METHODS

A retrospective review of Ontario Workers' Compensation Board (WCB (now renamed Workplace Safety and Insurance Board)) claims for all causes of OA which was previously performed for claims allowed between l984 and l988,⁴ was extended to claims allowed for OA started in the 14 year period between l980 and l993 inclusive. The questionnaire completed from each file was shortened from the one previously used⁴ but included data on demographic features of claimants, work descriptors, symptoms of asthma, reported improvement on weekends and holidays, and duration of symptoms of asthma before the main medical assessment. Probable causative agent for each case of OA was determined from the clinical records. Spirometry results were recorded as well as results (when available) of allergy skin tests, peak flow readings, methacholine challenges during a working period, and off work and specific occupational challenges. As these tests were usually arranged by the claimant's attending physician in many different centres, the methods used for these tests were not uniform.

Linkage was made between the data from the WCB files and information from the health information discharge abstract data base (from the Ontario Ministry of Health), from the time of the WCB accident date—the date of onset of symptoms, up to 31 December l996 as previously described^8,9—to compare the number of hospital admissions for asthma among those with OA from diisocyanates with that among those with OA from other occupational causes.

Statistical analyses

Data were entered into an SAS programme (release 6.12) for personal computer, using t tests and χ² tests as appropriate for analyses. To compare the annual rates of onset of accepted claims for OA induced by diisocyanates versus OA induced by other causes, the proportion of accepted claims for OA due to diisocyanates as a proportion of all accepted claims for OA was assessed for the period of 1980–86 and separately for 1987–93, using the Mantel-Haenszel χ² test for trend for each period. Subject variables were compared for subjects with new accepted claims, for OA attributed to diisocyanates and OA attributed to other causes. To assess changes over time, variables were initially compared for those whose onset of OA occurred in the first 7 years of the study period, with those whose onset of symptoms was in the last 7 years of the study. Analysis of variance (ANOVA) was performed to assess differences over period for those with OA caused by diisocyanates versus those with OA induced by other causes. Analysis of data on hospital admissions for OA from diisocyanates versus OA from other causes was based on the subset (810 out of the total of 844 OA claimants) with data available for smoking history and causative agent for OA. Multivariate analyses were performed adjusting for age, sex, smoking history, year of onset of symptoms, and work exposure (same work exposure, versus same workplace with reduced but possible exposure, versus no exposure) as well as exposure group (diisocyanates v other OA).

RESULTS

A total of 844 new claims for OA were allowed by the WCB with onset of symptoms between l980 and l993 inclusive. The number of accepted claims per year of onset steadily increased during this 14 year period for OA attributed to causes other than diisocyanates (419 claims) (fig 1). By contrast, claims allowed for OA attributed to diisocyanates (425 claims) initially rose parallel to OA induced by other causes, then rose more steeply than for OA induced by other causes between l988 and l990, and subsequently fell sharply by 1992–3. Overall, OA induced by diisocyanates accounted for 50.2% of all accepted claims during this period: initially around 50% in 1980–87, increasing to a maximum of 64% by 1988, then falling significantly to 29% in l992 and 35% in l993. There was no significant difference in the annual proportion of OA caused by diisocyanates from OA induced by other causes over the period 1980–86, but in the second half of the study period, 1987–93, there was a significant difference in this proportion over time (p=0.001) (fig 1). The most common attributed causative agents among the 844 claims are shown in table 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Number of allowed claims for OA induced by diisocyanates and OA induced by other causes by year of onset. A significant change occurred in the proportion of OA induced by diisocyanates and OA induced by other causes in the years 1987–93 (p=0.001).

The characteristics of claimants with OA from diisocyanates and from other causes are compared in table 2. Those with OA attributed to diisocyanates included more women (40%), were significantly younger, were less likely to be atopic, and were more likely to have smoked. They were less likely to have reported associated nasal symptoms (42% v 53%), and the onset of symptoms of asthma after the onset of exposure at the workplace was significantly earlier than with OA from other causes (mean 5 years v 7 years, p=0 .0001). The mean duration of symptoms while continuing to work before the main medical assessment in the WCB file was significantly shorter in those with OA induced by diisocyanates (mean 2 years versus 3 years (p=0.009)), and they were more likely to report improvement in symptoms at the time of the main medical assessment.

The proportion of each group who were still exposed to the implicated causative workplace agent by the main assessment in the WCB records were not significantly different (23% with OA induced by diisocyanates v 31% with OA from other causes). However, spirometric variables of forced expiratory volume in 1 second (FEV₁), vital capacity (VC), and forced expiratory flow at 50% vital capacity (FEF₅₀) at that visit, expressed as a percentage of the predicted normal, were significantly better for those with OA attributed to diisocyanates and they were less likely to be receiving inhaled corticosteroid treatment (p=0.001, table 3).

The differences between the OA due to diisocyanates and the OA due to other causes shown in table 3 remained significant after adjustment (for age, sex, smoking history, work exposure, and year of onset of symptoms) for FEV₁ (p=0.035), and for inhaled steroid usage (p=0.0003). Differences also remained significant between these groups for years of workplace exposure to the implicated work agent before the onset of symptoms of OA (p=0.0001), and duration of symptoms before leaving work (p=0.0008), when adjusted for year of onset of symptoms.

Linkage for hospital admissions with a primary or “most responsible diagnosis” of asthma up to the end of 1996 was analyzed for 414 accepted claimants for OA due to diisocyanates and 396 accepted claimants from OA due to other causes, for whom data were available on smoking history and likely exposure to an agent causing OA. A significantly smaller proportion of those claimants with asthma induced by diisocyanates had hospital admissions for their asthma (5.3%, table 3), than did those who had other attributed causes of OA (11.1%, p=0.003).

Changes in characteristics of occupational asthma over time

Comparison of the variables for those cases with onset of symptoms of OA induced by diisocyanates in 1980–6 was then made with those for whom the onset was in 1987–93 (table 4). Those with the onset of symptoms in the second half of the study period were younger (p=0.002) and there were relatively more women than in the first half of the study period. In the second half of the study period accepted claimants had a significantly shorter duration of symptoms of asthma before diagnosis or before leaving exposure to diisocyanates (mean 2.1 (3.2) years v 3.0 (3.9) years; p=0.014, table 4).

However, similar changes were found in those with OA induced by other causes. The proportion of women increased from 20% to 36% (p=0.001), and the period between the onset of symptoms and leaving exposure to the agent responsible in the second half of the study among those with OA induced by other causes, was 2.9 years versus 4.3 years (p=0.01)in the first half of the study. These differences over the two periods were not significantly different between OA induced by diisocyanates and OA induced by other causes.

Indices of severity of asthma at the time of diagnosis suggested milder asthma in those diagnosed in the second period of the study. Both for the group accepted for OA induced by diisocyanates (table 4) and the group accepted for OA induced by other causes, (data not shown), there was less limitation to airflow on spirometry in claims accepted in the last half of the study and less airway hyperresponsiveness on methacholine challenge while off work during the initial investigations (table 4). Again, these changes were not significantly different for OA induced by diisocyanates versus OA induced by other causes.

The risk of hospital admissions for asthma was about 50% less among those with OA induced by diisocyanates compared with other claimants with OA. This reduced risk was found both for claims with an onset of OA during 1980–86 (6.8% v 13.6%; OR 0.46 (95% CI 0.20 to 1.07)) and during 1987–93 (4.6% v 9.9%; OR 0.44 (95% CI 0.22 to 0.88)). The reduced risk of a hospital admission among workers with asthma induced by diisocyanates remained significant when adjusted for differences in age, sex, smoking, or period of claim as previously reported.⁹

DISCUSSION

The increase in the absolute numbers and in the proportion of allowed claims for OA induced by diisocyanates relative to OA induced by other causes in the same period l986–90 (fig 1) in conjunction with a shorter duration of symptoms of asthma at the time of the claim, a younger mean age of subjects, and milder asthma in those diagnosed during the last half of the study period is consistent with an increase in case finding of OA induced by diisocyanates during this time, possibly attributable to the medical surveillance programme.

An earlier diagnosis was made during the second half of the study period as reflected by the shorter time between the onset of symptoms and the main medical assessment in the WCB file. There were similar findings in claims accepted for OA induced by other causes but the period before the main assessment and the submission of a claim remained shorter for OA induced by diisocyanates throughout the study time. This earlier diagnosis for OA in the second half of the study period may have been related to better general awareness of OA in this period by workers and physicians in Ontario, perhaps related to national recommendations for diagnosis.¹⁰ It is also possible that for OA induced by diisocyanates the earlier diagnosis may have been related to the diisocyanates medical surveillance programme as suggested by our earlier study.⁷ Although a significant difference was not detected in the time related changes between the group with OA induced by diisocyanates and the group with OA induced by other causes, we do not have information on the extent of medical surveillance and education programmes which might have been in place in companies whose employees were at risk for OA induced by substances other than diisocyanates.

If related to surveillance programmes, a benefit may have been directly attributed to the measures of symptom questionnaires and spirometry, or may have been related to associated health education of the workers relative to OA induced by diisocyanates or other sensitisers, to increased awareness of this entity by their physicians in conjunction with the medical surveillance programme, or to industrial hygiene measures associated with the companies' control programmes.

It is less likely that the increase in accepted claims for OA induced by diisocyanates during the central period of the study, and subsequent fall, were related to a change in acceptance practices of claims by the adjudicators and the WCB, as claims for OA from other causes continued to increase (with some fluctuations) on an annual basis.

The lower risk of hospital admissions for those with OA induced by diisocyanates than with OA from other causes was reported by us⁹ in a fuller analysis of hospital admissions in this cohort of compensated claimants with OA compared with compensated injured workers and patients with non-occupational asthma. The reduced risk of a hospital admission among people with asthma induced by diisocyanates remained significant when adjusted for differences in age, sex, smoking, or period of claim, and therefore does seem to reflect milder asthma in these subjects, consistent with our findings from the other markers of severity of asthma in this study—such as spirometric variables, prescribed inhaled steroid therapy, and the trend to milder responses on methacholine challenge.

The decline in accepted claims for OA induced by diisocyanates after 1991 (during 1992 and 1993) down to lower numbers than those for OA induced by other causes cannot be explained only on the basis of a transient left shift of the incidence curve due to earlier finding of cases. The fall in annual numbers relative to the number of cases of OA induced by other causes may indicate a true fall in incidence of cases induced by diisocyanates, perhaps due to improved occupational hygiene measures in the workplace such as the use of robots in high exposure areas and routine monitoring of diisocyanate concentrations to conform with exposure guidelines. Another possible explanation for the fall in annual numbers of accepted claims might be random fluctuation, and a longer follow up period would be necessary to exclude this. Finally, it is possible that this apparent fall in incidence may reflect a decline in quality or availability of surveillance of diisocyanates in the last few years of the study. In the early years of the programme, the Ontario Ministry of Labour provided surveillance for many Ontario companies, but this was phased out in the early 1990s. Also, in 1991 participation by workers in the medical surveillance programme became voluntary rather than mandatory, as it had been before. Potentially this could result in delays in detection and a transient fall in accepted claims in the last few years of the study. In that event, the numbers would be expected to rise again in subsequent years, with an increase in duration of symptoms before diagnosis, and indices of severity of asthma.

There is no evidence of a reduction of use of diisocyanates in Ontario to account for a declining incidence of OA, and imports of diisocyanates into Canada have continued to rise.¹¹ However, there is no direct information available as to the number of workplaces or workers using diisocyanates, and whether this may have changed over time. A previous study by us⁶ reviewed data from 223 worksites that used diisocyanates, for which the Ministry of Labour of Ontario had exposure data on diisocyanates, but such monitoring after 1990 was not maintained by this Ministry in a systematic manner and there is no ongoing comprehensive record as to the compliance of companies that use diisocyanates with the medical surveillance and air monitoring requirements.

In conclusion, our findings indicate that OA due to diisocyanates in Ontario has been diagnosed earlier and has been associated with better pulmonary function and a better outcome as assessed by hospital admissions than OA induced by other causes. There has been an earlier diagnosis of OA, both induced by diisocyanates and by other causes especially in the second half of this study period. There is an apparent decline in incidence of OA induced by diisocyanates in the last few years of this study, which might reflect improved occupational hygiene measures at work or might reflect early identification of very mildly symptomatic workers and transfer of these workers to unexposed areas without submission of WCB claims, or without objective confirmation of the diagnosis. This would be consistent with our review of one company.¹² However, further follow up is needed to confirm this trend to a reduced incidence of OA induced by diisocyanates.