) as a whole was compared with administrative work (unexposed) in the age-adjusted model, the relative risk was 1.45 (1.37-1.53) in men and 1.27 (1.23-1.31) in women. These risk estimates did not change when adjustment for the follow-up period was included in the model. Based on the above risk estimates and the exposure prevalences among cases, the overall attributable fraction of exposure was 29% (95% CI 25- 33%) for men and 17% (95% CI 15-19%) for women.

As individual smoking habits of the entire employed Finnish population were not known, we used occupation-specific aggregate data on smoking habits to estimate confounding by smoking (see METHODS). The prevalence of regular smokers in men was 27% in administrative work and 35% in nonadministrative work and in women was 22% in administrative work and 18% in nonadministrative work. To estimate the maximum confounding smoking could introduce to the exposure effect, we used the extreme assumption that smoking would cause a 2-fold risk of asthma. By using this assumption and the above smoking prevalences it was found that the relative risk estimates based on dichotomized comparison of nonadministrative and administrative work would change only slightly: a 6% reduction for men and a 3% increase for women.

Assessment of the Reliability of the Diagnoses

For the assessment of the reliability of the asthma diagnoses, a complete medical certificate or the case history from the hospital records was available for all those identified 205 cases of the representative population sample, which belonged to our study population. The diagnostic criteria of Table 1 were fulfilled in 183 (89%) of the cases and in another 20 cases (10%) the diagnosis of asthma had been made by a chest physician on the basis of other clinical data (most often a history of typical asthma symptoms together with observed bronchial hyperresponsiveness). In addition, there were two cases of posttuberculotic respiratory problems in which asthma was clearly not present. Of the 203 patients with asthma, 20 also had another chronic pulmonary disease (18 patients had chronic obstructive pulmonary disease [COPD] and two had chronic symptomatic bronchiectasis).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We followed the employed population, which is the fraction of the adult population in which the health effects of workplace exposures should be the clearest. As the study covered the entire national workforce, all the relevant workplace exposures in Finland were covered. The case ascertainment was done through two national registries that rely on diagnoses made by chest physicians based on clinical or physiological criteria and representing a persistent form of asthma. This definition of asthma (i.e., clinical assessment) is the standard that is used to validate the other definitions used in epidemiological studies (13). The examination of the asthma diagnoses indicates that the positive predictive value of our definition is very high as compared with the definitions of asthma that have been used in most epidemiological studies. As the level of reimbursement of medication costs is relatively high (75% of total costs) it is also likely that all those who fulfil the criteria are registered. On the other hand the sensitivity of our definition is lower than for most other definitions of asthma, as it includes only persistent cases of asthma. Consequently our results apply to occupational risk of persistent asthma and the work-related excess may represent two types of causation. Factors at work may have induced a new-onset persistent asthma in an individual without any history of asthma or they may have aggravated a preexisting mild disease to fulfill the SII criteria for granted reimbursement. We could use the first date of granted reimbursement rights of asthma medication to determine the timing of asthma diagnosis (whether during childhood or as adult), whereas we did not have information on the preexisting occurrence of milder asthma or asthma-like symptoms in these individuals. This is very unlikely to have caused any differential bias between administrative workers and nonadministrative workers, although it does not allow estimation of the relative roles of the above-mentioned two types of work-related causation of persistent asthma.

The attributable fraction of occupational factors in adult-onset asthma in the employed population was 29% for men and 17% for women. This is much more than the 9% median of previously published estimates (6). In Finland, it was previously estimated that 6% of adult-onset asthma in men and 4% in women would be attributable to work-related factors (10). These estimates were based on occupational asthma as defined by the national compensation scheme. Also, the previous estimates on occupation-specific incidence rates of work-related asthma were based on the same source of data (14). These types of data have acknowledged weaknesses in coverage and case ascertainment (15). Our present results take into account the total work-related excess of adult-onset asthma incidence and indicate that the overall importance of factors at work (attributable fraction of 29% for men and 17% for women) might be five times greater than previously thought and the number of risk occupations much larger than previously identified. Our estimate is also slightly higher than those recently reported for Canada (18%), the United States (21%), or the 15% median of the studies with the highest quality ranking in a review article (6, 7, 16). Due to their smaller size, these studies may not have covered all relevant exposures of a national population and were not able to estimate the risks for occupations with a relatively small workforce. Nevertheless, underestimation of the risk and the attributable fraction may still have occurred in our study because a complete lack of harmful occupational exposure in our reference occupation is not tenable (e.g., exposure to paper dust, emissions from office equipment, or other impurities due to poor ventilation). Change of job due to respiratory symptoms is frequent before the diagnosis of asthma and may also have biased the risk estimates toward unity (17). Exposure to workplace air pollution occurs more often in occupations associated with a low socioeconomic status. Such an association could introduce confounding due to lifestyle factors and possibly result in overestimation of the work-related risk. A separate analysis using lower-level administrative workers as a reference group did not indicate such confounding of the work-related risk estimates.

There is conflicting evidence as to whether smoking is a risk factor for adult asthma, although there are more studies showing no association (18) Recent case-control studies of work-related risks of asthma have reported that adjustment for smoking had only a minimal effect or no effect on the risk estimates (3, 17, 23). We could not adjust for smoking at an individual level, but the differences in the prevalence of smoking between administrative and nonadministrative work were small and could not likely cause any important bias to our risk estimates.

The work-related excess of asthma was mainly due to an increased incidence in agricultural, manufacturing, and service work whereas those in technical, scientific, and humanistic work and those in sales work showed an incidence similar to administrative workers (Table 2). Agricultural work, various manufacturing occupations, and cleaners were identified as risk occupations for asthma also in a recent multinational cross-sectional study that reported an attributable fraction of 5- 10% for occupation in a population aged 20-44 yr (24).

In Finland, most recognized cases of occupational asthma are caused by sensitizing agents, animal epithelia, and flour dust, accounting for more than 50% of all reported cases (14). The primary prevention of occupational asthma caused by sensitizing agents has not so far been very successful. Even among the well-known causative agents of occupational asthma, the annual reported number of cases has decreased only for isocyanates during the past 10-15 yr in Finland (25). In our study, occupational asthma accounted for only 5% of all cases of asthma, which is clearly less than the attributable fraction estimate of 17-29% for all occupational factors, that is, the majority of the work-related excess may have not been the target for any preventive actions, which have mainly been guided by surveillance data concerning recognized occupational asthma. Therefore it is perhaps not surprising that adjustment for the follow-up period did not influence the work-related risk estimates in our study. This does not, however, mean that preventive actions could not have been effective in specific settings that could not be properly analyzed with our registry based design, for example, prevention of latex-induced asthma in health care departments at risk. But overall, the prevention of work-related excess of adult-onset asthma still appears to be a future challenge. The observed incidence rates and the attributable fraction estimates indicate that each year four or five Finns per 10,000 develop work-related persistent asthma. Whichever the mechanism, the adverse public health impact of occupational factors in adult-onset asthma and the potential for prevention seem much larger than previously assumed.