Document oeODmZVw8qYEJMj920nj6nXZD

''V Biffraciiim t I dispersion Siain' i ( Quartz in ' ' '^2 (1955), ;e 1r-: QuanliI) armination of ( hem. 40:358 >ussET ' 11^a in RespiraK. 8521. Pliable Inversion "f. 41:532 bon Hazards incept and the . '(.-ciive Sampling. ipects of Pneu" .Voc. 32:7 I - i'als on the id Respirable Criteria rjept. HEW, t ise Control, "it Conditions Hyg. Assoc. 1 '(-'termination 1 nal. 4:169 1 mination of 1 Hyg. Toxicol. in the I'ectroscopy, i) distances and "onmnt iciican Conf tv P.O. Box October, 1975 A mail-out questionnaire examined symptoms previously associated with vinyl chloride exposure. Assessment of exposure levels was made by analytical measurement and observation. The resulting statistical analysis revealed a dose-response type of relationship between exposure and certain morbidity symptoms. The association of vinyl chloride exposures with morbidity symptoms ROBERT SPIRTAS, M.S., ANTHONY J. McMICHAEL, M.D., Ph.D., JOHN GAMBLE, Ph.D. and MARK VAN ERT, Ph.O. Occupational Health Studies Group, University of North Carolina, Chapel Hill, North Carolina Introduction Before its recent implication in the etiology of angiosarcoma, vinyl chloride (VC) had been given a relatively clean bill of health. First synthesized in 1833, VC has come into in creased use since World War II in the produc tion of polyvinyl chloride (PVC) resins. Until 1960, no industrial fatalities associated with VC use had been reported. However, two fatalities then occurred in a PVC polymeriza tion plant--one worker died cleaning out a polymerization vessel, and the other while working in a pit.1 These two deaths under scored the potential seriousness of the acute effects of VC upon the central nervous system. Relatively few effects have been reported as a result of acute exposure to VC. Lehmann and Flury2 reported that VC was highly nar cotic, but had a wide margin between its nar cotic and lethal concentrations. Its local irritat ing effect, however, was described as very slight. It was estimated that VC concentrations of 7-10 percent in air (70,000-100,000 ppm) would be narcotic to humans in a short time. Mr. Spirtas is a candidate for the Dr. P. H. degree in Biostatistics at the Uni versity of North Carolina. He has been a biostatistician with the Occupational Health Studies Group since 1972 and has had prior experience with the Na tional Air Pollution Control Administration. He is cur rently serving on subcom mittees of the Intersociety Commitee and the A5TM. Dr. McMichael is an Assist ant Professor of Epidemiol ogy at the University of North Carolina. His recent research has focussed on work-related health haz ards, particularly cancer mortality excesses, within the rubber industry. His previous doctoral research, carried out in Australia, was in the field of social epidemiology and mental health. Dr. Gamble holds a M.A. degree in Biology from Oberlin College, and an M.S.P.H. and Ph.D. in In dustrial Hygiene and Tox icology from the University of North Carolina at Chapel Hill. His doctoral research was an environmental-epi demiologic study of respi ratory function and symp toms in selected rubber workers. Dr. Van Ert is currently serving as an industrial hygienist and laboratory di rector for the Occupational Health Studies Group in Chapel Hill. Dr. Van Ert did his graduate work at the University of North Car olina at Chapel Hill and holds a Master of Science in environmental chemistry and biology and a Ph.D. in industrial hygiene and tox icology. American Industrial Hygiene Association Journal 779 SL f Two cases of VC gassing were reported in Great Britain.1 A process worker, standing outside a polymerization vessel and washing it with a water stream, suddenly collapsed across the open manhole. Subsequently, he experienced tightness of the chest, nausea, abdominal pain and headache. The second case, a maintenance worker, was overcome while repairing a VC leak. A study of clinical manifestations of 186 male PVC workers identified the presence of the "narcotic syndrome", asthenic nervous symptoms, Raynaud's syndrome and hepato megaly.4 However, the absence of a control (unexposed) population, or of comparisons between PVC workers experiencing different exposures, hampered the interpretation of these findings. On studies of VC as a possible anesthetic agent, using dogs as subjects, incoordinated muscular activity of the extremities and serious cardiac arrythmias were observed.5 Angioneurosis (spasm of peripheral blood vessels) in PVC polymerization workers has also been reported.6 The association of vascular spasms (manifested as Raynaud's phemonenon) with acroosteolysis has been reported by several investigators.7'8 The pathogenesis of these peripheral vascular manifestations has been variously interpreted as disordered vasomotor control via the hypothalamus,4 as a general central nervous system dysfunction,9 or by localized physico-chemical action in conjunc tion with specific individual susceptibility.7 Overall, the literature of the recent dec ades has suggested that most of the toxic effects of exposure to VC act via the central nervous system occurring most obviously at very high exposure concentration, and that both acute and chronic toxic effects occur. A recent cross-sectional health survey of 354 VC workers related the proportion of workers who had experienced acute central nervous system symptoms at work (headache, light-headedness, dizziness and loss of con sciousness) with their duration of employ ment.10 Although a dose-response relationship was implied in the report, this relationship could be partly or wholly explained by differ ences between the several employment-duration categories either in age or in duration of opportunity to experience these acute symptoms. Methods In 1974, a mail questionnaire survey was con ducted on a population of VC polymerization workers, seeking information about the fre quency of occurrence of job-related symptoms. These self-reported symptoms were related, by the respondent, to the specific type of job he was engaged in within the polymerization plant, and to the calendar year of their occur rence. (See Appendix 1.) Further, equivalent information was collected, in identical fashion, from a sample of rubber workers in the adjoin ing plant. (See Appendix 2.) These two aspects of the study allowed internal compari sons between polymerization workers in different jobs and at different time-periods (i.e., different exposure situations), and also an external (in the sense of vinyl chloride exposure) comparison of polymeriza tion workers with rubber workers. In this way, the frequency of self-reported symptoms could be more meaningfully interpreted. The validity of self-reported work history information has been tested in another study being done on the workers in this plant and found to be accept able. In the previous study, a sample of work ers were asked to indicate the dates and job titles for all jobs held. This information was compared with company work histories. The results showed that workers were able to recall most jobs they had had and that their memory as to dates was fairly reliable within the last ten year period, but worsened in trying to go back further than ten years. Information was sought on the frequency of eight symptoms. The first four (dizziness or light-headedness, passing out, nausea or "sick feeling" and headache) were symptoms of central nervous system disturbance, similar to those examined in the study of Lilis, et al.* The next two sought symptom information for peripheral neurovascular and neuromuscular disturbance ("pins and needles" and muscle weakness). The seventh symptom referred to local irritation (of eyes, nose and throat), and the eighth symptom referred to the generalized asthenia ("run-down" feeling) originally reported by Suciu et al.4 Questionnaires were mailed to all workers involved in the vinyl chloride plant, and to a sample of the workers in the adjoining rubber plant. The latter population consisted predom inantly of rubber workers, with a small minor ity (five percent) of polyvinyl chloride (PVC) fabrication workers. The statistical sampling 780 October, 1975 SL 041436 SUF The company have worked cent evider be health V experience!7 1. Clock Card 2. List the tn. Dept. No. (Example: 093 &t .093. b. . c. . 093 093 d. . 393 3. If you exP related to for each s Symptom Dizziness, lif. headedness Passing out Nausea, "sick' Headache "Pius and nee. in arms or le Muscle weakne Irritation of nose or throa General fatig "run-down" fe Other symptot . 10^*3^ 4, If you ur on the rf diagnosi' relation! American Industrial is i DiiMliitii Ml- 'plums. k:alI jb "ion air;ilcnt 'sliion, nJjoin- 'U' ipari- ls iLii/away, could lidity 1 has n (lie pti >rkh is 1 lie ftall 'lory d icy or iek i to lor ir lo mtl veil ers i i M1- n SUPPLEMENTARY QUESTIONNAIRE: CHEMICALS (RESIN) DIVISION Appendix 1 The company has provided the university with the names of all persons who have worked in the chemicals division (i.e. depts. 093, 836 or 733). Re cent evidence suggests that some chemicals involved in PVC manufacture may be health hazards. We would therefore like to ask you about your health experiences during your employment in that area. 1. Clock Card Number . .^....................................... 2. List the main types of job that you have done in the chemicals division: Dept. No. (Example: 093 013 . b. ><13. c. 0^ "d. e. QH Type of job e.g. Dryer operator .&. . Svrvcxu + UH.lih . 7. Centy'fiqQ'* 10. L-J n e . S .TJeJiej-ib. . I+U Time period e*g. 1954-59) . . J9&. . . .. .. . . ... . . . iWritf. . . . l%7rW 3, If you experienced any Illness or adverse health effects that might have been related to working in any of these above jobs, please describe the symptoms For each symptom that you experienced, record an answer in Columns A and B. Symptom Dizziness, light headedness Passing out A. In whiefi one of the above lobs did this symptom occur most? (Circle one letter to identify the job, as iisted by you in Question 2.) 0b ab Qc e cde Nausea, "sick" feeling Headache "Pins and needles" in arms or legs 0b c Oe Q O <D e abcde B. How often did ithe symptom occur? Oncei Occasionally Often 1 (S) 3 123 1 3 1 20 1 23 Muscle weakness abcde 123 Irritation of eyes, nose or throat Qa b e o1 2 General fatigue, "run-down" feeling Oa 0 d> e 1 2 Other symptom: . Jl .01 Vi) a c d e 1 2 bcde 123 4. If you underwent medical examination for any of these symptoms, please describe, on the reverse side of this page, the following: medical tests carried out, diagnosis or explanation given to you, type of treatment, any known or suspected relationship of symptoms to factors in your work environment. nLEAS|aBE_SURE^TO=ENCLOSE_THIS_PAGE_Al,SO, WHEN RETURNING THE MAIN QUESTIONNAIRE. ' 075 American Industrial Hygiene Association Journal j SJ~ 041437 781 SUPPLEMENTARY QUESTIONNAIRE: OTHER TYPES OF SYMPTOMS Appendix 2 Recent evidence within the PVC-manufacturing industry has emphasized the potential health hazards of exposure to chemicals in the work environment- In order to conduct an initial survey of adverse health effects due to chemical exposure, we are asking a sample of workers to answer these extra questions, in addition to the main questionnaire. 1. Clock Card Number 2. List the main types of jobs that you have done while employed in the work areas that you recorded on page 5 of the main questionnaire: Dept. No. Type of job Time period (Example: 115 a. . na> e.g. Warm-up millman e.g. 1954-59) . tfkuiaMAM .OpMOiftU........................... IQMrbtf b. //.A* /ay. LJcUaJ..................... 7/r7f d e 3. If you experienced any illness or adverse health effects that might have been related to working in any of these above jobs, please describe the symptoms. For each symptom that you experienced, record an anwer in Columns A and B. Symptom A. In which one of the above lobs did this symptom occur most? (Circle one letter to identify the job, as listed by you in Question 2.) Dizziness, light headedness 0a b de B. How often did the symptom occur? Once Occasionally Often 1o 3 Passing out ab c de 12 3 Nausea, "sick" feeling Headache 0a. b o c de de 2 12 3 Q "Pins and needles" in arms and legs abc de 12 3 Muscle weakness 0a b de 1Q 3 Irritation of eyes, nose or throat 0 c de 12 O General fatigue, "run-down" feeling 0a b de 1O 3 Other symptom: .a b c de 12 3 ,a b c de 12 3 4. If you underwent medical examination for any of these symptoms, please describe, on the reverse side of this page, the following: medical tests carried out, diagnosis or explanation given to you, type of treatment, any known or suspected relationship of symptoms to factors in your work environment. SLEASE^BE^SU^^ENCLOSE^THIS PAGE ALSOa iWttHaEsNaBRaEaTsUaReNsINsGBsTsHeEiisMaAeIkNiiSQBUnEaSsTcIOwNMNiAsIaRsEs. 782 October, 197S si 4U*3*8 the potential order to exposure, we Edition to 1II work areas ' period 1954-59) otrTO,. Hr7f. _ Liquid Monomer V.C. Polymerize mix with copolymer High (N=226) Low (N-245) Control (N-227) ve heen 1't.oms; 'd B. ten did the " occur? ionall-y Often 03 Service & Utility (06) (10) Trucking, Loading (11) 3 O D "scribe, "'it, ispected I BE, October. 1975 Maintenance (07) ( ) Denotes O.T. from Table I. Figure 1 - Schematic flow diagram of vinyl chloride polymerization. procedure for choosing which rubber workers to include in the study entailed taking a sys tematic one-in-twelve sample with a random starting place from an alphabetical listing, thereby assuring a representative sample of the rubber plant workers. The response rate for current vinyl chloride workers was 200/298 (67 % ) while the rate for the current rubber plant workers was 89/212 (42% ). This dif ference in response rates may be due to the adverse publicity on vinyl chloride exposure. In order to evaluate the exposure experi ences of each worker, the individual jobs reported on the questionnaire were grouped into 12 categories (sec Figure 1). The princi ples and procedures of regrouping production- oriented categories into health-oriented cate gories has been recently described.11 Industrial hygiene analytic sampling re sults are present in Tables I, II and III. Table I is based upon data taken by company personnel before major changes were made in controlling levels of vinyl chloride exposure, and should be representative of the relative levels that workers were exposed to during their earlier work history. Table II is based upon company data taken during the installation of environ mental controls. Table III contains data taken by University of North Carolina personnel in 1975, after major environmental controls had been installed. By comparing these three tables, it can be seen that the absolute levels of vinyl American Industrial Hygiene Association Journal 783 su TABLE I VC Exposure Index Company Data Grab Sample-Gas Chromatograph (Prior to 1974) O.T. 01 03 04 05 08 06 06 02 07 7 09 09 09 11 DESCRIPTION Tank Farm Operators Reactor Operators High Pressure H,0 Stripper Operators Relief Operators Supply Men Service & Utility Janitors Maintenance Pilot Plant Janitor Centrifuge Operators Dryer Operators Bagger Operators Truck Operators EXPOSURE INDEX H H H H H* H` M M M M L L L L (L) tow 0-50 ppm (M) Medium 50-200 ppm (H) High 200 & Above chloride have decreased, yet the gradient of exposure between job categories appears to have remained the same. Based on these sampling results, and on general industrial hygiene evaluation of the plant operation, it was possible to create a further grouping of the 12 occupational cate gories in Figure 1 into three exposure cate gories representing high, low, and essentially no exposuie to VC. Figure 1 illustrates this grouping in the context of a flow diagram of the vinyl chloride manufacturing process. By defining three exposure categories as noted, several purposes were achieved. If, in fact, VC exposure was causative for some or all of the given symptoms, then it would be natural to expect a dose-response type of pattern between exposure categories and symp tom response. Because of the way exposure categories subdivide the population, any poten tially confounding effect of other chemicals should be accounted for. Thus, if rubber work ers, for example, had more headaches because of solvent exposure, the effect would be to lessen the chances of a dose-response pattern (as defined in the study) emerging. The ma jority of workers spent time in more than one exposure category. In addition, within the VC plant, the respondents had no knowledge of which exposure category they worked in. Thus, many of the workers served as "controls" on themselves for different jobs, and in addition, some of the VC workers will serve essentially as "controls" on other vinyl chloride workers. It is possible that no real difference in symp toms existed and the workers who answered the questionnaire gave biased answers because of their apprehension of working with vinyl chloride. Because of the internal consistency of the results as well as the external consistency with the findings in previous studies, this possibility is unlikely. Results Since the symptoms under consideration are all transitory and reversible, an acute effects model has been employed. Thus, each reported person-job held, rather than each person, is treated as the unit of observation. Each person has been able to contribute to the analysis his experience in more than one job (the average is approximately three jobs per person, and often a person has experienced jobs in more than one O.T. group). By treating the job as the unit of observation, there is a tendency to dampen the effect of individual variability in symptom experience since workers are often acting, in effect, as their own controls. Any TABLE II Exposure Data For VC and PVC Workers in a VC and PVC Plant Company Data (Second quarter 1974) O.T. 01 04 03 08 06 03 11 08 09 09 09 11 07 OTG High High High Low High Htgh Low Low Low Low Low Low High VINYL CHLORIDE MONOMER EXPOSURE # WORKERS HIGH EXPOSURE (most readings--20-30 ppm) 3 Tank Farm Attendant 6 High Pressure Water Cleaner MEDIUM EXPOSURE most readings--10-20 ppm 10 Reactor Operator^Bldg. 2 9 Relief Operator--Bldg. 2 36 Service Utility LOW EXPOSURE most readings--0-10 ppm 4 Reactor Operator--Bldg. 1 12 Shipping Operator 3 Relief Operator--Bldg. 1 12 Centrifuge Operators 4 Dryer Operators 8 Baggers 9 Truckers 48 Maintenance MAXIMUM READING <40 >50 >50 50 >50 <20 <30 <20 <40 <40 <10 <10 >50 i 784 October, 1975 SL 041440 05 08 09 09 09 02 02 03 05 08 07 10 12 (1) Ai (2) P' (3) N' CDhaerrei )i;r Gmb in)> O.i Mini Low Ar ImJ i F American Industrial TABLE III Vinyl Chloride Sampling Results Chemical Plant (Buildings I & II) and Rubber Plant (3-24 to 3-26-75) O.T. 03.04 location BUILDING 1 Reactor Room Operator's Desk Desk Cluster (4 tubes) Mezzanine General Room Typical Peak VINYL CHLORIDE LEVELS (PPM) AS DETECTED BY:* V.C. INDICATOR GRAB SAMPLE CHARCOAL MIRAN TUBE BY G.C. TUBE 152 155 21 2 15 2 93 2 20 5-6 ---- j-,- 4*5 --_ _ 11.7 7.0. 11.7, 10.9, 7.4 47.4 11.7 05 Stripper Area Operator's Desk 33 5 24 7 200 (PVC & CoPolymer I Side) _ ___General Area 24-32 08 Relief Area Operator's Desk General Area 28-36 3 ---- 1.43 -- 09 Second Floor Spray Driers General Room 3 __ -- 09 Bagging Area Lines 4,6.7 9-163 ---- ------- 09 Bulk Drying Area (Rotary Driers) Panel Board Area Bagger: Line 13 1.5 3.5 02 Transfer Pump House No Pumping of V.C. Pumping of V.C. -- 02 Purification Room -- 3 ---- 8 -- *-- 20 -- 6.1 -- 5.8 300u> 26.1 BUILDING II 03 Reactor Room Operator 5,2I!> 05 Stripper Level Operator ___ __ . -- 8.8TM 08 Relief Area Operator .... __ -- 5.0 07 Maintenance Pipe Fitter -- __ 5.9' RUBBER PLANT 10 Film and Sheeting Mezzanine Area Banbury Mill #4 Area 12 General Room 1.8-4.0 3.2 --6.5 Not Detectable I0' w -- 5.0< ___ (1) Air-supplied respirator required. (2) Personnel sample (3) Not breathing zone 'Definitions for V.C. Analytical Techniques Charcoal Tube Samples: Rases and vapors adsorbed (q activated charcoal and subsequently analyzed bj gas chromatography by OHSG laboratory personnel; reflect time weighed averages. Grab Samples: gases and vapors drawn into sampling bottles, bottles sealed, and contents analyzed by injecting portion of contents into gas sampling port of AID portable gas chromatograph. Data obtained in field low due lo V C. toss from sampling bottles Miran Infra-Red Analyzer: continuous monitoring of V.C. by the direct reading instrument. Lower detectable limit of instrument: O.H-l.O ppm. Area samples minimum of 10 minutes. Indicator Tubes: Gastec indicator lubes with range of 0.5-14.0 ppm. Easy to read and correlate well with other methods (i.e., charcoal tubes and Miran). American Industrial Hygiene Association Journal SL 041441 785 differences found with this type of model can then be more easily ascribed to real differences in environmental exposure. Tables IV-1 through IV-8 present the basic data used in the analysis. Limiting the statistical analysis to white males (who make up 95 percent of the plant population) was necessary in order to provide adequate cell sizes. In Tables IV-1 through IV-8, the cate gories of response are labelled 0, 1, 2 and 3 to correspond with the questionnaire replies as to frequency of symptoms: never, once, occasionally and often. By examining Tables IV-1 through IV-8, it can be seen that a certain type of pattern is evident for many of the symptoms. This pattern is most evident for Table IV-1 (corresponding to symptom 1). For the 255 high exposure group jobs recorded on individuals' question naires, 94 (37 percent) were reportedly asso ciated with an episode of dizziness or light headedness at least once. The comparable fig ures for the low exposure group were 28 out of 236 jobs (12 percent) and the values for the rubber worker jobs were 21 out of 214 jobs (10 percent). The purpose of presenting TABLE IV-1 Exposure Category by Frequency of Dizziness, Light-headedness O.T. GROUP NEVER ONCE High Exposure Low Exposure Rubber Worker Column Total 161 (63.1) 20S (68.1) 193 (90.2) 562 (79.7) 8 (3.1) 1 (0.4) 5 (2.3) 14 (2.0) Row percentage applies to Tables IV-1 through IV-8. OCCASIONALLY 72 (28.2) 25 (10.6) 14 ( 6.5) 111 (15.7) OFTEN 14 (5.5) 2 (0.8) 2 (0.9) 18 (2.6) ROW total 255 (100.0) 236 (100.0) 214 (100.0) 705 (100.0) O.T. GROUP High Exposure Low Exposure Rubber Worker Column Total TABLE IV-2 Exposure Category by Frequency of ___________ Passing Out___________ NEVER 244 (95.7) 231 (97.9) 212 (99.1) 687 (97.4) ONCE 8 (3.1) 3 (1.3) 2 (0.9) 13 (1.8) OCCASIONALLY 2 ( 0,8) 2 < 0.8) 0 ( 0.0) 4 ( 0.6) OFTEN 1 (0.4) 0 (0.0) 0 (0.0) 1 (0.1) ROW total 255 (100.0) 236 (100.0) 214 (100.0) 705 (100.0) O.T. GROUP High Exposure Low Exposure Rubber Worker Column Total TABLE 1V-3 Exposure Category by Frequency of Nausea, "Sick" Feeling NEVER 193 (75.7) 212 (89.8) 190 (88.8) 595 (84.4) ONCE 5 (2.0) 0 (0,0) 3 (1 4) 8 (l.D OCCASIONALLY 42 (16.5) 20 ( 8.5) 16 ( 7.5) 78 (li.D OFTEN 15 (5.9) 4 (1.7) 5 (2.3) 24 (3.4) ROW TOTAL 255 (100.0) 236 (100.0) 214 (100.0) 705 (100.0) O.T. GROUP High Exposure Low Exposure Rubber Worker Column Total TABLE IV-4 Exposure Category by Frequency of Headache NEVER 187 (73 3) 194 (82.2) 183 (85.5) 564 (80.0) ONCE 3 (1.2) 0 (0.0) 0 (0.0) 3 (0.4) OCCASIONALLY 42 (16.5) 30 (12.7) 25 (11.7) 97 (13.8) OFTEN 23 (9.0) 12 (5.1) 6 (2.8) 41 (5.8) ROW TOTAL 255 (100.0) 236 (100.0) 214 (loo.O) 705 (100.0) 786 October, 1975 these spons impot follov weigh respei swere given twice Figui perfo quest work Ante these tables is to describe the patterns of re sponse for each symptom. A measure of the importance of these results is given in the following statistical analysis, which employs a weighting scheme with values 0, 1,2 and 3 respectively. (For example, if a person an swered "occasionally" to the frequency of a given symptom, this response was treated as twice as important as an answer of "once".) Using the Occupational Title Grouping in Figure l, a series of statistical analyses were performed to determine the relationship of the questionnaire responses of these groups of workers to their relative exposures to VC. The differential exposures to VC between the three study groups account for only one of the many possible hazards. Thus, the con trols in this study must be viewed as "internal controls" who may be exposed to other chemi cal exposures. However, the consistency and strength of the results of this study in conjunc tion with similar results from previous studies, tend to confirm the suspicion that the causative agent in this study is vinyl chloride. Since questionnaire answers are recorded in categories rather than a continuum, it was therefore appropriate to use statistical analysis techniques designed for categorical data. The O.T. GROUP High Exposure Low Exposure Rubber Worker Column Total O.T. GROUP High Exposure Low Exposure Rubber Worker Column Total O.T. GROUP High Exposure Low Exposure Rubber Worker Column Total O.T. GROUP High Exposure Low Exposure Rubber Worker Column Total TABLE IV-5 Exposure Category by Frequency of "Pins and Needles" in Arms or Legs NEVER 219 (85.9) 216 (91.5) 201 (93.9) 636 (90.2) ONCE 2 (0.8) 0 (0.0) 1 (0.5) 3 (0.4) OCCASIONALLY 21 ( 8.2) 12 ( 5.1) 7 ( 3.3) 40 ( 5.7) OFTEN 13 (5.1) 8 (3.4) 5 (2.3) 26 (3.7) TABLE IV-6 Exposure Category by Frequency of Muscle Weakness NEVER 211 (86.7) 224 (94.9) 194 (90.7) 639 (90.6) ONCE 6 (2.4) 0 (0.0) 2 (0 9) 8 d.i) OCCASIONALLY 19 ( 7.5) 11 ( 4.7) 13 ( 6.1) 43 ( 6.1) OFTEN 9 (3.5) 1 (0.4) 5 (2.3) 15 (2 1) TABLE IV-7 Exposure Category by Frequency of Irritation of Eyes, Nose or Throat NEVER 166 (65.1) 158 (66.9) 159 (74,3) 483 (68.5) ONCE 2 (0.8) 1 (0.4) 0 (0.0) 3 (0.4) OCCASIONALLY 49 (19.2) 47 (19.9) 43 (20.1) 139 (19.7) OFTEN 38 (14.9) 30 (127) 12 ( 5.6) 80 (11.3) TABLE IV-8 Exposure Category by Frequency of General Fatigue, "Run-Down" Feeling NEVER 177 (69.4) 181 (76.7) 171 (79.9) 529 (75.0) ONCE 1 (0.4) 0 (0.0) 2 (0 9) 3 (0.4) OCCASIONALLY 39 (15.3) 29 (12.3) 28 (13.1) 96 (13.6) OFTEN 38 (14.9) 26 (11.0) 13 ( 6.1) 77 (10.9) ROW TOTAL 255 (100.0) 236 (100.0) 214 (100.0) 705 (100.0) ROW TOTAL 255 (100.0) 236 (100.0) 214 (100.0) 705 (100.0) ROW TOTAL 255 (100.0) 236 (100.0) 214 (100.0) 705 (100.0) ROW TOTAL 255 (100.0) 236 (100.0) 214 (100.0) 705 (100.0) American Industrial Hygiene Association Journal S1 0414*3 787 TABLE V Results of Statistical Analysis of Eight Symptoms SYMPTOM 1. Dizzyness, lightheadedness 2. Passing out 3. Nausea, "sick" feeling 4, Headache 5. "Pins and needles" in arms and legs 6. Muscle weakness 7. Irritation of eyes, nose or throat 8. General fatigue, "run down" feeling dose-response pattern yes borderline yes yes yes no borderline yes procedure used in this analysis was the Cate gorical Linear Data Model.12 (This procedure makes use of the cell proportions in a weighted least squares analysis of complex contingency tables.) The information on frequencies of re sponse in Table 1V-1 through IV-8 was input into the categorical data computer program (LINCAT). The statistical model is of the form: F(ir) = X/3 where 7T is the vector of empirical proba bilities derived from each of the frequency Tables IV-1 through IV-8 X is an arbitrarily chosen design matrix /3 is the (unknown) parameter vector. Although it is beyond the scope of this paper to discuss the statistical procedure, it is suffic ient to indicate that this approach to categorical data analysis is analagous to the General Linear Model approach13 to the statistical analysis of continuous data. The first set of analyses examined which of the symptoms, if any, exhibited a significant dose-response type of relationship with pre sumed VC exposure. A dose-response pattern was deemed to exist if the (weighted) symptom response rate for O.T. Group 1 was significant ly* greater than for O.T. Group 2 and, in addition, the response rate for O.T. Group 2 was greater than for O.T. Group 3. This takes account of both the direction and magnitude of the respective differences. It was next decided to determine the importance of age at time of exposure to each job as a possible confounding variable. The study population was subdivided by age at exposure into ^ 36 and 37 at time of entry into job. For those symptoms where adjustment by age produced a better fitting model (in the statistical sense), the adjustment Significance here is defined as: probability of a Type 1 error (it) ^.05. for age was made in Table V. Only the statisti cal models for symptoms 1, 7 and 8 were enchanced by age adjustment. In each of these three cases age was not a significant variable in its own right, but its inclusion resulted in a better fitting model. (The direction of the dif ference in age groups for all three of these symptoms indicated that younger people tended to exhibit these symptoms more fre quently than older people, although this dif ference was not statistically significant. This may be due to a possible adaptation by older workers or to self-selection by the more sensi tive workers.) The results presented in Table V show a clear dose-response relationship for symptoms 1, 3, 4, 5 and 8. This means that in addition to there being a significant difference in the re sponse rate for all VC workers when compared with rubber workers, there is a further differ ence between VC workers when they are cate gorized into high and low exposure groups. For symptoms 2 and 7, the direction of the response gradient was of a nature to indicate a dose-response pattern but the magnitude of the differences was not quite great enough to meet the criteria for statistical significance. These two symptoms have consequently been labelled as "borderline". Symptom 6 (muscle weak ness) did not exhbit a dose-response gradient. Preliminary attempts at utilizing informa tion on smoking, length of service and calendar period of exposure have yielded little to date. Further work is being done in these areas. In summary, there appears to be a doseresponse relationship between certain acute symptoms (predominantly neurological) and level of vinyl chloride exposure. The responses of the workers may have been subject to bias due to prior knowledge of the hazard of vinyl chloride exposure. However, the cxistance of the dose-response pattern for certain symptoms, as well as the absence of such a pattern for one symptom tend to support the validity of this study. 788 October, 1975 * i *A * istiesc . i r- HS vl4 ns lit ed j Acknowledgment * j We arc indebted to Mr. Earl Arp for his help j in obtaining environmental samples and for | discussions concerning the industrial hygiene f aspects of the paper, to Professor Lawrence j Kupper for his review of. the statistical meth< odology, to the data processing and laboratory staffs of OHSG for their steadfast efforts in our ; behalf and to Mrs. Wanda Dodson for manu script preparation. Special thanks are due to Mr. Louis Beliczky, Director of Industrial Hygiene, United Rubber Workers Union, and to the Corporate Medical Director of the com pany under study for their advice and constructive criticism. References 1. Danziger, H: Accidental Poisoning by Vinyl Chloride: Report of Two Cases. Canadian Med. Assoc. J. 82:828 (1960). 2. Lehmann, K. B,, F. Flury: Toxicology and Hygiene of Industrial Solvents. Williams and Wilkins, Baltimore (1943), 3. Annual Report of the Chief Inspector of Factories for the Year 1951. London: HMSO Cmd. 8772, March, (1953). 4. Suciu, I,, J. Drejman, M. Valaski: Contribu tions to the Study of Affections Caused by Vinyl Chloride. Medicina Interne 15:967 (1963). 9. Oster, R, H., C. J. Carr, J. C. Krantz, M. J. Sauerwald: Anesthesia XXIV. Chemical Con tribution of Hydrocarbons and Cardiac Automaticity. J. Pliant, and Exper. Therap. 97:1 (1949). 6. Filatova, V. S., E. S. Gronsberg: Hygienic Working Conditions in Polyvinyl Chloride Tar Plants. Cigi Saint. 38 (1957). 7. Wilson, R. E., W. E. McCormick, C. F. Tatum, et al.: Occupational Acroosteolysis. JAMA. 201: 577 (1967). 8. Dinman, B. D., W. A. Cook, et al.: Occupation al Acroosteolysis. 1. An Epidemiological Study. Arch. Environ. Health. 22:61 (1971). 9. Filatova, V. S., E. S. Gronsberg. Cited in: Anghelescu, F., et al.: Clinico-Pathogenic Con siderations of Raynaud's Phenomenon in Work ers of the PVC Industry. Medicina Interna. 21: 473 (1969). 10. Lilis, R. et al.: Prevalence of disease among vinyl chloride and polyvinyl chloride workers. Paper presented to New York Academy of Sci ences Conference on Vinyl Chloride, New York, May, (1974). 11. Gamble, J. F., R. Spirtas: Job Classification and Utilization of Complete Work Histories in Occupational Epidemiology. J. Occup. Med. In press. 12. Grizzle, J. E., C. F. Starmer, G. G. Koch: Analysis of Categorical Data by Linear Models. Biom. 52:489 (1969). 13. Starmer, C. F. and J. E. Grizzle: A Computer Program for Analysis of Data by General Linear Models. Inst. Stat. Mimeo Series No. 560 (1968). he et eel it. i;ir I if ms. 'vs American Industrial Hygiene Association Journal SL 041445 789