Document b226e5zndwjVENVwpbVx8Z1D

British Journal of Industrial Medicine 1984;4E:328-333 f)/0 nrf Epidemiological study of the lung function of workers at a factory manufacturing polyvinylchloride M H LLOYD, S GAULD, L COPLAND AND C A SOUTAR From the Institute of Occupational Medicine, Edinburgh EH8 9SU, UK abstract A preliminary epidemiological study has been carried out to investigate a report that some men working in a factory manufacturing polyvinylchloride (PVQ had abnormally low values of the single breath diffusing capacity for carbon monoxide (TlCO). All 265 present and past employees of the PVC factory were studied, together with 219 men from the workforce of a nearby foundry. Each man's TlCO was measured and a smoking history and detailed occupa tional history obtained. The distribution of standardised TlCO results from all persons examined was symmetrical and did not indicate an unexpectedly high proportion of men with clinically important impairment of TlCO. The results of a case-control study showed that, having allowed for age, height, weight, and smoking habit, TlCO was associated with a history of working in the PVC factory before 1975 (when levels of vinylchloride monomers (VCM) were much higher than subsequently), and slightly associated with working in jobs where exposure to VCM was likely to have been highest. The men with low TlCO also tended to have smoked more heavily than controls. The relative importance of occupational factors and smoking in relation to low TlCO is not clear, but the results give some support to the hypothesis that work in the PVC factory before 1975 entailed exposure to a substance that caused impairment of lung function in a small number of men. The management and occupational physician of a Methods factory making polyvinylchloride (PVQ had become concerned over an apparent excess in their selection of survey population , workforce of men with impairment of the single All 177 men currently employed at the PVC factory breath transfer factor for carbon monoxide (TlCO). were invited to take part and attempts were made to The present study was planned to determine trace and invite all the 263 former employees. At whether there was such an excess of men with low the foundry, the study was confined to 285 men values of TlCO in the workforce, and to make a selected randomly from the 1085 current male em- preliminary assessment of whether low TlCO might ployees. The number of men at the foundry aged be related to occupational factors. under' 25 was limited to 15 to take account of a The factory had been in operation since 1969 and disproportionately large number of young men in produced PVC powder from vinylchloride monomer the foundry population compared with the PVC gas (VCM). All men who had worked at the factory, workers and to provide a group of similar size to the whether or not they were currently employed there, 13 PVC workers of this age. Apart from this, sclec- were invited to participate in the study. A sample of tion did not take account of age. men working in a nearby foundry were also invited to participate; they had bfcen exposed to the same medical survey environmental conditions outside the workplace and The medical survey team visited the factories and were included in order to augment the number of recorded a full occupational history since leaving men in the study who had not been exposed to poss- school for each man. Trained personnel using a ible hazards in the PVC factory. questionnaire which was a modified version of the Medical Research Council questionnaire of respiratory symptoms' recorded details of each Received 28 February 1983 Accepted 9 May 1983 man's smoking history. Additional questions were related to age of starting smoking and the number of 328 BFG38690 Epidemiological study of the lung function of workers at a factory manufacturing polyvinylchloride 329 cigarettes smoked during a man's maximum smok ing period. Measurements of single breath gas transfer factor for carbon monoxide were carried out on all men using the breathholding method2 based on the mod ified Krogh technique (transfer test B; P K Morgan Ltd, Chatham, Kent). The mean of two measure ments was used in the subsequent analysis. AH PVC factory employees had forced expiratory volume in one second (FEV,) and forced vital capacity (FVC) measured using an electronic dry rolling seal spirometer (Ohio 800). The best of three technically satisfactory measurements was used for the analysis. Approval was not obtained for FEV, and FVC measurements in the foundry workers. METHODS OF ANALYSIS Analysis of data from the total survey population The data were studied initially by summary descrip tive statistics and by multiple linear regression analysis of lung function taking into account age, height, weight, and smoking habit. Men were assigned to one of three smoking categories; lifelong non-smoker, current smoker, and ex-smoker. The two smoking groups included smokers of pipes, cigars, and hand-rolled cigarettes but the majority smoked manufactured cigarettes. In the regression model used for both T^co and FEV, different intercepts and age slopes were Mowed for the three smoking groups, non-smokers, .Trent smokers, and ex-smokers. This means that the estimated effects of age were not constrained to be the same necessarily in the three smoking habit groups. The relation between a man's observed and pre dicted lung function values was expressed as a stan dardised residual (SR). This is defined by the equa tion: (observed value--value predicted from fitted equation) (estimated standard deviation of the prediction) _ distribution of FEV, and TlCO SR values of the whole population was examined. Case-control stwfiy For each lung function variable, those men with the largest negative SR value were chosen as cases. Two groups of controls were selected randomly, firstly, from those men whose SRs were clustered around zero ("average" controls) and, secondly, from those with largest positive SRs ("healthiest" controls), groups were chosen because men with average js may have included some whose lung function had been slightly affected by noxious influences, whereas those with very high values were least likely to have been affected by noxious influences but Table 1 Classification ofall jobs at PVC factory into six broad occupational groups. Broad occupational group Jobs included in group All men working in reactors 1 and 2* Dryer room B operatorst Dryer room Lead, A and A/B operators Warehousemen, maintenance staff, laboratory staff, and drivers^ General staff who visit plant, such as shift supervisors and foremen Management and office staff Men most likely to have been exposed to high concentrations of VCM in the past. tMen engaged in the bagging of PVC and probably exposed to higher concentrations of dust than other workers in the dryer buildings. tMen potentially exposed to all raw materials and products, might include men who were highly resistant to these influences. Cases were initially compared with the rest of the survey population in order to identify any systematic differences in age, smoking habit, and employment status. Then the occupational and smoking histories of cases and controls were examined and compared. Smoking histories were classified as they were for the total population, but additional information about the number of cigarettes smoked and the age of starting smoking was analysed in the case-control study. Broad occupational categories were defined for the PVC factory (table 1) and for the foundry, and the total number of years spent by cases and controls in each occupational category was calcu lated. Additional information was available for PVC employees in relation to exposure to emulsion polymer (which has a smaller particle size than sus pension polymer and which was manufactured at the factory between 1971 and 1974) and employment in the factory before 1975 when levels of VCM were much higher than the currently permitted levels. From the full occupational history the number of years spent in potentially "noxious?' industries (such as coal, steef, chemicals, and a miscellaneous group of industries in which exposure to noxious com pounds might have occurred) was recorded. In the analysis Fisher's exact test, the chi-squared test for 2 x 2 contingency tables, and the chisquared test for linear trend were used where appropriate. Results A total of 488 men was seen, consisting of 165 cur rent PVC factory employees (93% of those who were asked to participate), 102 PVC former em ployees (39% of those invited), and 221 foundry employees (78% of those invited). Four men (two PVC and two foundry) were excluded from the BFG38691 ,330 Lloyd, Gauld, Copland Soutar Table 2 Features ofT^CO cases, controls, and total survey population. Mean (and standard deviations) ofage, heightt weight and Tjjco No of men T--^-C-O--g-r-o-u-p-s----------------- Total survey Cases Controls Controls population (31) (average) (healthiest) (484) (31) (31) Fig 1 Number of men by ranges of T^CO standardised residuals. Age (y): Mean (SD) Height (cm): Mean (SD) Weight (kg): Mean (SD) Ti CO (ifil/min mm Hg): Mean (SD) 41-8 (10-2) 176-1 (7-1) 82-7 (13-3) 22-2 (4-1) 42-3 (12-5) 174 0 (6-6) 84-9 (18-5) 31-8 (2-9) 40-3 (11"3) 173 9 (5-6) 79-2 (9-5) 37-3 (4-3) 42-2 (11-9) 173-8 (7-0) 80-2 (12-6) 30-5 (3-9) TLCO analysis because of incomplete data, and to the final analysis was based on 484 men. As FEV, and FVC measurements were not obtained from three PVC factory workers and all 221 foundrymen this analysis was based on 264 men. The distributions of standardised residual values trols). Similarly, 30 men with an FEV, SR value of less than -1* 1 were selected as FEV, cases and two groups of controls (FEV, average controls and FEV, healthiest controls) were chosen from those men with SR value between -0-25 and +0-75 and in excess of +0-75 respectively. of TlOO and FEV, for the total survey population after allowing for age, height, weight, and smoking habit were fairly symmetrical and conformed approximately to Normal distributions (figs 1 and 2), although for FEV, there were slightly more extreme negative values than extreme positive val ues. CLINICAL FEATURES OF CASES, CONTROLS, ANIJ THE TOTAL POPULATION The general features of TlCO and FEV, cases, con trols, and the total population are shown in tables 2 and 3. The mean age of the TlCO case and control groups were similar. Observed TlCO and FEV, values of the cases SELECTION OF CASES AND CONTROLS were also expressed as a percentage of values pre ' A threshold SR value for TlCO of -1- 5 supplied 31 dicted on the basis of Cotes' regression equation.' men, a number that was considered large enough to Fifteen of the 31 TlCO cases had TlCO values form a study group. A control group of 31 men was below 70% of their predicted values; four were selected randomly from those men with SR values below 60%. Twelve of these 15 men were PVC between -0-25 and +0-75 inclusive (TlCO average workers and three were foundrymen. Nineteen of controls). A second control group of 31 "extra heal the 31 cases were from the PVC factory and four of thy" men was selected randomly from those with SR these.had evidence of airflow obstruction (FEV, < values in excess of +0-75 (TlCO healthiest con- 80% and FEV/FVC < 70%); examination of forced expiratory flow-volume loops suggested that a further seven might have mild airflow obstruction. In the remaining eight the low TlCO appeared to be an isolated abnormality. Of the 30 FEV, cases, 13 had observed FEV, values below 80% of their predicted values and 10 of these men had FEV/FVC ratios below 70% sug gesting airflow obstruction. Seven of the FEV, cases also qualified for inclusion as Tlco cases. Fig 2 Number of men by ranges of FEV, standardised residuals. OCCUPATIONAL HISTORIES Analysis of employment status showed that 19 of the Tlco cases worked or had worked at the PVC factory and 12 were foundrymen, but this difference could easily have arisen by chance. Most workers 332 Lloyd, Gauld, Copland, Soutar be related to a mild reduction of lung function and chest x ray abnormalities,4 s less is known about the pulmonary effects of exposure to VCM. There have been some case reports of men exposed to VCM who had impairment of lung function, particularly TlCO,*' and changes in lung histology have been reported .** The present study was designed as a preliminary examination of whether an undue number of men at this PVC factory had impaired Tlco, and whether such impairment was likely to be related to their occupation. We have found no evidence of a sub stantial excess of men with impairment of the Tlco, although low TlCO values were found in a few men, Fig 3 Number ofcigarettes smoked a day at weekends among currentsmoking Tfj:o cases and healthiest controls. as might be expected in any large population of workers. After allowing for other factors,, however, those men with the lowest Tlco tended to have worked more often than would be expected at the PVC factory before 1975, a date before which con centrations of VCM were known to have been much higher than subsequently; these men also tended slightly to have worked in the jobs where the highest <14St7B>21 Ag*(y) %L JZL Cos <I4 66T7B2021 Age(y) concentrations of VCM were found. This evidence gives some support to the hypothesis that work in the PVC factory entailed exposure to one or more substances that caused impairment of lung function in a small number of men. The most likely substance is VCM, on which Fig 4 (a) Age as start ofcigarette smoking among Tj_CO cases and healthiest controls who were current smokers, (b) Age at start ofcigarette smoking among Tlco cases and average controls who were current smokers. suspicion has previously been cast by case reports of vinylchloride workers with impairment of gas trans fer factor, but in which proof of association has been lacking.*1 It is also possible that exposure to respir able PVC dust may have contributed to the low gas to emulsion polymer at the PVC factory. transfer factors although there was little positive \ No evidence was obtained that the observed dif evidence for this in our study. PVC dust has been ference between cases and controls could be shown in other studies to be associated with impair explained on the basis of previous employment in ment of other aspects of lung function (though not industries other than the PVC factory or foundry. the gas transfer factor) and slight chest x ray abnor malities.4 5 SMOKING HABITS OF CASES AND CONTROLS There was little evidence to support, as an alterna Whereas the procedure for selection of cases and tive^hypothesis, the suggestion that the workforce controls had taken account of whether a man recruited in the early years of production included smoked, it was found that the TlCO cases had an excess of men with poor respiratory health at the tended to smoke more than controls. A slightly time of entry to the factory. The FEV,, a more sensi higher? proportion of TlCO cases (65%) were cur tive indication of the common types of chronic rent cigarette smokers than either group of controls respiratory illness, was not found to be particularly (on average 48%) (table 5), and more of these cases adversely affected in men working in the factory smoked 25 cigarettes pr more at the weekend than before 1975. the healthiest controls (fig 3). Weekday smoking fol The selection of cases took into account whether a lowed a similar but statistically insignificant pattern. man smoked, but even after allowing for this, the The TlCO cases tended to have started smoking at cases with IowTlCO tended to have smoked slightly an earlier age than controls (fig 4). more than the controls, and notably the age at which a man started to smoke appeared to be related to Discussion low Tlco. The relative contributions towards impairment of TlCO of smoking and of occupational Whereas exposure to PVC dust has been shown to factors are not clear from this study, but it is possible Epidemiological study of the lung function of workers at a factory manufacturing polyvinylchloride 333 that both were important. The men with the lowest TlCO values selected as cases were not all necessar ily suffering from lung disease but among them were 15 men whose Tlco was less than 70% of the value predicted by published normal values, a level which probably indicates the presence of disease. In eight of the 19 cases who had worked at the PVC factory the low Tlco was an isolated abnormality, a pattern unlikely to be the result of damage from smoking and which is usually associated with interstitial lung disease or disturbance of pulmonary perfusion. Reports of both animal and clinical studies give further support to the suggestion that exposure to VCM might have been an important causative fac tor. Animal studies have shown that exposure to high levels of VCM leads to pulmonary fibrosis, the changes being most severe in the animals which had been exposed for the longest time.* There have been isolated case reports of individuals with interstitial lung disease who had been employed in the man ufacture of PVC and who had, therefore, been exposed to VCM.*10 Hunter etal described six such workers, three of whom had died from liver disease related to VCM exposure*; none of the six had respiratory symptoms, three had abnormal chest x ray films, and one an abnormal Tlco, but the lung histology from all six men showed a similar picture with the presence of an inflammatory interstitial infiltrate. Diligent attempts were made to contact all men who had worked at the PVC factory but response ..as low (39%). 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 dis tribution of values among the population seen would be unrepresentative to that extent. We conclude that this preliminary study has shown no evidence that current working conditions in either factory have caused impairment of lung function but it does give some support to the hypothesis that work in the PVC factory before 1975 entailed exposure to a substance that caused impairment of lung function in a small number of men. Further work including clinical studies of affected men would be necessary to characterise this impairment and to explore more fully the relations with occupational and other factors. This study was carried out with the full cooperation of the management and workforce of Vinatex Limited to whom we are grateful for financial sup port. We thank the management and employees of the foundry for their generous cooperation in the study and the medical officers at Vinatex and British Steel for their help. We also thank Dr A Seaton and Dr M Jacobson for their advice. Further details of the results of the work are recorded in Institute of Occupational Medicine report TM/82/4. References 1 Medical Research Council working party on research into chronic bronchitis. Report. London: Medical Research Coun cil. 1976. Meade F. Saunders MJ. Hyett F. Reynolds JA. Pearl N. Cotes JE. Automatic measurement of lung function. Lancet 1965;ii:573-5. 1 Cotes JE. Lung function. Oxford: Blackwell Scientific Publica tions, 1979:369-85. * Soutar CA, Copland LH. Thomley PE, et at. Epidemiological study of respiratory disease in workers exposed to polyvinyl chloride dust. Thorax 1980;35:644-52. ' Mastrangelo G, Manno M, Marcer G, etal. Polyvinyl chloride pneumoconiosis: epidemiological study of exposed workers. JOM 1979:21:540-2. Darke CS, Vinyl chloride and the production of PVC. Proceed ings of the Royal Society of Medicine 1976. ' Lange CE. Juhe S. Stein G, Veltman G. Die sogenannte Vinylchlorid-Krankheit eine berufsbedingte Systemsklerose? International Archiv Arbeitzmedizin 1974:32:1-32. Hunter AM, Wagner JC, Johnston N, Seaton A. Lung fibrosis in polyvinylchloride workers. Thorax 1981;36:711-2. " Prodan L, Suciu I, Pislaru V, Ilea E, Pasou L. Experimental chronic poisoning with vinyl chloride (monochloroethane). Ann NY Acad Sci 1975;246:159-63. Amaud A. Pommier de Santi P, Garbe L, Payan H, Charpin J. Polyvinyl chloride 'pneumoconiosis. Thorax 1978;33:19-25. BFG38695