METHODS

We undertook a cross sectional questionnaire mail survey to ascertain the prevalence of asthma and respiratory symptoms, among MRTs compared with a control group of physiotherapists. The proposal received ethics approval from the Office of Research Services, University of Toronto.

Subjects

Subjects were recruited from members of the two professional associations, the OAMRT and the Ontario Physiotherapy Association (OPA). Our prestudy power calculations were based on an estimated target population for each profession of approximately 3000 members, 90% currently working, an anticipated response rate to the questionnaire of at least 50%, and estimated prevalence of outcomes of 5–7% in the comparison group. This would provide groups greater than that required (about 400–550) to detect a doubling of respiratory symptom outcomes, at an alpha level of 5% and 80% power. Of the four subgroups within the OAMRT, we included in our sampling frame only the 2700 radiological technologists (excluding radiation therapists, nuclear medicine technologists, and magnetic resonance imaging technologists). For the OPA, we excluded from the sampling frame a minority (about 10%) who had elected not to be approached for unsolicited mailings. This left more than 3000 members, so a random sample of 3000 was selected for mailing.

Questionnaire

The questionnaire was adapted from those used in the previous surveys by Smedley and colleagues⁶ and Wymer and colleagues,⁸ who kindly provided their instruments, plus input from the OAMRT. We included questions common for both professions regarding demographics, dates of training and starting work, shifts, hours, respiratory and other symptoms, questions regarding chemical sensitivity, psychosocial factors, and smoking history. The questionnaire for MRTs included questions (mostly for the past 12 months) regarding ventilation conditions, x ray processing tasks, leaking processors, clogged drains, cleanup activities, and personal protective equipment (PPE). The questionnaire also included items dealing with workplace psychosocial factors and musculoskeletal symptoms, which are currently being analysed. The questionnaire was pilot tested with 15 MRTs, following which several awkward questions were revised. The covering letter indicated that this was a study of health and safety of various health professions and did not indicate a focus on respiratory problems or on MRTs. We conducted an initial mailing in late October 2000 with a reminder card one month later, followed by a second mailing in January 2001 to non-respondents with another reminder card one month after that. Questionnaire printing, distribution, and data entry were conducted by the Institute for Social Research, York University, Toronto.

Respiratory system outcomes

Physician diagnosed asthma

This was defined as positive responses to both “Have you ever had asthma” and “Was it confirmed by a doctor?”.

New onset asthma

This was defined as meeting the conditions for physician diagnosed asthma and reported onset after starting practice as an MRT or physiotherapist.

Respiratory symptoms in past 12 months

We incorporated questions for the prevalence during the past 12 months based on the nine respiratory symptoms for which two or three positive responses were previously validated by Venables and colleagues¹⁸ to be sensitive and specific indices for the presence of self reported asthma and bronchial hyperresponsiveness (see Appendix). For those answering affirmatively to a given symptom, we assessed whether the symptom was work related, based on the response to the additional question “while away from work how often does [the symptom] occur? ...”. We considered a symptom to be work related (WR) if the respondent indicated that this occurred less frequently than on workdays or not at all, as adapted from questionnaires used in previous hazard evaluations in Ontario and at NIOSH.^19–21 We determined the prevalence of three respiratory symptom outcome measures: three or more respiratory symptoms; two or more WR symptoms; and three or more WR symptoms. Current smokers answered affirmatively to “Have you ever smoked cigarettes (at least one cigarette a day for one year)?” and “Do you currently smoke cigarettes?”; ex-smokers answered affirmatively to the first question and negatively to the second.

Statistical analysis

We compared proportions of categorical variables between the two exposure groups with χ² tests, and continuous variables with t tests. We initially conducted contingency table analyses and considered interactions by assessing the Breslow-Day test for the homogeneity of the odds ratios. Where interactions occurred, stratified results are presented. To compare the prevalence of physician diagnosed asthma between MRTs and physiotherapists, we stratified by age (in three groups) and gender, while analyses of respiratory symptoms were also adjusted for childhood asthma (defined as physical diagnosed asthma with onset less than 16 years of age). The relative risk was estimated by the Mantel-Haenszel odds ratio (OR). We subsequently used the OR from unconditional multiple logistic regression analyses to estimate the risk for respiratory system outcomes between the professions, adjusting our models simultaneously for potential confounders such as age, gender, and childhood asthma. We also used logistic regression analysis to identify, among MRTs, possible workplace/exposure factors associated with respiratory outcomes, while adjusting for the same confounders. Including the age term in either three groups or as a continuous variable yielded almost identical results; the latter are presented. Given that multiple comparisons with exposures and tasks were made, we arbitrarily considered associations with exposures to be “of note” or clinically important if the ORs were 1.8 or greater and the 95% confidence interval (CI) excluded unity. As the workplace and task exposure variables are all dichotomous, the regression coefficients in tables 4 and 5 represent the comparison between “yes” and “no” in all cases. In additional analyses, we further adjusted the above comparisons between the professions for latex glove use, and compared latex associated symptoms among MRTs versus physiotherapists. Most analyses were confined to never smokers who comprised the majority of both professions. For comparisons of asthma and respiratory outcomes between the professions, we also examined the former and current smoking groups separately. All statistical analyses were conducted using SAS version 8 (SAS, 2000).

RESULTS

Questionnaire responses

Of the 2761 and 3000 questionnaires distributed to the MRTs and physiotherapists, respectively, there were “bad addresses/returned to sender” responses for 15 and 24; and “bad sample units” (deceased, not working, out of country) for 0 and 6, respectively. Of the remaining 2746 and 2970 members, completed questionnaires were received from 1754 and 1875, yielding response rates of 63.9% among MRTs and 63.1% among physiotherapists, respectively. We then excluded respondents who indicated that their work was totally administrative (34 MRTs and 27 physiotherapists) or that they were currently on sick leave (one MRT), leaving 1719 and 1848, respectively. We confined most analyses to the 1110 (65.1%) MRTs and 1523 (82.9%) physiotherapists who never smoked and did not spend all their time in administrative positions. Among the never smokers, the MRTs were older and had been in practice longer, were slightly more likely to be female, and were less likely to work on day shift (table 1). The MRTs were somewhat less likely to have had childhood asthma. Age and years in practice were highly correlated among both MRTs (r = 0.90) and physiotherapists (r = 0.94).

Comparison of main outcomes between MRTs and physiotherapists

Physician diagnosed asthma

Never smokers

Although the prevalence of physician diagnosed asthma was similar in both groups (table 2), the prevalence of physician diagnosed asthma with new onset since starting work in the profession was greater among MRTs than physiotherapists (6.4% v 3.95%). However, there was an interaction across gender (table 2). The age adjusted prevalence of new onset asthma was about 30% greater among females, but was fivefold greater among males (age adjusted OR 5.3; 95% CI 1.4 to 20.2; p < 0.05).

Ex- and current smokers

There were few respondents for analysis (table 3). Physician diagnosed asthma was less frequent among MRTs than physiotherapists among both ex-smokers (10% v 14%; OR 0.6) and current smokers (9.0% v 16.3%; OR 0.5). New onset asthma could be analysed only among females, since there were no cases among the 47 male MRTs and 28 physiotherapists who were ex-smokers, and among the eight male physiotherapists currently smoking; there was one (4.6%) case among the 22 MRTs currently smoking. After adjusting for age, there were similar proportions of females in the two professions with new onset asthma, both in former (OR 1.1) and current (OR 1.0) smokers (table 3), indicating lower proportions (by subtraction) of MRTs with asthma of onset prior to starting work.

Prevalence of respiratory symptoms in past 12 months

Among never smokers, after adjusting for age, gender, and childhood asthma, MRTs were twice as likely to report three or more of the nine marker respiratory symptoms, and were over three times more likely to report two or more, or three or more WR respiratory symptoms (table 2). These associations were observed among both those with and those without childhood asthma (data not shown). Similar patterns of increased prevalence of respiratory symptoms among MRTs were also shown among ex-smokers and current smokers (table 3).

Latex glove use and symptoms

MRTs were slightly more likely than physiotherapists to report ever using latex gloves in the past 12 months (75.2% v 71.3%; p = 0.03) and much more likely to use them always or usually than rarely or never (49.2% v 16.5%; p < 0.0001). However, the increased risks for reporting respiratory symptom outcomes were little changed when further adjusted for latex glove use (table 2); the relative risks were somewhat greater among those not using than among those using these gloves (data not shown). MRTs were also more than twice as likely to report latex associated symptoms than physiotherapists; however, the increase in all three respiratory symptom outcomes persisted among MRTs when the analysis was limited to those not reporting latex associated wheeze, shortness of breath, or chest tightness (table 2).

Work tasks/exposures and respiratory symptoms among MRTs

As the prevalence of asthma and respiratory symptoms was shown to be increased among the MRTs, we then explored among the never smokers, the associations of workplace factors, tasks, and exposures of the MRTs during the past 12 months with two of these outcomes: three or more respiratory (Respsx3m), and two or more WR respiratory symptoms (WRresp2m).

Workplace factors and exposures not associated or only weakly associated with respiratory outcomes

Workplace factors

There was no association with working as an MRT outside the primary workplace, or reporting the presence of air supply inlets or air exhaust outlets, or a silver recovery unit (table 4).

Tasks and exposures (table 4)

There were only weak or no associations (OR <1.8) with the frequency of loading film into a daylight processor or in a darkroom, removing processed film from an automatic processing machine, pouring processing chemicals, processing film by hand (for Respsx3m), removing used chemicals, mixing fixer/developer chemicals, routine processor cleaning (racks, tanks—periodic servicing), and cleaning the crossover rollers of a processor (usual quality control). There was only a weak association with the number of films processed per day, dichotomised at the median. Of note, many of these are routine tasks or activities that are unlikely to generate aerosol or chemical exposures, although pouring and mixing chemicals might be considered to be exceptions.

PPE

The respiratory outcomes were not associated with using various types of PPE (table 4).

Workplace factors and exposures associated with respiratory outcomes

Workplace factors

It was reported that all processing machines (compared with some or none) having local exhaust ventilation (LEV), and the perception that the ventilation in the area where films were processed was adequate (versus inadequate) were associated with a significantly reduced risk of respiratory symptoms (table 5).

Tasks and exposures (table 5)

Although there were only weak associations with reporting that a processor leaked in the past 12 months, a delay in the clean up for greater than one day was associated with more than a doubling of these outcomes. The outcomes were also significantly associated with the floor drain being clogged, and the frequency of freeing a film jam, unblocking a blocked processor drain, processing film by hand or wet developing (for WRresp2m), and detecting the odour of processing chemicals in the work area. While the frequency of routine cleaning of processors and cleaning crossover rollers was not associated with these outcomes (table 4), cleaning up a processor chemical spill was strongly associated (table 5). For a number of these outcomes, the associations were stronger for the work related outcome than for simple respiratory symptoms, and many of these activities may generate chemical exposures in vapour or aerosol form. A number of gender specific ORs were greater among males than females, but the interaction terms were not statistically significant (data not shown).

Frequency of exposure by gender

Because gender modified the prevalence of new onset asthma, we also examined how the reported frequency of exposures or tasks varied by gender, during two time windows: in the past 12 months and “in 1995 or earlier” as a snapshot of conditions “five years ago or more”. Males and females reported similar frequencies for mixing chemicals, removing processed film, loading film in a daylight processor or a darkroom, and cleaning crossover rollers (usual quality control of processor) during 1995 or earlier. However, males were significantly more likely to report freeing a film jam, periodic servicing of processor involving racks and tanks, cleaning processor chemical spills, unblocking processor drain, removing used chemicals, processing film by hand, and pouring chemicals (fig 1). Similar reporting patterns were observed for responses related to the past 12 months (data not shown). Moreover, among males, those with new onset asthma were more likely than those without respiratory symptoms to report freeing a film jam (50% v 34.5%), removing used chemicals (40% v 24.1%), wet film developing or processing film by hand (62.5% v 31.5%), and cleaning processing chemical spill (70% v 52.6%), referring to 1995 or earlier. Those with new onset asthma were also more likely to report periodic servicing of processor in the past 12 months (50% v 34%) and were less likely to perceive that the ventilation in the area where films were processed was adequate (42.9% v 70.9%).

Occurrence of new onset asthma over time

Using five year groupings, subjects in each profession were classified according to the time period since starting work. We then determined the proportion of respondents with new onset asthma in relation to all respondents starting work in a given time period (fig 2). Among those starting prior to 1985, there was a tendency for the proportion of new onset asthma to be greater among the MRTs (exceeding 7 per 100 in all four time periods and 8 per 100 in two periods) than among physiotherapists (exceeding 7 per 100 in only one period and 8 per 100 in none). On the other hand, the proportion was approximately 4 per 100 or lower in both professions among those starting in recent time periods. Similarly, the ratio of the proportions of new onset asthma across professional groups (MRTs/physiotherapists) was greater (⩾1.5) in three of the four periods prior to 1985, but was close to unity in both of the recent time periods in 1985 or after (fig 2).

DISCUSSION

The present study, which is comparable in size to that of Smedley and colleagues,⁶ represents one of the largest population based surveys among radiographers to date, and is the first to incorporate immunological analyses, which are in progress. In this report, we have highlighted self reported asthma and respiratory symptom outcomes among MRTs.

Asthma

Among never smokers, who comprised the majority of respondents, the two professions did not differ in the overall prevalence of physician diagnosed asthma; however, the prevalence of new onset asthma since starting work was greater among the MRTs, and gender modified this association. The association among females was weak and not statistically significant, although if causal, this may be of public health significance given the very large numbers in the profession. On the other hand, among males, after adjustment for age, the prevalence of new onset asthma was five times greater among MRTs than among physiotherapists. The reasons for this are not known; however, in the past, male MRTs were more frequently assigned to work with processing chemicals and carrying the chemical tubs, which may have played a role in the increased risk observed (R Hesler, Executive Director, OAMRT; personal communication, 2001). Moreover, in our survey, a number of exposures and tasks were reported significantly more frequently by males than females (fig 1), and also among males, more frequently by those with new onset asthma than those without respiratory symptoms.

Among the smaller numbers of current and ex-smokers, the overall prevalence of physician diagnosed asthma among MRTs was only about half of that among physiotherapists, and the prevalence of new onset asthma was similar in the two professions. Thus, by subtraction, there were far fewer MRTs who had asthma with onset in childhood or prior to work. This suggests that MRTs with asthma may have left the profession prior to this survey as a result of workplace conditions such as respiratory irritants which are known to be present in their work setting but not in that of physiotherapists. Monso and colleagues²² recently reported health related self selection in a prospective study of apprentices exposed to high molecular weight agents, supporting the hypothesis of attrition bias in the natural history of OA.²³ In addition, our observation that there was a higher proportion with new onset asthma among MRTs than among physiotherapists prior to 1985 (fig 2), suggests that working conditions were worse in the earlier years, with some improvement in the past 15 years.

Respiratory symptoms in past 12 months

We observed a moderately strong association among MRTs compared with physiotherapists for reporting respiratory symptoms previously shown to denote asthma, and these increased in strength to more than threefold greater risk for work related symptoms. Latex associated symptoms were reported more frequently among MRTs (table 2). In fact, Wymer and colleagues⁸ observed an identical OR of 2.7 for “latex allergies”. However, these latex related symptoms were reported by less than 10% of the entire group (⩽2.5% for lower respiratory symptoms; table 2), and the increased risks among MRTs persisted after further adjustment of the regression models for reported latex glove use, or when examined among those not reporting latex associated lower respiratory symptoms, suggesting that latex allergy cannot explain these findings.

Associations of symptoms with tasks during past 12 months among MRTs

Among the MRTs, respiratory outcomes were associated with a number of workplace factors and tasks that may generate or exacerbate aerosol or chemical exposures (table 5). They were also associated with characteristics related to the adequacy of darkroom or processor ventilation, such as whether all processors had LEV, the perception of adequacy of ventilation, and the frequency of detecting odour of processing chemicals, factors that may play a role in their origin. This also suggests the possibility of overlap with sick building syndrome, which we have observed in separate analyses of non-respiratory symptoms (Tarlo et al, in preparation). In their examination of exposure determinants in radiographic film processing, Teschke and colleagues¹⁷ found that the presence and flow rate of LEV were significant predictors of GA and SO₂ concentrations, but general room ventilation was not associated with contaminant reductions. Our observation that latex glove use in the past 12 months was not associated with respiratory symptoms (table 4) may reflect the fact that respondents who were sensitised may have already changed to non-latex gloves.

The strongest associations (ORs exceeding 2.3) that we observed were for delayed clean up of leaks, freeing a film jam, unblocking blocked processor, clean up of process spill, and detecting odour. On the other hand, there were weak or no associations with workplace factors or activities that were unlikely to generate (or be markers for) aerosol or chemical exposures, such as merely working in a second location outside of primary workplace, the number of films processed per day, or the frequency of loading film into a daylight loader, removing processed film, or the routine cleaning of processors. Moreover, the fact that MRTs did not link their respiratory symptoms with all workplace factors provides some specificity and reassurance against recall bias and a tendency to “over report” symptoms. It should be emphasised that the increased symptoms (table 2) and associations with workplace and exposure factors (table 5) refer to the past 12 months. Thus, despite the indication of recent improvement in conditions (fig 2), there is still substantial evidence for current problems.

Comparison to previous surveys

Smedley and colleagues⁶ observed that the one year period prevalence for two or more asthma marker symptoms, based on the same questionnaire used here,¹⁸ was 22.9% among radiographers versus 18.4% among physiotherapists (period prevalence ratio of 1.23 (1.09–1.40)), and that the lifetime prevalence of one or more of respiratory symptoms worse at work was 2.1 (1.6–2.4) times more frequent among radiographers. Wymer and colleagues⁸ reported that among never smokers, several respiratory symptoms (woken by shortness of breath, regular breathing problem, and daytime shortness of breath) were 2.2–2.5 times more frequent among radiographers than physiotherapists, after adjustment for age, gender, and childhood asthma.

With respect to associations with exposures/tasks in previous surveys, Smedley and colleagues⁶ examined associations of activities in the past year among radiographers with work related non-respiratory but not respiratory symptoms. The tasks that most consistently entered models for these symptoms were working for at least 20 hours a week in a room with an automatic processor, and working in a room where a machine was obviously leaking. Wymer and colleagues⁸ found the strongest (two- to threefold) and most consistent associations of respiratory symptoms among never smoking radiographers with detecting odours and cleaning spills, similar to our findings.

Might the associations observed in our study be causal? Considering the Bradford Hill criteria,²⁴ the associations that we observed (both between the professions and among the MRTs) are moderately strong, and consistent with findings in the two previous surveys^6,8 as well as with prior case reports. These associations are also biologically plausible given the presence of irritants and sensitisers in the darkroom environment. There is some indirect evidence for dose-response relations in that (a) the associations were much stronger for work related symptom clusters than merely three or more respiratory symptoms; (b) there were much stronger associations with delayed clean up of leak compared with simply reporting a leak; and (c) there was a pattern of links with tasks or factors likely to generate chemical or aerosol exposures, but an absence of associations with tasks unlikely to generate exposures. Although temporality cannot be established in cross sectional studies, it is suggested by the greater prevalence among MRTs for new onset asthma but not for simple physician diagnosed (or childhood) asthma (this evidence was limited to never smokers).

This investigation is limited by the cross sectional design in which one cannot confirm temporality, and the subjective nature of the questionnaire (asthma and symptoms; exposures or tasks, which have not been confirmed). However, of those reporting a diagnosis of physician diagnosed asthma who underwent confirmation, at least half of both MRTs (six of 12) and physiotherapists (seven of 10) had positive methacholine challenge tests (PC₂₀ ⩽ 8 mg/ml),²⁵ confirming current bronchial hyperresponsiveness. The self reported survey data are also limited by the potential for recall bias; participants with symptoms may overestimate their exposures. On the other hand, as in cross sectional studies in general, if symptomatic MRTs tended to leave work, the magnitude of risks may be underestimated. This is suggested by the marked deficit of asthma with onset prior to starting work among ever smoking MRTs. Confounding due to smoking is possible when examining respiratory symptoms. We addressed this by looking at the asthma and respiratory symptom outcomes separately by smoking status, and by limiting most analyses to the much larger group of never smokers. The strengths of our investigation include the large sample size, the reasonable response rate for a mail survey which was comparable between the groups, the utilisation of a previously validated asthma symptom questionnaire, and the incorporation of a second laboratory and field component with objective confirmation of asthma and its work relatedness in a sample of respondents, and the characterisation of a sample of darkroom environments.

In summary, the findings of increased new onset asthma among never smokers suggest that there may be an increase in work related asthma among MRTs. The increase of respiratory symptoms shown to connote asthma among MRTs, plus the associations with tasks and scenarios likely to generate aerosol or chemical exposures but not with a number of routine tasks, suggests a work related aetiology. However, the mechanism remains unknown despite the presence of both irritants and sensitisers. Our preliminary findings of low percentages of eosinophils in induced sputum among MRTs²⁵ suggest that the mechanism may involve, at least in part, a role for respiratory irritants, and/or other non-IgE mediated mechanisms.

APPENDIX 1: RESPIRATORY QUESTIONS FROM VENABLES AND COLLEAGUES¹⁸

1. When you have run or climbed stairs fast, have you:

   • coughed?

   • wheezed?

   • got tightness in the chest?

2. Has your sleep been broken by:

   • wheeze?

   • difficulty in breathing?

3. Have you woken up in the morning with:

   • wheeze?

   • difficulty in breathing?

4. Have you wheezed:

   • when you were in a smoky room?

   • when you were in a very dusty place?