Tobacco Documents Online

Limitations to the Use of Employee Exposure Data on Air Contaminants 295 4 Addjtional Factors Influencing Exposure to a Substrate 4.1 Part-time Exposure It is rather unusual that an employee is occupied with a single operation during the whole workday or shift. As an example, in metal arc welding of nonalloy steel in workshops, the geometric mean of the arc time factorwas 22%[31]. This applies for employees engaged full-time in welding, i.e. almost 80% of the time is used up for preparations before the welding or grinding, etc. after the welding. Other causes of limited exposure which should be considered are the use of respirators and rotation of personnel, practised in e.g. the control of lead ex- posure in many countries [7]. If uptake is estimated from concentration data of air contaminants without considering limited exposure an appreciable positive bias will result. 4.21nf1uence on Uptake of a Sttbstauce of Short-ternt Variations in Exposure It is quite obvious that variations in concentrations ofsubstances in the airwithin a day or shorter periods is important to the results. A single breath ofair contain- ing a very poisonous gas in a sufficiently high peak concentration may be fatal, although the average concentration of this gas over one day may be tolerable. This is an area of concern for accident prevention, however, and has very little to do with the monitoring of gases and vapours in order to check compliance with exposure limits. In the simplest exposure model, response or effect is studied as a function ofa single dose. This model of course is far from reality. A more complicated but still unrealistic model implies repeated doses ofthe same size. In reality the exposure will be composed of a complicated pattern of episodes with repeated doses of varying magnitude interrupted by breaks of varying lengths without exposure. Furthermore all exposed subjects in reality have an individual exposure pattern. As has already been pointed out by Roach [23, 24], the durations of peak con- centrations in relation to the biological half-life of the substance should be con- sidered in judging exposure to air contaminants. The critical durations of peak concentrations and interruptions in the exposures are of the same magnitude as the biological half-life of the substance in the body. If the peak durations and the interruptions between peaks are much shorter than the biological half-life, the substance eventually will reach a concentration level in the body corresponding to the equilibrium at the average concentration in the air. If the durations of peaks are much longer than the biological half-life of the substance in the body, there.will be enough time for the substance to accumulate to a concentration level in the body corresponding to the height of the peaks. The uptake will come into "resonance" with the environment, cC Fig. 5, based on models suggested by others [13, 15, 22], further discussed in reference [27]. The implication of Fig. 5 is that although the average concentration is the same in the different exposure cases the uptake will be very different. It is possible that different organs in the body will respond differently to the exemplified exposure cases, e.g. one organ may respond to the area under the concentration-time curve, while a second organ may respond to the peak heights. Very little is known about this, but it is

296 concaninanc concentration in air !relauve scale) ~nair ul ~ort ntam~naot n tnr Pody t.elnti~c scale) ~, 10 15 20 25 (unit = halt-tme) U. Ulfvarson Fig. 5. Examples of "resonance" with environment, i.e. unfavourable exposure situations. Simple lirsl order kinetic with main uptake and excretion via the lungs is assumed [22, 271. In all exentplified exposure situations the average air concentration of the contaminant is the same (= I in the relative scale). The peaks of the body concentration will corresponded more and more closely to the peaks in the concentration in the airwhen the exposure episodes have a lenght at least-as long as the biological halt'-Iife of the substance in the body observe that the diagrams should be read simultaneously and show what is assumed to happened at the same time in the inspiration air and the body! suggested that when available data allow an investigation, the distribution of exposure over time should be considered. The possible bias is negative since the measured concentration is lower than the effective concentration during short periods. 5. Conclusions Sampling strategies have been discussed almost exclusively with the view in mind of checking compliance with exposure limits or to some extent finding a basis for or checking control measures to decrease the exposure. Except for tt few recent contributions [11,12,32j almost no eti'orts have been made to develop sampling strategies in order to describe the true uptake pattern of substances in the bodies of exposed employees and for obvious reasons: the ethical problem involved in prospective epidemiological studies [l6] and the prohibitively high costs in making measurements when the future use of the measurements are uncertain.

Limitations to the Use of Employee Exposure Data on Air Contaminants 297 Table 4. Bias in the estimation of uptake ofa substance in agroup ofemployees when uptake is deduced uncritically from various sources of information. +=means that uptake is overrated, -=means that uptake is underrated in comparison with probable true uptake Premises of data Cause of bias Probable Validity sign of of sign bias of bias Measurements and uptake in the same period and work place i) Identification of the sub- stance Biased sampling in the locality to get above detection limit Biased sampling among employees Biased sampling among employees Biased sampling in the locality to Gnd "hot spots" Biased sampling among employees + 1 2 + 3 + or - + 2 + I 0 2 - 2 + 2 i?) "Worst Case" 3) Monitoring daily averages t~) General work area sampling Biologic sampling .5) 6,I Unconditioned, regular No additional bias check-ups No repeated measurement are made when the first result shows compliance Measurements and uptake in the same period and "sinritar" work place Biased sampling (the first result may have been un- usually low) Biased sampling among enterprises - - - - - - - - - - - - - - - - - - - - - - - - - - - Measurernents and uptake in di(ferent periods ol'time Measurements made in a period before uptake Measurements made in a period after uptake Measurements and uptake in different seasons Other circumstances High exposure limit during measurements Rotation of employees to unexposed work Use ot' effective respirator 3 Unfavourable exposure ~ ; pattern ll~~ Hard physical labor Technical development + 2 Technical development - 2 Regular variations + or -. Few data. Biased - 1 interpretation Invalidation of data + 3 Invalidation of data + 3 "Resonance" (cf. text) 2 Increased lung ventilation - 2 Code referring to the estimated validity M'the suggested sign of bias: 3 =self evident; 2=a conclusion with some reservation; I =an educated guess

298 U. Ulfvarson Table 5. The consequence of the sign of bias in measurements on the error in standard setting Sign of Interpretation Effect on standard setting Consequence bias + Observed conc. too high Standart too high "Health error" - Observed conc. too low Standard too low "Economic error" To use the limited data available the investigator must have a reasonable idea of the sign of the bias in the estimated uptake. The bias may be due to a lack of representativeness or to additional circumstances, in the work situation. In Table 4 the probable bias of estimated uptake deduced from exposure data in an uncritical way is summarized. Some of the conclusions in Tab,le 4 are self evident, others must be regarded more or less with reservations as discussed in some details in Section 2. The opinion ofthe author about the validity of the sug- gested signs of bias in Table 4 is expressed in the form of a code in the table. As has already been stated, ifthe exposure is overestimated the risk will be under- estimated and vice versa. An inspection ofthe summaries ofsings in the errors in Table 4 seems to suggest that overestimation of the uptake will be the most com- mon outcome of judging the exposure from old data. The epidemiologist using old exposure data may use Table 4 as a checklist and try to find out the premises of his data and thus the most probable sign of bias. The implication of an over- rated uptake is that exposure limits set will tend to be too high and the risk will be underrated ("health error"), cf. Table 5. It may be possible to some extent to counteract this simply by applying saier (=lower) exposure limits, but this may be possible only when the technical feasibility is obvious. In the long run there is no natural "safe side," since an exposure limit which is too low will cause un- necessary costs "economical error") affecting the possibilities of limiting more critical exposures. In the future, tiled exposure data should be accompanied by all information necessary to judge their validity. The following factors should be considered. (a) The name and nature of the operation(s) going on, products used and manuftctured (declaration of content), contaminants formed. (b) The average proportion of time used for the operation per day, week, year. (c) Regular use of respirators, rotation of employees, notation of hard physical labour of the employees. (d) Why, when, where and how the sampling was performed. (e) Analytical method. (f) Exposure limit at the time of sampling. References 1. Ahlmark A, Gerhardsson L(1981) The silicosis in Sweden since 1930 ( in Swedish). ,lrhete och I liilsa 1981 : 15. :\rbetarskveldsverket, Stockholm. 55 iid, I hilaga 2. Aitchiron J, Brown JAC (1976) The lognormal distribution with special reference to its uses in economics. Cambridge University Press, Cambridge, p 17b

Limitations to the Use of Employee Exposure Data on Air Contaminants 299 3. Andersson A (1957) Gesundheitliche Gefahren in der Industrie bei Exposition fiir Trich- lorathylen. Acta Med Scand [Suppl] 323:220 4. Andersson I-M, Rosen G (1979) Solvent exposure in paint manufactoring. An investiga- tion of 51 employees at 8 paint factories. Undersokningsrapport 18: Arbetarskydds- styrelsen, arbetsmedicinska avdelningen, Stockholm (29 sidor) (in Swedish) 5. Anonymous (1977) NIOSH manual ofanalytical methods, vol 1, 2nd Ed. DHEW (NIOSH) Publication No. 77-157-A. U.S. Dept of Health, Education and Welfare, Public Health Service. Cincinnati, Ohio 6. Anonymous (1982) Threshold limit values for chemical substances in workroom air adopted by American Conference of Governmental Industrial Hygienists. ACGIH 6500 Glenway Ave, Bldg D-5 Cincinnati, OH 45211 7. Anonymous (1979) Criteria documentation on inorganic lead. Arbete och H5lsa 1:24: Arbetarskyddsverket. Stockholm, p 55 (in Swedish) 8. Anonymous (1981) Permissible exposure limits (in Swedish). AFS 1981 :8. Arbetarskydds- verket, Stockholm. 39 sid 9. Armstrong BG (1982) Effects of approximation in exposure assessments on estimates of exposure response relationships. Scand J Work Environ Health [Suppl] 8:20-23 10. Corn M, Esmen NA (1979) Workplace exposure zones for classilication of employee exposures to physical and chemical agents. Am Ind Hyg Assoc J 40 :47-57 11. Esmen N (1979) Retrospective industrial hygiene surveys. Am Ind Hyg AssocJ 40:58-65 12. Gamble J, Spirtas R (1976) Job classification and utilization of complete work histories in occupational epidemiology. J Occup Med 18:399-404 13. Haggard HW (1924) The absorption, distribution and elimination of ethyl ether. J Biol Chem 59:753-770 14. Hald A (1960) Statistical theory with engineering applications,. John WilPy & Sons, Inc, New York London, p 783 15. Henderson Y, Haggard HW (1924) Noxius gases. Reihold Publ Comp, New York 16. I-lernberg S (1974) Epidemiologic methods in occupational health research. Work, Environ, Health 11 :59-6R 17. Hernberg S, Tola S, Vaaranen V (1978) The occupational exposure to lead has decreased in Finland. 27th meeting on occupational hygiene in Denmark. 20-22 November 1978. Arbejdsmiljoinstituttet, Arbejdstilsynet, Danmark (in Swedish) 18. Krantz S. Lindstedt S, Lundgren L, Palmqvist U, Tillman C, Ulfvarson U (1983) Inter- laboratory control of airanalyses in occupational hygiene. To be published in Arbete och Halsa. Arbetarskvddsverket, Stockholm (in Swedish) 19. Leidel NA, Busch KA, Lynch JR (1977) Occupational exposure sampling stategy manual. U..S. Department of Health, Education and Weltare. DHEW (NIOSH) Publ No. 77-173 20. Leidel NA, Busch KA, Crouse WE (1975) Exposure measurement action level and occu- pational environment variability. HEW Publication No.76-131 (NIOSH). Dept of Health, Education and Welfare, Public Health Service. NIOSH, Cincinnati. Ohio, p 38 21. MacMahon B, Pugh TF (1970) Epidemiology. Little & Co., Boston 22. Riggs DS (1963) The mathematical approach to physiological problems. Williams & Wilins Co. Baltimore 23. Roach SA (1966) A more rational basis lorair sampling programs. Am Ind Hyg Ass J 27 :1- 12 24. Roach SA (1977) A most rational basis for air sampling programmes. Am Occup I-lyg 20: 65-84 25. Sherwood RJ, Greenhalgh DMS (1960) A personal air sampler. Ann Occup Hyg 2:127-132 26. Ulfvarson U, Rosen G, Cardfeit M, Ekholm U (1975) Chemical hazards in the paint industry. dnr 4979/75 (64-sidor). Fiirgindustrins Arbetsmiljo, Branschutredning med stod fran arbetarskyddsfonden. Del l. Kemiska halsorisker (in Swedish) 27. Ullvarson U. Ovrum P (1976) Distribution of organic solvents between blood and air. Arbete och H51sa 1976: 7. Arbetskyddsverket, Stockholm, p 20 (in Swedish) 28. Ulfvarson U (1977) Statistical evaluation of the results of ineasurements of occupational exposure to air contaminants. Scand J Work Environ ltealth 3:109-115

300 U. Ulfvarson ~ 29. Ulfvarson U (1977) Chemical hazards in the paint industry. International symposium on the control of a'tr pollution in the working environment. Stockhotm 6-8 Sept. 1977. Part 11. solvents - welding, pp 62-75. International Labour Office, Geneva and Worker's Protec- tion Fund. Stockholm 30. Ulfvarson U, llallne U, Bellander T (1978) Welding problems connected with work environment. 5. Welding in stainless steel with metal arc-welding with covered electrodes and gas-shielded welding. 1. Survey of air contaminants (in Swedish). Arbetarskydds- verket, Stockholm 1978, 1-52b Arbete ocli Hiilsa 1978:6 31. Ulfvarson U (1981) Survey of air contaminants from welding. Scand J Work Environ ' Health [Suppl) 7:2-28 32. Vihma T (1981) Health hazards and stress factors in small industry - prevalence study in the province of Uusimaa with special reference to the type of industry and the occupati- onal title as classifications for the description of occupational health problems. Scand 1 Work Environ Health [Suppl) 7:3-149 Received August 26, 1982 / Accepted May 5, 1983