Can serial PEF measurements separate occupational asthma from allergic alveolitis?

Occupational Medicine 2015;65:251 255 doi:10.1093/occmed/kqv013 Can serial PEF measurements separate occupational asthma from allergic alveolitis? P. S. Burge 1, V. C. Moore 1, C. B. S. G. Burge 1, A. D. Vellore 1, A. S. Robertson 1 and W. Robertson 2 1 Occupational Lung Disease Unit, Birmingham Heartlands Hospital, Birmingham B9 5SS, UK, 2 Department of Public Health, Warwick Medical School, University of Warwick, Coventry CV4 7AL, UK. Correspondence to: P. S. Burge, Occupational Lung Disease Unit, Birmingham Heartlands Hospital, Bordesley Green East, Birmingham B9 5SS, UK. Tel: +44 (0)121 424 0734; fax: +44 (0)121 772 0292; e-mail: sherwood.burge@heartofengland.nhs.uk Background Occupational asthma commonly results in work-related changes in serial peak expiratory flow (PEF) measurements. Whether alveolitis can result in similar changes is unknown. Aims Methods Results To identify differences and similarities of serial PEF between workers with occupational alveolitis and asthma seen during an outbreak investigation in a factory with metal-working fluid exposure. Workers with respiratory symptoms and rest-day improvement were identified by questionnaire. Each was asked to measure PEF 8 times daily for 4 weeks at home and work. Alveolitis was subsequently diagnosed from a validated scoring system including radiological changes, carbon monoxide diffusing capacity, bronchoalveolar lavage and biopsy results. Occupational asthma was confirmed with a positive Oasys score >2.5 and a mean rest-work PEF >16 l/min from serial 2-hourly PEF measurements. The Oasys PEF plotter calculated differences between rest and workdays for mean PEF, diurnal variation and the scores were used to confirm an occupational effect (Oasys, area between curve and time point). Records were compared between the alveolitis group and the group with occupational asthma without alveolitis. Forty workers with occupational asthma and 16 with alveolitis had indistinguishable PEF changes on workdays in terms of magnitude (median reduction 18.5 and 16.1 l/min, respectively) and diurnal variation. Immediate reactions were more common with occupational asthma and late reactions more common with alveolitis. Conclusions PEF responses to metal-working fluid aerosols do not distinguish occupational asthma from alveolitis except in timing. They can be used to identify the workplace as the cause of asthma and also alveolitis. Key words Introduction Allergic alveolitis; hypersensitivity pneumonitis; meta-working fluid; Oasys; occupational asthma; peak expiratory flow. Serial measurements of peak expiratory flow (PEF) are an established method for physiological confirmation of occupational asthma and are recommended as the first confirmatory test in guidelines [1 3]. The Oasys analysis program is the best validated method for assessing workrelationship from PEF records [4]. So far, analysis has been confined to workers who might have occupational asthma. Other diseases that may vary over fairly short periods of time might also modify PEF on a daily basis. Although alveolitis is predominantly a restrictive lung disease, PEF would be expected to decline in parallel with forced vital capacity (FVC) and forced expiratory volume in 1 s ( ) making it likely that the temporal changes related to work exposure may be detected by serial PEF measurements when the alveolitis is due to work exposure. We had the opportunity to compare PEF changes in workers with occupational asthma and alveolitis from the same outbreak of respiratory illness due to metalworking fluid aerosols [5]. The purpose of this study was to answer the question: Can serial measurements of PEF differentiate occupational asthma from alveolitis in workers exposed to the same metal-working fluid aerosols? The Author 2015. Published by Oxford University Press on behalf of the Society of Occupational Medicine. All rights reserved. For Permissions, please email: journals.permissions@oup.com

252 OCCUPATIONAL MEDICINE Methods This article comes from a systematic study of workers with work-related respiratory symptoms in a factory manufacturing car engines and transmissions with general exposures to metal-work fluid within the then exposure standards of 1 mg/m 3 [5]. The investigation followed the identification of 12 workers with occupational allergic alveolitis referred to a specialist occupational lung disease clinic. They were asked to complete PEF records before being seen in the clinic as is our usual practice in those in whom occupational asthma seems possible from the referral letter. Following this, the investigation was carried out in an epidemiological way. A screening questionnaire was completed by all exposed workers and those with any respiratory symptom or weight loss were invited to complete a detailed respiratory questionnaire, spirometry and blood tests, together with randomly selected asymptomatic controls. All workers who gave a history of work-related respiratory symptoms were asked to complete peak flow records irrespective of the likelihood of specific diagnosis. They were later invited to a hospital clinic where chest X-rays, full lung function including gas transfer and methacholine reactivity, exhaled breath nitric oxide, allergy skin prick tests and a full occupational history and examination were performed. The factory replaced all the metal-working fluid and steam cleaned all the machines during this investigation. This study of serial peak flow records was confined to the time before cleaning or the introduction of respiratory protective equipment (RPE). Workers were asked to measure PEF eight times daily for 4 weeks at home and at work using linear mini-wright peak flow meters. Alveolitis was diagnosed from a scoring system based on combinations of systemic symptoms with breathlessness, weight loss or fever, audible crackles in the lungs, chest x-ray or computerized tomogram scan showing compatible interstitial changes, reduced carbon monoxide diffusing capacity (DLCO) and biopsy results following hospital referral using the Fox criteria [6]. This was modified in a minor way allowing replacement of open biopsy with a bronchoalveolar lavage lymphocytosis in workers without open biopsy histology. The presence of definite or probable alveolitis excluded the worker from the occupational asthma group. Occupational asthma was diagnosed from work-related wheeze or breathlessness and required confirmation with physiological tests. All had serial PEF records with an Oasys score >2.5 and a mean work-rest PEF >16 l/min. The qualifying PEF record could have been obtained after machine cleaning and replacement of metal-working fluid, which helped but did not eliminate the problem. The Oasys plotter calculates the difference between rest and workdays by comparing mean diurnal variation, the area between the 2-hourly mean PEF measurements on work and rest days and the overall mean PEF from all readings on rest and work days. Oasys has four analyses for detecting work-relationship: the Oasys score uses a discriminant analysis based on pattern recognition of the plot of daily maximum, mean and minimum PEF, a score >2.5 has a specificity of 94% for occupational effect [7]. The area between curves (ABC) score compares the area between the mean 2-hourly PEF on rest days and work days, a score 15 l/min/h has a specificity of 96% for occupational effect [8]. Time point analysis is a statistical method based on the ABC plot, mean workday values outside the 95% confidence interval (CI) for rest days are counted, 1 non-waking time point has a specificity of 93% for occupational effect [9]. Finally, the whole record mean difference between PEF on rest and workdays has an upper 95% CI of 16 l/min in non-occupational asthmatics and asymptomatic workers occupationally exposed to high levels of irritants [10]. The Oasys PEF plotter was used to calculate differences between rest and workdays for mean PEF, diurnal variation and the four scores used to confirm occupational asthma: Oasys [7], ABC [8], time point [9], restworkdays mean PEF 16 l/min [10]. All days were adjusted to start with the first reading at work and to stop at the last reading before work on the following day (day interpretation). In addition, the mean 2-hourly plot was classified visually into immediate reactions (where the fall in PEF started within 2 h of starting work), flat reactions (where the PEF failed to increase during the work day but did so on rest days), late reactions (where the fall started later and continued after leaving work) and unclassifiable records (Figure 1). PEF records were classified into those fulfilling and failing to fulfil the quality standards for assessing work causation using the ABC criteria, which are likely to be the most appropriate for accurate calculation of diurnal variation. The ABC criteria require at least eight readings a day on at least eight workdays and three rest days. The study was approved by a local research ethics committee (IRAS 129653). Statistical analysis was with t-tests, Mann Whitney U and χ 2 using IBM SPSS Statistics version 21. A stepwise logistic regression (forward) was conducted to predict membership of occupational asthma (OA) and alveolitis groups using variables from Table 1 and the PEF variables from Table 2 as predictors. Results The screening questionnaire was completed by 808/836 (97%) exposed workers. Ninety-four percent (458/481) of symptomatic workers and 96% (48/50) of controls completed a detailed respiratory questionnaire, spirometry and blood tests. A total of 164 workers who gave a history of work-related respiratory symptoms were invited for full clinical investigation; 97% (159/164) completed this stage of whom 141 returned peak flow

P. S. BURGE ET AL.: RESPIRATORY ILLNESS IN workers exposed to METAL-WORKING FLUID 253 Figure 1. Plots of mean 2-hourly PEF in workers with an immediate (top), flat (middle) and late (bottom) reactions. The squares show the mean values for all days away from work, the crosses for the mean of all workdays on morning shifts. The intermediate lines are the 95% CI for the rest-day readings from the time point analysis. Working times have a shaded background. Mode times of starting and stopping work are shown by the black vertical lines. Pale grey shading shows the latest starting time (left side) and earliest stopping time (right side). The lower panel shows the time, the number of readings within each mean (days away from exposure on the right, workdays on the left). Below this are the differences between rest and workdays in litres/minute/hour and at the bottom the ABC score which is a mean of the 2-hourly measurements. records. Following the full clinical investigation, 19 fulfilled the case definition for alveolitis and 66 the case definition for occupational asthma without alveolitis. Serial PEF records taken before cleaning and RPE were available for 40/66 of the total cohort of workers with occupational asthma and without allergic alveolitis, and 16/19 with alveolitis. There were no statistical differences between the total cohort and the subgroup studied (Table 1). The, FVC and DLCO were statistically lower in studied patients with occupational alveolitis compared with those with occupational asthma. Workers with alveolitis were less atopic than the occupational asthma group, but had similar PEF measurements (asthma mean 97% predicted, alveolitis mean 101%). Table 2 shows the indices from the PEF records. The magnitude of the difference in PEF between rest and workdays was similar for those with occupational asthma or alveolitis (median reduction 18.5 and 16.1 l/min). The diurnal variation on work and rest days was also similar between groups, 25% and 31% for the occupational asthmatics and alveolitics, respectively, having a mean diurnal variation >20% predicted. Immediate reactions were more common in those with occupational asthma (40% versus 6%), flat reactions were only seen in the occupational asthma group and late reactions more common in the alveolitic group (69% versus 25%; P < 0.01 for a 4 2 χ 2 comparison). The scores from the Oasys plotter identifying differences between work and rest days were similar for those with occupational asthma and alveolitis. Many records did not reach the quality standards optimal for the diagnosis of occupational effect, requiring further PEF records after the start of remedial action. Quality standards for individual diagnoses of occupational asthma were reached for the Oasys score in 66%, the ABC score in 54% and time point in 59%. The similarity in workrelated PEF parameters persisted when records were separated in those of good and less good quality. The mean differences between rest and work were 13 and 14 l/min, respectively, for occupational asthmatics and alveolitis for good-quality records, and 27 and 24 l/min for those with less good quality. Mean workday diurnal variation was 16% and 17% for those with good quality and 13% and 16% for those with less good quality (all non-significant). The stepwise logistic regression model identified percentpredicted DLCO, a late PEF reaction and percent-predicted FVC as significant predictors separating the occupational asthma and occupational alveolitis groups (χ 2 = 29.8, P < 0.001, df = 3). The model excluded all the other PEF variables. Prediction success overall was 80% (88% for OA and 63% for the alveolitics) between prediction and grouping. Discussion Workers with occupational alveolitis due to metal-working fluid aerosols showed a similar magnitude of workrelated change in PEF and similar changes in diurnal variation to those with occupational asthma from the same exposures. Serial PEF records could not differentiate workers with occupational asthma from occupational alveolitis, unless the timing of the reaction was immediate or flat when occupational asthma was the likely diagnosis. Age, sex and smoking were similar between

254 OCCUPATIONAL MEDICINE Table 1. Demographics of workers studied compared with the total cohort from the original survey Occupational asthma Occupational alveolitis Included OA versus alveolitis, P value Total cohort (n = 66) Included (n = 40) Total cohort (n = 19) Included (n = 16) Age, years (mean, SD) 43.4 (7.6) 45 (7.8) 47 (7.9) 44 (7.8) NS Male/female 59/7 37/3 17/2 14/2 NS Current smokers, n (%) 15 (23) 4 (10) 3 (16) 2 (12) NS Atopy, n/n (%) 33/61 (54) 21/40 (52) 3/13 (23) 4/14 (29) NS % predicted (mean, SD) 95 (14.9) 96 (12.2) 84 (15.9) 84 (11.5) <0.001 FVC % predicted (mean, SD) 103 (13.2) 103 (13.2) 85 (15.9) 88 (13.1) <0.001 /FVC % (mean, SD) 79 (7.1) 79 (7.0) 81 (7.9) 79 (8.4) NS DLCO % predicted (mean, SD) 99 (16.6) 80 (19.7) <0.01 The P values refer to the difference between the workers studied with occupational asthma or alveolitis. There were no statistical differences between the workers studied and the total cohort (NS, non-significant; P > 0.05). Table 2. Differences between PEF records in occupational alveolitis and occupational asthma the groups making these unlikely confounders. In practice, the cause of allergic alveolitis is often unclear; Oasys analysis of serial PEF records is a suitable method to identify those due to exposures at work. As work-related change in PEF was an inclusion criteria for the occupational asthma group, it could be argued that some of this group had diseases other than asthma. All had symptoms of asthma with rest day improvement. Features of alveolitis unrelated to PEF changes excluded a worker from the occupational asthma group, excluding alveolitis as a plausible cause for the changes in the asthma group. Work-related changes in PEF are not known to occur in any other condition, and the mean magnitude of workday changes seen were outside the 95% CIs for individuals with non-occupational asthma or asymptomatic workers exposed to high levels of respiratory irritants [8,10]. Clearly, PEF is not an adequate method for the diagnosis of allergic alveolitis. We defined allergic alveolitis using validated criteria independent of any changes in PEF [6]. The logistic regression identified lower values of FVC and DLCO as predictors of allergic alveolitis as expected, all measures of PEF variability were excluded from the final model. It could be argued that Occupational asthma (n = 40) Alveolitis (n = 16) P Median rest-workday mean PEF, l/min (interquartile range) 18.5 (30.6) 16.1 (31.1) NS Median diurnal variation (% mean) on workdays (interquartile range) 15.4 (11.0) 16.3 (16.9) NS Median diurnal variation (% mean) on rest days (interquartile range) 9.7 (9.4) 8.3 (10.2) NS Mean workday diurnal variation (% mean) 20% predicted 25.0% 31.3% NS Immediate PEF reaction, n (%) 16 (40) 1 (6) <0.01 Flat PEF reaction, n (%) 4 (10) 0 Late PEF reaction, n (%) 10 (25) 11 (69) Unclassifiable PEF reaction, n (%) 4 (10) 0 Unclassifiable, n (%) 6 (15) 4 (25) Oasys score >2.5, n (%) 29 (72) 10 (62) NS ABC dayshift score 15 l/min/h, n (%) 18 (45) 8 (50) NS Positive time point analysis, n (%) 27 (67) 12 (75) NS Rest-workdays mean PEF 16 l/min, n (%) 23 (57) 8 (50) NS NS, non-significant, P > 0.05. the PEF changes were due to co-existing asthma, which would need to have been present in nearly all those with alveolitis to account for the changes seen. The finding of a predominance of late reactions in the alveolitis group and immediate and flat reactions in the asthma group makes asthma an unlikely cause for the PEF changes in the alveolitis group (P < 0.01). Specific challenges in allergic alveolitis have a similar delayed onset to the late reactions seen in PEF in this study [11]. The PEF would be expected to fall in parallel with FVC and in any restrictive lung diseases such as alveolitis. The quality requirements in terms of numbers of readings, duration and days away from exposure required for diagnostic analysis were not an inclusion criteria for this study, explaining the relatively low diagnostic sensitivity of the individual parameters seen. They are, however, likely to be sufficient for calculating the differences between work and rest days shown in Table 2 and not biased by improvements following steam cleaning and the introduction of RPE. Separation of records into those with at least eight readings a day on at least eight workdays and three rest days, and those with fewer readings, showed no differences between the occupational asthmatics and alveolitics,

P. S. BURGE ET AL.: RESPIRATORY ILLNESS IN workers exposed to METAL-WORKING FLUID 255 although those with less good quality had larger differences between rest and work days than those with high-quality records for both the asthmatic and alveolitic groups. Including records made after cleaning and the availability of RPE, all workers in the occupational asthma group had a PEF record fulfilling the criteria for occupational asthma. Oasys has only been validated against workers with independent diagnoses of occupational asthma, this is the first report in workers with occupational alveolitis. The inclusion of all available records increases the likelihood that those studied were representative of the whole population, originally including 97% of the total exposed workforce [5]. The workers studied in this report are a subset of those included in the original outbreak, being all those keeping records before the confounding factors of RPE or cleaning occurred. There were no statistical differences between the workers studied and the original cohort. Many of the diagnostic PEF records included in the original report were obtained after the institution of RPE and the replacement of the metal-working fluids, which reduced the magnitude of PEF changes in workers keeping repeat records. The original report included all workers with alveolitis and a positive PEF record as having both alveolitis and asthma; this in retrospect was incorrect. Previous case reports have shown changes in PEF in a worker with alveolitis from moulds in a sawmill [12] and a worker with alveolitis from anhydrides in a polyester paint [13]. There are others with good evidence of asthma (from changes in non-specific reactivity) and alveolitis [14]. To our knowledge, this is the first report comparing PEF changes in asthma and allergic alveolitis in a comparable group showing a similar magnitude of change in both groups. We suggest that serial measures of PEF can be used to identify the workplace as a cause for allergic alveolitis as well as occupational asthma. Key points Work-related changes in peak expiratory flow are seen in occupational alveolitis as well as occupational asthma. Occupational asthma and alveolitis cannot be differentiated by the magnitude of peak expiratory flow decline at work nor changes in diurnal variation. Immediate reactions suggest asthma rather than alveolitis. Serial peak expiratory flow records are a suitable method for identifying the workplace as a cause of allergic alveolitis. Funding W.R. is part funded by the National Institute for Health Research Collaborations for Leadership in Applied Health Research and Care West Midlands initiative. Conflicts of interest None declared. References 1. Nicholson PJ, Cullinan P, Taylor AJ, Burge PS, Boyle C. Evidence based guidelines for the prevention, identification, and management of occupational asthma. Occup Environ Med 2005;62:290 299. 2. Fishwick D, Barber CM, Bradshaw LM et al. Standards of care for occupational asthma: an update. Thorax 2012;67:278 280. 3. Baur X, Sigsgaard T, Aasen TB et al. Guidelines for the management of work-related asthma. Eur Respir J 2012;39:529 545. 4. Moore VC, Jaakkola MS, Burge PS. A systematic review of serial peak expiratory flow measurements in the diagnosis of occupational asthma. Ann Respir Med 2010;1:31 44. 5. Robertson W, Robertson AS, Burge CB et al. Clinical investigation of an outbreak of alveolitis and asthma in a car engine manufacturing plant. Thorax 2007;62:981 990. 6. Fox J, Anderson H, Moen T, Gruetzmacher G, Hanrahan L, Fink J. Metal working fluid-associated hypersensitivity pneumonitis: an outbreak investigation and case-control study. Am J Ind Med 1999;35:58 67. 7. Gannon PF, Newton DT, Belcher J, Pantin CF, Burge PS. Development of OASYS-2: a system for the analysis of serial measurement of peak expiratory flow in workers with suspected occupational asthma. Thorax 1996;51:484 489. 8. Moore VC, Jaakkola MS, Burge CB et al. A new diagnostic score for occupational asthma: the area between the curves (ABC score) of peak expiratory flow on days at and away from work. Chest 2009;135:307 314. 9. Burge CB, Moore VC, Pantin CF, Robertson AS, Burge PS. Diagnosis of occupational asthma from time point differences in serial PEF measurements. Thorax 2009;64:1032 1036. 10. Anees W, Blainey D, Moore VC, Robertson K, Burge PS. Differentiating occupational asthmatics from non-occupational asthmatics and irritant-exposed workers. Occup Med (Lond) 2011;61:190 195. 11. Hendrick DJ, Marshall R, Faux JA, Krall JM. Positive alveolar responses to antigen inhalation provocation tests: their validity and recognition. Thorax 1980;35:415 427. 12. Halpin DM, Graneek BJ, Turner-Warwick M, Newman Taylor AJ. Extrinsic allergic alveolitis and asthma in a sawmill worker: case report and review of the literature. Occup Environ Med 1994;51:160 164. 13. Piirilä P, Keskinen H, Anttila S et al. Allergic alveolitis following exposure to epoxy polyester powder paint containing low amounts (<1%) of acid anhydrides. Eur Respir J 1997;10:948 951. 14. Malo JL, Ouimet G, Cartier A, Levitz D, Zeiss CR. Combined alveolitis and asthma due to hexamethylene diisocyanate (HDI), with demonstration of crossed respiratory and immunologic reactivities to diphenylmethane diisocyanate (MDI). J Allergy Clin Immunol 1983;72:413 419.