Relative importance of cigarette smokingin occupational lung … · British JournalofIndustrial Medicine 1981;38:1-13 Relative importance ofcigarette smokingin occupational lung disease*

British Journal of Industrial Medicine 1981;38:1-13

Relative importance of cigarette smoking inoccupational lung disease*P C ELMES

From the MRC Pneumoconiosis Unit, Llandough Hospital, Penarth, S Glamorgan CF6 IXW, UK

ABSTRACT Since 1900 respiratory disease has remained a constant serious cause of chronic ill healthand premature death in Britain. The falling importance of tuberculosis and pneumonia has beenoff-set by the rise in lung cancer. Bronchitis morbidity and mortality have fallen only slightly since1935. To produce any real improvement in the future existing information as to cause must bestudied. The relative contribution of occupational exposure is compared with the importance ofcigarette smoking. Relevant information is scanty and has been produced to emphasise the existenceof occupational diseases rather than assess their importance to the community as a whole. In Britainthe evidence is that within the coal mining and iron and steel industries conditions are now suchthat dust exposure contributes little to the morbidity or mortality compared with the workers'smoking habits. Similar results have been shown by a cross-sectional survey of many dusty occupa-

tions in Western Germany. Only in the disappearing Welsh slate industry has dust disease been atleast as important as smoking. Until the current regulations were introduced conditions existedamong asbestos workers such that the combined effect of cigarette smoking and dust exposure led toa loss of life expectation of over 10 years in moderate smokers. Since the new regulations were

introduced the risk for asbestos workers should approximate to that for other industrial workers.While control of occupational exposure to respiratory hazards remains important, a far greaterimprovement to respiratory health would be produced by controlling tobacco smoking.

Respiratory disease places a heavy burden on oursociety. In 1977 it caused 59m days' sickness absenceand 118 000 deaths in England and Wales. It was themost important cause of sickness absence and came

second only to circulatory disease as a cause ofdeath in that year (tables 1 and 2). During the pastcentury the pattern of serious disease has changed,but its importance has not diminished. The reductionin disease due to infection has been offset by theincrease in lung cancer. For those concerned withoccupational diseases, respiratory disease is clearlyimportant because of sickness absence. But it is alsoimportant as a cause of death before retiral age.Figure 1 shows the pattern of mortality in the 50-59age group since 1911. There has been a 20-foldreduction in deaths from tuberculosis and a 10-fold

*The Ernestine Henry Lecture delivered at the Royal Collegeof Physicians London on 6 November 1979.

Received 26 March 1980Accepted 1 May 1980

Table 1 Principal causes of deaths in England andWales 1977-all ages: number of deaths. (Per-centages inparentheses)

Cause Men Women Total

Circulatory disease 143 201 (49) 149 395 (52) 292 596 (51)Respiratory diseases(including cancer) 70 960 (24) 47 260 (17) 118 220 (21)Non-respiratory cancer 40 112 (14) 51 098 (18) 91 210 (16)Allcauses 289773 286 155 575928

Mortality Statistics 1977. OPCS DH2 No 4.

Table 2 Principal causes of sickness absence in Britain1977 (millions of days). (Percentages inparentheses)

Cause Men Wonzen Total

Respiratory diseases 48-4 (19) 10-8 (17) 59-2 (18)Circulatory diseases 44-3 (17) 4-6 (7) 48-9 (15)Musculoskeletal disorders 32-2 (13) 6-8 (10) 39 0 (12)Mental disorders 22-2 (9) 9-5 (15) 31-7 (10)All causes 256 2 65 3 321-5

Social Security Statistics 1977, DHSS.1

on Decem

ber 31, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.38.1.1 on 1 F

ebruary 1981. Dow

nloaded from

Elnmes

1911-5 21-5 31- 5 41-5 51-5 61-5 71-5 0(4'

Fig 1 Pattern of deaths from respiratory disease since1911 for men aged 55-59. For clarity line forbronchopneumonia has been omitted as it remainedconstant at about 0.3/1000 living a year throughoutperiod. (From statistical reviews issued by Registrar-General's Office and OPCS, HMSO.)

reduction for lobar pneumonia. Deaths from bron-chitis fell dramatically between 1911 and 1935 butsince then have fallen more gradually. Deaths fromasthma seem to have fallen slightly (owing toa change in classification recently these figuresfor asthma are unreliable), while deaths frombronchopneumonia (not shown on the graph) haveremained unchanged throughout the period at about0 3 deaths per thousand living a year. The onlyrespiratory disease to show an increase is cancer ofthe lung. The increase is not too dramatic in this agegroup but nevertheless lung cancer is now the mostimportant cause of death, more than twice as im-portant as bronchitis. Although the other respiratorydiseases are unimportant causes of death in this agegroup, they are important causes of sicknessabsence from work. Lung cancer runs a compara-tively short and fatal course so that it is under-represented on sickness absence certificates and doesnot appear in fig 2. Since 1953 the main causes ofsickness absence (other than influenza) have beenbronchitis and tuberculosis. (Again because ofchanges in classification consistent figures forasthma are not available.)To reduce the burden of respiratory disease we

must consider the factors that may have influencedthe disease in recent years. The causes may be listedas follows: (1) exposure to infective agents; (2)social factors, such as population density and

0 PI

fA--00,,0s ,6 0.a.

6 0 -oO--O~~~I ,O.

Bronchitis 0----0

Respiratory TB C-

Y I

Year _o 9'

Fig 2 Pattern of sickness absence due to respiratorydisease since 1953 for all ages. Figure for "allneoplasms" is well below Im days each year. Statisticsfor asthma alone are incomplete but appear to start atabout 3m days in 1953-4 and fall to about 2m in1969-70. (From annual reports ofMOH DHSS, andSocial Security Statistics, HMSO.)

nutrition; (3) exposure to non-infective toxic agentsin the air; and (4) inherited susceptibility.Apart from genetic counselling little can be done

about inherited susceptibility, and indeed apart fromrare conditions such as ot-antitrypsin deficiency thereis little clear-cut evidence that inheritance is of greatimportance except in asthma. Infections may now betreated effectively with antibiotics, which may havecontributed to the fall in mortality and sicknessabsence due to tuberculosis in the very recent past.But it is noticeable, both for lobar pneumonia andtuberculosis, that the initial introduction of effectivetreatment did not influence mortality immediately.For both these diseases social factors and possiblythe isolation of infective cases were the main factorsleading to the change observed, which should con-tinue. Asthma is not due to infection, and broncho-pneumonia is usually the terminal event in manyillnesses not primarily due to infection. Socialchanges do not appear to have affected eitherasthma or bronchopneumonia.

Bronchitis presents a more complex picture. Themajor improvement occurred during a period whenno specific treatment was available. Since 1935another factor or factors has interfered with theimprovement, and sickness absence rose between

2

IV

on Decem

ber 31, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.38.1.1 on 1 F

ebruary 1981. Dow

nloaded from

Relative importance of cigarette smoking in occupational lung disease

1955 and 1968. Both criteria show a slight improve-ment since then. These changes have occurreddespite an increasingly wide range of antibacterialagents being available since the late 1930s. Indeed,the failure of the improvement in bronchitis statis-tics to be sustained could be interpreted as coincidingwith the availability of these agents. Chronicbronchitis is more frequent in cigarette smokers andin dwellers in large conurbations suggesting thatcigarette smoke and environmental atmosphericpollution are important. The recent improvementcould be due to the effects of the clean air legislationor to a fall in smoking by older men or to bothcombined.

It is the view of many, expressed in the reportsfrom the Royal College of Physiciansl-3 thatcigarette smoking is the main cause of the veryserious increase in deaths from lung cancer that hasoccurred since 1920. They would attribute the recentimprovement to a tendency in older men to stopsmoking, although it is also possible that cleaning ofurban air has helped.The two diseases of the chest that create the

heaviest burden on our society are bronchitis andlung cancer. Asthma and related diseases are lessimportant. Unless some new initiative is taken cur-rent evidence does not suggest that the situation inrelation to these three diseases will improve. Forboth bronchitis and lung cancer cigarette smoking isa contributory cause, and atmospheric pollution bythe products of fossil fuel combustion may also be afactor. Atmospheric pollution by organic materialsin industry, agriculture, hobbies, and in the home is afactor in asthma and related diseases.

Sufficient evidence is available to recommendmeasures leading to a drastic reduction in cigarettesmoking to relieve society of part of the burden ofbronchitis and lung cancer. Complex arguments,however, have been brought forward that haveblocked effective action so we must consider theimportance of other factors to see whether othermeasures can be recommended that might be aseffective as a reduction in cigarette smoking. Asbefits an Ernestine Henry lecture, "diseases due tooccupation" will be considered to determine whetherthey contribute significantly to the burden of res-piratory disease. A comparison will be made of therelative importance of toxic agents inhaled at workwith the importance of inhaled cigarette smoke.

Detection of occupational factors

The industrial revolution created occupationaldiseases such as potter's asthma and miner's pneumo-coniosis. It also created the social conditions thatencouraged tuberculosis and atmospheric pollution

which caused bronchitis. These two diseases haveprovided the main burden of lung disease, and theirprevalence has obscured the contribution of occupa-tional lung disease to the total burden. The Clean AirActs and the present level of success in controllingtuberculosis have made it easier to recognise andmeasure the occupational contribution. An increas-ing flow of research has been published analysing theeffects of various occupations on health. The re-search has been aimed at identifying occupationalhazards. It has been assumed that these hazardsshould be removed regardless of the relative import-ance of the diseases they caused or the effects on theindustry. Populations have been studied in relationto a particular hazard and compared either with thegeneral population, with similar populations inother occupations, or with sub-populations withinthe same work force whose exposure is different. Inall good studies allowance is made for age, sex, andsmoking habits. In practice this can prevent a truecomparison of the relative importance of the occu-pational hazard with that produced by tobaccosmoke on the one hand or general levels of atmos-pheric pollution on the other. These researchmethods are forced on us because the lungs deterio-rate with age."Aging" is a combined result of infective illnesses

and time-weighted exposure to other non-occupa-tional environmental factors as well as the agingprocess itself. As most of these factors affect thepopulation as a whole the identification of anoccupational factor is easiest to achieve by standard-ising for all the other factors in combination ratherthan by considering each separately.

Studies of occupational disease are seldom com-plete but are carried out in stages. It is usual to startby examining men currently at work for evidence ofa particular disease whose existence has been sug-gested by case reports. This stage is the "cross-sectional survey" and once identified the populationmay be followed up to determine the progress ofexisting cases of disease and to measure the attackrate of new cases. At this stage it becomes a "cohort"study, whose value may be increased by identifyingthe working population at some time in the past andcombining a retrospective with a prospective pro-cedure. The incidence and severity of the diseasemay be identified on medical grounds by question-naires, examinations, and tests during life. Alterna-tively, it may be done by sickness-absence certifica-tion or studying the reasons for premature retiral.Finally, the population may be followed until deathand the effect measured in terms of lost life expecta-tion or an unexpected distribution of causes ofdeath.

Evidence about occupational respiratory disease

I

on Decem

ber 31, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.38.1.1 on 1 F

ebruary 1981. Dow

nloaded from

4

xa 1

c

1G

do0

c:0.

0

Fig 3 Relation between age, dust exposure (as measuredby x-ray category), and lung function in coal miners inBr itain 1955. Dyspnoeic index =

standardised ventilation x 100maximum voluntary ventilation

and was used as a measure ofbreathlessness or disability.It is seen to increase slightly with age in non-miners (N)and in miners with none or simple pneumoconiosis (NMto category 3). These categories are all slightly morebreathless than non-miners in all age groups. A seriousincrease in breathlessness occurs in all age groups withcomplicated disease (categories B and C). Gilson andHugh-Jones4 give reasons for disregarding part of columnabove dotted line joining B to B'.

has been accumulated over many centuries. But,apart from descriptive studies establishing theexistence of these diseases, little quantitative workhas been done until recently. The most extensivelystudied group is coal miners.

Coal workers

A series of cross-sectional studies carried out inBritain before and after the 1939-45 war showedthat coal workers' pneumoconiosis was dust-dose related. Unexpectedly, the frequency andseverity of simple pneumoconiosis was closely relatedto the cumulative dose of total respirable dust ratherthan to the dose of crystalline silica that occurs invarying proportions in the dust. The second unex-pected finding was that while the x-ray shadows ofdust disease remained discrete (simple) there waslittle disability, but as soon as the shadows coalesced(became "complicated") disability and pronounceddisturbance of lung function occurred (fig 3). Com-plicated disease, called progressive massive fibrosis,

Elmes

did not usually appear until a certain profusion ofdiscrete shadows was present. Up to that stage thedisease did not progress without further dustexposure, but once the massive shadows appeared,they and the deterioration in lung function pro-gressed after the men had left the pit. At that time,in the 1950s, the analysis did not show any relationbetween cumulative dust exposure or x-ray changesand the symptoms and physiological changes ofbronchitis.4 The effect of cigarette smoking was notconsidered. These findings led to the present systemof health supervision in the mines by which men areadvised to stop doing dusty jobs when the radio-graphic changes reach category 2. At this stage theyare entitled to compensation in the absence of detect-able disability and are extremely unlikely to progressto disabling disease if they stop dusty work. Inanother cross-sectional study based on the year 1962,Liddell5 showed that miners took more time off inmore spells due to sickness even than men in otherarduous occupations. The time off increased withage and was associated with various diagnoses butmainly respiratory illness and injuries. One couldconclude that the high sickness absence was not dueto dust because the worst affected groups under-ground were not those with the highest exposure.This finding showed a major source of error in cross-sectional surveys due to selection within the popula-tion. The earlier surveys had already shown thatyoung men at the coal face had better respiratoryfunction than average for men of that age becausefitness is a prerequisite for this arduous work, whichis also the most dusty. As time goes by some faceworkers become less fit (more often ill) and moveaway from dusty arduous work taking their highmorbidity with them. This source of error may beavoided by carefully planned longitudinal studiesusing cumulative dust exposure rather than currentexposure category as an index of exposure.A thorough cross-sectional study of symptoms and

respiratory function in miners has been carried outin Germany as part of an industry-wide study on thecauses of bronchitis.6 In this study exposure to dustand other respiratory "noxae" was measured andused to classify workers into grades. Respiratorysymptoms were recorded using a Medical ResearchCouncil-type questionnaire and careful measure-ments of respiratory function carried out aimed atdetecting both airways disease and damage to thealveoli (emphysema). Chest radiographs were takento detect pneumoconiosis as well as unrelated lungdiseases. They were not used as a method of measur-ing dust exposure or its effects. The findings arereported with a standard format for the workers inall the industries. Examples of the figures and tablesare given in fig 4 and table 3. Seventeen hundred

on Decem

ber 31, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.38.1.1 on 1 F

ebruary 1981. Dow

nloaded from

Relative importance of cigarette smoking in occupational lung disease

Non- smoker ---

Intensity of smoking <lOg/d * -..Intensity of smoking >10g/d .-.

Fig 4 Coal mine, active staff. Example of datapresentation in German cross-sectional study of coalminers. In left-hand section prevalence of bronchiticsymptoms in age group 20-35 is shown separately fromnon-smokers, those smoking up to 10 g of tobacco a day,and those smoking over 10 glday. Dot at beginning ofeach line represents prevalence ofsymptoms for menwith a low level of dust exposure (categories 1 and 2).Arrow at end of line represents prevalance in heavilyexposed group (category 3). Numbers by dots and arrowsindicate number of workers in each category. Similarinformation is displayed for 36-55 age group in centralsection. Information for over-55 age group (right handsection) is incomplete because numbers of workers insub-categories were inadequate. Figure shows thatsymptoms of bronchitis in coal miners are more frequentin men with high dust exposure who smoke heavily.Frequency increases with age but the separate effect ofsmoking and dust exposure ceases to be obvious.

Table 3 Unweighted mean excess prevalence ofbronchial reaction in relation to smoking, age, and dust6

Decision function Dust Smoking Age

1-10 g > 10 gtobacco/day tobacco/day

Anamnestic and clinical(ABR) 4-7 34-3 37-7 24-7Function analysisindicators (Fov) 4-3 < I < 1 9 3Combination (Kov) 5-3 12 7 11-3 25Emphysema (FLU) 10-3 3-3 < 1 2

miners were selected at random from the generalpopulation of coal miners in a large number of pits,also nearly 400 pensioners and 820 men with varyingdegrees of "silicosis" (pneumoconiosis).Smoking and age were the most important causes

of the symptoms of bronchitis, although dust oftencontributed to a minor but statistically significantdegree. In pensioners and in men still working afterthe age of 55 the incidence ofsymptoms was frequent,but it was difficult to separate the influence of dustand smoking from that of age, partly because there

_-. Low (< 100gh/ m 3)c---e Medium (100-199gh/m3) dust exposureX x High (>200gh/m3)

4.01

3.5.

_ 30-

U-

25-

20-

W.\No-smokers\

25 30 35 40 45 50 55 60 6525 30 35 40 45 50 55 60 65Age (yr) Age (yr)

Fig 5 Falling fast expiratory volume (FEV1) with agein British miners related to cumulative dust exposure andcigarette smoking. Those with high dust exposure havelower initial FEV1 but rate of decline is same as forthose with low dust exposure. Smokers show a morerapid rate of decline than non-smokers.7

were so few non-smokers. There was little evidencethat smoking or dust led to airways obstruction upto the age of 55. Only in men with severe pneumo-coniosis was there an obvious increase in airwaysobstruction not attributable to age or smokinghabits. The authors showed that coal mining pro-duced a higher incidence of bronchitis than the otherindustries and postulate that this is due to the highdust levels in all mining jobs, which obscures theeffect of dust in the "within industry" comparisons.

In the 1950s a longitudinal study in 20 pits scat-teredover British coalfields was set up by the NationalCoal Board and is now yielding information that can-not be gained from cross-sectional studies. It has con-firmed the close relation between cumulative respir-able dust exposure and the presence and profusionof the shadows of simple pneumoconiosis. Roganet a17 showed that there was a progressive decline inventilatory capacity with age that was greater insmokers than non-smokers (fig 5). Although in eachage group the men in the dustiest jobs had a slightlylower FEV1 than men in less dusty jobs, their rateof decline was the same. The age-related regressioncoefficients were the same for each dust group, whichsuggests that the differences were due to some changein the airways related to current dust exposure,which did not get worse with age. The group studied,however, was a survivor population from which mendeveloping progressive massive fibrosis were alsoexcluded. This study also showed a slight fall in FEVwith increasing profusion of small opacities, but itwas not possible to separate this from the dust-related effect already discussed, which occurredequally in men with and without x-ray changes.This is presumably the dust effect detected in the

5

on Decem

ber 31, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.38.1.1 on 1 F

ebruary 1981. Dow

nloaded from

Elmes

German study where, being non-progressive, it wasnot obvious in the older men.The effect of cigarette smoking on British coal

miners has been studied in more detail by Jacobsen8using the National Coal Board data. The earlierstudy7 had shown that the FEV declined morerapidly in smokers than non-smokers irrespective ofdust exposure (fig 5). In the more detailed study norelation was found between smoking and the size orprofusion of simple pneumoconiosis shadows.Smoking appeared to be the main, if not the only,factor causing respiratory symptoms and disabilityin miners without massive fibrosis.The ultimate test of occupational disease lies in

mortality studies. Permanent damage to the lungscauses a reduction in life expectation and an increasein the proportion of respiratory compared with non-respiratory causes of death. Liddell9 studied deathsrecorded by the Registrar-General occurring in coalminers aged under 65. The standardised mortalityrate for miners was below that of the general popula-tion, lowest in the dustiest jobs, and approximating tothe rest of the population in surface workers. Deathsfrom coal worker's pneumoconiosis occurred moreoften in surface than in coal-face workers, who hadexceptionally low rates for pneumonia, bronchitis,and lung cancer. These anomalous findings were dueto the "fit-worker" effect and difficult to interpretbecause no attempt was made to trace workers whohad left the industry more than tor to extend the study past the re

Cochrane's exemplary study'0 (

Leigh populations, however, inwho had left the mine early a

LO1958r

1;2- T

1-1

ol

o

u 1-0-

0-1

C I

2 0

-1 5

0

- I10 >~2

0-62 _- -.1

0-05 _ - __-_ _ = _

=_ _:

0 10 20 30Age (yr)

Fig 6 Estimated profusion ofsimplfoundry workers related to age. Eaciaverage reading on ILO scale for meat 15. After 40 years' work foundryaverage a score of 0/1 whereas fettleILO scale.12

retired at the normal age. The Rhondda miners had araised standardised mortality ratio of between 115and 120 except those who had moderately advancedor advanced massive fibrosis at the beginning of thestudy. The men with massive fibrosis had a ratio of195. The ratios for bronchitic deaths were high(circa 200) and for lung cancer low (circa 70). In theLeigh population the mortality for miners withoutmassive fibrosis was the same as that for non-miners. The lack of correlation between the level ofdust damage detected by radiographs and the deathsfrom respiratory disease in all miners without severemassive fibrosis in both these populations suggeststhat factors other than dust are dominant. Unfor-tunately Cochrane did not include the effects ofsmoking in these studies.Rook et all' suggest that smoking is the determin-

ing factor for lung cancer in Lancashire coal miners.In analysis of necropsy findings they found carci-noma of the lung in 11-4 , not significantly differentfrom other male workers in the area. Thirteen percent of miners without pneumoconiosis died of lungcancer whereas only 8-4% of those with progressivemassive fibrosis died of cancer, although in thisseries they lived to an average of 72 years, about ayear longer than those with simple pneumoconiosis.

Iron and steel workers

ten years previously Iron and steel foundries used to be dirty, dangeroustiring age. places. The Registrar-General's tables for 195112of the Rhondda and indicated an excess mortality from non-malignanticluded both those (SMR 210) and malignant (SMR 150) respiratoryLnd those who had disease. A cross-sectional survey was carried out on

a 1:40 sample of foundries in Britain in 1964-5.13There was a simple pneumoconiosis that was

exposure related (fig 6). There was more x-raychange in fettlers than in foundry floor workers, andit was enhanced by smoking in the latter. None de-veloped progressive massive fibrosis. Symptoms of

Fettlers bronchitis increased with age and exposure (fig 7).The effect of smoking 25 or more cigarettes a day innon-foundry workers seemed to equal the effect offoundry dust in non-smokers and the combined

Foundy effects seem additive. Surprisingly, although 2000workers were studied the contribution of dust to the

ZAge' only incidence of bronchitic symptoms was not found tobe statistically significant, and there was no evidencethat dust contributed to the fall in lung function

40 50 60 65 usually associated with bronchitis.Fe pneumoconiosis in Lowe et al'4 studied the population of two largeh curve predicts steel works in South Wales in the 1960s and wereen who started work unable to detect any factors other than cigaretteworkers would smoking as causing respiratory disease or loss of?rs average 1/1 on ventilatory capacity. They also found that the levels

of dust and especially sulphur dioxide in these plants

6

on Decem

ber 31, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.38.1.1 on 1 F

ebruary 1981. Dow

nloaded from

Relative importance of cigarette smoking in occupational lung disease

25-

E0

U)>1

20-U)U)U)

-S

un

aI,

-c lo.

0~

fl10

cL

5-

0aI,

01

No foundry work

Foundry work fromage 15 years

30 40 50Age ( yr )

60

Fig 7 Estimated prevalence of bronchitic symptonfoundrIvmen Malated tn age andi cioarettP xmnkina IJ"<"ItLifrfcft ^rcuic u "'V Uf"4u/ gurultv affuivirt. ift1rt

were five categories ofsmoker (NS = non-smoker,Ex = ex-smoker, I = 1-14 cigs a day, l = 15-24, andIII = 25 or over.) By retiral age smoking produces morethan a two-fold increase in bronchitic symptoms in bothfoundry men and non-foundry men. Foundry work hassame effect on non-smokers as category 3 cigarettesmoking has on non-foundry workers. Combined effect isabout a five-fold increase.12

were lower than those prevailing in the streets ofcentral London at the time.The German cross-sectional study6 of about 6500

workers in steel works and foundries included someheavy engineering workers. Both smoking and agewere considerably more important than dust incausing bronchitic symptoms in all the groups.Smoking was more important than age in some worksand less important in others. Smoking and age con-tributed to the level of airways obstruction andemphysema in some of the older men, dust did not.

It is difficult to reconcile the results of these cross-sectional studies with the mortality findings listed bythe Registrar-General'2 already quoted. It is truethat the mortality relates to conditions that mayhave ceased to exist by the time the surveys werecarried out. But the figures for 1971 quoted in theofficial report for 1970-2 lists an overall mortalityratio of 122 with 155 for lung cancer and 167 for non-malignant respiratory disease.15 Grundy16 in a

lII partial follow-up of the South Wales steel workersconfirmed that the high death rate for lung cancer isstill present and is not accounted for by smokinghabits. More cohort studies are needed.

Slate workers

Ex A study of slate workers in North Wales has finallyshown that they develop dust-related disease.17Previous studies had failed, perhaps because of the

NS high incidence of pulmonary tuberculosis and be-Inl cause measurements of dust exposure and respiratory

function were not carried out. Even for this recentstudy, accurate dust measurements were not avail-able, and the workers had to be classified into job

,Ex groups according to the relative level of past andpresent exposure. The level of simple pneumoconio-

NS sis was found to be related to the duration andseverity ofexposure.

Figure 8 shows the relation between x-ray changeand breathlessness. The prevalence of breathlessnessincreases most rapidly in those with pronouncedx-ray changes (final category 3). The difference

70 between smokers and non-smokers is greatest wherethe dust has had least effect (final category 0). The

ns in importance of smoking decreases with increasingThprp dust effect. Figure 9 shows the relation between x-ray

70

60

50-

40-o/.

30

20

10

0

3SmokersNon-smokers .

10 20 30 40Age (yr)

50 60 70

Fig 8 Prevalence of dyspnoea in slate workers.Prevalence ofMRC questionnaire breathlessness (grade 2)is plotted against age. Dust effect is shown as x-raychange on ILO classification finally achieved at age 65.Increase is greatest in those who finally show most dustdisease. Smoking increases prevalence ofsymptoms mostin those with least severe x-ray changes.'7

7

1

on Decem

ber 31, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.38.1.1 on 1 F

ebruary 1981. Dow

nloaded from

8

5-

4.

w

uL 3.

2-

Non-smokersSmokers

10 20 30 40 50 60 70Age (yr)

Fig 9 Decline in FEV1 in smoking and non-smokingslate workers. Rate of decline has been shown

separately for smokers and non-smokers according tox-ray category reached at age 65.17

changes, age, and an objective measurement ofrespiratory disability. The fast expiratory volumedeclines more rapidly in those with the most x-ray

change. The steepest decline is seen in non-smokerswho developed category 3 disease. The decline insmokers of the same x-ray category was not as steep.This implies that the effect of dust is greater than theeffect of smoking or that their combined effect is lessthan additive due to interference between the twofactors. Without measurements of dust levels it is notpossible to equate the airways effects of any level ofdust exposure with any level of cigarette smoking.The pattern of disease in slate workers differs from

that in coal workers in two respects. Firstly, there isan obvious relation between dust exposure andmeasured disability before massive fibrosis occurs.Secondly, the simple pneumoconiosis progresses tomassive fibrosis long after exposure ceases. Thefollow-up study of the effect of this dust on lifeexpectation and the causes of death has not beencarried out.

Asbestos workers

Asbestos releases a dust that consists of elongatedparticles (fibres) that are more readily retained in thelung than the dusts considered up to now. Dustlevels that are not perceptible to the naked eye areassociated with an increased risk of pulmonary

Elmes

disease. The danger of lung fibrosis was well docu-mented by HM's Factory Inspectorate in 1930 whenMerewether and Price18 showed that fatalitiescould occur within 15 years of first exposure. Regu-lations introduced in 1933 reduced this risk con-siderably, and premature deaths due to uncompli-cated asbestosis are now rare. Asbestos dust is muchmore difficult to measure than dusts composed ofisometric particles, and measurements by weight ornumber of fibres may be inappropriate without de-tailed size distribution measurements including bothlengths and diameters. Because exposure measure-ments have been inadequate, cross-sectional andlongitudinal studies of asbsetos workers have so farfailed to produce reliable information on the cumula-tive dose needed to cause detectable, certifiable, orlife-threatening asbestos-related disease. Theshadows on radiographs are unlike those in otherpneumoconioses in that they are linear and notrounded, affect the lower zones first, and are associ-ated with shrinkage of the lung tissue. There may beboth diffuse and discrete thickening of the pleura.

While the 1933 regulations reduced the incidenceand severity of asbestosis in those sections of theindustry to which they were applied, they also leadto the detection of the second and most importantfatal asbestos-related disease-primary lung cancer.By 1960 over half of the men certified as havingasbestosis by the pneumoconiosis panels were dyingwith lung cancer.19 Surveys of working populationshad not detected this hazard because men seldomremain at work for long with detectable lung cancer.It was because the pneumoconiosis compensationscheme included necropsy examination before thepayment of widows' benefit that this association wasdocumented. The much rarer cancer, mesotheliomaof the pleura or peritoneum, was first associated withasbestos exposure around the crocidolite mines ofthe North-western Cape Province of South Africa.20

McDonald2' has summarised the epidemiologicalevidence concerning asbestos-induced mesothelioma.Of the three types of asbestos in commercial use,crocidolite is most likely to cause mesotheliomawhen mined or when being used in industry. But inthe United States amosite may be as important,22although it has caused few mesotheliomas where it ismined in South Africa. The incidence of mesotheli-oma in an industrial population has been shown tobe dose related,23 but the long latent period and lackof dust measurements over the relevant period hasmade it impossible to put the relationship on aquantitative basis.

In comparing the asbestos hazards with otherenvironmental factors, such as cigarette smoking, thelack of information on past exposure is only part ofthe difficulty. There seems to be an interaction be-

on Decem

ber 31, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.38.1.1 on 1 F

ebruary 1981. Dow

nloaded from

Relative importance of cigarette smokinig in occupational lung disease

Table 4 Men working in asbestos textile factory inEngland

No Percentages with signsofmnen Crackles Possible Certified Smnall

asbestosis asbestosis opacities*

Non,light smokers 55 0 0 0 4Medium/heavysmokers 116 19 9 7 19Ex-smokers 24 17 13 8 17

>1 /O.

tween smoking and dust in the cause of asbestosisand lung cancer but not mesothelioma. The evidenceconcerning this has built up gradually, starting withthe cross-sectional surveys. These surveys haveshown that increasing dust exposure is associatedwith x-ray changes and reduction in vital capacity,and FEV1, and basal rales. A reduction in gas transferwithout evidence of airways obstruction is typical inworkers with relatively advanced disease under theage of 50. But in older men the symptoms of bron-chitis and evidence of airways obstruction are alsopresent, and it is difficult to separate these effectsfrom the effects of smoking. The German study6included about 500 asbestos workers. Bronchiticsymptoms were seen in smokers with high dustexposure but not in non-smokers. It is the onlygroup of workers in this large study where no rela-tion between smoking alone and bronchitic symp-toms was shown. This study, like many others, alsofailed to show evidence of a risk of cancer.The more detailed studies of two populations in

Britain have been analysed to investigate theoriginal suggestion made by Weiss24 that smokingincreased lung fibrosis in asbestos workers.25 Bothin an asbestos textile mill and in the dockyard workersin Plymouth the cross-sectional studies showed thatthe prevalence ofearly signs of asbestosis (crepitationsand small opacities) was higher in smokers than non-smokers allowing for age and exposure. Advanced(certifiable) asbestosis was also more common insmokers. Table 4 shows the analysis for the textileworkers.

Longitudinal studies show that established asbes-tosis may progress, slowly, after exposure has ceasedor under the reduced dust conditions laid down bypresent legislation. Analyses of this progression inrelation to smoking have not been published, but thecohort studies do show a significant fall out ofdeaths from asbestosis, cancer, and mesothelioma.As the cohorts are followed the populations' age andthe pattern of deaths change to include an increasingproportion of deaths unrelated to exposure. The factthat the premature, occupationally related deathstend to be grouped into the first half of the totalmortality of a cohort has led to an exaggeration of

10-

8- i

E~~~~~~~~~J4-a

1940-4 1945-9 1950-4 1955-9 1960-4 1965-9 1970-4

Mesothelioma m Expected respiratory

Bronchus or Observed gastrointestinal Dmesothelioma L

Bronchus Expected total for gastrointestinalcancers is less than 01 in each *five-year period

Fig 10 Cancer deaths in cohort of Belfast insulationworkers. If half of "bronchus or mesothelioma" deathsal-e included ratio of observed to expected lung cancerdeaths in 1955-9 is 20:1 and in 1970-4 less than 3:1.

the importance of hazards. The well-studied cohortsof asbestos workers have been analysed by the mostsophisticated statistical techniques with standardis-ing for age, sex, and social class, and the results areexpressed first as standardised mortality rates andthen as ratios comparing the exposed groups withthe general population. By this sort of method in-sulation workers show a very high risk.26 27

In the early years of follow-up the Belfast cohorthad a 20-fold increase in risk of lung cancer com-pared with the general population in Belfast (fig 10).As 10°o of men in Belfast die eventually of lungcancer the 20-fold increase for working with asbestoscannot be maintained, and indeed the long-termfollow-up confirms an apparent reduction in risk.In the larger American study of insulation workersthere appeared to have been risk ratios as high as87-fold (heavy smokers) or 36-fold (ex-smokers).

Saracci28 has pointed out that all the populationsso far studied are lacking in two important respects:(1) the non-smokers exposed to asbestos are toosmall in number; and (2) there were no non-smokingnon-exposed groups within the studies. Controlswere usually derived from the general populationstatistics including smokers.

In most of the earlier reports no deaths from lungcancer in non-smokers were recorded, and this wasinterpreted as indicating that asbestos did not in-crease the risk of cancer in non-smokers. Lungcancers, however, have now been recorded in non-smoking women who worked with asbestos inBritain. The figures available have been reanalysed25(table 5) to determine the relation among smoking,asbestos, and lung cancer. There is an obviousexcess of lung cancer in the smoking groups in themale insulation and factory workers and in thefemale factory workers. The number of cancersobserved correlated closely to that calculated using

9

on Decem

ber 31, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.38.1.1 on 1 F

ebruary 1981. Dow

nloaded from

Elmes

Table 5 Distribution of lung cancers in asbestosworkers by smoking habits

Observed Expected Expected deathsdeaths non-occupational (including occupational)

deathsAdditive Multiplicativemodel model

Cigarettesmokers 24 3 0 18-9 22-6Others 0 0-2 5-1 1 4MenSmokers 32 12 0 31-4 33-1Non-smokers 0 0.0 1-2 0-1WomenSmokers 18 1-9 14 6 17-8Non-smokers 2 0-2 5-4 2-2

Table 6 Insulation workers lung cancer29

Group Exposure History Death Mortality Mortalityto cigarette rate difference ratioasbestos ? smoking?

Control No No 11-3 0.0 1-00Asbestos workers Yes No 58-4 +47-1 5-17Control No Yes 122-6 +111 3 10-85Asbestos workers Yes Yes 601 6 +590 3 53-24

a multiplicative model. But the lack of adequatenumbers in the non-smoking groups means that thisconclusion can only be tentative.As time passes and the cohorts grow older the

situation becomes clearer. The latest analysis of theAmerican insulation workers29 compares them witha carefully selected control group from the generalpopulation (table 6). The risk of cigarettes aloneappears to be a 10-fold increase in lung cancer. Forasbestos in non-smokers the risk is about five-fold,and the difference in risk between no asbestos ex-posure and asbestos exposure in smokers is also five-fold. The combined risk is almost exactly that pre-dicted for arithmetic model (53-fold). This risk ratiocannot be used to predict the ultimate proportion ofdeaths from smoking or asbestos or both in thispopulation, but it seems likely that smoking is twiceas important as asbestos for lung cancer. Deathsfrom mesothelioma (not attributable to smoking)were estimated as being between a quarter and halfof those from lung cancer and were about the samein number as those attributed to asbestosis (to whichsmoking does contribute). On the aggregate itappears that in this relatively heavily exposed groupcigarette smoking is rather more important in caus-ing deaths from respiratory disease than asbestosdust.Other methods of comparing mortality experience

may be applied to this type of cohort material.27When analysed in the usual way an excess ofobserved over expected respiratory cancers was seenin each successive five-year period. The cohort con-

160

140 -

120 -

100-c

E 80-0

0z 60-

40 -

20 -

Expected *-

Observed o---o

1939 1344 1949 1954 1959 1964 1969 1974

100

80

60

40

20

Fig 11 Survival of 162 Belfast insulation workerscompared with men of same age and social class oversame period in N Ireland. From 1949 to 1969 asignificantly greater number died due to all causes thanwas expected in each five-year period.

tained men of all ages when it was first identified in1940, and the older men tended to die first. Never-theless, the mean age of the cohort rose and the ratioof observed over-expected for primary lung cancerfell from 20:1 to less than 3:1 between 1950-4 and1970-4. The mean age at death rose from 55 to 66in the same period. If the experience is expressed as asurvival curve (fig 11) the mortality for all causesmay be compared with the rest of the population.Death rates are higher (the slope of the curve issteeper) than for the expected until the last five yearswhen they run parallel. In this cohort it was not

100-

80

60CA

(5

20-

0 Non. Ex,and Pipe o .o0-14 Cigarettes ..15-24 Cigarettes A----->25 Cigarettes o-.-o

0J -

40 50Age (yr)

Fig 12 Survival of Belfast insulation workers dividedinto smoking categories compared with all N Irelandmen (smokers and non-smokers) ofsame age and socialclass,

10

on Decem

ber 31, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.38.1.1 on 1 F

ebruary 1981. Dow

nloaded from

Relative importance of cigarette smoking in occupational lung disease

possible to show a relation to cumulative dust ex-posure because no measurements were made. Dura-tion of exposure appeared to have no influence, inthat the mortality was as high in the men with shortexposure as it was in those with long. Therefore onecan only compare the effect of different levels ofsmoking with a single level of asbestos exposure(fig 12).

Bearing in mind the small total numbers in thiscohort and that smoking history was not availablefor all of them there is a remarkably clear separationof the risk by smoking habit. The 5000 death pointfor the control population is 72 years and for all theinsulators is 61, indicating an average loss of lifeexpectation of 11 years. For the heaviest smokersthe 500% death point is 53 (loss of 19 years), themoderate smokers 61 (loss 11 years), light smokers 62(loss of 10 years), and the non-, ex-, and pipesmokers 66. Taking the entire curve instead of the50% death point the experience of the non-, ex-, andpipe smokers appears to approximate to the controlpopulation, indicating that asbestos by itself carriesabout the same risk to life expectancy as smoking inthe population of Belfast.No analysis of the effect of different levels of

cigarette smoking on life expectation in Britain or NIreland has been published in this form. The effect ofasbestos alone (or cigarette smoking alone) can becalculated only indirectly. The multiplicative modelthat fits the figures for lung cancer in Americaninsulation workers may be applied to this smallcohort. Assuming the mean effect of smoking in thegeneral population is the same as the effect of asbes-tos in non-smoking insulation workers then the lossof life expectation for each alone is three to fouryears.* This would mean that the life expectationfor non-smoking men in Belfast is about 75 years.Without the results of a direct analysis of the effectof cigarette smoking alone it is manifestly difficult tocalculate the relative importance of other hazards,such as asbestos, on life expectation. Hammond atthe World Conference on Smoking and Health of theAmerican Cancer Society in New York in 1968calculated that for American men cigarette smokingalone reduced the life expectation ofyoung Americanmen by four and a half to six years depending on thenumber of cigarettes smoked. Doll and Peto30 havecalculated that British doctors lose five years of lifeby smoking 20 cigarettes a day.

Prevention

The dusty industries considered (coal, steel, and

*If x = loss of life expectation due to smoking alone andasbestos exposure alone then the combined risk = 11 + x =x2.

asbestos) could make a major contribution to res-piratory disease in the community. There may bemany smaller industries that cause respiratory illnesseither due to dust (as in the slate workers) or to achemical as in the polyurethane plastics industry.Many of these have not been fully investigated, butthese hazards of disease will probably be avoided bycarefully controlling dust or fumes or by substitutingsafer materials. This will undoubtedly lead to thedisappearance of certain traditional products, suchas the roofing slate, because substitutes are cheaperand safer. As a result of removing many small con-tributions to society's burden of respiratory diseasethere may be a noticeable reduction in the total.

But in considering industries like coal and steeland to a lesser extent asbestos the importance ofthese industries in our economy must be borne inmind. The decision to increase the preventivemeasures to protect the health of these large num-bers of workers must be weighed against the possi-bility that the product will be made too expensive,thereby causing employment to fall both in the in-dustry itself and in the dependent industries. In thecase of coal it seems unlikely that we will be able tosubstitute other sources of energy in the foreseeablefuture and therefore the cost of more efficientworker protection must be weighed against their gainin health. The question that must be answered is,"Is the cost-benefit relation such that this method ofimproving the community's respiratory healthshould be chosen in preference to the spending of anequivalent amount in reducing general atmosphericpollution or cigarette smoking?" In the iron andsteel industry it seems that the environment hasalready been improved to the point at which nosignificant amount of chronic respiratory disease isproduced. But research is needed to identify andeliminate the cause of the lung cancer in this in-dustry.

Asbestos, although widely used, is not so essentialas coal or steel. Some countries are attempting to dowithout it altogether. But this policy is open to twoserious questions that have not yet been answeredby scientific study. Firstly, are the substitutes aseffective? One does not know, for instance, whethercement products reinforced with asbestos substituteswill be as efficient and will not fail structurally aftertime, stress, and weathering. Nor is it yet knownwhether brake linings will be as efficient or fire pro-tection as effective. Secondly, the substitutes are notfree from respiratory hazard, although work in handindicates that up to the present time the man-mademineral fibre and ceramic fibre industry has notcreated as high a health risk as the asbestos industry.Those making policy decisions on how to improve

the health of the community have to consider these

11

on Decem

ber 31, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.38.1.1 on 1 F

ebruary 1981. Dow

nloaded from

Elmes

various factors. It is the responsibility of the medicalprofession to give accurate information on the rela-tive importance of the various factors affectinghealth.

Summary and conclusions

Since the turn of the century the burden of respira-tory disease on our community has changed. Someof the benefits have been due to improved socialconditions and some to more effective medical care.There remains a serious level of disease and prema-ture death due to chronic bronchitis and asthmathat is unlikely to respond further to social changeor to treatments available at present. The incidence oflung cancer has increased alarmingly. These threediseases, bronchitis, asthma, and lung cancer are atleast in part due to materials inhaled into the lungsand to this extent are preventable.Much has been done as a result of the Clean Air

Act to improve the general level of atmosphericpollution, and this may be the reason for the recentimprovement in morbidity and mortality for bron-chitis. But the other two important sources of pollu-tion of inhaled air (the work place and tobaccosmoking) could be reduced further. In this age offinancial stringency cost effectiveness is often theultimate determinant as to which factor we shouldtackle first.

I have reviewed the evidence of the relative im-portance of smoking and dust in coal miners, slatequarry workers, iron and steel workers, and thoseworking with asbestos. The evidence is scanty and inmany instances unreliable. But it indicates thatalthough dust may cause the symptoms of bronchitisin young men, these symptoms may persist but areunlikely to cause serious airways obstruction orpremature aging of the lung (emphysema). In mostcases smoking is the main cause of bronchiticsymptoms in all age groups and may lead to airwaysobstruction and emphysema. Age by itself, rep-resenting the general environment, is less importantthan smoking but is usually more important thandust in causing disability. Heavy exposure to dustsmay lead eventually to scarring of the lung and pro-gressive disability. In the industries where a scienti-fically based effort has been made to control the dustexposure these serious consequences are being pre-vented. This appears to be true for the coal miners inBritain and may also be true for the lung fibrosis inasbestos workers. But whereas coal miners may nowexpect to live as long as other workers we are not ina position to be certain about steel workers orasbestos workers. Steel workers have an excess riskof lung cancer, the cause of which is not known orfully investigated. It is not as great as that due to

cigarette smoking.Lung cancer and mesothelioma of the pleura and

peritoneum can result from exposure to asbestos.There is a long interval between critical exposure andthe development of this tumour so that preventiontakes a long time to show results. Measurements ofpast exposure in this industry have not been verymeaningful so that the comparisons between therisks due to asbestos and the risks due to smokingare difficult. The situation is complicated by asynergism between tobacco smoke and asbestos incausing both lung cancer and disabling lung fibrosis(smoking does not contribute to mesothelioma). Butit appears that in some of the groups who sufferedfrom heaviest exposure before the current regulationsthe overall risk from asbestos was approximatelyequal to the overall risk from smoking.Only in the archaic conditions of the slate industry

did dust exposure exceed cigarette smoking inimportance as a cause of disabling disease. In all theother situations reviewed cigarette smoking was thedominant cause of symptoms, of disability and ofpremature death from respiratory disease. Bearingin mind that smoking also causes serious cardio-vascular disease there is no question that even inthese dusty industries stopping smoking would havea far greater effect on the burden of disease than thecomplete suppression of all the dusts and fumes.

References

Royal College of Physicians. Smoking and health. London:Pitman Medical, 1962.

2 Royal College of Physicians. Smoking and health now.London: Pitman Medical, 1971.

3Royal College of Physicians. Smoking or health. TunbridgeWells: Pitman Medical, 1977.

4Gilson JC, Hugh-Jones P. Lung function in coalworkers'pneumoconiosis. London: H MSO, 1955. (MedicalResearch Council Special Report Series No 290.)

Liddell FDK. Morbidity of British coal miners in 1961-2.Br J Ind Med 1973;30:1-14.

6 Deutsche Forschungsgemeinschaft. Research report: chronicbronchitis and occupational dust exposure. Boppard:Boldt, 1978.

7Rogan JM, Attfield MD, Jacobsen M, Rae S, Walker DD,Walton WH. Role of dust in the working environmentin development of chronic bronchitis in British coalminers. Br JInd Med 1973;30:217-26.

8Jacobsen M. Dust exposure, lung diseases, and coalminers'mortality. Edinburgh: University of Edinburgh, 1976.(PhD thesis.)

9Liddell FDK. Mortality of British coal miners in 1961.Br J Ind Med 1973 ;30:15-24.

10 Cochrane AL, Haley TJL, Moore F, Hole D. The mor-tality of men in the Rhondda Fach, 1950-70. Br J ndMed 1979;36:15-22.

Rooke GB, Ward FG, Dempsy AN, Dowler JB, WhitakerCJ. Carcinoma of the lung in Lancashire coalminers.Thorax 1979 ;34:229-33.

12

on Decem

ber 31, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.38.1.1 on 1 F

ebruary 1981. Dow

nloaded from

Relative importance of cigarette smoking in occupational lung disease

12 Registrar General. Decennial slupplement England andWales 1951. Occupational mortality. London: HMSO,1958.

13 Lloyd Davies TA. Respiratory disease in foundrymen.Report ofa survey. London: HMSO, 1971.

14 Lowe CR, Campbell H, Khosla T. Bronchitis in twointegrated steel works. III Respiratory symptoms andventilatory capacity related to atmospheric pollution.Br J Ind Med 1970;27:121-9.

15 Registrar General. Decennial supplement for England andWales, 1970-2. Occupational mortality. London: HMSO,1978. (Series DS No 1.)

16 Grundy PF. Mortality amongst 5000 steelworkers in rela-tion to smoking obesity lung function and respiratorysymptoms. Cardiff: University of Cardiff, 1979. (MDthesis.)

17 Glover JR, Bevan C, Cotes JE, et al. Effects of exposure toslate dust in N Wales. Br J Ind Med 1980;37:152-62.

18 Merewether ERA, Price CW. Report on the effects ofasbestos duist on the lungs and dust suppression in theasbestos industry. London: HMSO, 1930.

19 Buchanan WD. Asbestosis and primary intrathoracicneoplasms. Ann NYAcad Sci 1965;132:507-18.

20 Wagner JC, Sleggs CA, Marchand P. Diffuse pleuralmesothelioma and asbestos exposure in the North-western Cape Province. Br JInd Med 1960;17:260-71.

21 McDonald JC. Epidemiology of asbestos related diseases.Biological effects of mineral fibres. Lyon: InternationalAgency for Research in Cancer. (In press.)

Vancouver styleAll manuscripts submitted to the Br J Ind Medshould conform to the uniform requirements formanuscripts submitted to biomedical journals(known as the Vancouver style).The Br J Ind Med, together with many other

international biomedical journals, has agreed toaccept articles prepared in accordance with theVancouver style. The style (described in full inBr Med.J 24 February 1979, p 532) is intended tostandardise requirements for authors.

References should be numbered consecutivelyin the order in which they are first mentioned inthe text by Arabic numerals above the line oneach occasion the reference is cited (Manson'confirmed other reports2-5 . . . ). In future refer-ences to papers submitted to the Br J Ind Medshould include: the names of all authors if there

22 McDonald AD. Mineral content of lung in mesothelialtumours. Preliminary report. Biological effects ofmineral fibres. Lyon: InternationalAgency for Research inCancer. (In press.)

23 Newhouse ML, Berry G. Predictions of mortality frommesothelial tumours in asbestos factory workers. Br JInd Med 1976;33:147-51.

24 Weiss W. Cigarette smoking, asbestos and pulmonaryfibrosis. Am Rev Respir Dis 1971 ;104:223-7.

2S Rossiter CE, Berry G. The interaction of asbestos exposureand smoking on respiratory health. Bull Eur PhysiopatholRespir 1978;14:197-204.

26 Selikoff IJ, Hammond EC, Seidman H. Cancer risk ofinsulation workers in the United States. In: Bogovski P,Gilson JC, Timbrell V, Wagner JC, eds. Biological effectsof asbestos. Lyon: International Agency for Research inCancer, 1973:209-16.

27 Elmes PC, Simpson MJC. Insulation workers in Belfast. Afurther study of mortality due to asbestos exposure (1940-75). Br J Ind Med 1977;34:174-80.

28 Saracci R. Asbestos and lung cancer: an analysis of theepidemiological evidence on the asbestos-smoking inter-action. Int J Cancer 1977 ;20:323-3 1.

29 Hammond EC, Selikoff IJ, Seidman H. Asbestos exposure,cigarette smoking and death rates. Ann NY Acad Sci.(In press.)

30 Doll R, Peto R. Mortality in relation to smoking: 20 years'observations on male British doctors. Br Med J 1976ii:1525-36.

are six or less or, if there are more, the first threefollowed by et al; the title of journal articles orbook chapters; the titles of journals abbreviatedaccording to the style of Index Medicus; and thefirst and final page numbers of the article orchapter.Examples ofcommon forms of references are:

International Steering Committee of Medical Editors.Uniform requirements for manuscripts submitted tobiomedical journals. Br MedJ 1979;1 :532-5.

2 Soter NA, Wasserman SI, Austen KF. Cold urticaria:release into the circulation of histamine and eosino-phil chemotactic factor of anaphylaxis during coldchallenge. N EnglJ Med 1976;294:687-90.

3Weinstein L, Swartz MN. Pathogenic properties ofinvading micro-organisms. In: Sodeman WA Jr,Sodeman WA. eds. Pathologic physiology: mechanisnmsof disease. Philadelphia: W B Saunders, 1974:457-72.

13

on Decem

ber 31, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Br J Ind M

ed: first published as 10.1136/oem.38.1.1 on 1 F

ebruary 1981. Dow

nloaded from