INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Occupational asthma (OA) is the most prevalent respiratory occupational illness (1, 2). High-molecular-weight (HMW) agents are related to the development of OA through an IgE-mediated mechanism; and atopy is a predisposing factor (3). Among the different HMW agents present in the workplace, exposure to animal proteins for laboratory technicians, to flour for bakers, and to latex for health workers, are among the most frequent causes of specific immunologic sensitization (SIS) and OA (7, 8).

Most epidemiologic studies of OA have used a cross-sectional design that likely resulted in a survival bias, i.e., nonaffected workers were more likely to stay on the job than affected ones. This bias was hypothesized to be more noticeable in the case of OA than for mineral dust-induced occupational lung diseases, owing to a shorter latency period between first exposure and onset of disease in the first case (9). In order to avoid this problem, cohort studies performed in workers exposed to HMW agents have enrolled the population to be studied at the time they joined the workforce at risk for OA (10), or shortly thereafter (11, 12). These studies, however, did not consider that subjects who join a workforce may have been exposed to agents related to OA, during their apprenticeship or in their current occupation. Because SIS to HMW agents and work-related symptoms in most cases appear during the first years of exposure (13), one can hypothesize an early health-related selection during apprenticeship that will affect the population joining the workforce.

Hypothesizing that individual characteristics related to SIS and/or OA can be associated with early quitting of apprenticeships with HMW-agent exposure, we studied a cohort of students who joined career programs in animal health, pastry-making, and dental hygiene, to determine if health-related factors influenced the completion of the apprenticeship.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Context of the Study

A cohort of students entering career programs that involve regular exposure to known HMW sensitizers was identified and followed prospectively from 8 to 44 mo to determine the incidence of work-specific immunological sensitization (SIS) and work-related symptoms. Baseline characteristics of this cohort at entry in the study have been described elsewhere (16). As an ancillary study (attrition study), subjects who quit their career program during follow-up were examined and compared with subjects who pursued the career, to measure the association between individual characteristics assessed at baseline values and during the training period, and quitting. All cohort members answered a questionnaire and had allergy skin tests, spirometry, and measurement of nonspecific bronchial responsiveness as a preexposure assessment. Follow-up visits were scheduled every year, and the same measurements were repeated. All subjects gave written consent to their participation in the study. The protocol was accepted by the ethics committee of the Hôpital du Sacré-Coeur de Montréal.

Subjects

Students enrolled during academic years 1993-94, 1994-95, and 1995- 96 at teaching institutions offering a career program in animal health technology (AHT), pastry-making (PM), or dental hygiene technology (DHT) in Quebec and located within 300 miles of Montreal, and the only school of veterinary medicine in the province of Quebec, were the sampling frame. Because the exposure to laboratory animals was similar in AHT and veterinary medicine, the last group of students was considered part of the AHT program for the purpose of the study. Details of the enrollment procedure have been reported elsewhere (16). A total of 769 subjects, representing 77% of those eligible, were enrolled in the study (participation rates: AHT n = 417, 81.3%; PM n = 230, 73.1%; DHT n = 122, 70.5%). Most participants were female (85.0%), less than 20 yr old (51.2%), and nonsmokers (77.4%), with a similar prevalence of positive skin tests to common and specific allergens in the three programs at the initial visit. At baseline, 57 students of AHT were sensitized to at least one of the laboratory-animal-derived allergens used for skin-prick testing (13.8%), 11 students in PM were sensitized to flour (5.0%), and 1 DHT student was sensitized to latex (0.8%).

Study Design

Subjects who quit their program were identified from periodic school reports and scheduled in the 3 mo after quitting for a visit including a questionnaire and allergy skin-tests, with the same standardized procedure followed at baseline and by the same research personnel including a nurse and a technician. Quitters who did not attend their scheduled visit were recontacted and asked to answer the questionnaire by phone interview. Subjects who followed their career program had their regular follow-up visits after an average of 8 mo into their apprenticeship and yearly up to 44 mo; the follow-up visits included the same questionnaire and complementary tests.

Study Measures

Questionnaire. A questionnaire derived from the International Union against Tuberculosis and Lung Disease (IUATLD) questionnaire (17) was administered at baseline by a trained nurse. Presence of asthma symptoms was defined as at least two positive answers to questions about wheezing, chest tightness, or shortness of breath under usual conditions or under such conditions as exercise, exposure to cold air, heavy smells, smoke, and dusts. Smoking habits and self-reported hay fever were assessed at baseline; nasal, eye, respiratory, and skin symptoms while exposed to the specific agents present at work, before quitting or currently, were assessed at the follow-up visits in quitters and stayers, respectively. Additionally, subjects who quit their career program were asked to give their reasons for quitting.

Skin testing. Skin tests were performed at baseline by the prick method (18). Histamine phosphate (1/200 mg/ml) was used as a positive control and diluent (glycerine, 50%) as the negative control. The largest weal diameter was assessed 10 to 15 min after the introduction of the antigen extract. A positive reaction was defined as a weal equal to or higher than 3 mm in the absence of reaction to the diluent and in the presence of a positive reaction to histamine phosphate. Atopy was defined as at least one positive reaction to common inhalants. In addition, skin-prick tests were performed with extracts from aeroallergens present in the working areas of the students: urinary proteins from rat, mouse, and rabbit (Pharmacia Allergon AB, Angelholm, Sweden) and hamster, guinea pig, and rabbit dander (Omega, Montreal, PQ, Canada) for AHT; mixed flours, wheat and bran flour (Omega) for PM; and latex (Omega; and Stallergènes, Institut Pasteur, France) for DHT. Skin-prick tests with the previously mentioned specific allergens were administered to all participants regardless of their career program.

Spirometry with the assessment of FEV₁ and FVC was carried out using a Collins apparatus (Survey/1 Plus; Collins, Braintree, MA) according to published standards (19). The methacholine challenge test was selected for the assessment of nonspecific bronchial responsiveness, and was performed with a Wright's nebulizer (output = 0.14 ml/ min), breathing at tidal volume for 2 min according to a standardized procedure (20). The maximal concentration of methacholine used was 32 mg/ml. The provocative concentration causing a 20% fall in FEV₁ (PC₂₀) was interpolated from individual dose-response curves drawn on a semilogarithmic scale. Reference values for FEV₁ and FEV₁/ FVC were taken from Knudson and coworkers (21). Significant bronchial hyperresponsiveness was set at a PC₂₀  8 mg/ml (20). FEV₁ was dichotomized into < 85% and  85% of predicted value as this was judged as being a reasonable cutoff point for normal and abnormal values in this young age group.

Analysis of Results

Permanence in the career program was the binary outcome variable. Baseline subject characteristics, including gender, age (two categories, above and below the median: > 20 versus  20 yr), atopy, SIS, rhinitis on contact with pets, hay fever, physician-diagnosed asthma, shortness of breath, asthma symptoms, cough, airway hyperresponsiveness (AHR), as well as career program (transformed into dummy variables: AHT versus others, and PM versus others) were the first group of independent variables examined. Work-related symptoms and symptoms of asthma in general during attendance in the program, constituted the other set of independent variables examined.

Univariate logistic regression analyses were used to estimate odds ratios (OR) and 95% confidence intervals (CI) for quitting the career program in relation to the baseline variables listed above. Analyses were performed for the whole cohort, and for each program group separately, in which case only factors with a significance level of p  0.20 from the analysis in the whole cohort were examined. Although the study focus was on development of sensitization from exposure to HMW agents, the different proportions of incident cases of sensitization and of subjects quitting their training according to training programs justified the examination of factors in the three programs separately. Multivariate logistic regression analyses were performed, both for the whole cohort and for the AHT subgroup using the same criteria (p  0.20) for inclusion of independent variables in the models. Descriptive statistics were computed for symptoms reported during attendance in the program before quitting or at the first follow-up visit, accordingly.

Analyses were performed using SPSS for Windows, version 7.5 (Chicago, IL) and StatXact 3 (Cytel Software Corp. Cambridge, MA). All statistical tests were two-sided, and a p value less than 0.05 was considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Eighty-nine of 769 subjects (11.6%) quit their career program during the follow-up. The attrition rate was different in the three programs studied: 36 of 417 in AHT (8.6%), 43 of 230 in PM (18.7%), and 10 of 122 in DHT (8.2%). Seventy-four subjects (82%) were traced and agreed to participate in the attrition study after quitting; the average time since entry in the career program was 9.6 ± 6.8 mo. From the AHT program, 26 of 36 (75%) answered the questionnaire and 23 of 36 (63.9%) accepted skin testing; from PM, these proportions were 36 of 43 (83.7%) and 10 of 43 (23.3%), respectively, and from DHT, 10 of 10 (100%) and 7 of 10 (70%).

Baseline characteristics among quitters who participated in the attrition study and those who did not participate were examined. The OR for responders versus nonresponders, for age  20 yr, hay fever, shortness of breath, and symptoms of asthma were, respectively, 0.97 (95% CI 0.61 to 1.53), 1.14 (95% CI 0.52 to 2.46), 1.35 (0.46 to 3.98), and 0.90 (0.67 to 1.20), suggesting that the two groups were relatively comparable for these factors. For baseline SIS, the proportions were 10 of 71 (14.1%) in responders and 1 of 15 (6.7%) in nonresponders; the small number of subjects with baseline SIS among nonparticipants made the interpretation of the OR impossible.

Only two subjects (2.3%) reported quitting due to rhinoconjunctivitis symptoms at work. Table 1 lists the relevant baseline variables examined in relation to quitting in the whole cohort. Hay fever was significantly associated with quitting, as was age. Shortness of breath was of borderline significance. The program was a strong determinant. The proportion of subjects with FEV₁ < 85% predicted was 3.7% in quitters and 5.4% in stayers. Only one subject among those who quitted had an FEV₁/FVC < 85% of predicted. In a multivariate logistic regression analysis with the following variables: program attended, age group, history of hay fever, shortness of breath, symptoms of asthma, cough, the only significant determinants for quitting were a history of hay fever (OR = 1.66, 95% CI = 1.0 to 2.75) and attending the PM program (OR = 2.33, 95% CI = 1.11 to 4.91).

Table 2 gives the same information after subgrouping by program. In AHT, immediate skin reactivity to laboratory animals, hay fever, and shortness of breath on entry in the program were all significant determinants for quitting. In the PM program, age was the only significant determinant. Hay fever was significantly associated with quitting in DHT. The multivariate logistic regression analysis carried out in the AHT program, including variables that were significant at a p  0.2 level in the univariate analyses, showed that risk of quitting was increased, although not significantly so, with immediate skin reactivity to laboratory animals (OR = 1.97, 95% CI = 0.78 to 4.98), a history of shortness of breath (OR = 2.70, 95% CI = 0.73 to 10.06), and hay fever (OR = 1.73, 95% CI = 0.76 to 3.92). No multivariate analysis was performed in the PM program because only one factor (age) was significant; nor was it performed in the DHT program because of the small number of quitters.

For the whole cohort, symptoms of asthma in general and work-related symptoms present before quitting were not significantly associated with the likelihood of quitting (Table 3). In the AHT group, the occurrence of asthma symptoms in general during the follow-up period was a determinant for quitting (Table 4). Among the 40 quitters who underwent skin testing at follow-up (before and/or after quitting), 3 of 23 (13.0%) acquired SIS in AHT, compared with 52 of 395 (13.2%) in stayers who attended their first follow-up visit. In the DHT program, 1 of 7 (14.3%) quitters compared with 4 of 110 (3.6%) stayers acquired SIS to latex. None of the quitters in PM developed SIS.

Finally, in a multivariate model for the AHT group, including characteristics at baseline (hay fever, SIS) and at follow-up (symptoms of asthma in general), baseline SIS (OR = 2.8, 95% CI = 0.94 to 8.38) and symptoms of asthma during follow-up (OR = 2.2, 95% CI = 0.94 to 5.38) appeared as the main determinants of quitting, though of borderline statistical significance. Equivalent models were not developed for the other two programs for reasons mentioned previously.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study shows that some characteristics at baseline, namely attending the PM program and reporting hay fever, are associated with the risk of quitting an apprenticeship program in which students are exposed to HMW inhalants. Although an attrition bias has been suspected in these careers (9), this is the first study, to our knowledge, that demonstrates such an association in a prospective epidemiological design.

In the subgroup of students exposed to laboratory animals, we found in the univariate analysis that baseline SIS, a history of hay fever and shortness of breath were associated with quitting, as were symptoms of asthma occurring during the first year of training. SIS to laboratory animals is a frequent condition. Ten to 15% of exposed subjects will develop work-related symptoms (rhinoconjunctivitis, urticaria, or asthma) (4, 13) in their first year of exposure and more will do so in the following years (14, 15, 22). Work-related symptoms are more prevalent in subjects with SIS to laboratory animals (11, 13). By multivariate analysis, in the same group, there was also a tendency, though not statistically significant, for baseline SIS, a history of shortness of breath and hay fever to be associated with quitting.

A self-selection at entry for laboratory-animal-exposed workers according to health characteristics has been suggested in some cross-sectional studies. Renström and colleagues performed a survey on 155 laboratory technology students and determined that subjects who entered the workforce as laboratory technicians had a 4.7% prevalence of SIS to laboratory animals at baseline, compared with a 13.4% prevalence in subjects who decided not to work with laboratory animals, a difference that was marginally significant (23). A survival bias in laboratory animal workers was also suggested by the work by Newill and colleagues who found a low prevalence of nonspecific bronchial responsiveness in workers with an exposure to animals longer than 2 yr (24). Baseline atopy was not associated with quitting in our cohort, a finding that has also been observed by others (23) nor were work-related symptoms (rhinoconjunctivitis, urticaria, and/or asthma) at follow-up. We also previously showed that sensitization to cats and dogs as well as to common allergens did not differ in the three groups of students, suggesting that immunological sensitization to these common allergens did not represent a selection bias on entry (16). These results apply for apprenticeships of 10 to 44 mo duration, which does not rule out the possibility that these apprentices will fail to join workforces with animal exposure or leave the workplace at a later stage depending on the severity of their work-associated symptoms.

Age was the only factor associated with quitting the PM apprenticeship, as a greater proportion of older subjects (86.0%) than of younger ones (66.8%) left this career program before the end. Socioeconomic reasons, not assessed in the present study, can be hypothesized for the observed effect, but we cannot discard other age-related causes for this association. Baseline hay fever was the only characteristic associated with quitting in the DHT program. It is possible that the lack of association between SIS and quitting these two programs was related to the lower prevalence of baseline SIS to flour (5.0%) and to latex (0.8%), in the respective programs, by comparison with laboratory animal workers in whom the prevalence of SIS to laboratory animals was 13.8%.

It is interesting to note that no incident work-related symptoms or SIS were found to be associated with quitting. In the case of work-related symptoms, the explanation may well be that subjects tended to conceal from themselves and/or from the interviewer that quitting could be associated with health repercussions. On the other hand, quitters were traced within 3 mo of leaving their program, thus they may have tended to forget or deny symptoms at work and emphasize symptoms in the general environment instead. Therefore, a recall bias is possible in answers to the questionnaire. However, baseline atopic status and SIS were assessed through objective skin-prick testing. Another possible explanation is the short duration of exposure before quitting the apprenticeship, which, combined with the fact that no participants had been exposed to the specific allergens for a significant duration (< 3 mo) before entering the program (this was a criterion on entry), may have precluded development of work-related symptoms. In case of incident SIS, the low participation rate (55.1%) and the low number of incident cases (n = 4) precluded any interpretation of results.

We have previously shown (16) that students entering the AHT program did not show a higher prevalence of IgE-mediated sensitization to pets than students entering PM and DHT programs. Moreover, we found that prevalence of SIS to laboratory animal urinary proteins was similar in the three groups of students, before entering their career program. Therefore, it is unlikely that prior specific sensitization or self-selection occurred here.

Those lost to follow-up among quitters had no impact on the interpretation of the association between baseline characteristics and quitting, because baseline questionnaire data were available for all subjects, and skin-test results for 96.6% of them (86/89). A bias resulting from those lost to follow-up, when symptoms acquired during the follow-up for the whole cohort together are considered, is unlikely given the high participation rate in the attrition study (82%) and the comparability in the proportion of quitters and stayers with the relevant risk factors at baseline. However, we cannot exclude that some self-selection to participate in the cohort study (77%) may have occurred.

In summary, besides attending the PM program, reporting hay fever is a significant risk factor for quitting a career program in which students are exposed to HMW allergens. Our findings support the hypothesis of the existence of an attrition bias in the study of the natural history of the development of OA (9).