Document r6kRpbYpzbMVJ23KXv9n2ZEd7

I REPORT :;c. T>t/*2/4 UK 612.2 : 678.7^3 AN EPIDEMIOLOGICAL STUDY OF THE LUNG FUNCTION OF '-'CKXERS AT A FACTCRT MANUFACTURING polyvinylchloride fey HH Lloyd S Gauld L Copland CA Scutar JANUARY 1932 1 URL 15054 INSTITUTE OF OCCUPATIONAL MEDICINE AN EPIDEMIOLOGICAL STUDY OF THE LUNG FUNCTION OF WORKERS AT A FACTORY MANUFACTURING POLYVINYLCHLORIDE by MH Lloyd, S Gauld, L Copland, CA Soutar ^ 15055 Medical Branch Institute of Occupational Medicine, Roxburgh Place, EDINBURGH EE8 9SU. (Tel. 031 667 513D JANUARY 1982 (ii) CONTENTS Page No. ABSTRACT tir) 1. INTRODUCTION.................................................................. 1.1 Respiratory effects of VCM and PVC 1.1.1 Animal studies 1.1.2 Clinical studies . 1.1.3 Epidemiological studies 1.2 The present study 1.2.1 The PVC factory 1.2.2 The foundry 1 1 2 2 3 3 4 5 URL 15056 2. METHODS ........... 2.1 Selection of the surrey population ... 2.2 The medical surrey ....... 2.3 Methods of analysis 2.3.1 Classification of smoking histories . 2.3*2 Analysis of data from total surrey population ....... 2.3.3 Selection of cases and controls ... 2.3.4 Case-control study ...... 6 6 6 7 7 7 8 8 3. RESULTS................................................................................................... 3*1 Description of total survey population 3*2 Analysis of lung function data from total Surrey population 3.2.1 Calculation of predicted raiues and standardised residuals ... 3.2.2 3.2.3 Distribution of standardised residual raiues of T-co and FEV^ for total surrey population ...... Selection of eases and controls . 3*3 General features of cases, controls and the total surrey population ...... 3*3*1 Age, height and weight .... 3.3.2 Employment status and smoking habit 3*3*3 Lung function . 10 10 10 10 10 10 11 11 11 11 (iii) Pag* No. 3*^ 3*5 ^Clinical description of Tcco cases and FEVt cases . . ....... Case-control study 3.5*1 Occupational histories of case3 and controls 3*5.2 Smoking habits of cases and controls . 12 12 12 14 *. DISCUSSION................................................................................. 16 5*CONCLUSIONS AND RECOMMENDATIGNS.........................................................19 ACKNOWLEDGEMENTS RETESTCES TABLES FIGURES APPENDICES URL 15057 4 Report No TM/82/4 Civ) INSTITUTE OF OCCUPATIONAL M5DICINS AN EPIDZttlOLCGICAL STUDY OF THE LUNG FUNCTION OF WORKERS AT A FACTORY MANUFACTURING POLYVINYLCHLORIDE by HH Lloyd, S Gauld, L Copland, CA Soutar 8Qt Ictfl ABSTRACT V have conducted an epidemiological study to investigate a report that some men working in a factory manufacturing polyvinylchloride (PVC) had abnormally low values of one 1'mg function test, namely the single breath transfer factor for carbon monoxide (Tueo). We studied 265 present and past employees of the PVC factory, and 219 sen from the workforce of a nearby foundry. The fouadrymen were included to increase the number of men with little or no exposure to possible respiratory hazards at the PVC factory. Each man's Tlco was measured, and smoking history and detailed occupational history obtained. More detailed lung function tests were carried out on the PVC workers only (forced expiratory flow-volume curves). The study consisted of twe parts; measurement of the Tlco of current and past employees at the factory and current foundry employees and examination of the distribution of residuals after allowing for age, height, weight and smoking habit, to search for evidence of an excess of men with low values of Tlco; and a case-control study. In this, men with the lowest Tlco after allowing for age, height, weight and smoking habit were selected as eases, and their occupat ional history compared with that of two groups of controls; one group selected from those with near average gas transfer values, and another group from those with the highest values. The Tlco results among this combined population conformed to a Normal distribution after allowing for age, height, weight and smoking habit, and did not suggest that a substantial number of men had clinically important impairment of their Tlco. The 31 men selected as 'eases' included 15 men whose Tlco was abnormally low (less than 7C& of published predicted values), of whom ii5&cwere PVC workers and were foundryaen. Among the 19 cases working at the PVC factory the pattern of lung function abnormality suggested that the low Tlco was <v) accompanied by evidence of airflow obstruction in 11 men, suggesting the presence of emphysema, hut was an isolated functional abnormality in the other eight, possibly suggesting the presence of pulmonary fibrosis or abnormalities of lung perfusion* PVC workers ware slightly over-represented among the cases, although this difference could easily have arisen by chance* Men who had worked on jobs at the PVC factory which were close to the pressure vessels where exposure to vinyl chloride monomer was most likely to have occurred in the past, wen also slightly over-represented among the cases, and this difference approached statistical significance at conventional levels. There was, furthermore, a statistically significant association between low Tl-co and a history of having worked at the PVC factory before 1975* This date is important because it was at this time that exposure of the workforce to vinyl chlcride monomer was drastically reduced, in the light of information about some of the potential hazards of this gas*. The cases included slightly trore smokers than the controls, and those cases who smoked, smoked slightly more heavily than controls* The relative importance of these end other factors in causing the lung functional abnormalities is not'clear, but the3e findings give some support to the hypothesis that work in tho PVC factory before 1975 involved exposure to a substance that caused impairment of lung function in a small number of men. There was no evidence of a relationship between low Tlco values and current working conditions at the PVC factory or with past or present conditions at the foundry* The results suggest that more detailed studies of the relation ships with vinyl chloride monomer, polyvinylchloride, smoking habit and other causative factors would be desirable, together with a clinical assessment of the lung damage in selected men* URL 15059 1. INTRODUCTION The health of workers employed in the manufacture of polyvirrylch 1 a ride (PVG) has been a eause of concern during resent years. Some adverse effects of exposure to vinyl chloride monomer (VCH; are well recognized, but lung disease in sen has net been shewn, to be related to exposure to VCM. Exposure to airborne ?VC dust, however, has been shown to cause slight, impairment of lung function and slight chest radiographic abnormalities* Vinyl chloride is the basic raw material used in the production of PVC. Although the manufacturing process was first developed in 1923 in the USA it was not until 50 years later that evidence of the toxic effects of vinyl chloride exposure began to accumulate. CCRDII23 vis the first to describe scro-osteolysis in reactor cleaners in Belgium and in the following year SUCItJ (1967) described Raynaud's phenomenon among workers in a PVC factory* Changes in the skin and internal organs were subsequently recognised nd in I97tv the term 'vinyl chloride disease1 was used to describe the 3yrdrcme which has cany features in cocr.cn with scleroderma (VELTMAN -t al., 1975). URL 15060 The hepatic effects of exposure to VCM have aroused the most interest and include ncn-cal*.scant changes. It was the discovery cf the association between vinyl chloride exposure and angiosarcoma of the liver (CREECH and JOHNSON, 197*) which caused the greatest concern and affirmed the need for stricter control of exposure levels. In 175 regulations were introduced which, established a ceiling exposure value of 30 parts per mill'ica (ppm) and an eight-hour tiee weighted average (TVA) of 10 ppm. Iu 198O the permitted maximum exposure was reduced to an, annual average of 3 ppm. 1.1 Respiratory effects of VCM and PVC The work described ir. this report was confined to a study of the lung function of workers in a factory producing PVC where sen were exposed potentially to VCM, PVC and the other compounds used in the manufacturing process. Previous studies involving the use of animals, observation on patients and epidemiological surveys, have provided information on the potential respiratory effects of these coe.co'mds. 1.1.1 Animal studies The toxicity of both VC and PYC has been investigated in aaisal experiments. PSODAN et al. (1975) exposed guinea pigs to 1Cg6 vinyl chloride concentration for two hours daily for periods of up to three months. At autopsy the lungs of all exposed animals showed the presence of fibrosis yhich was most severe in the animals which had had the longest exposure time. ACARYAL et al. (1978) studied the effect of intratracheal injections of PYC dust in rats and found that inflammatory changes were induced in the terminal and respiratory alveoli and that granulomatous lesions associated with multinucleated giant cells developed. Similar giant cell3 and histiocytes were identified in the alveoli of the lungs of small mammals which were kept in the bagging area of a PVC factory (FBOKCIA at al.. 197*0* 1.1.2 Clinical studies There have been isolated case reports of pulmonary disease in workers employed in factoH.es manufacturing PVC. ABT1AIQ et al. (1Q7$) studied a 53-year-old man who had worked for 25 years in the bagging plant of a PVC factory. His chest X-ray showed ctieroncdular shadowing and lung biopsy revealed lArge numbers of macrophages containing particulate material and moderate diffuse fibrosis and a few alveolar granulomata. However, his symptoms were few and apart from a alight reduction in. vital capacity, his lung function tests including his transfer factor for carbon monoxide were normal. His disease was probably related to PVC dust rather than VCK. DAEXE (1976) investigated 1** men who nad been employed in a PVC manufacturing plant exposed to VCK and ell of whom complained of breathlessness. He found that their chest radiographs were normal. In six of these men transfer factor for carbon monoxide is reported as being impaired but no measurements or details of other lung function tests are given. Lung biopsy from one of the men showed focal alveolar wall thickening and macrophages present in alveolar spaces. LANGE et al. (1973) described some cases of VC workers with impairment of transfer factor for carbon monoxide, but the relationship with occupation was not" clear. 1*1*3 3 Epidemiological studies Studies of pulmonary function of workers employed in PVC manufacturing plants have produced conflicting results, and it has not always been possible to separate the effects of FVC dust from potential effects of VCM. A high prevalence of abnormal lung function, mainly obstructive in type, was found by MILLER et al. (1975) and LILIS et al* (1975; 1976) among workers in two FVC manufacturing plants in the United States of America but smoking was not taken into account in the analysis. Smoking effects were allowed for in the study by GAMBLE et al. (1976) which compared the lung function of workers in a FVC plant with that of rubber workers and chemical workers exposed to VCM: no significant differences were found in the lung function of these three groups but no measured personal exposures were available. VERTKIN ^ al. (1970) studied 96 FVC workers in the USSR and found a high prevalence of chest radiograph abnormalities but the majority of the workers had normal lung function testa. CHIVZR2 et al* (1980) similarly found no apparent impairment of lung function among 509 employees in a FVC factory of whom I0*t were exposed to FVC dust only, 112 to non-chloriaated solvents only and the remaining 293 were exposed to a mixture of both* Ho difference between the three groups was identified and the observed impairment of lung function in some men was related entirely to the effects of smoking. However, measured personal exposures were not available for these men* URL 15062 A study of 8l8 men currently and previously employed at a FVC manufacturing plant (SCUTAR et al. 1980) in which estimates of personal exposure to FVC dust were based on detailed occupational histories and current measurements of dust levels in the factory showed a statistically significant reduction in tests of lung function among man with the highest PVC dust exposure, and some of these men had small rounded opacities on their chest radiographs* The men with these radiological changes had significantly lower mean lung function test values than men whose films were oonaidered to be normal* Ho relationship was found between exposure to PVC dust and impaired transfer factor for carbon monoxide* 1*2 The uresent study The present atudy was requested by Vinatex Ltd*, and was carried out with the full agreement of management and workforce, and arose from their concern about some employees whose lung function, particularly the 4 transfer factor for carbon monoxide, appeared to be abnormal* The transfer factor for carbon monoxide measures tie ability of the lung to exchange gases between the inspired air and the blood* The study was designed to describe the distributions of various measures of lung function, including the transfer factor, in current and former employees at the PVC factory and to identify men with low lung function values. These men would be regarded as 'cases' and would be compared with two groups of controls, in order to examine possible associations between abnormally low lung function values and personal, occuoational and environmental factors. Tho total survey population included the workforce of two factories; one was the PVC factory of interest and the other was a large iron foundry situated a few hundred yards away. URL 15063 1.2.1 The PVC factory The plant had been in operation since 1969 and for most of that time had produced suspension polymer. For a pericd of three years (1971 - 1974) emulsion polymer, which has a smaller particle size than suspension polymer, had been manufactured in one of the two reactor buildings. During the manufacturing process vinyl chloride is pumped from the tank farm storage area via another storage tank to two reactor buildings. Coder conditions of increased temperature and pressure and in the presence of other substances acting as catalysts and surfactants, polymerisation of VCM takes place. The resulting PVC is piped as a slurry to blending tasks and is then dried by centrifuging and rotary drying. The dried polymer is sifted and either removed to storage silos or bagged and stored in the warehouse before being dispatched. The manufacturing process is operated in batches throughout the 24 hours and the workforce is divided into four shifts. In the reactor and dryer buildings men usually start as rB* operators and can then proceed to be 'A* operators and sometimes 'Lead* operators. , Available data suggest that prior to 1975 levels of VCM in the factory were considerably higher than the current low levels* During the period 1969 - 1975 men working in the reactor buildings, and particularly those involved, in the manual cleaning of reaetors, are likely to have been exposed to the highest levels of VCM* 5 1.2.2 The foundry The foundry was situated near the PVC factory in the same industrial alley and manufactured moulded steel products. The manufacturing process involved the production of molten steel which was cast Into appropriate moulds to which various finishing techniques were applied. URL 15064 6 2. METHODS 2*1 Selection of the survey population All 177 men currently employed at the FVC factory were invited to take part* The names of another 263 men who had worked at the plant since it started production in 1969 vere identified from factory records and attempts were made to trace and invite all these sen using information supplied by the management, union representatives and local records* At the foundry the study was confined to 283 men selected randomly from the 1,083 current male employees* 2*2 The medical survey The factories were visited by the medical survey teas and men were 3een by appointment* Details of each man's smoking history were recorded by a trained clerk using a questionnaire which was a modified version of the smoking section of the Medical Research Counci] Questionnaire of Respiratory Symptoms (MRC, 1976) (see Appendix 1)* A full occupational history was recorded for each man by a trained clerk who had previously visited both factories and was familiar with their layout and the manufacturing processes. Details of all .jobs which a man had done since leaving school were recorded and the time spent in each of them noted* All FVC factory employees had measurements of forced expiratory volume in one second (FEVj), forced vital capacity (FVC) and gas transfer factor for carbon monoxide (Tico)* For foundry workers, approval was ootained for measurement only of Tuco. Measurements of FEV^ and FVC were made using an electronic dry rolling-seal spirometer (Ohio 800)* After a practice expiration, three forced expirations which were considered to be technically satisfactory from a maximum of six were recorded* The maximum values of FEVt sod FVC, not necessarily in the same breath, were need for the analysis, end the ratio of FEV^ to FVC was derived from these figures* C XI <=5 7, Single breath gas transfer factor for carbon monoxide (Tvco) was measured by the breathholding method of HEADS et_ al. (1965) based on the modified Krogh technique. An automated spirooeter/gos sampling system (Transfer Test B, P.K. Morgan Ltd., Chatham, Kent) was used and duplicate estimations were made for each man at least 10 minutes apart. Technically unsatisfactory tests were rejected and repeated (i.e. if the inspired volumes of the two tests differed by more than 1C0G, if the spirograms of the manoeuvres were of different shape or if the rates of inspiration and expiration were too elow). The mean of the results of the two satisfactory testa was used in the analysis. Eaeh man was weighed fully clothed and bis standing height measured. 2*3 Methods of analysis The data were studied initially by summary descriptive statistics and multiple linear regression analyses of lung function taking into account age, height, weight and smoking habit. Subsequent examination of the standardised residuals* was carried out and on the basis of this oases and controls were selected. 2*3*1 Classification of smoking histories Men were assigned to one of three smoking categories; lifelong nonsmoker, current smoker and ex-smoker* No distinction was made in the analysis between type of smoking and the two 'smoker' groups included smokers of manufactured cigarettes, hand-rolled cigarettes, pipes and cigars, since the majority of men smoked manufactured cigarettes* 2*3*2 Analysis of data from total survey copulation The distributions of lung function and various explanatory variables URL 15066 * A standardised residual, SR, is the difference between an observed lung function value and its predicted value from multiple linear regression analysis, divided by the estimated standard deviation of the unexplained variation from this model* i.e. SR (observed value - value predicted from fitted equation) (estimated standard deviation of the prediction) 8 were examined for systematic patterns amongst current and former employees of the PVC factory and current employees at the foundry. This waa further explored for fuco and FEVj by multiple linear regression analysis taking account of age, height, weight and smoking. The relationship between a man's observed and predicted values was expressed as a standardised residual (SR). Thus men whose observed values were elose to the predicted values would have SR values clustered around zero and correspondingly sen with impaired lung function would have the largest negative SR values. The distribution of SR values was examined to see if there was evidence of an unexpectedly large number of men with impaired lung function, and if eot to examine if either PVC factory or foundry eoploye-s were over, represented amongst these men. l90SV TdC 2.3.3 Selection of cases and controls For each lung function variable those men with the largest negative SR values were ehesen as cases. Two groups of controls were selected randomly, firstly from those men whose SRs were clustered around zero and secondly from those with the largest positive SRs. 2.3.^ Case-control study Casas were initially compared with the rest of the survey population in order to identify systematic differences amongst related factors, e.g. age, smoking habit, employment status. 'The occupational and smoking histories of eases and controls were studied for systematic differences. Fisher's exact test, chi-squared tost for 2x2 contingency tables and chi-squared test for linear trend were used where appropriate. Classification of smoking and employment histories Smoking histories were classified as they were for the total population (oee Section 2.3.1). Additional information about the number of cigarettes smoked and age of starting smoking was analysed in the easecontrol study. Bread occupational categories were defined for both PVC 9 factory and the foundry (Tables 1 and 2) and the total number of years spent by all eases and eontrols in each occupational category was calculated* Additional information was available for PVC employees in relation to exposure to emulsion polymer and employment in the factory prior to 175 Proa the* full occupational history, the number of years in 'noxious1 industries was recorded* 'Noxious* industries were classified into four groups; the coal and chemical industries have certain well recognized hazards, while the lung function of men in the steel industry has been reported to be lower than oen in non-dusty employment (HIGGINS, 1970); the fourth, designated 'other1, included industries iu which hazards could possibly arise although they would be less likely than those of the otner groups (see Appendix 2 for a list of 'other noxious' jobs)* Employment of sen in the manufacturing of PVC other than at the PVC factory was also noted* URL 15068 URL 15069 10. 3* RESULTS 3.1 Description of total survey population A total of 486 men was seen consisting of 163 current PVC factory employees (93# of those who were asked to participate), 102 PVC factory leavers (39# of those who were invited) and 221 foundry employees (7& of those invited)* Three men (one from each employment category) did not complete the Tuco measurement and one foundryman did not have height and weight values recorded; the analysis was based on the remaining 484 men. FEVt and FVC measure ments were not obtained from three PVC factory leavers and were not carried out on the 221 fo-andrynen and the analysis waa based therefore on 264 men* 3*2 Analysis cf lung function data from total survey population 3.2*1 Calculation of predicted values and standardised residuals The effects of age, height, weight and smoking habit on the Tuco and FEVj. were studied in tills population by multiple linear regression. The development of the regression model used is described in Appendix 3* Standardised residuals (SR) for Tuco and FEVX were calculated for each man from the final regression model. Thus a man's standardised residual Tuco represents the amount by which his Tuco value differs from that predicted by his age, height, weight and smoking habit after standard isation* 3*2*2 .Distribution of standardised residual values of Tuco and FEV, for total survey population After allowing for the above factors, the standardised residual values of Tuco for the total survey population showed a symmetrical distribution, with no greater number of men in the negative tail of the distribution than would be expected to have occurred by chance. This indicated that there was no evidence that a substantial number of men had suffered a clinically Important reduction of their Tuco* However, for the FEVX SR distribution there were slightly sore extreme negative values than extreme positive values (Figures 1 and 2)* 3*2*3 Selection of cases and controls It was decided that about 30 cases would form a study group of appropriate size* The threshold S3 value for Tuco below which there was approximately this number of men, was -- 1*5, giving 31 men* A 11 control group of 31 men wan selected randomly from those sen with SB values between - 0,25 and + 0-75 inclusive (Tuco 'average* controls)* A second control group of 31 'extra-healthy* teen was selected randomly from those with SB values in excess of + 0,75 (Tuco 'healthiest* controls)* Similarly 30 men with an FEVj SB value of less than - 1*1 were selected as FEY* cases and two control groups, each of 30 men, were selected randomly from those with sn SR value between - 0,25 and + 0.75 inclusive (FEVt 'average* controls) and from those with as SR value in exceaa of + 0.75 (TtV^ 'healthiest* controls), URL 15070 3*3 General features of cases, controls and the total survey population 3,3.1 Aye, height and weight The seen age, height and weight of Tj-co cases and controls, FEV4 ease3 and controls and the total population are given in Tables 3 and 4. Tieo eases and controls were of similar age but cases tended to be taller than the control groups. Although the mean ages of Tt-co cases and control groups were similar, study of the age distributions (Table $) showed an & excess of current PVC factory employees in the 45 - $b age group when I* compared with both groups of controls and the rest of the population. FBVj cases were slightly older and taller than the controls and the rest * of the population but there was no difference in the age distribution betwecr employment groups or cases and controls. 3*5*2 Snroloyment status and smoking habit The employment status of eases, controls and the total population is shown in Tables 6 and 7* PVC factory employees were slightly over represented amongst Tt-co cases but this eould easily have occurred by chance. Differences in smoking habit are discussed in a later section, 3*3*3 Lung function The mean values of Tuco for T-eo cases and controls and the total population are shewn in Table 3 and mean values of TLco, FEVt and FVC for FEVi eases and controls are shown in Table 12 Observed Tuco and FEVj values were expressed as a percentage of values predicted on the basis of Cotes's regression model (COTSS, 1979) and the distribution of these values for cases, controls and the total population are shewn in Tables 6 and 9* 3.4 Clinical description of Tuco cases and FEV, cases Fifteen sen (nine FVC workers and six foundrymen) of the 31 Two cases had Tuco values below 706 of their predicted values and four of these were below 6CE (Table 8)* Nineteen of the 31 cases were from the PVC factory and of those, four men had F2Vt values of less than 8CS of their predicted value and low FEVj/TVC ratios indicating airflow obstruction. The remaining 13 men had FEVt values within the normal range although further examination ef their lung function data (forced expiratory flow-volume loops) suggested that seven of these might have slid airflow obstruction. Details of the 10 PVC workers with the lowest Tuco standardised residual values are given in Table 10. URL 15071 Of the 30 FEVj eases, 13 had observed FEVj^ values below 8cgs of their predicted valuer and 10 of these were less than 703* (Table 9). The majority (10) of the 13 men with reduced F2Vt had FEVj/FVC ratic3 below 7C5 suggesting airflow obstruction. Seven of the FSV^ cases were also selected as Tuco eases.. 3.5 Case-control study The occupational histories and smoking1 habits of cases and controls were compared in order to explore the observed differences in their lung function. 3*5.1 Occupational histories of cases and controls Analyses of the current employment status of eases and controls showed that PVC factory-workers were slightly over-represented amongst the T'-co cases (Table 5), though this difference could easily have arisen by chance. The majority of PVC workers had worked in more than one type of job and in more than one broad occupational category during their employment with the company. Analysis of the mean duration of employment in each broad category failed to show any differences between Tuco cases snd their URL 15072 13. controls (Table 11), though more cases than controls had worked in jobs in category 1, those occupations close to the pressure vessels where exposures to VCM were mostly likely to have occurred in the past. This difference could easily have occurred by chance, if ?VC workers alone are considered but if the foundrymen are included, 11 of 31 cases (35) had worked in category 1 jobs-in the PVC factory, whereas only 10 of 62 controls (163) had done the same. The difference is significant at the 73 level. These occupational differences were not apparent for the 75Vx cases and controls (Table 13). There was also movement between jobs at the foundry but there were no differences in the pattern of employment cf eases and controls (Table 12). Prc-l?*7? employment Following the recognition of the hazards of VCM exposure in 1973, levels of VCH in the factory were progressively reduced during 197^ and by January 1972 the exposure levels were low. The numbers of cases and controls who had been employed at the FVC fsstory before 1st January' 1975 were compared. Seventeen of the 19 FVC worker Twco cases had been employed before 1975 compared with five of the 12 `healthiest* controls and 11 of the 13 average controls (Table 1*0. Using a X* test for trend, there was strong evidence 60 suggest a significant increase in the proportion of men working before 1975 over 'healthiest* controls, 'average* controls and cases. Looking at comparisons between groups only the difference between cases and 'healthiest' controls.was significant. Analysis of the duration of employment of men within different age groups showed that pre-1975 working was seen in younger as well as in older men (Tables 15, 16, 17). No differences were seen in the number of men employed before 1975 between FVC factory FEVj cases and controls (Table 18). A similar analysis of foundry Twco cases failed to show any relationship with pre-1975 employment (Table 19). In conclusion there is evidence to suggest that employment at the PVC factory before 1975 was associated with impairment of T\_eo. 14 Exposure to emulsion polymer The employment of eases and controls in the manufacturing of emulsion polymer was studied and compared. Bo man had been exclusively involved in the production of emulsion polymer so men were divided into three groups; those who had worked with suspension polymer only, those who had worked with both suspension and emulsion, and those who had not been exposed to either. Similar proportions of Tlco cases and controls (Table 20) sed FEV4 cases and controls (Table 21) are represented in each SSlof these three groups. There is no evidence that exposure to emulsion felfpolymer adversely affected the lung function of those who worked with aosnan Previous employment in other industries There was no evidence that the observed differences between cases and controls could be explained on the basis of previous employment in industries other than the PVC factory or the foundry* Mo differences were found either in the numbers of cases ana controls or in the mean numbers of years spent by them working in other industries which might be considered noxious coal* steel* chemicals and 'others') (Tables 22, 23* 24). No sen had been employed in the manufacturing of PVC outside this PVC factory. 3*5*2 Smoking habits of cases and controls A higher proportion of Teco cases (63#) were current cigarette smokers than were either group of controls (43$) or total survey population (43) (Table 25). This difference was not seen between the FEVj cases and controls but there were more ex-smokers among the FEVj eases than there' were among the controls (Table 26) The smoking histories were examined in greater detail by analysing the number of cigarettes which current cigarette smokers smoked daily both on weekdays and during the weekend* the number smoked by each man during the maximum smoking period of bis life and the age at which each man started smoking. There was no significant difference in the number of cigarettes smoked on weekdays between both groups of cases and the'healthiest' controls (see Figures 3 and 4), and similar results were found for 'average' controls. URL 1SU74 15 Analysis of weekend smoking showed that more Tlco eases smoked 25 or fmore cigarettes per day than did the Tlco healthiest controls (Figure 5)* No such difference was found to exist with average controls. Comparing the weekend smoking habits of FVC factory and foundry Tlco cases, it was found that a larger proportion of foundrymea smoked mere than 25 cigarettes per day (Figure 5)* For the FSV4 men, more cases than average controls smoked 25 or more cigarettes per day at weekends. This difference was not seen between cases and healthiest controls. Consideration of the number of cigarettes smoked during a man's maximum smoking period showed that of the men at the PVC factory, five of the 13 Tlco cases, compared with none of the six Tlco'healthiest' controls had smoked a maximum of 25 or more cigarettes per day (Figure 6). but this difference in proportions could easily have arisen by chance. The age at which men started smoking was also examined. There were significant differences between cases and controls. Of the 20 Tlco cases who were current smokers, eirnt had started smoking by the age of 15 compared with one out of the 14 'healthiest' controls and 16 had started by the ace of 13 compared with five out of the 14 'healthiest' controls (Figure Fa), while the figures for average controls were that six out of 15 had started smoking b7 the age of 16 (Figure Fb). Of the 13 FEV1 cases who were current smokers, seven had started smoking by the age of 16 compared with two of the 11 'healthiest' controls (Figure aa). These differences were not seen between cases and 'average' controls (Figure 9b). In conclusion Tlco cases and 'healthiest' controls and FEVl eases and 'average' controls differed in the number of men smoking 25 or more cigarettes per day at weekends. A similar difference was found between Tlco cases at the PVC factory and the foundry. In general, cases started smoking at a younger age than controls. URL 15075 16 4, DISCUSSION The present study had two aims* First, to describe the lung function, particularly the Tuco, of current and past employees of a factory manufacturing PVC, and seek evidence of a possible excess of men whose TLeo was abnormally low. Second, to examine the relationship between Tuco and occupation and other personal and environmental factors by a ease-control study, in which the cases were men with lowest gas transfer factor after allowing for age, height, weight and smoking habit. Men from a nearby foundry were included in the study in order to Increase the number of men with low or no exposure to possible respiratory hazards at the PVC factory, but who were exposed to similar environmental conditions outside their workplace. In all 265 men from the PVC factory and 219 sen from the foundry were included in the Tuco analysis. he Tlco residuals among this combined population conformed approximately o a Normal distribution after allowing for age, height, weight and smoking abit, and did not suggest an undue excess of extremely low values. This ndieated that it was unlikely that a substantial number of men in the hole population were suffering from a clinically important reduction of he Tlco. In the case-control study the 31 men with the lowest Tuco after allowing for age, height, weight and smoking habit were selected as 'eases'. This does not imply that all 31 were suffering from lung disease, but among them were 15 men whose Tuco was less than 7G& of the value predicted by published normal values, a level which would generally be regarded as abnormally low. Among the 19 cases at the PVC factory the pattern of lung function accompanying the low Tuco was that of definite airflow obstruction in four men, and possible airflow obstruction in another seven men, suggesting the presence of emphysema. The remaining eight men did not have evidence of airflow obstruction or reduction of their lung capacity Coital capacity). This pattern is unusual in clinical practice, and may possibly reflect pulmonary fibrosis or disturbances of the blood flow in the lungs. The degree of reduction of transfer factor of earbon monoxide and associated airflow obstruction might have been sufficient to cause breathlessness on exertion in some of these men. 17 The possible contribution of occupation in causing these lung function abnormalities was examined by comparing the occupational histories of the eases with those of controls. Two groups of controls were selected; one group from men with values of gas transfer near the mean for this population; the other group from those with values towards the high end of the range ('average* and 'healthiest' controls). The reasons for choosing two such groups were that sen with near average values may have included some whose lung function had been slightly affected by noxious influences, while those with very high values were least likely to have been affected by noxious influences, but also could conceivably include sen who were highly resistant to these influences, and would therefore not be suitable for the comparisons jw.dc in this study. The PVC workers were slightly over-represented among the cases with low Tl.co, though this difference eould easily have arisen by chance. Hen who had worked on jobs at the PVC factory which were close to the pressure vessels where exposure to VCH was most likely to have occurred in the past, were also slightly ovex*-represent*d among the cases, and this difference approached statistical significance at conventional levels. There was, furthermore, an association between low gas transfer factor and a history of having worked at the PVC factory before 1975* This date is important because it was at this time that, in the light of information about some of the potential hazards of VCH, exposure to this gas was drastically reduced. This association between low gas transfer factor and working before 1973 was demonstrated by comparison with the 'healthiest' controls, the 'average' controls showing an intermediate pattern of association not statistically significantly different from the eases. This association with working before 1973 was not found among foundry workers. This conjunction of findings suggests that an unknown environmental or personal factor or factors associated with working at the PVC factory before 1973 might have been responsible for some impairment of the gas transfer factor in some of these men. The most obvious candidate for suspicion is VCH, exposures to which are known to have H 1=0 been ouch, higher than post-1975 levels. Some support for this is provided by published case reports of VCK .workers with impairment of gas transfer factor, but proof of association has been lacking though it is possible that many workers throughout the factory were exposed to a lesser degree. It is also possible that exposure to respirable PVC dust may have contributed to the low gas transfer factors, for this dust has been shown in other studies to be associated with impairment of other aspects of lung function (though not the gas trarsfer factor) and slight cheat radiographic abnormalities (KASTkAHCELO _et al., 19^9^ SGUTAR *t al., 1980). There was no evidence to suggest that a current risk to respiratory health exists at either the PVC factory or foundry. Some differences m smoking habit were found between cases and controls, non-smokers being slightly commoner amonr controls, and heavier smokers slightly commoner among the cases. The eases tended to start smoking at a slightly younger age. We do not know if these small differences in smoking habit were sufficient to account for the reductions of gas transfer factor in the cases, but they would seem unlikely to account for the whole of the differences between cases and controls. While the response rate among currently employed workers was excellent, that for those men who had left the PVC factory waa low (59K). If the men not examined had included either an excess or deficiency of men with impaired gas transfer factor compared with the population examined', then the analysis of the distribution of values among the population seen could be unrepresentative to that extent. A similar analysis for another lung function test, forced expired volume in one second, was carried out for the PVC workers only. While some men did have impairment of this test result as would be expected in an industrial population of this kind, no associations with occupation could be demonstrated* 19. 5* CONCLUSIONS AND RECOMMENDATIONS We found no evidence of an unexpectedly large number of men with impairment of their lung gas transfer factor among either PVC factory or foundry employees. There was, however, a snail number of men in both factories with impairment of their transfer factor. This could not be shown to be associated with current working conditions at either factory, but the pattern of results among PVC workers suggested that there was an association between impairment of the gas transfer factor and a history of working in the PVC factory before 1975* There was alao evidence that heavy smoking was associated with impairment of gas transfer. These findings give seme support to the hypothesis that work in the PVC factory before 1975 involved exposure to a substance that caused impairment of lung function in a small number of men. The results suggest that more detailed studies of the importance of exposure to vinyl chloride monomer, polyvinylchloride, smoking habit and other possible factors in causing impairment of lung function would be desirable, tog*ther with a clinical assessment of the extent of lung damage in selected men. URL 15078 ACKNOWLEDGEMENTS This study was carried out with the full co-operation, of the management and workforce of Vinatex Ltd., to whom we are grateful for financial support. We also thank the management and workforce of the foundry for their generous co-operation in the study. fiiO S l Ib ll 21 REFERENCES AGARWAL DK, KAW JL, SRIVASTAVA SP, SETH PE (1978) Some biochemical and histopatfcological changes induced by polyvinyl chloride dust in rat lung. Environmental Research; 16: 333 - 341. URL 15080 AHNAUD A, PCMMIER de SANTI P, GARBE L, PAYAN H, CHASPIN J (1978) Polyvinyl chloride pneumoconiosis. Thorax; 33: 19 - 25. CHIVERS CP, LAWRENCE-JONES C, PADDLE GK (1930) Lung function in workers exposed to polyvinyl chloride dust. British Journal of Industrial Medicine; 37: 1*7 - 151. COBDIER JM, FIEVSZ C, IXFSVBE M-T, SEVRIN A (1966) Acroosteolyse et lesions cutanees assccieea ches deuxouvriers affecteea au nettoyagc d*autoclaves. Cahiers de Kedecine du Travail; 4: 14 - 19. COTES JE (1979) Lung function. Oxford: Scientific Publications: 369 - 385. Blackwell CREECH JL, JOHNSON MN (1974) Angiosarcoma of liver in the manufacture of polyvinyl chloride. Journal of Occupational Hedicine; 16: 150 - 151. DARKS CS (1976) Vinyl chloride and the production of PVC (Discussion). Proceedings of Royal Society of Medicine; 69i 280. FRONGIA N, SPINAZZGLA A, BUCAREILI A 0974) Lesion! polaonari sperimentali da inalazione prolungata di PVC in ambiente di lavoro. Medicica del Lavoro; 6p: 321 - 342. GAMBLE J, LI3 S, McMICHAEL AJ, WAXWEILER RJ 0976) Effect of occupational and nonoceupational factors on the respiratory system of vinyl chloride and other workers. Journal of Occupational Medicine; 18: 659 - 670. HIGGINS ITT 0970) Occupational Factors in Chronic Bronchitis and Emphysema. Review. In: Orie NGH, Van Der Lende R, eds. Bronchitis IIT. Proceedings of the Third International Symposium on Bronchitis at. Groningen, The Netherlands. Springfield Charles C. Thomas, 1970 : 83 - 99. URL 15081 22 LANGE CE, JT/HE S, STEIN G, VELTMAN G 0973) Die eogenannte Vinylcblorid-KrarJcheit - eine herufsbedingte Systemsklerose? International Archives of Occupational Health; 32: 1 - 32. LIUS R, ANDERSON H, NICHOLSON WJt DADM S, FTSCHREIN AS, SELIKOFF IJ (1975) Prevalence of disease among vinylchloride and polyvinyl chloride workers. Annals of New York Academy of Science; 246: 22-41. r.TT.TS r, ANDERSON H, HILLER At SEUKOFF IJ (1976) Pulmonary changes among vinyl chloride polymerisation workers. Chest; 69: (Suppl. 2 Feb.), 29? * 30?. MASTSANGELO G, MAKNC H, MARCER G, BAETCLUCCI GB, GEHIGN \NI C, SALADINO G, SIMONAIO L, SAIA 3 (1979) Polyvinyl chloride pneumoconiosis: epidemiological study of exposed workers. Journal of Cceupational Medicine; 21: 540 - 542. MEADS F, SAUNDERS MJ, KYETT F, REYNOLDS JA, PEARL N, COTES JE (1965) Automatic measurement of lung function. Lancet; 2: 573 - 575- HRC (1976) Medical Research Council Working P"ty on Research into Chronic Bronchitis. Publications Group, Medical Research Council, 20 Park Crescent, London WIN 4AL. KILLER A, TEIRSTEIN AS, CHUANG M, SELIKOFF IJ (1975) Changes in pulmonary function in workers exposed to vinyl chloride and polyvinyl chloride. Annals of New York Academy of Science; 245: 42 -- 52. PRODAN L, SUCIU I, PISLARU V, ILEA E, PA30U L (1975) Experimental chronic poisoning with vinyl chloride (monochloroethane). Annals of New York Academy of Science; 246: 159 - 163- SOUTAR CA, COPLAND LH, THORNLEY PE, BOSLEY JF, OTTERY J, ADAMS WGF, BENNETT B (1980) Epidemiological study of respiratory disease in workers exposed to polyvinylchloride dust. Thorax; 35: 644 - 652. SICIU I, DREJMAN I, VELA5KAI M O967) Study of disease caused by vinyl chloride. Medic ina. del Lavoro; 58: 261 -- 271. 23. VELTMAN G, LANGE CE, JTJHE S', STEIN G, BACHNER U (1975) Clinical manj festations and course of vinyl chloride disease. Annala of New fork Academy of Science; 6 - 17. VEHTKIN YI, MAMONTOV YS (1970) The atate of the bronchi and lungs in workers employed in the manufacture of polyvinyl chloride articles. Gigiena Truda i Professional r nye Zaholevaniya; l4: 29 - 32. URL 15082 TABLE 1 25 Classification of all jobs at Vinatex into six broad occupational groups. Broad occupational group Jobs included in group 1 All men working in reactors 1 and 2 * 2 Dryer room B operators 3 Dryer room Lead, A and A/B operators . 4 Warehousenen, maintenance staff, laboratory staff and drivers 5 General staff who visit plant, e.g. shift supervisors and foremen j6 Management and office staff * Thea* men are mo.*t likely to have bee.t exposed to high concentrations of VCM in the case, ** These men are involved in the bagging of PVC and are probably exposed to higher concentrations of dust than other workers in the dryer buildings.*** *** These men are potentially exposed to all raw materials and products. URL 15083 TABLE 2 26* Classification of foundry jobs into seven broad occupational groups. URL 15084 Occupational j^roup No* Description Jobs included in group 1 Furnace and melt: furnace operator, burner, charger, weirher, skinner, smelter, sl&gman, ladle repairer, concrete liner. 2 Sand and blasting: coremaker, sand tester, sand mixer, shot blaster, core stripper, knock-out, brick maker. 3 Metalwork: driller, grinder, blacksmith, tinsmith, wagon repairer, rope splicer, cupola repairer, maintenance fitter's sate, maintenance fitter, turner, trainee in the training centre, borer, saw repairer, wagon repairer, milling operator, inspector in machine shop, welder. 4 General and other: labourer, patternmaker, plater, platelayer, pine loader, crane driver, driver, cleanor, technician, trainee, pipe roller, elinger, quality control, joiner, greaser, inspector, saddler, bricklayer, core inspector, despatcher, hydraulic attendant, shunter, plumber, pipe recorder, runner operator, sawyer, staff, power house attendant, canteen, tool room attendant, electrician, storeman, ore crusher, ram driver, safety officer. 3 Paint and dipping: dipper, paint sprayer, painter and decorator. 6 Casting: caster, manhole maker, socket man, machine moulder.. 7 Fattier: fettler TABLE 3 27. Features of Tico cases, controls and total survey population. Uean (and standard deviations) of age, height, weight and Tcco. Humber of men Cases 31 Tuco Groues ControIs (average) 31 Controls (healthiest) 31 Total survey population 484 Ae (yr) Uean 41*8 (SD) (10.2) 42-3 (12.5) 40-3 (11.3) 42.2 (11.9) Height (caa) Uean 176-1 (SB) (7.1) 174-0 (6.6) 173-9 (5.6) 173-8 (7.0) URL15U85 Weight (less) Mean 82.7 (SD) (13-3) 34-9 (18.5) Tlco {ml/min, ma Eg) Uean (SD) 1 22.2 (4-1) 31-3 (2.9) 79-2 (9.5) 37-3 (4.3) 30.2 (12.6) 30.5 (5.9) TABLE *+ 28 Features of FE7t eases, coatrols and total survey population* Mean (and standard deviations) of age, height, weight* Ttco, ?E7t and FVC. Kiaber of men Cases 30 Tlco Croups Controls (average) 30 Controls (healthiest) 30 Total survey population 264 Iff* (yrs) Mean (3D) 43*3 (11.Q) 41.3 (12.5) Height (cas) freight 0=ff*) Mean (3S) 176.0 (7.4) Mean (SD) 82.2 (13-3) 174*5 (6-6) 78*4 (10*1) FETj (litres) Mean (SD) 2*7 (0.7) 3.9 (0.7) FVC (litres) Mean (SD) 4-1 (0.9) 5-1 (0-8) Tuco Mean (al/iain*.iiniHg) (SD) 29.1* (6-3)* 32.3 (6.0) 42.1 (12.0) 175.9 (8.0) 30.4 (11.5) 4-5 (0.8) 5-9 (0.9) 31.3 (S.5) 42.2 (11.8) 174.3 (7.1) 30.4 (12.3) 3.7 (O.S) 4.9 (1.0) 30.5** (6.0** "M Ol * Talus based on 29 observations* ** Value based on 262 observations* TABLE 5 Age distribution of Tujo eases and eontrole and the total survey population by employaent status* CASES PVC currant san PVC leavers Foundrysan AVERAGE CONTROLS PVC ourrant pan PVC laavara Foundrysan HEALTHIEST CONTROLS PVC currant nan PVC leavers Foundrysan TOTAL POPULATION PVC current sen PVC leavers Foundrysan Age ranges In years 0 - 2*1 25 - 34 35 - 44 45 - 54 55 - 64 65* Total 1 4e 11 222 1 1 8 1 434 12 1 41 3 1 12 3 2544 9 4 18 2 324 1 12 1 2643 12 3 16 9 37 57 35 26 0 164 4 24 33 19 11 10 101 17 48 50 60 44 0 219 i80snan TABLE 6 30 Employment status of Tuco cases, controls and total survey population* Number of men in each employment group with percentages in brackets. Four men. excluded, see text* PVC Current Workers Total survey population 164 PVC Leavers 101 Foundry 219 Total 484 Cases Controls (average) Controls (healthiest) All 11 ( 34) 9 < 28) 8 ( 5?) 12 ( 26) 4 ( 27) 18 ( 39) 12 ( 33) 3 ( 20) 16 C 35) 32 (ICO) 15 CiOO) 46 (100) 31 31 31 93 URL 15083 TABLE 7 Employment status of FEV* eases, controls and total survey population* Number of men in each employment group with percentages in braeketa* Three men excluded, see text* Sounding error of percentages net adjusted to justify totals* - Current Workers Total survey population 165 PVC Leavers 99 Cases Controls (average) Controls (healthiest) AH 14 ( 26) IS C 34) 21 < 40) 53 (100) 16 ( 43) 12 ( 32) 9 ( 24) 37 (100) Total 264 30 30 30 90-