Respiratory health among bleachery workers exposed to ozone and chlorine dioxide

among bleachery workers exposed to ozone and chlorine dioxide. Objectives This study investigated the possibility of occupational exposure to ozone increasing the risk of obstructive airway disease among bleachery workers. Methods Bleachery workers (N=129) from two Swedish pulp mills using ozone for bleaching were studied together with referents (N=80) from adjacent paper mills. The pulp mills had previously used chlorine dioxide as the bleaching agent. Testings included spirometry, methacholine challenge testing, and questionnaires. Area samplings showed sporadic ozone levels exceeding 0.9 ppm. Results There was a greater prevalence of wheezing (25%) among the bleachery workers with a history of gassings (from ozone, chlorine, or sulfur dioxide) than among those without gassings (18%) and among the referents (13%). Among the current smokers the fraction with a slightly increased bronchial responsiveness to methacholine was greater among the bleachery workers reporting gassings than among those that had not been gassed. For the period from 1992 to 1996, when the mills were using ozone, there was an increased incidence rate of wheezing among the workers in the bleachery (incidence rate ratio 2.3, 95% confidence interval 1.6–5.8). Conclusions Repeated exposure to irritants increases the risk of asthma-like symptoms. This finding reinforces the view that repeated peak exposures to irritants must be

Pulpmill workers can be exposed to a wide variety of irritant gases, especially chlorine or chlorine dioxide (1) and sulfur dioxide (2). Some epidemiologic studies have shown an increased prevalence of wheezing and decreased lung function (3)(4)(5), especially among bleachery workers. In that group a history of gassing events seems to be a major risk factor (3), especially for smokers (5). Exposure to irritants is also of importance in other occupations (6). In a longitudinal study of 211 metalworkers, self-reported gassings with chlorine were associated with increased bronchial hyperresponsiveness and impaired lung function (7).
Ozone was introduced as a bleaching agent in the Swedish pulp industry in 1992 (8). Process disturbances caused bleachery workers to be accidentally exposed to high concentrations of ozone (over 1 ppm). In ambient air the concentrations of ozone can vary, but in Swedish cities they seldom exceed 0.15 ppm. Ozone is a respiratory irritant, and respiratory health effects had been described already at the end of the 19th century. The symptoms following acute exposure to high concentrations that have been described include nose and throat irritation, coughing, dyspnea, chest pain, and severe fatigue (9). It has been concluded in two reviews that short-term exposure to ozone at ambient concentrations will result in airway inflammation (10,11). It is however unclear whether repeated exposure to high levels contribute to the development of a new onset of asthma among adults (10).
We report on a cross-sectional study of a large group of bleachery workers with long-term occupational exposure to chlorine dioxide, and more recent exposure to ozone following its introduction as a bleaching agent. The specific aim of the study was to investigate the extent to which occupational ozone exposure was associated with respiratory health effects, especially with the occurrence of obstructive airway disease.

Subjects and methods
All process workers, maintenance workers, and laboratory workers (N=139) in the bleaching department from two sulfate pulp mills (A and B) were selected as exposed subjects. Both mills had been using chlorine or chlorine dioxide as bleaching agents since the 1950s. In mill A, chlorine dioxide was replaced by ozone in 1992, and in mill B the replacement took place in 1993. Chlorine dioxide had however been used periodically up to 1995.
The study was carried out in 1996. Ten of the exposed subjects refused to participate. All the process workers (N=93) from two adjacent paper mills were recruited as unexposed referents. The paper mills made printing paper. Dust measurements in similar paper mills have shown low levels (around 0.5 mg/m 3 ) of paper dust (12). Thirteen refused to participate. In all, 129 bleachery workers and 80 referents were included in the final study (table 1).
All the subjects were sent a respiratory questionnaire, containing items similar to previous questionnaires (13)(14)(15). The key items regarding symptoms and diseases are shown in the appendix. All the subjects underwent baseline spirometry with a dry wedge spirometer (Vitalograph®). The subjects performed at least three technically acceptable trials, and the largest value for the vital capacity (VC), forced vital capacity (FVC), and forced expiratory volume in 1 second (FEV 1.0 ) were registered and compared with predicted values (16). The methacholine challenge test was performed according to published guidelines (17). Briefly, the test commenced with the inhalation of saline diluent, and the first technically acceptable post-diluent FEV 1.0 recorded 2 minutes later was used as the control value. The methacholine was delivered with continuously doubling concentrations, starting with 0.5 mg/ml and stopping at 32 mg/ml, unless the provocative concentration causing a 20% decrease in FEV 1.0 (PC 20 ) had been reached earlier. The subjects known to have asthma started with a methacholine concentration of 0.125 mg/ml. Three different cut-off points were used, PC 20 ≤4 mg/ml, PC 20 ≤8 mg/ml and PC 20 ≤32 mg/ml. Blood samples were obtained from all the subjects, and the sera were frozen immediately at -20°C. The sera were analyzed with respect to Phadiatop® using commercial kits in accordance to the manufacturer's instructions (Pharmacia Upjohn Diagnostics, Uppsala, Sweden).
Atopy was defined as a positive Phadiatop® test (18), whereby class 0 was regarded as nonatopic and class 1 as atopic.
The items used from the questionnaire are shown in the appendix. Current asthma symptoms were defined as an affirmative answer to item 11. Physician-diagnosed asthma was defined as an affirmative answer to item 7.
Asthma was defined as PC 20 ≤ 8 mg/ml and a positive response to the questions about wheezing (item 9) or dyspnea (item 8) with symptom-free intervals in between (item 10), current asthma symptoms (item 11) or physician-diagnosed asthma (item 7).
The local ethics committee approved the study, and an informed written consent was obtained from all the subjects prior to their inclusion in the study

Exposure
All the mills had been using chlorine dioxide as a bleaching agent since the 1950s, and chlorine dioxide was replaced by ozone as the bleaching agent between Table 1. Basic data about the ozone-exposed workers and referents. a Atopy was defined as positive Phadiatop ® test. tration of ozone in a pulp mill is low (< 40 ppb). In the process of bleaching, ozone gas is used at a concentration of 9%. During accidents, for example, leakages or breaks of pipes, it is our estimation that the ozone concentration can briefly be very high (>10 000 ppb) in positions where ozone is still concentrated. Stationary measurements were performed in the mixing and ozonator rooms (the critical areas for high ozone exposure) of the mills. We used a continuous analyzing ultraviolet-photometric instrument (Environics Series 300 Ozone Analyser Computerized, Environics Inc, United States). These measurements register the variation in ozone concentrations by continuous analyses. The response time for the analyzer is 10 seconds. The measurement equipment was positioned in the center of the rooms, which were all >100 m 2 , and the distance from the actual leak influenced the levels registered, at least in cases of smaller leaks. Table 2 presents the results of measurements on 366 days. The table indicates the percentage of days (24 hours) when the instrument indicated ozone levels exceeding 50 ppb and 300 ppb. The ozone levels exceeded 900 ppb one or more times on six of these days. In 1994 a surveillance system initiating an alarm and reinforced ventilation when the ozone concentration reached 300 ppb were introduced.
The estimation of the personal exposure, however, was based on the subjects' self-reports. In the respiratory questionnaire the following exposure-related questions were put to the subjects: (i) "Have you been exposed to ozone resulting in coughing, wheezing, breathlessness or pain in the thorax? If 'yes', when did this occur?"; (ii) "Have you been exposed to chlorine dioxide resulting in coughing, breathlessness or wheezing? If 'yes', when did this occur?"; (iii) "Have you been exposed to sulfur dioxide resulting in coughing, breathlessness or wheezing? If 'yes', when did this occur?" "Ozone gassing" was defined as a positive response to question 1, "chlorine gassing" as a positive response to question 2, and "sulfur dioxide gassing" as a positive response to question 3. In the main analyses the workers were placed into the following three groups: (i) those who reported "gassings" either from ozone, chlorine or sulfur dioxide, (ii) bleachery workers not reporting of any gassings, (iii) unexposed referents. The subgroups of workers reporting ozone gassings were treated separately in additional analyses.

Statistical methods
The statistical analyses were performed with the statistical software package SAS version 8 (SAS Institute, Cary, NC, USA). The univariate statistical analyses of the basic data for the subjects were based on the Student's test and the chi-square test. In general the P-values have been reported. Prevalences have either been expressed as prevalence ratios or as fractions. The significance testing was based on the Mantel-Haenszel test. The incidence rate of new onset of wheezing after the introduction of ozone bleaching (1992-1995, from 1993 for mill B) was calculated as new-onset wheeze/100 person-years among bleachery workers and referents. The risk for new-onset wheezing was calculated as the incidence rate ratio.
In the analyses of categorical variables, univariate analyses were performed, mostly followed by stratification for smoking, atopy, or gender. In the final analyses of the prevalent respiratory symptoms, stratification for smoking and atopy was used. The 95% confidence intervals (95% CI) relied on the test-based method (19). Trends were analyzed with the Kruskall-Wallis test.
A proportional hazards model (Cox's) assessed the association between the exposure estimates and the outcome variables, after adjustment for smoking habits, age, and atopy. The referents were used as reference, and the significance of the slope in the multivariate regression model was based on Wald statistics.

Exposure
Ozone gassings were reported by 55 (43%) of the bleachery workers during the period from 1992 to 1996. During the same period chlorine gassings were reported by 20 (16%) workers and sulfur dioxide gassings by 18 (14%) workers. Ten bleachery workers reported gassings from ozone, chlorine, and sulfur dioxide. In a logistic regression model controlling for atopy, gender, and smoking among the exposed workers, the strongest predictor for reporting ozone gassings was previous chlorine gassing [odds ratio (OR) 2.4, 95% CI 0.96-5.9). No relation was found between smoking habits, employment time or atopy and the reporting of gassing events.

Respiratory health effects
There was an increased prevalence of wheezing and exertional dyspnea among the bleachery workers, and the increase still remained after stratification for smoking and atopy (table 3). The findings were the same with stratification for atopy, gender, or smoking. The prevalence ratios obtained from the Cox regression modeling are shown in table 4. When the bleachery workers were divided into those reporting or not reporting gassings, the prevalences of respiratory symptoms were highest among the gassed workers (table 5). A similar pattern was seen after stratification for smoking habits (not shown in the table), the most exceptional trend being found for ever smokers (referents 18%, bleachery workers without gassings 19%, and "gassed" bleachery workers 37%; test for trend P=0.05). Fifteen subjects were classified as having asthma [11 among the bleachery workers and 4 among the referents (prevalence ratio 1.5, 95% CI 0.7-3.2)]. The highest prevalence of asthma occurred for the gassed workers (10%, N=9), followed by 6% (N=2) among the ungassed bleachery workers and 5% (N=4) among the referents (P-value for trend 0.25).
The prevalences of respiratory symptoms, with the workers divided into those with and without ozone gassings (other gassings being disregarded) are shown in table 6. The differences diminished between the workers reporting ozone gassings and those that did not, mainly because workers gassed by chlorine dioxide were accumulated among the bleachery workers not reporting ozone gassings.
Eleven subjects reported an onset of wheeze during the period from 1992 to 1995, when the mills were   There were only marginal differences in regard to lung function; among the never smokers, the FEV 1.0 was 101% for the exposed workers as compared with 104% for the referents (P=0.19). The methacholine challenge test was performed by 203 subjects (97%); 36 reached PC 20 ≤4 mg/ml, 46 reached PC 20 ≤8 mg/ml, and 94 reached PC 20 ≤32 mg/ml. The fraction reaching PC 20 at different concentrations was not connected with exposure status, although there was a tendency towards lower PC 20 among the gassed workers. For the currently smoking bleachery workers, the prevalence of PC 20 ≤32 mg/ ml was more frequent (P=0.05, Fisher's exact test) among those reporting gassings (7 of 10, 70%) than among those not reporting them (1 of 7, 13%). This phenomenon was not seen among the never smokers (52% versus 42%) or among the ex-smokers (46% versus 50%).

Discussion
Several important observations can be made from this study. The ozone measurements indicated that ozone levels around at least 1 ppm can occur; during our sampling period they occurred during 2% of the days. The bleachery workers reported more asthma symptoms than did the referents, and for workers that had been subjected to gassings the positive response rate was still higher. The incidence of new-onset wheezing was significantly increased among the bleachery workers when ozone was being used as a bleaching agent.
There were several potential biases in this study. As in most populations of industrial workers, selection bias is of importance. Bleachery workers with respiratory problems may leave the pulp mill or obtain a new job in the adjacent paper mill. A few papermill workers reported chlorine gassings, but excluding them did not change the results. The prevalence of atopy was 25% for the referents, compared with 20% for the bleachery workers, the result indicating a slight selection bias.
We estimated that the workers with the highest exposure were those who had been present when an acute leak had occurred, and next came the workers who searched for leaks. However, it was not feasible to measure the exposure for these workers, and hence the exposure classification was based on self-reports of gassings. The use of this classification may have introduced a potential source of information bias. This method has however been used by others (3,4). Kennedy et al (3) used the question "Have you ever been gassed?", and Henneberger et al (4) used "Have you ever been gassed by chlorine or chlorine dioxide?" Since ozone does not have such an obvious smell as does, for instance, chlorine dioxide, we had to alter the wording: "Have you been exposed to ozone with consequent coughing, wheezing, breathlessness or pain in the thorax?" Such wording may be more sensitive to recall bias, since workers with preexisting respiratory symptoms may be more liable to recall gassings. If so, one would expect the prevalence of atopy to be increased among those reporting gassings. The prevalence of atopy among the gassed bleachery workers was 20% versus 21% among those not reporting gassings, and in the logistic regression modeling atopy was not a predictor for reporting gassings.
Bleachery workers may also be more eager to report symptoms such as wheezing and exertional dyspnea than are referents. Such bias was impossible to estimate in this study, and the lack of such an estimation may, to Table 6. Prevalence of current asthma symptoms, physician-diagnosed asthma, ever wheezing, nocturnal dyspnea, and cough with phlegm among the bleachery workers with and without reported gassings to ozone and among the referents. trend some extent, explain the increased prevalence of reported respiratory symptoms among bleachery workers. Questions about respiratory symptoms are probably more influenced by recall bias than are questions about specific conditions, such as physician-diagnosed asthma. In this study we observed the same exposure-dependent pattern for physician-diagnosed asthma as for reported symptoms, although the low prevalence of physiciandiagnosed asthma meant it had no formal statistical significance. This is to some extent an argument against recall bias in respect to the reporting of symptoms. Chlorine dioxide and sulfur dioxide are confounders, since they may also cause asthma-like symptoms and are strongly associated with ozone gassings. Previous chlorine gassings may also act as effect modifiers. The risk of developing respiratory symptoms as a result of ozone gassings may increase among previously gassed workers, since the inflammatory response due to repeated chlorine peaks may not have completely subsided. There are also data indicating that accidental occupational exposure to chlorine can lead to long-lasting impairment of airway caliber (7).
We chose to use the old Mantel-Haenszel method for analyzing prevalence ratios, instead of calculating prevalence odds ratios. Calculating prevalence ratios is an adequate way of analyzing cross-sectional data (20). In these simple models we were able to control for smoking and atopy. In addition, we also analyzed the prevalence ratios with Cox regression models, which contained more covariates. These models gave somewhat higher risk estimates, but without altering the main findings.
It was consistently observed that bleachery workers, especially if gassed, showed an increased prevalence of wheezing and that smokers seemed to be at still greater risk. We also found an association between reported gassings and an increased prevalence of methacholine responsiveness (PC 20 <32 mg/ml) for the current smokers, but not for the nonsmokers (never + former smokers). That smokers emerge as a susceptible subgroup agrees with the results of other studies (4,7).
After 1991, 55 workers reported ozone gassings, and during this period there was an increased risk for newonset wheezing among the exposed workers. Nine bleachery workers reported new-onset wheezing, and one of them reported chlorine gassings after 1992. This reconstruction of incidence data has some limitations. Workers who left the mills in between the start of the bleaching process and the present investigation would not be included, the result being selection bias. There was probably an overrepresentation of diseased workers among the not-included workers with a resulting underestimation of the observed assessed risk estimate.
In a previous study we also found that gassed bleachery workers had an increased output of nitric oxide in their exhaled air and that ozone gassings were an independent predictor of an increased output of nitric oxide (21). Hence occupational exposure to ozone gassings may initiate an airway inflammation that can cause respiratory symptoms and asthma.
Respiratory health effects due to occupational exposure to ozone have been reported for many other occupational groups, such as welders (9), those engaged in the production of plastic bags (22), different kinds of outdoor workers (23)(24), aircraft personnel (25), cement-kiln workers (26), and persons working with equipment inducing high-voltage discharge (27)(28). The acute symptoms that were reported have mainly been cough, headache, and dyspnea. These studies are mainly case reports, and controlled studies of workers with long-term occupational exposure to ozone are lacking.
Our study confirms what has been described by other groups: that repeated exposure to irritants increases the risk for asthma symptoms. Even if our results indicate adverse respiratory health effects as a result of occupational exposure to ozone, the cross-sectional design of the study makes it impossible to determine whether the increased risk of asthma symptoms is due to gassings from ozone, chlorine dioxide, or sulfur dioxide. Because of the limited power of the study, we were unable to find a statistically significant increased risk of asthma, even though an increased prevalence was observed among the exposed workers. The study population will be followed by a prospective longitudinal study with improved ability to distinguish the effect of ozone from that of other irritants.
The results underscore that repeated peak exposures to irritants must be controlled in pulp mills and also better prevented. Most important would be to control accidental exposures.