Thorax 1996;51:947-955 947

Occupational lung disease * 9 Series editor: P S Burge Thorax: first published as 10.1136/thx.51.9.947 on 1 September 1996. Downloaded from

Occupation and chronic obstructive pulmonary disease (COPD)

D J Hendrick

Historical background and current claims and controversies which have followed. controversies Particular attention may be drawn to three Chronic obstructive pulmonary disease evolving issues. (COPD) is defined by the presence of a diffuse The first is that many (perhaps most) reduction in airway calibre relative to the authorities would now regard grades degree of lung inflation which cannot be 1 and 2 as occupational attributable to reversed by treatment. The "fixed" nature of cotton dust. The physiological correlates this obstruction is the cardinal feature of (acute but reversible episodes of airway ob- COPD. Obstruction which responds partially struction) are diagnostic of asthma, and the but not fully to treatment implies that there curious work-related periodicity of the symp- may be an asthmatic (or acute bronchitic) toms (which had been considered particularly component to the disorder, of which the fixed characteristic of byssinosis) has since been component alone should be identified as observed with many other types of occupa- COPD. Its importance lies with its tendency to tional asthma. As with active asthma of any progress and to cause a disabling, even life aetiology, airway hyperresponsiveness can be threatening, loss of lung function. demonstrated to a variety of bronchoconstric- tor agents at times when byssinosis grade 1 or http://thorax.bmj.com/ EXPERIENCE WITH COT1ON DUST grade 2 is active. If; then, byssinosis grades 1 That COPD might arise as a consequence of and 2 are examples of occupational asthma, is the occupational environment is a matter of occupational COPD (byssinosis grade 3) a evolving interest and importance, and not a consequence of occupational asthma? little controversy. The possibility of COPD of The relation of byssinosis grade 3 to occupational origin, unassociated with compli- byssinosis grades 1 and 2 may be looked at cated , first gained widespread from a different viewpoint, and this constitutes acceptance as recently as the 1960s as a result the second issue. The investigations which led on September 28, 2021 by guest. Protected copyright. of a series of publications of investigations in to the Schilling grading classification were cotton workers by Schilling and Bouhuys and cross sectional not longitudinal in structure, their colleagues.' 2 They suggested that byssi- and this may pose problems over assumptions nosis could be usefully classified into three dis- of causality. Grade 1 byssinosis was not, in fact, tinct clinical grades. Workers who developed observed to progress to grade 3 byssinosis, and chest tightness and breathing difficulty only on so it is entirely conceivable that byssinosis the first working day of each week were said to grade 3 could be a fundamentally different dis- have byssinosis grade 1; those who had similar order from byssinosis grades 1 and 2, albeit symptoms on additional working days of the one which is also induced by occupational week, but who recovered fully away from the exposure to cotton dust. Could byssinosis workplace, were said to have byssinosis grade grade 3 after all be the prototype for occupa- 2; and those who had symptoms of persisting tional COPD, even in the absence of occupa- respiratory disability were said to have byssino- tional asthma? If so, is it likely that cotton dust sis grade 3. Physiological studies indicated that alone among occupational agents would have byssinosis grades 1 and 2 were associated with this effect on the airways? reversible airways obstruction, while byssinosis The third issue highlights the most impor- Chest Unit and grade 3 was associated with fixed airways tant problem in assessing occupational COPD Regional Unit for Occupational Lung obstruction. It was assumed that affected in the 1 990s. Most of the cotton workers of the Disease, workers followed an orderly progression 1960s had been heavy cigarette smokers for Newcastle General through these escalating grades. Cotton dust many years, and there are now doubts whether Hospital, Westgate Road, was consequently a potential cause of COPD, the confounding effect of this on the develop- Newcasde upon Tyne and byssinosis grade 3 was occupational ment of COPD (which was barely recognised NE4 6BE, UK COPD. in the 1 960s) was adequately taken into D J Hendrick These were landmark publications, and they account.' Thus, the conclusion that occupa- Correspondence to: provide a convenient and highly relevant tional exposure to cotton dust causes COPD Dr D J Hendrick. historical focus from which to review -the appears less secure, and the confounding role 948 Hendrick

of cigarette smoking has come to lie at the through smoking alone and through the occu- centre of current controversies. pational environment alone - poses great diffi- culty for the investigator, largely because Thorax: first published as 10.1136/thx.51.9.947 on 1 September 1996. Downloaded from ROLE OF CIGARETTE SMOKING doubts over the accuracy of past smoking The relative effects of smoking and occupation histories may invalidate, or at least weaken, the on COPD can only be assessed from meticu- statistical analyses involved. There may be lous epidemiological investigations where strong justification for such doubts. One study study populations include smokers and non- concluded that up to 7% of smokers and smokers, and workers with and without the former smokers had described themselves relevant occupational exposure. There may be incorrectly as never smokers.4 great difficulty in finding such a balanced population, and there may be great difficulty in ROLE OF ASTHMA avoiding bias in the exposure histories which There are two further sources of confusion. are obtained. Most investigations of occupa- Firstly, there is the realisation that fixed airflow tionally induced disease lead eventually to limitation may also arise as a consequence of anxiety, anger, and claims for compensation in long standing asthma, and secondly there is the at least some of the study population; and once knowledge that regular occupational exposure these arise memory tends to become unduly to respirable irritants commonly leads to stimulated for the noxious nature of former mucus hypersecretion and to chronic produc- workplace environments but paradoxically tive (that is, chronic bronchitis)which, in suppressed for estimates of former smoking an occupational setting, is often termed consumption. An important additional ten- "industrial bronchitis".5 When chronic bron- dency is for those who are best able to tolerate chitis develops coincidentally alongside the heaviest consumption levels of to asthma, the emergence of obstructive symp- take on jobs with the most noxious occupa- toms (, chest tightness, undue shortness tional exposures, and vice versa. Conversely, of breath) is often attributed, not unreason- workers may leave employment causing sur- ably, to COPD and to the cause of the produc- vival bias if they are less able to tolerate such tive cough. Chronic bronchitis does not itself exposures, or if the work is physically too cause airway obstruction, however, and so demanding and they develop coincidentally obstruction when it does occur with chronic smoking-induced COPD. These influences on bronchitis has to be attributed to other patho- worker selection make it particularly difficult physiological processes - that is, to asthma to conduct epidemiological studies among when obstruction is reversible, to COPD when occupational populations, and to allow fully for obstruction is or to both when the contributory effects to COPD from smok- fixed, there are ing. both reversible and fixed components.6 http://thorax.bmj.com/ The greatest risk for the development of The possible presence of coincidental "cryp- COPD in the population at large is undoubt- togenic" asthma (an increasingly common dis- edly posed by cigarette smoking, and cigarette ease at all ages in the population at large) may smoking remains disproportionately common consequently simulate occupational COPD, among the very working populations most sus- just as asthma in smokers in a chest clinic set- pected of developing COPD from occupa- ting may simulate smoking induced COPD. tional exposures. Cigarette smoking therefore This is particularly likely if "industrial bronch-

itis" (or "smokers' bronchitis") coexists. Crypto- on September 28, 2021 by guest. Protected copyright. provides a major potential source of confound- genic asthma may also simulate occupational ing in investigations which seek to identify asthma because it is likely to worsen transito- excessive longitudinal losses in ventilatory rily in the workplace if there are respirable irri- function in working populations. Thus investi- tants, and in some occupational settings such gators, regulating bodies, employers, and em- as cotton mills the development of occupa- ployees may be hard pressed to recognise when tional asthma is as plausible as the develop- "excessive" loss is truly excessive once the idio- ment of occupational COPD. If it is true that syncratic effect of cigarette smoking is taken there are substantial differences in the preva- into account. Full smoking histories are conse- lence of cryptogenic asthma from region to quently essential in research, and should region, from population to population, and include a quantitative measure of cumulative from birth cohort to birth cohort, as is now consumption. suggested, formidable difficulties must be The issue is unduly complicated because expected in allowing for this if a given working only a minority of smokers (15-20%) actually population appears to show an excess of develop COPD (hence the idiosyncrasy), and COPD. because COPD in never smokers is extremely rare even in working populations with regular exposure to noxious occupational agents. This CURRENT CONTROVERSIES assumes the workers do not suffer from The experience with the cotton industry asthma. There is evidence, however, that when consequently provides invaluable lessons for smokers additionally work with noxious respir- the assessment of occupational COPD, and for able agents, COPD occurs with unusual evaluating today's inevitable controversies. In frequency and/or severity. This implies an essence they centre on whether an excess of adverse interaction between smoking and the COPD in a given working population can be working environment. Recognising such an attributed to smoking or cryptogenic asthma interaction - that is, the occurrence of COPD rather than to the working environment; and in smokers in excess of that to be expected whether, if the workplace is responsible, Occupation and COPD 949

COPD arises independently of occupational function'- or interact with environmental ex- asthma or as a direct consequence of it - or posures of relevance. Otherwise there is the both. risk that these factors - for example, age, Thorax: first published as 10.1136/thx.51.9.947 on 1 September 1996. Downloaded from height, race, sex, changing body mass, smok- Pathophysiology ing, viral infections, atopy, airway Fixed airway obstruction may be a consequence responsiveness - might be distributed un- ofeither intrinsic inflammatory/fibrotic disease of evenly between subgroups of the study popula- the intrathoracic airways, or ofdestructive disease tion which differ also in the levels of exposure of the parenchyma and interstitium of the lung to the agent suspected of inducing COPD. If (emphysema). With emphysema, airway obstruc- such unevenness occurs, it might explain or tion and air trapping are associated additionally exaggerate significant differences which appear with impairment ofgas transfer and loss ofelastic otherwise to be a consequence of the occupa- recoil. It is this loss of elastic support to the tional agent. Conversely, it might mask a true smaller airways, whose walls are not supported occupational effect in investigations which by cartilage, which leads to airway collapse and appear to give "negative" results. The potential obstruction. This mechanism is seen in occupa- for such confounding is critical to the planning tional COPD attributable to complicated pneu- and analysis of investigations of this type. moconiosis. With the exception of cadmium From a series of pioneering epidemiological induced emphysema, it appears to be an unusual investigations of COPD attributable to ciga- mechanism for other types of COPD attributed rette smoking, Fletcher and colleagues showed to occupation, though it should be said that during the 1960s that the majority of smokers histological research into the matter is very experienced no adverse effect on ventilatory limited. function.7 In the minority who were adversely With most types of COPD attributed to affected, a steady excess annual decline in occupation lung function tests reveal no forced expiratory volume in one second significant impairment of gas transfer com- (FEVI) was noted from study year to study pared with control data, implying that intrinsic year, the magnitude of which (FEVy slope) was disease of the airways is a more probable related to the level of smoking consumption. mechanism than emphysema. Extensive work The mean annual decline among symptomatic with COPD attributable to smoking has smokers was approximately twice that of the suggested that the smaller unsupported air- non-smokers. The excess decline could be ways again provide the most probable primary detected after as little as five years of regular site for the obstruction to airflow, there being a smoking, and its cumulative dose related effect mixed pathological picture of inflammation was readily quantified by the degree of fixed and fibrosis chiefly in and around the bronchi- airway obstruction evident at the time of the oles. The disorder is more accurately described initial measurement of ventilatory function. http://thorax.bmj.com/ as a bronchiolitis or bronchiolitis obliterans The excess rate of decline ceased when smok- than as a chronic obstructive bronchitis. How- ers gave up the habit, but the damage already ever, bronchial inflammation is often seen, sustained was permanent. A similar natural especially when there is chronic productive history is to be expected from any cause of cough. This is best regarded as an independent COPD, though there will be much variability phenomenon, unrelated to airflow limitation.6 from individual to individual, and in some of The mechanism and primary site underlying those affected many exposure years will pass COPD attributable to chronic asthma remain before the disease becomes evident. on September 28, 2021 by guest. Protected copyright. to be clarified, and so there is much to be said A cohort of the men investigated by Fletcher for using the diagnostic term COPD in a and colleagues (male transport workers in purely functional sense-that is, to indicate London) were followed longitudinally so that fixed airway obstruction attributable to a the actual annual changes in FEVy could be variety of possible pathological processes. measured. The mean decline in FEVy (when standardised for height) was about 30 ml/year. Recognition of excessive longitudinal The actual rates of decline were related to decline in ventilatory function height, smoking, and symptoms and appeared An epidemiological approach to estimate the to increase a little with increasing age. rate of decline in ventilatory function offers the only practical means of identifying occupa- DISCREPANCY BETWEEN CROSS SECTIONAL AND tional causes of COPD in living populations. If LONGITUDINAL DATA the occupational contribution to COPD in a When data from the transport workers were given workforce is relatively small, and if analysed from the initial cross sectional study, COPD itself is uncommon (often the case in regression of FEV1 on age suggested a rather populations subjected to "healthy worker" and greater mean annual decline of 45 ml/year. "survivor" biases), the investigation of large That is to say, the differences in FEV1 between numbers of subjects will be necessary if an the youngest and oldest participants suggested occupational effect is to be shown convinc- that the mean annual decline in FEVI, after ingly. Such investigations are necessarily time allowing for differences in height, was about consuming and expensive. They must also be 1.5 times the rate actually measured during the complex if modern statistical analyses using subsequent longitudinal phase of the investiga- multiple regression techniques are to take tion. The discrepancy was not readily ex- adequate account of all the various factors plained at the time, but has since been which may independently influence the meas- recognised by a number ofinvestigators. Those ured end point - that is, the level of ventilatory from Tulane University, New Orleans pro- 950 Hendrick

Table 1 Age related changes in ventilatory function OTHER FACTORS OF POSSIBLE RELEVANCE The study of London transport workers Cross sectional regressions of mean FEVy Thorax: first published as 10.1136/thx.51.9.947 on 1 September 1996. Downloaded from against age: usefully showed that, although intercurrent 1974 - 42.6 ml/year episodes of respiratory tract viral infection- 1975 - 47.2 that is, episodes of acute bronchitis and 1976 - 44.9 1977 - 50.5 exacerbations of COPD - caused significant 1979 - 44.6 reductions in ventilatory function, the effect Mean of longitudinal regressions of was only temporary. Full recovery to former FEV1 against time: 1974-9 (5 years) - 12.4 ml/year levels ofventilatory function were noted within 1975-9 (4 years) - 17.4 a few weeks, after which the rate of longitudi- nal decline returned to its usual level. Repeated The investigation involved a 'normal' population of 52 adult caucasian men aged 30-58 years of whom nine were current viral infections do not therefore influence the smokers, 16 former smokers, and 27 never smokers. There assessment of occupational COPD from either were no relevant occupational exposures. Three subjects reported chronic cough, and seven undue breathlessness. The cross sectional or longitudinal studies, pro- initial 1974 measurements were slightly less than those of vided there are no acute infections at the time 1975, possibly reflecting inexperience with spirometric manoeuvres, and so the period 1975-9 may provide a more of study. Whether intercurrent episodes of accurate estimate of longitudinal change than 1974-9. acute bronchitis following brief exposure to Reproduced from Glindmeyer et at' with permission. toxic chemicals at work, such as gassing accidents, will prove to be equally benign is duced similar mean figures for FEVy from a currently a matter of much speculation, five year longitudinal study of a normal popu- concern, and ongoing investigation. Major lation of men aged 30-58 years.8 The cross accidents which produce life threatening pul- sectional data from each survey year indicated monary toxicity are recognised to cause a mean decline of 40-50 ml/year with increas- bronchiolitis obliterans in a small proportion of ing age, once the effect ofheight was taken into survivors, though most appear to recover fully. account, while the mean of the actual longitu- A few develop asthma which may persist dinal declines measured in each participant indefinitely and so pose a further pathway for was of the order of 15-20 ml/year (table 1). the emergence of COPD. This complication of Thus, cross sectional data cannot be ex- acute pulmonary toxicity has been termed trapolated to predict longitudinal change with "reactive airways dysfunction syndrome" any precision, and should not be compared (RADS) in North America.'° The designation directly with longitudinal data from other is possibly unhelpful since it suggests incor- investigations. The discrepancy probably arises rectly that RADS is a disorder which is distinct because subjects at the younger end of an age from asthma. 20-60 from a current Age, height, race and sex, however, all exert spectrum of, say, years http://thorax.bmj.com/ cross sectional investigation do not necessarily important influences on the measurement of have the same mean FEV1 as did those at the ventilatory function, and it has recently been older end ofthe spectrum some 40 years earlier suggested that air pollution and increasing when they were of a similar young age. In fact, body mass during the course of a longitudinal the currently young subjects are likely to have investigation may also exert a potentially higher values of FEVy at a standardised young important (adverse) effect. The wide range of age, partly because they have larger lungs and apparent individual susceptibility to COPD partly because they are less likely to have suggests the possible dependence on genetic sustained respiratory diseases such as tubercu- factors, and there is supportive evidence for on September 28, 2021 by guest. Protected copyright. losis and bronchiectasis which impair ventila- this especially from twin studies.". 12 Geneti- tory function. The difference in FEVy between cally determined atopy has not generally been younger and older subjects at the time of a found to influence COPD, but this depends on cross sectional study therefore exaggerates the how COPD is defined and how vigorously an true declines which were experienced by the asthmatic contribution is excluded. older subjects during the preceding years. Fur- A preliminary analysis of the American thermore, the currently young subjects are multicentre Lung Study suggests that likely to encounter less lung damaging insults as they age to 60 years than those already aged 5 60. Thus, mean longitudinal decline in lung function in populations is not wholly a 4 Al consequence of increasing age, but depends 3 also on the cumulative burden of damaging U- A2 .1~ B1 the pulmonary insults during ageing period. LL 2 This is illustrated in fig 1, which offers an explanation for the discrepant results from 1 B2 cross sectional investigation. By using study subjects as their own 0 20 40 60 "controls" and avoiding the problems which Age (years) arise from mismatching in cross sectional - Figure 1 Discrepant mean declines in FEVI from cross studies that is, by eliminating between sectional and longitudinal data as workers agefrom 20 to worker variability - longitudinal studies are 60 years. A,-B, = true (future) longitudinal decline in inherently more robust.9 They may be unduly subjects currently aged 20years;A2-B2 = true (completed) longitudinal decline in subjects currently aged 60 years; vulnerable to diminishing participation rates, A,-B2 = apparent (current) longitudinal decline using however, and this may introduce new risks of cross sectional data from subjects currently aged 20-60 bias. years. Occupation and COPD 951

airway responsiveness (to methacholine) may With wood dust, a further common cause of prove to be almost as important as cigarette occupational asthma, evidence for a COPD smoking in exerting an adverse influence on effect is less strong, but again smokers appear Thorax: first published as 10.1136/thx.51.9.947 on 1 September 1996. Downloaded from the rate of decline in ventilatory function. If it unduly susceptible to it. This is noteworthy in is confirmed that a measure of asthmatic activ- view of the curious "protective" effect of smok- ity is relevant to predicting the development of ing observed with occupational asthma attrib- COPD (a relationship enshrined in what utable to western red cedar.'8 A similar protec- became known as the "Dutch hypothesis" of tive effect in smokers, for which there are the 1970s and 1980s but was disputed there- plausible immunological explanations, has after), then factors of aetiological relevance to been observed with extrinsic allergic alveolitis asthma may also need to be taken into account and sarcoidosis, and so should not be dis- when studying COPD.'3 14 It may be useful missed too hastily."9 20 therefore to consider the occupational agents which have been associated with COPD in two CHEMICAL INDUCERS OF OCCUPATIONAL ASTHMA categories - those which are also believed to In Britain at present seem to be the cause asthma and those which appear to have most common cause of occupational asthma, no asthmagenic properties. and asthma continues to provide a typical example of asthma attributable to occu- pational chemicals. Not unnaturally, isocy- COPD attributable to agents known to anate workers have provided a focus for inves- induce occupational asthma tigations, of which there have been many, of NATURALLY OCCURRING INDUCERS OF possible occupational COPD. The results have OCCUPATIONAL ASTHMA been conflicting and have stimulated much Respiratory disease in grain workers has been controversy. A major multicentre surveillance widely recorded since the 18th century and the programme in Britain revealed no hint of a time of Ramazzini. Grain dust has become a COPD effect, while the study of a single isocy- particularly well recognised cause of occupa- anate producing plant in the USA suggested a tional asthma, though the precise causative crippling occupationally induced mean annual agent (or agents) remains unclear. Storage decline in FEVy exceeding 100 ml/year.2' 22 mites, microbial contaminants, pesticides and Both investigations may have been flawed - the fungicides, and even rodent urinary proteins one because there was an implausibly low have all been incriminated, together with aller- prevalence of occupational asthma (probably a genic material derived from the grain itself. A consequence of survivor bias), the other number of investigators have produced impres- because the levels of ventilatory function sive evidence that occupational exposure to observed initially were too well preserved for grain dust may also lead to COPD, though an excessive decline of such a degree to have http://thorax.bmj.com/ none have suggested that this is a direct conse- been occurring before the investigation com- quence of occupational asthma and most have menced. found some inconsistencies among their Not surprisingly, reports of these investiga- data.'5 16 tions stimulated many others, and these have The experience of Chan-Yeung and her col- produced a spectrum of conclusions which is leagues in the port of Vancouver provides a almost as wide. The five year investigation of useful illustrative example. 17 They followed Diem and colleagues of workers in a new tolu- port grain workers and control subjects work- ene diisocyanate manufacturing plant included on September 28, 2021 by guest. Protected copyright. ing in civic centre posts between 1978 and an extensive series of isocyanate exposure 1990. After six years the annual rates for the measurements, from which the workforce was decline in FEVy were significantly greater usefully separated into categories of low to among the grain workers (-31 ml/year) than average cumulative exposure (% 68.2 ppb- the controls ( + 4 ml/year), but when the months) and high cumulative exposure smokers alone were compared little difference (> 68.2 ppb-months).23 The high exposure was noted between grain workers and controls group did show a significantly greater annual (-36 ml/year versus -31 ml/year). In this decline in FEVy (-37 ml/year) than the particular investigation, therefore, the "grain non-smokers of the low exposure group (+1 COPD effect" was seen largely in the non- ml/year), but it was not influenced by smoking. smokers (in whom asthma was not satis- Smoking did, however, lead to a similar exces- factorily excluded though atopy proved to be sive decline in those with low levels of exposure irrelevant), and it appeared to be of similar to isocyanate. Thus, the investigation sug- magnitude to the "smoking COPD effect" in gested a modest excessive decline in FEVy as a the control workers. Furthermore, its demon- consequence of either isocyanate exposure or stration depended on there being an unusually smoking, without there being any additive or small annual decline in FEV1 (in fact there was multiplicative effects. These conclusions differ a trivial increase) among the controls. After 12 from the general consensus, which favours an years significant differences between grain interaction between smoking and the occupa- workers and controls were no longer evident, tional environment, but they mirror those but this may have been a consequence of survi- derived from the grain workers in the port of vor bias or of greatly reduced exposure levels. Vancouver. Interestingly, Diem and colleagues Not all investigators have found this COPD did attempt to identify asthmatic subjects from effect, and when it was observed it was their study population and showed no weaken- generally more clear among the smokers than ing of the COPD effect when these workers non-smokers. were excluded from the analysis. 952 Hendrick

COPD attributable to agents not known retired miners, and protagonists have argued to induce asthma that this is to be expected in a job that did not Thorax: first published as 10.1136/thx.51.9.947 on 1 September 1996. Downloaded from CADMIUM readily tolerate any loss of physical fitness. Cadmium occupies an unusual place among Thus, earlier reassuring epidemiological inves- occupational agents reported to cause COPD tigations ofworking coal miners were flawed by in that it appears to do so by causing the industry's unusual susceptibility to the emphysema. Occupational COPD attributable "survivor effect". Furthermore, a small mean to non-focal emphysema seems otherwise to be loss in ventilatory function, which has not confined to complicated pneumoconiosis. attracted much dispute, might be a conse- Cadmium induced emphysema has not been quence of large and disabling effects in a few without controversy, however, from the time of miners rather than of small and clinically its first description in 1952.24 For some years inconsequential losses in many. Antagonists doubts persisted as to whether smoking alone have responded that smoking histories are was responsible for the apparently excess notoriously unreliable in coal mining com- prevalences of COPD noted among some, but munities (in one study 20% of miners who not all, cadmium exposed working popula- described themselves as light smokers were tions.25 More recent investigations involving after death described by their relatives as heavy long term surveillance have provided more smokers 36), and that if disabling losses of ven- convincing evidence in support.26 Although tilatory function do occur, they are more likely cadmium is a trace component of cigarette to be due to smoking (active or passive) or smoke, cumulative exposures from smoking other non-occupational factors such as asthma alone are not likely to approach those sustained or the miners' penchant for breeding pigeons. occupationally. It seems improbable therefore The most recent detailed analysis of longitu- that smoking-induced emphysema could be dinal data from British coal mines did, like ear- attributed to cadmium. lier investigations, provide evidence for a small excess longitudinal decline in FEVI, independ- MINERAL DUSTS ent of pneumoconiosis, which was closely The mineral dusts have provided perhaps the related to measured levels of exposure to coal greatest opportunity for COPD controversy, dust.37-39 The risk of a disabling loss of ventila- even conflict.27 Although airway obstruction tory function (defined by the mean loss in was quickly recognised to be a feature of com- FEVy of 942 ml shown by the miners with plicated pneumoconiosis and of its accompa- disabling symptoms) was less than 5% among nying emphysema, COPD in the absence of non-smoking miners with low cumulative complicated pneumoconiosis has generally levels of coal dust exposure (% 100 g.hour/m3) been attributed to other coincidental disorders. who had worked to the age of 55 years, but the High mortality rates from respiratory disease in confidence intervals were wide. The risk http://thorax.bmj.com/ some groups of miners led to early and increased exponentially to about 20% in those intensive investigations of large numbers in with unduly heavy levels of cumulative expo- many countries.28-31 The outcome was the rec- sure (500 g.hour/m3). Most interestingly, it was ognition that, although complicated pneumo- not only greater in smokers, but the smokers coniosis and occupational tuberculosis were showed a significantly steeper slope relating associated with excess respiratory mortality increasing exposure level to increasing risk. and morbidity, the pattern of respiratory This study therefore supports the notion of an on September 28, 2021 by guest. Protected copyright. disease in miners was otherwise closely dupli- adverse interactive effect between smoking and cated in their families and in suitably matched occupational exposure on the development of control populations.32 34 It appeared to be COPD. The magnitude of the adverse effect of related more to social circumstances (and par- smoking was, however, some three times that ticularly to smoking) than to the working envi- of coal dust, and may have been influenced ronment.35 (underestimated) more than the coal dust The largest investigations have involved coal effect by survivor bias. The accumulating miners, principally because of the enormous evidence for COPD attributable to coal dust populations employed over the years in the coal has recently led to it becoming a compensatory mining industry and partly because of political disease in Britain, providing the claimant has influence. In general, mean losses in ventilatory worked underground for at least 20 years, there function attributable to coal dust exposure is radiographic evidence of dust deposition (rather than cigarette smoking) have been (that is, pneumoconiosis category 1 or more), small or trivial in the absence of complicated and the severity is sufficient to reduce the pneumoconiosis, and because of the large FEV1 by one litre from its predicted value. numbers in many of these investigations these A similar situation has arisen in South essentially negative findings have attracted high Africa, though strengthening evidence that levels of confidence. Some investigations, how- respirable silica in gold mines might cause ever, have suggested that the COPD effect COPD was not published until after the intro- from coal dust might approach that of cigarette duction of state compensation in 1952.9 40 41 smoke. Again there has been controversy, some later The controversy has arisen because dis- investigations failing to provide clear confirma- abling levels of airway obstruction have been tory evidence of a silica-related loss of ventila- noted in a small minority of miners who tory function.42 43 Other recent investigations claimed never to have smoked and who did not (cross sectional and longitudinal) have shown show evidence of complicated pneumoconio- an excess of COPD in workers exposed to silica sis. They were found more prominently among which could not be attributed to smoking, but Occupation and COPD 953

there was no clear relationship with intensity of fume exposure were 17.7 ml/year and 16.4 exposure."'6 In one of these investigations ml/year, respectively. Interactions were noted silica and smoking appeared to exert effects of between welding fume and smoking, and Thorax: first published as 10.1136/thx.51.9.947 on 1 September 1996. Downloaded from similar magnitude, while in another the influ- between welding fume and atopy. There was ence of smoking on the decline in FEVy was no effect from welding fume on gas transfer. approximately twice that of silica exposure. In The longitudinal study consequently suggested the severe cases which led to death, an inter- that a mild COPD effect attributable to action between the two environmental factors welding fume might also occur in non- seemed likely. smokers, but it confirmed that this effect was disproportionate in smokers. The further of the effect in subjects WELDING FUME enhancement atopic Welding fume is conveniently considered sepa- provides a hint of an asthmatic component, rately from other mineral dusts, partly because and the lack of any effect on gas transfer the circumstances of exposure are rather suggests that COPD was not occurring as a different, partly because a possible COPD consequence of emphysema. effect in welders is a matter of considerable A further hint that welding fume from mild topical interest, and partly because evidence of steel might influence asthma (asthma is an additional asthmagenic effect is now emerg- increasingly reported in stainless steel welders ing. Exposure to welding fume consequently as a consequence of presumed chromium brings the discussion full circle. hypersensitivity) has emerged from work of my Consistent with the investigation of numer- own colleagues in a further shipyard.52 We ous work forces exposed regularly to respirable examined cross sectionally apprentices in vari- irritant or noxious dusts, vapours or fumes, ous trades from the age of 15-17 years (when many investigations of welders have demon- they left school) to 27 years (when they had strated a clear excess prevalence of chronic completed nine years at work) and we focused productive cough (industrial bronchitis). Until the investigation on methacholine tests to pro- recently none have found convincing evidence vide objective measurements of airway respon- of an excess of airway obstruction, implying, siveness. These were carried out using a locally perhaps, that if there is a COPD effect it must designed microprocessor-controlled dosimeter be of relatively small degree.47"9 and a locally developed protocol.53 The results A small but significant COPD effect was the were expressed conventionally as PD20 (the conclusion of a cross sectional study of 607 cumulative dose of methacholine estimated to shipyard workers aged 17-69 years carried out provoke a decrement in FEVy of 20%). Airway by Cotes and colleagues.50 After allowing for responsiveness improved steadily with age in apprentices without exposure to welding fume, age and height, the trades associated with http://thorax.bmj.com/ welding fume exposure showed a mean and it changed little in those with ambient exposure significant loss in FEVy of 250 ml compared (shop floor trades not associated directly with with the unexposed trades. This effect was fume generation), and it showed a small noted only among the smokers. Of the original increase in the regular welders. The odds ratios study population, 487 were re-examined seven for the risk of having a positive methacholine years later.5' Multiple regression analyses test after five years work compared with the suggested that was declining at an risk at 0 years are illustrated in fig 2. In essence, FEV, the observations suggest that, after five years of

annual rate of 16.2 ml for a 50 year old on September 28, 2021 by guest. Protected copyright. non-smoking worker without occupational occupational exposure during the 1980s when exposure to welding fume. The additional exposure levels were rigorously controlled, losses attributable to smoking and welding welding fume exposure led to a mean PD20 which was 50% of expected. This implies that, Regular of the regular welders with a PD20 < 1000 ig welders (the range within which subjects with active 15 H asthma are usually found53), some 15-30%, Shop floor depending on atopic status and smoking workers habits, represent an excess attributable to -a _, Ds welding fume exposure. U1) 1.0 H- 0 CO) Office coCO workers Perspective in the population at large .U) The relative risks posed to the general popula- tion by common allergens, outdoor air pollu- tion, air pollution within the home, and respir- co able agents at work is a matter of increasing CO 0-5 public concern, particularly with regard to the 0 aetiology and progression of COPD and asthma. Both disorders are clearly dependent on changing patterns of life within developed countries, and so environmental factors must 0-3 be of great relevance. It may be that many dif- 0-05 0.1 1*00 2 0 ferent respirable agents are capable of exerting Mean exposure to welding fume particulates (mg/mr3) an influence, and that in different individuals Figure 2 Odds ratios (with 95% confidence intervals) for a positive methacholine test the measured COPD effects are due to after workingforfive years.52 different combinations of these agents (a 954 Hendrick

"multiple hit" hypothesis), together with some conclusions reflect a changing scene and a per- inherent susceptibility. sonal current interpretation of it. They also Recent investigations in the former East and depend on additional, but often controversial, Thorax: first published as 10.1136/thx.51.9.947 on 1 September 1996. Downloaded from West regions of Germany provide interesting conclusions that COPD occurs with excessive comparisons within the developed world be- prevalence in a number of other occupational tween an area of high smoking prevalence and settings, including paper pulping mills, print- relatively high outdoor air pollution from ing, firefighting, and flour mills.58... industrial emissions, and one of high domestic 1. Some occupational environments are affluence.54" COPD appears to dominate in likely to exert a COPD effect. the former and asthma in the latter, though in 2. Its impact is likely to be less than that of the more affluent societies there is more rather smoking (perhaps much less), but will vary than less outdoor pollution from vehicle from industry to industry depending on exhausts. This is thought to be of relevance to potency and exposure level of the agent asthma in Los Angeles where such pollution is involved. persistently high and intermittently dramati- 3. Complex adverse interactions probably cally so, but a recent comparative study using exist with smoking, and (presumably) with measurements of airway responsiveness be- other environmental agents. tween an urban area with more modest 4. Many environmental factors may be of exposure to vehicle exhaust (Newcastle upon relevance to its aetiology - the so called "mul- Tyne) and a rural area with low exposure tiple hit" hypothesis. (Cumbria and the English lake district) 5. It is plausible that both asthmatic and showed no difference.56 non-asthmatic pathways could play a role. A further perspective regarding the possible 6. It will be found rarely in the absence of role of occupational exposures in the aetiology both smoking and asthma. of COPD in the population at large can be seen from the results of an epidemiological investi- I am grateful for helpful comments and suggestions from Dr M of urban and rural (Horda- G Pearson, Consultant Physician, Fazakerley Hospital, Liver- gation (Bergen) pool and Dr S C Stenton, Senior Lecturer, University of New- land) communities in western Norway.57 A castle upon Tyne. postal survey first sampled 5000 individuals from a total population of 250 000 and, from 1 Bouhuys A, Heaphy U, Schilling RSF, Welborn JW. in the United States. N Engl Med 1967; Byssinosis _ the respondents, an age stratified sample of 227:170-5. 1512 was invited to undertake a more detailed 2 Schilling RSF, Vigliani EC, Lammers B, Valic F, Gilson JC. included respiratory symp- A report on a conference on byssinosis (14th International investigation which Conference on Occupational Health, Madrid, 1963). toms, smoking history, work history, and spiro- International Congress Series No 62. Amsterdam, Ex- 1275 individuals participated. cerpta Medica: 1963, 137-44. metric tests; http://thorax.bmj.com/ 3 Parkes WR. Occupational asthma (including byssinosis). In: When COPD was defined from strict spiro- Occupational lung disorders. London: Butterworths, metric parameters alone < 70%, 1982:445. (FEV,/FVC 4 Wald NJ, Nanchahal K, Thompson SG, Cuckle HS. Does and FEV1 < 80% of predicted), smoking breathing other people's tobacco smoke cause lung cancer? proved to be the only explanatory variable of BMJ 1986;293:1217-21. 5 Morgan WKC, Seaton A. Occupational lung disease. clear relevance from multiple regression analy- Philadelphia:W B Saunders, 1984. ses. When COPD was defined less stringently 6 Ciba Guest Symposium. Thorax 1959;14:286. 7 Fletcher CM, Peto R, Tinker C, Speizer F. The natural but more clinically from both objective spiro- history of chronic bronchitis and emphysema. Oxford:Oxford metry (FEV1/FVC < 70%) and subjective University Press,1976. 8 Glindmeyer HW, Diem JE, Jones RN, Weill H. Noncompa- on September 28, 2021 by guest. Protected copyright. symptoms (chronic productive cough together rability of longitudinally and cross sectionally determined with undue breathlessness or wheeze), the annual change in . Am Rev Respir Dis 1982; 125:544-8. odds ratios for participants reporting specific 9 Beckl ake MR. Chronic airflow limitation: its relationship to occupational exposure to aluminium, welding/ work in dusty occupations. Chest 1985;88:608-17. 10 Brooks SM, Weiss MA, Bernstein IL. Reactive airways dys- metal fume, quartz, or all increased to function syndrome (RADS): persistent asthma syndrome significant levels, as did the odds ratio for heavy after high level irritant exposures. Chest 1985;88:376-84. 11 Cohen BH, Diamond EL, Graves CG, Kreiss P, Levy DA, exposure to any source of occupational dust. Menkes HA, et al. A common familial component in lung An increased prevalence of COPD was also cancer and chronic obstructive airways disease. Lancet 1977;ii:523-6. noted in the urban compared with the rural 12 Redline S, Tishler PV, Lewitter FI, Tager IB, Munoz A, communities, but this was attributable to the Speizer FE. Assessment of genetic and non-genetic influences on pulmonary function: a twin study. Am Rev greater number of elderly individuals living in RespirDis 1987;135:217-22. Bergen, who were affected disproportionately. 13 Van der Lende R. The epidemiology of chronic nonspecific lung disease. Vol 1. A critical analysis of three field surveys of The study population was consequently suffi- CNSLD carried out in the Netherlands. Assen:Van Gor- ciently large for subjective evidence of occupa- cum, 1969. to be 14 Potsma DS, de Vries K, Koeter GH, Sluiter HJ. Independ- tional productive cough demonstrated, ent influence of reversibility of air-flow obstruction and but the excess prevalences of airway obstruc- nonspecific hyperreactivity on the long-term course of certain of lung function in chronic airflow obstruction. Am Rev tion which were noted among groups Respir Dis 1986;134:276-80. workers did not reach conventional levels of 15 doPico GA, Reddan W, Flaherty D, Tsiatis A, Peters ME, Rao P, et al. Respiratory abnormalities among grain statistical significance. handlers: a clinical, physiologic and immunologic study. Am Rev Respir Dis 1977;115:915-27. Conclusions 16 Becklake MR. Grain dust and health. In: Dosman JA, Cot- ton DJ, eds Occupational pulmonary disease in grain workers: Evidence relating occupational environments focus on grain dust and health. New York: Academic to COPD is extensive but conflicting, and this Press,1980:189-200. 17 Chan-Yeung M, Enarson DA, Kennedy SM. The impact of inevitably stimulates controversy. Any sum- grain dust on respiratory health. State of the Art. Am Rev mary is necessarily influenced by personal, Respir Dis 1992;145:476-87. 18 Chan-Yeung M, Lam S, Koerner S. Clinical features and biased, and possibly preconceived views, so it natural history of occupational asthma due to Western red should be emphasised that the following cedar (Thaja plicata). Am J Med 1982;72:411-5. Occupation and COPD 955

19 Warren CPW. Extrinsic allergic alveolitis: a disease com- 43 Manfreda J, Sidwall G, Maini K, West P, Cherniack RM. moner in non-smokers Thorax 1977;32:567-9. Respiratory abnormalities in employee of the hard P, C, Pre J, Battesti JP, Georges R, et rock mining industry. Am Rev Respir Dis 1982;126: 20 Valeyre D, Soler Clerici Thorax: first published as 10.1136/thx.51.9.947 on 1 September 1996. Downloaded from al. Smoking and pulmonary sarcoidosis: effect of cigarette 629-34. smoking on prevalence, clinical manifestations, alveolitis, 44 Hnizdo E. Combined effect of silica dust and tobacco and evolution of the disease. Thorax 1988;43:516-24. smoking on mortality from chronic obstructive lung 21 Adams WGF. Long-term effects on the health of men disease in gold miners. BrJtInd Med 1990;47:656-64. engaged in the manufacture of toluene di-isocyanate. BrJ 45 Cowie RL, Mabena SK. , chronic airflow limitation, Ind Med 1975;32:72-8. and chronic bronchitis in South African gold miners. Am 22 Peters JM, Murphy RLH, Pagnotto LD, Whittenberger JL. Rev Respir Dis 1991;143:80-4. Respiratory impairment in workers exposed to "safe" lev- els of toluene diisocyanate (TDI). Arch Environ Health 46 Hnizdo E. Loss of lung function associated with exposure to 1970;20:364-7. silica dust and with smoking and its relation to disability 23 Diem JE, Jones RN, Hendrick DJ, Glindmeyer HW, and mortality in South African gold miners. BrJt Ind Med Dharmarajan V, Butcher BT, et al. Five-year longitudinal 1992;49:472-9. study of workers employed in a new toluene diisocyanate 47 Barhad B, Teculescu D, Craciun 0. Respiratory symptoms, manufacturing plant. Am Rev Respir Dis 1982;126:420-8. chronic bronchitis, and ventilatory function in shipyard 24 Baader EW. Chronic cadmium poisoning. Ind Med Surg welders. Int Arch Occup Environ Health 1975;36: 137-50. 1952;21:427-30. 48 McMillan GHG, Pethybridge RJ. A clinical, radiological 25 Leading article. Cadmium and the lung. Lancet 1973; and pulmonary function case control study of 135 ii:1134-5. dockyard welders aged 45 years and over. 7 Soc Occup Med 26 Davison AG, Fayers PM, Newman Taylor AJ, Venables KM, 1984;34:3-23. Darbyshire J, Pickering CAC, et al. Cadmium fume inha- 49 Simonato L, Fletcher AC, Andersen A, Anderson K, Becker lation and emphysema. Lancet 1988;i:663-7. N, Chang-Claude J, et al. A historical prospective study of 27 Stenton SC, Hendrick DJ. Airflow obstruction and mining. European stainless steel, mild steel, and shipyard welders. In: Banks DE, ed. The mining industry. Occupational medi- BrJrInd Med 1991;48:145-54. cine: state of the art reviews 1993;8:155-70. 50 Cotes JE, Feinmann EL, Male VJ, Rennie FS, Wickham 28 The Registrar General's Decennial Supplement, England CAC. Respiratory symptoms and impairment in shipyard and Wales, 1951. Occupational mortality, Pt II, Vol I, welders and caulker/burners. Br J Ind Med 1989;46:292- Commentary. London:HMSO,1958. 301. 29 Enterline PK. Mortality rates among coal miners. Am J 51 Chinn DJ, Stevenson IC, Cotes JE. Longitudinal respiratory 1964;54:758-68. survey of shipyard workers: effects of trade and atopic sta- 30 Atuhaire LK, Campbell MJ, Cochrane AL, Jones M, Moore tus. Br3tInd Med 1990;47:83-90. F. Mortality of men in the Rhondda Fach 1950-80. Br J 52 Beach JR, Dennis JH, Avery AJ, Bromly CL, Ward RJ, Wal- Ind Med 1985;42:741-5. ters EH, et al. An epidemiologic investigation of asthma in 31 Finkelstein M, Kusiak R, Suranyi G. Mortality among min- welders. Am J Care Med 1996 (in press). ers receiving compensation for silicosis in Ontario: 1940- Respir Crit 75. J Occup Med 1982;24:663-7. 53 Beach JR, Young CL, Avery AJ, Stenton SC, Dennis JH, 32 Medical Research Council. Chronic bronchitis and occupa- Walters EH, et al. Measurement of airway responsiveness tion. BMJ 1966;1:101-2. to methacholine: relative importance of the precision of 33 Miller BG, Jacobson M. Dust exposure, pneumoconiosis, drug delivery and the method of assessing response. and mortality of coal miners. BrJr Ind Med 1985;42:723- Thorax 1993;48:239-43. 33. 54 von Mutius E, Fritzsch C, Weiland SK, Roll G. Magnussen 34 Foxman B, Higgins ITT, Oh MS. The effects of occupation H. Prevalence of asthma and allergic disorders among and smoking on respiratory disease mortality. Am Rev children in united Germany: a descriptive comparison. Respir Dis 1986;134:649-52. BMJt 1992;305:1395-9. 35 Ortmeyer CE, Costello J, Morgan WKC, Swecker S, Peter- 55 Schlipkopter HW, Stiller-Winkler R, Ring J, Willer HJ. son M. The mortality ofAppalachian coal miners, 1963 to Impact of air pollution on children's health: results from 1971. Arch Environ Health 1974;29:67-72. Saxony-Anhalt and Saxony as compared to Northrhine- 36 Cockcroft A, Seal RME, Wagner JC, Lyons JP, Ryder R. Westphalia. In: Health and ecological effects. Papers from the Anderson N. Postmortem study of emphysema in coal 9th World Mean Air Congress, Montreal, Quebec, Canada workers and non coal workers. Lancet 1982;ii:600-1. 1992. Pittsburgh: Air and Waste Management Association, 37 Love RG, Miller BG. Longitudinal study of lung function in 1992. http://thorax.bmj.com/ coal-miners. Thorax 1982;37:193-7. 56 Devereux G, Ayatollahi SMT, Ward R, Bromly C, Bourke 38 Marine WM, Gurr D, Jacobson M. Clinically important SJ, Stenton SC, Hendrick DJ. Asthma airways responsive- respiratory effects of dust exposure and smoking in British ness and air pollution in two contrasting districts of north- coal miners. Am Rev Respir Dis 1988;137:106-12. ern England. Thorax 1996;51:169-74. 39 Soutar C, Campbell S, Gurr D, Uoyd M, Love R, Cowie H, 57 Bakke PS, Batse V, Hanoa R, Gulsvik A. Prevalence of et al. Important deficits of lung function in three modern obstructive lung disease in a general population: relation to colliery populations. Relations with dust exposure. Am Rev occupational title and exposure to some airborne agents. Respir Dis 1993;147:797-803. Thorax 1991;46:863-70. 40 Wiles FJ, Faure MH. Chronic obstructive lung disease in gold miners. In: Walton WH, ed. Inhaled particles: IV 58 Chan-Yeung M, Wong R, Maclean L, Tan F, Dorken E, the Schulzer M, et al. Respiratory survey of workers in a pulp Proceedings of an international symposium organised by and paper mill in Powell River, British Columbia. Am Rev British Society, Edinburgh, 22-26 Sep- on September 28, 2021 by guest. Protected copyright. tember 1975.Oxford: Pergamon Press, 1977:727-34. Respir Dis 1980;122:249-57. 41 Irwig L, Rocks P. Lung function and respiratory symptoms 59 Sparrow D, Bosse R, Rosner B, Weiss ST. The effect of in silicotic and non silicotic gold miners. Am Rev Respir Dis occupational exposure on pulmonary function: a longitu- 1978;117:429-35. dinal evaluation of fire fighters and non fire fighters. Am 42 Sluis-Cremer GK, Walers IG, Sichel HS. Ventilatory Rev Respir Dis 1982;125:319-22. function in relation to mining experience and smoking in a 60 Kauffman F, Drouet D, Lellouch J, Brille D. Occupational random sample of miners and non-miners in a Witwaters- exposure and 12-year spirometric changes among Paris rand town. Br3tInd Med 1967;24:13-25. area workers. BrJfInd Med 1982;39:221-32.