Document yraE9xJxKm6xq9KQEzMjJxMX2

SI39IOOOO AAA Am. Ind. Hyg. Assoc. J. 45< I0):681-688 (1984) Workplace Protection Factor Measurements on Powered Air-Purifying Respirators at a Secondary Lead Smelter: Results and Discussion W.R. MYERS'. MJ. PEACH. lll'\ K. CUTRIGHT" and \V, ISKANDER8 A Department of Health and Human Services. National Institute for Occupational Safety and Health. Division of Safety Research. Testing and Certification Branch. Morgantown. WV: industrial Engineering Dept.. West Virginia University. Morgantown. \VV A study was conducted at a secondary lead smelter to evaluate the workplace performance of the 3M W-344 and Racai AH3 powered air-purifying respirators equipped with helmets and high efficiency filters. The research protocol developed for the study has been described in a companion paper. The results of the study indicate that the mean lead concentrations, measured inside the facepiece of both PAPRs, were significantly less than the OSHA lead exposure limit of 50 gg/m'J. The means of the workplace protection factor measurements on both PAPRs were significantly less than the PAPR selection guide protection factor classification of 1000. Correlation analysis of preshift quantitative fit factors and corresponding workplace protection factors indicated no linear association between these two measures of performance. This finding suggests that for PAPRs equipped with helmets and high efficiency filters quantitative fit factors as presently determined are not indicative of the workplace protection which the respirators provide. Since the PAPR protection factor classification of 1000 was originally based on quantitative fit factors, the lack of a demonstrated association between quantitative fit factors and workplace protection as found in this study may explain why their performance was significantly less than expected. Introduction The workplace performance of the 3M W-344 and Racai AH3 powered air-purifying respirators (PAPR) equipped with helmets and high efficiency filters has been evaluated in a secondary lead smelter while worn by workers and while mounted on stationary formed heads. The specific job tasks in which the PAPRs were to be used were selected on the basis of retrospective data about personal exposure to lead. Blast furnace and reverberatory furnace operator or helper and lead ingot caster were the jobs with the highest potential lead exposure which still complied with the OSHA/ NIOSH protection factor limitation of 1000 for a PAPR.01 The blast furnace and reverberatory furnace operators and helpers were responsible for monitoring the operation of the fur naces. tapping lead or slag when needed and for fluxing the molten lead after it has been tapped from the furnaces. Fluxing is done by adding various chemicals to the molten lead which form a variety of metal salts resulting in nonmolten dross on top of the lead. The dross is "skimmed" from the top of the molten lead by hand with shovel ladles. The skimming activity is done mainly by the furnace helper. Once fluxed, alloying agents are added to the molten lead as desired. The pig caster operates a casting machine which takes the refined lead and meters it into 30 kg ingots. Test Group and Data Collection Based on exposure history, the job classifications selected for this study were furnace operators or helpers and pig casters. Two workers from each of these threejob classifica- Mention of company names or products does not constitute endorse ment by the National Institute for Occupational Safety and Health. tions during both the 7 a.m. and 3 p.m. shifts were selected to participate. No attempt was made to control the facial in dices of the workers. Two workers per job classification per shift were chosen so that the PAPRs could be tested together. A total of twelve workers were used in the study. Each worker wore both models of PAPR twice, however, for the second wearing, he was not given the same specific 3Mor Racai PAPR that he wore the first time. A complete descrip tion of the 3M Model W-344and Racai Model AH3 used in the study is given in a companion paper.<J> Because it was anticipated that relatively small, inboard leakages would occur, the inside and outside facepiece sam ples were collected over the full shift. Statistical Analysis These data were analyzed using the t-test, analysis of var iance (ANOVA) and correlation analysis. These statistical tests all require that the data being analyzed must be nor mally distributed. However, it is generally believed that environmental occupational exposure data approximately follows a lognormal distribution. Testing of the data col lected in this study showed it to be lognormally distributed as expected. A Kolmogorov-Smirnov test was run on ran domly selected cells of the ANOV A to check the assumption of normality. Those results indicated that the log trans formed data were far superior to the untransformed data in meeting this assumption. In addition to the KolmogorovSmirnov test for the analysis of variance, skewness and kurtosis of each variable distribution were calculated. The skewness is a description of the asymmetry of a distribution while kurtosis is a measure of the degree of peakedness of a American Industrial HvK'ene Association JOURNAL Copyright 1984. American industrial hygiene Association (45) 10/84 0000l621fa 681 distribution. For a normal distribution, the skewness and kurtosisare zero. The calculated values of these moments for various distributions are given in Table I. In ail cases, the log-transformed data best approximate the normal distribu tion. Accordingly, the raw PAPR field data were log trans formed and all statistical analyses were performed on the transformed data unless otherwise noted. In addition to geometric mean, geometric standard deviation and confi dence intervals, certain performance aspects of each PAPR are also characterized by tolerance limits. The statistical tolerance limits furnish limits above or below that in which it can be confidently expected to find a prescribed percentage of individual items of the population.13' Methods and Materials The protocol used in this study has been described in a companion paper.12' Within the context of this report facepiece refers to that part of the helmet enclosure that covers the face, often times called the visor, shield or faceshield, Results Workplace Performance Tests Sampling data collected on the Racal AH3 and 3M W-344 PAPRs, while worn by workers are presented in Tables II and 111. The lead concentrations measured inside the facepiece of the Racal AH3 were found to range from 0.6 to 33.6 Mg/m3 with a geometric mean (GM) of 4.9 Mg/m3 and a geometric standard deviation (GSD) of 2.89. The 3M W-344 had lead concentrations inside the facepiece ranging from 0.2 to 66 Mg/ m3 with a geometric mean of 5.5 Mg/ m3 and a -- geometric standard deviation of 3.2. The 3M and Raca mean lead concentrations were not significantly different (P > 0.05). Within the range of ambient lead concentrations experienced during this study, both means were significantly (PC0.01) less than the OSH A lead exposure standard of 50 Mg/ rn . The two-sided 95% confidence interval limits on the means are respectively for the Racal AH3,3.land7.8 Mg/ m3 and for the 3M W-344, 3.3 and 9.3 pgj m3. No significant difference (P > 0.05) was found between the mean lead concentrations measured outside the lacepiece of either PAPR from which it is concluded that both experienced similar challenges of ambient lead concentra tion during the study. Analysis also indicated that the mean ambient lead concentration was significantly (P < 0.05) greater at the PAPR facepiece than at the inlet to the high efficiency filter(s) of the PAPRs. This resuit is not surprising since the worker participants spent most of their time facing their respective work stations from which lead emissions were occurring. The lead concentrations measured outside the PAPR facepiece represent an estimate of the lead expo sure a worker would experience were he not wearing the respirator, therefore, it has been used in the calculation of workplace protection factor. Associations between lead concentrations measured inside the facepiece and lead concentrations measured outside the facepiece or lead concentration challenging the filter(s) of the PAPRs were evaluated by linear regression analysis. With a 5 percent level of significance, a linearassociation was found on the Racal AH3 PAPR between the lead concentra- TABLEI Skewness and Kurtosis of Various Parameter Distributions Parameter Racal 3M Grouiped Skewness Kurtosis Skewness Kurtosis Skewness Kurtosis Untransformed Inside Facepiece Concentration Log-transformed Inside Facepiece Concentration 1.8.7 2.96 3.29 12.36 3.14 12.35 0.04 -0.41 -0.S5 2.52 -0.27 0.97 Untransformed Outside Facepiece Front Concentration Log-transformed Outside Facepiece Front Concentration 1.38 0.30 1.73 -0.96 3.07 11.18 1.15 1.13 2.40 7.82 0.69 -0.19 Untransformed Workplace Protection Factor Log-transformed Workplace Protection Factor 2.97 10.07 0.67 -0.01 4.39 19.94 0.87 1.32 5.11 29.31 0.71 0.68 Untransformed Quantitative Fit Factors Log-transformed Quantitative Fit Factors 0.71 -0.41 0.13 -1.31 1.78 0.17 3.14 -0.85 1.12 -0.08 0.74 VVV 000016217 -1.26 682 Am ind Hyg Assoc, i (45) October. 1984 . TABLE II Lead Concentrations, Workplace Protection Factors and Quantitative Fit Factors Obtained With The Racai Model AH3 PAPR Day/Shift Work Activity11 Concentration ^g/m3 Workplace' Inside Outside Facepiece Protection Facepiece Front Rear Factor Fit Factor 1 - a.m. 1 - p.m. 2 - a.m. 2 - p.m. 3 - a.m. 3 - p.m. 4 - a.m. 4 - p.m. Geometric Mean Geometric Std. Dev. 95% C.L. on Mean A B C A B C A B C A B C A B C A B C A B C A B C 8.0 5.7 7.1 1.7 11.5 7.5 5.1 15.2 17.0C 30.5 3.1 20.0 33.6 9.7 14.6C 2.0 1.6 4.3 3.1 1.9 3.6 3.9 1.0 0.6 4.9 2.89 3.1 - 7.8 1095 567 1018 366 952 - 1727 732 1165 3621 434 2000 1424 847 471 1984 459 409 2725 563 750 2221 579 1443 806 571 - 375 422 414 972 1440 1847 313 1733 1205 992 609 508 939 637 1340 676 668 . 701 464 137 99 143 211 83 - 341 48 68c 119 139 100 42 87 32c 1016 292 95 867 293 206 571 559 2323 205 2.83 128 - 325 13500 3000 21500 7700 2200 11500 15400 2200 17400c 20000 8300 10700 21700 2000 1500c 10500 31500 13500 6300 3100 14200 2800 2000 17000 7900 2.5 5300-11700 ^Calculated before inside facepiece concentration values were rounded to the nearest tenth and the outside facepiece concentration rounded to the nearest whole number. UA - reverbatory furnace operator/helper; B - ingot caster; C - blast furnace operator/helper. lNot used in data analysis due to sampling pump failure. lion inside the facepiece and the lead concentration challeng ing the PAPR filters. While significant, the coefficient of determination (r-) for the regression was only 0.23, which indicates that only 23 percent of the variation in the inside tacepiece concentrations has been explained with the help of the lead concentrations challenging the filters. No such asso ciation was found for the 3M W-344 PAPR. A somewhat surprising finding for both PAPRs was that no linear associa tion existed between lead concentration inside the facepiece and lead concentration outside the facepiece. However, if these same data were pooled, a linear association was found at a 5 percent level of significance. But even though signifi cant, the regression model yielded a very weak coefficient of determination of only 0.12. The workplace protection factors measured on the Racai AH3 were found to range from 42 to 2323 with a geometric mean of 205 and a geometric standard deviation of 2.83, and on the 3M W-344 the workplace protection factors were found to range from 28 to 5500 with a geometric mean of 165 and a geometric standard deviation of 3.57. Statistical t-test "analysis indicated there was no significant difference (P > 0.05) between the means of the Racai and 3M workplace American Industrial Hygiene Association JOURNAL (45) 10/84 protection factors. Both means were also found to be signifi cantly less (P < 0.01) than the current PAPR protection factor classification of 1000. The two-sided 95% confidence interval limits around the means of the workplace protection factor were respectively for the Racai AH3, 128 and 325, and for the 3M W-344, 94 and 292. The lognormal probability plot of the grouped workplace protection factor data is presented in Figure 1. The pre-shift quantitative fit factors measured on the Racai AH3 were found to range from 2000 to 31 500 with a geometric standard deviation of 2.5 and on the 3M W-344 they were found to range from 1200 to 24 600 with a geomet ric mean of 5 100 and a geometric standard deviation of 2.4. Analysis of the quantitative fit factor data indicated no significant difference (P > 0.05) between the means of the quantitative fit factor observed on both PAPRs. The twosided 95% confidence interval limits about the means are as indicated in Tables 11 and 111. The lognormal probability plot of the grouped quantitative fit factor data is presented in Figure I. It was thought that, in general, the better the quantitative fit factor achieved on a worker/ PAPR combination, the *vv oooo16218 TABLE III Lead Concentrations, Workplace Protection Factors and Quantitative Fit Factors Obtained With The 3M Model W-344 PAPR Day/Shift Work Activity15 Concentration ,ug/mJ Workplace' Inside Outside Facepiece Protection Facepiece Front Rear Factor Fit Factor 1 - a.m. 1 - p.m. 2 - a.m. 2 - p m. 3 - a.m. 3 - p.m. 4 - a.m. 4 - p.m. Geometric Mean Geometric Std. Dev. 95% C.L. on Mean A B C A B C A B C A B C A B C A B C A B C A B C 4.1 10.1 19.1 2.7 4.2 34.2l 4.3 6.5 12.6 27.7 3.4 7.7 6.8 5.6 14.2C 3.0 1.8 2.3 3.2 14.4 13.2 0.2 2.6 66.0 5.5 3.2 3.3 - 9.3 697 652 656 2484 574 - 1018 506 519 1271 528 5464 1975 695 552 766 451 472 850 1227 439 1232 1625 1864 742 774 484 313 548 1028 359 545 724 569 - 951 612 323 292 438 237 740 392 459 354 829 170 65 34 913 136 235 78 41 46 153 714 291 123 39' 255 245 209 267 85 33 5500 619 28 165 3.57 94 - 292 4000 10300 5300 1200 11000 2100 24600 10900 2500 2000 6700 3900 1900 12200 5600' 1400 8600 2600 2800 6700 2000 6500 23500 8500 5100 2.4 3500-7500 ACalculated before inside facepiece concentration values were rounded to the nearest tenth and the outside facepiece concentration rounded to the nearest whole number. hA - reverbatory furnace operator/helper; B - ingot caster; C - blast furnace operator/helper. ` Not used in data analysis due to sampling pump failure. 000016219 better the workplace protection factor would be for that same combination. A correlation analysis was done to test for an association between the paired workplace protection factor and quantitative fit factor data. That analysis found no significant linear association in the data and resulted in coefficients of determination of 0.08 for the Racal AH3 and 0.01 for the 3M W-344. These associations were also evalu ated with curvilinear regressions, however, again no signifi cant regressions were found. The study design was such that relationships between other statistics of quantitative fit factor and workplace per formance could be investigated. As mentioned, statistical analysis indicated that the means of the workplace protec tion factor and means of the quantitative fit factor of both the Racal and 3M PAPRs were not significantly different. When these data are combined replicate workplace protec tion factor and quantitative fit factor data on the 12 workers involved in the study could be obtained. The geometric mean, geometric standard deviation and ranee of these data for each worker are shown in Table IV. The means of the -workplace protection factor ranged from 40 to 883 and the means of the quantitative fit factor ranged from 2800 to 16 100. Correlation analysis on the mean workplace protec tion factor and mean quantitative fit factor data indicated no significant linear or curvilinear association. The coefficient of determination was less than 0.1. Other possible relationships between quantitative fit fac tor and workplace protection factor would be to compare each worker's lowest observed quantitative fit factor to the mean of his workplace protection factor measurements or to his lowest observed workplace protection factor. That anal ysis, which was conducted on both the untransformed and log transformed data, found no significant associations. The coellicients of determination for these analyses were again less than 0.1. In addition to workplace protection factor, another pos sible parameter to correlate w ith the quantitative fit factor is the lead concentration measured inside the facepiece of the PAPR. The geometric mean, geometric standard deviation and range on these lead concentration data are shown Table V lor each worker. The mean lead concentrations measured inside the facepiece were found to range from 1.0 to 17.8 jig; mJ. Correlation analysis between each worker's mean inside facepiece concentration and his lowest quantita- 684 Am Ind Hyg. Assoc J/45) October. 1984 TABLE V Lead Concentrations Measured Inside the Facepiece of Powered Air-Purifying Respirators Used on Study Participants Worker 1 2 3 4 5 6 7 8 9 10 11 12 Inside Facepiece Concentration (^g/m() GM (n=4) GSD Range 5.2 4.1A 4.4 12.1 8.2 15.8 2.7a 1.0* 2.9 2.7 3.0 1 7.8a 1.54 1.27 1.77 1.27 1.88 1.17 1.53 4.15 2.83 1.51 3.35 3.94 3.1 - 8.0 3.2- 5.1 1.9-6.5 9.7- 15.2 3.6 - 14 2 13.2- 19.1 1.7-3.9 0.3 - 2.7 1.0 - 11.5 1.6-4.2 0.6 - 7.7 4.3 - 66 GM = Geometric Mean. GSD n Geometric Standard Deviation A(n = 3) PROTECTION FAC I OR PERCENTAGE F.gure 1 -- Lognormal probability plots of grouped workplace protection factor and quantitative fit factor data on the 3M W-344 and Racal AH3 powered air purifying respirators equipped with helmets and high efficiency filters. live fit factor found no significant linear association in the data with a coefficient of determination of 0.13. Initially, several factors other than fit were also consid ered as possible sources of variation in the workplace periormance of the PAPRs. These factors were: day of the week; shift; and w'orker activity. Correlation analysis and A NOVA indicated none of these factors or their interactions were significant sources of performance variation. Howe\er, there was some trend in the data to suggest that worker activity may be a possible source of variation which should be accounted for in field testing. Other factors which may be important but were not investigated are workrate and min ute volume, head and body movement, and air current veloc ity and direction at the test site. Formed Head Tests The mean lead concentrations measured inside the facepiece of both PAPRs during manikin testing were significantly (P < 0.01) less than the OSH A lead exposure standard of 50 Hgj m3. These mean lead concentrations were not found to be significantly (P > 0.05) different. T he geometric mean, geometric standard deviation and range of those lead con- TABLE IV Workplace Protection Factor and Quantitative Fit Factor Data for Study Participants Worker 1 2 3 4 5 6 7 8 9 10 11 12 Workplace Protection Factor GM (n=4) GSD Range 300 160 129 69 83 40 194 883 204 242 702A 64A 2.17 2.55 1.78 1.32 2.35 1 44 2.86 4.81 2.25 2.05 3.35 2.05 137 - 867 42 - 341 78 - 293 48 - 87 39 - 206 32 - 68 46 - 571 119 - 5500 83 - 559 136 - 619 209 2323 28 - 100 Quantitative Fit Factor GM (n=4) GSO Range 7900 7800 6000 4200 8100 4100 2800 6400 4000 16100 6700 7100 3.02 2.75 2.14 2.25 2.64 3.03 2.10 3.32 2.12 1.88 2.44 2.32 1 900 - 24600 2800 - 21700 3100 - 12200 2000 - 10300 2500 - 21500 1 500 - 1 7400 1400 - 7700 1200 - 20000 2000 - 8600 8300 - 31500 2600- 17000 2100- 13500 GM = Geometric Mean, GSD = Geometric Standard Deviation An o3 VVV 000016220 American Industrial Hygiene Association JOURNAL ('45; 10/84 685 TABLE VI Lead Concentrations Measured Inside the Facepiece of Powered Air-Purifying Respirators Used on Manikins Test Location Racal Inside Facepiece Cone, (^g/m ) GM (n = 4} GSD Range 3M Inside Facepiece Cone. |g/m3) GM (n = 4) GSD Range Reverbatory Furnace Caster Blast Furnace 3.9 17.0 3.0a 1.3 3.0 - 5.5 1.5 9.6 - 26.7 2.3 1.4 - 7.4 16 0.8 2.8 2.0 0.6-3.4 33 0.3 - 3.8 2.0 1.0 - 4.3 An 3 centration data obtained during manikin testing are sum marized in Table VI by test location. Particle Size Analysis Results Eight impact or samples were collected on each work activity for particle size analysis. Inspection of the impactor data indicated that the size range limits (< 0.68 /am to > 17 /im) were inappropriate to allow complete determination of the aerosol size distribution. However, the impactor data did show that around the reverbatory furnace and ingot caster approximately 35% of the lead aerosol size distribution was larger than 17 and approximately 30% was smaller than 0.68 *im. This implies that high proportions of both lead dust and lead fume were present which is consistent with the nature of these operations. At the blast furnace, approxi mately 60% of the lead aerosol size distribution was larger than 17 /zm and only about 8% was smaller than 0.68 indicating a large percentage of lead dust or agglomerated lead fume was present. Discussion Analysis of the lead concentration data collected from inside the facepiece of the 3M W-344 and Racal AH3 PAPR.s indicates that the respirators were able to reduce the average lead exposure in this population of workers to a level signifi cantly lower than the 50 /jg/m3 OSHA personnel exposure limit for lead. However, it would be prudent from the stand point of worker health to consider the efficacy of these PAPRs based upon consideration of the exposure level of individual workers rather than the average exposure level of the worker population. Considering the geometric mean and geometric standard deviation of the pooled lead exposure data, a probability estimate can be made of the percentage of workers having exposures below a certain level. Based upon such a consideration and selecting a 0.95 probability level it would be concluded that workers in this population or other populations studied under similar conditions would have lead exposures of less than 32 It is possible, by applying tolerance limits to these same data, to estimate the proportion (P) of lead exposures falling below a certain level at a specified level of confidence (y). Setting P = 0.95 and y0.95 it can be said with 95%: confidence that 95% of the workers in this population or other populations studied under similar conditions w ould experience lead exposures of -.Jess than 53 /ng/m3. In regard to these estimates of lead exposure, it is worth noting again that the PAPRs w'ere properly fitted, worn, used and maintained during the study and that ambient concentrations of lead ranged from 7 to 109 times the 50 /xg/ m3 OSH A personnel exposure limit for lead. It was assumed that lead concentrations measured inside the PAPR facepiece during manikin testing would reflect penetration through the PAPR filtcr(s) because PAPRs equipped with helmet enclosures having air supply rates greater than 170 Lpm are considered to operate as positive pressure devices. However, the lead concentrations observed appear to be of higher magnitude than what would be expected through the PAPR's high efficiency filter system, possibly implicating that other sources of leakage W'ere occurring. A prime source of such potential leakage with helmet type enclosures is around the helmet facepiece. Cecala et a/.11' recently published a report of wind tunned studies done to determine the effects of ambient air velocity and direction on the Racal AH21 PAPR. These studies were conducted with the PAPR operating at approximately 200 Lpm. They reported that the PAPR's efficiency dropped from levels close to 100%: in static dust chamber tests to levels close to 94 percent (i.e.. protection factors of around 17) in the wind tunnel test at air velocities under 400 fpm (4.5 mph). They also reported that when the PAPR was not directly facing the air stream, its efficiency w<as reduced even further because contaminated air was forced into the helmet on the windward side along the facepiece shields. Their results suggest that the 200 Lpm airflow to the helmet enclo sure was not sufficient to maintain "positive facepiece pres sure" during the wind tunnel studies. While the velocity and direction of air currents at the manikin test positions in this smelter were not measured, it is very likely that at times the air velocity was near or above 400 fpm. Therefore it is hypothesized that under the workplace conditions encountered the 170 Lpm or greater flow rate to the 3M and Racal helmet enclosures was not sufficient to prevent inboard leakage from occurring around the facepiece during the manikin tests. By extrapolating Cecala's laboratory results, the quantity of facepiece leakage would be, in part, a function of the air velocity and direction occurring at a particular test site on a particular day. There fore. some ot the variability observed in the lead concentra tions measured inside the PAPR facepiece may be due to differing quantities of facepiece leakage. The American National Standards Institute (ANSI) respi rator standard states that "A respirator protection factor is a 686 VVV 000016221 Am Ind. Hyg Assoc. J (45) October. 1984