Christopher C. Coffey

Performance of N95 Respirators: Filtration Efficiency for Airborne Microbial and Inert Particles

In 1995 the National Institute for Occupational Safety and Health issued new regulations for nonpowered particulate respirators (42 CFR Part 84). A new filter certification system also was created. Among the new particulate respirators that have entered the market, the N95 respirator is the most commonly used in industrial and health care environments. The filtration efficiencies of unloaded N95 particulate respirators have been compared with those of dust/mist (DM) and dust/fume/mist (DFM) respirators certified under the former regulations (30 CFR Part 11). Through laboratory tests with NaCl certification aerosols and measurements with particle-size spectrometers, N95 respirators were found to have higher filtration efficiencies than DM and DFM respirators and noncertified surgical masks. N95 respirators made by different companies were found to have different filtration efficiencies for the most penetrating particle size (0.1 to 0.3 micron), but all were at least 95% efficient at that size for NaCl particles. Above the most penetrating particle size the filtration efficiency increases with size; it reaches approximately 99.5% or higher at about 0.75 micron. Tests with bacteria of size and shape similar to Mycobacterium tuberculosis also showed filtration efficiencies of 99.5% or higher. Experimental data were used to calculate the aerosol mass concentrations inside the respirator when worn in representative work environments. The penetrated mass fractions, in the absence of face leakage, ranged from 0.02% for large particle distributions to 1.8% for submicrometer-size welding fumes. Thus, N95 respirators provide excellent protection against airborne particles when there is a good face seal.

Fitting Characteristics of Eighteen N95 Filtering-Facepiece Respirators

Four performance measures were used to evaluate the fitting characteristics of 18 models of N95 filtering-facepiece respirators: (1) the 5th percentile simulated workplace protection factor (SWPF) value, (2) the shift average SWPF value, (3) the h-value, and (4) the assignment error. The effect of fit-testing on the level of protection provided by the respirators was also evaluated. The respirators were tested on a panel of 25 subjects with various face sizes. Simulated workplace protection factor values, determined from six total penetration (face-seal leakage plus filter penetration) tests with re-donning between each test, were used to indicate respirator performance. Five fit-tests were used: Bitrex™, saccharin, generated aerosol corrected for filter penetration, PortaCount® Plus corrected for filter penetration, and the PortaCount Plus with the N95-Companion™ accessory. Without fit-testing, the 5th percentile SWPF for all models combined was 2.9 with individual model values ranging from 1.3 to 48.0. Passing a fit-test generally resulted in an increase in protection. In addition, the h-value of each respirator was computed. The h-value has been determined to be the population fraction of individuals who will obtain an adequate level of protection (i.e., SWPF ≥ 10, which is the expected level of protection for half-facepiece respirators) when a respirator is selected and donned (including a user seal check) in accordance with the manufacturers instructions without fit-testing. The h-value for all models combined was 0.74 (i.e., 74% of all donnings resulted in an adequate level of protection), with individual model h-values ranging from 0.31 to 0.99. Only three models had h-values above 0.95. Higher SWPF values were achieved by excluding SWPF values determined for test subject/respirator combinations that failed a fit-test. The improvement was greatest for respirator models with lower h-values. Using the concepts of shift average and assignment error to measure respirator performance yielded similar results. The highest level of protection was provided by passing a fit-test with a respirator having good fitting characteristics.

Simulated Workplace Performance of N95 Respirators

During July 1995 the National Institute for Occupational Safety and Health (NIOSH) began to certify nine new classes of particulate respirators. To determine the level of performance of these respirators, NIOSH researchers conducted a study to (1) measure the simulated workplace performance of 21 N95 respirator models, (2) determine whether fit-testing affected the performance, and (3) investigate the effect of varying fit-test pass/fail criteria on respirator performance. The performance of each respirator model was measured by conducting 100 total penetration tests. The performance of each respirator model was then estimated by determining the 95th percentile of the total penetration through the respirator (i.e., 95% of wearers of that respirator can expect to have a total penetration value below the 95th percentile penetration value). The 95th percentile of total penetrations for each respirator without fit-testing ranged from 6 to 88%. The 95th percentile of total penetrations for all the respirators combined was 33%, which exceeds the amount of total penetration (10%) normally expected of a half-mask respirator. When a surrogate fit test (1% criterion) was applied to the data, the 95th percentile of total penetrations for each respirator decreased to 1 to 16%. The 95th percentile of total penetrations for all the respirators combined was only 4%. Therefore, fit-testing of N95 respirators is necessary to ensure that the user receives the expected level of protection. The study also found that respirator performance was dependent on the value of the pass/fail criterion used in the surrogate fit-test.

The Effect of Subject Characteristics and Respirator Features on Respirator Fit

A recent study was conducted to compare five fit test methods for screening out poor-fitting N95 filtering-facepiece respirators. Eighteen models of NIOSH-certified, N95 filtering-facepiece respirators were used to assess the fit test methods by using a simulated workplace protection factor (SWPF) test. The purpose of this companion study was to investigate the effect of subject characteristics (gender and face dimensions) and respirator features on respirator fit. The respirator features studied were design style (folding and cup style) and number of sizes available (one size fits all, two sizes, and three sizes). Thirty-three subjects participated in this study. Each was measured for 12 face dimensions using traditional calipers and tape. From this group, 25 subjects with face size categories 1 to 10 tested each respirator. The SWPF test protocol entailed using the PortaCount Plus to determine a SWPF based on total penetration (face-seal leakage plus filter penetration) while the subject performed six simulated workplace movements. Six tests were conducted for each subject/respirator model combination with redonning between tests. The respirator design style (folding style and cup style) did not have a significant effect on respirator fit in this study. The number of respirator sizes available for a model had significant impact on respirator fit on the panel for cup-style respirators with one and two sizes available. There was no significant difference in the geometric mean fit factor between male and female subjects for 16 of the 18 respirator models. Subsets of one to six face dimensions were found to be significantly correlated with SWPFs (p < 0.05) in 16 of the 33 respirator model/respirator size combinations. Bigonial breadth, face width, face length, and nose protrusion appeared the most in subsets (five or six) of face dimensions and their multiple linear regression coefficients were significantly different from zero (p < 0.05). Lip length was found in only one subset. The use of face length and lip length as the criteria to define the current half-facepiece respirator fit test panel may need to be reconsidered when revising the panel. Based on the findings from this and previous studies, face length and face width are recommended measurements that should be used for defining the panel for half-facepiece respirators.

Comparison of Five Methods for Fit-Testing N95 Filtering-Facepiece Respirators

Five fit-testing methods (Bitrex, ambient aerosol condensation nuclei counter using the TSI PortaCount Plus, saccharin, modified ambient aerosol condensation nuclei counter using the TSI PortaCount Plus with the N95-Companion, and generated aerosol using corn oil) were evaluated for their ability to identify poorly fitting N95 filtering-facepiece respirators. Eighteen models of NIOSH-certified, N95 filtering-facepiece respirators were tested by a panel of 25 subjects using each fit-testing method. The penetration of the corn oil and the ambient aerosols through the filter media of each respirator was measured in order to adjust the corresponding generated and ambient aerosol overall fit factors, reflecting only face-seal leakage. Fit-testing results were compared to 5th percentiles of simulated workplace protection factors. Beta errors (the chance of passing a fit-test in error) ranged from 3 percent to 11 percent. Alpha errors (the chance of failing a fit-test in error) ranged from 51 percent to 84 percent. The ambient aerosol using the TSI PortaCount Plus and the generated aerosol methods identified poorly fitting respirators better than the saccharin, the Companion, and Bitrex methods. These errors rates should be considered when selecting a fit-testing method for fitting N95 filtering-facepieces. When both types of errors were combined as an assignment error, the ambient aerosol method using the TSI PortaCount Plus had the lowest percentage of wearers being assigned a poor-fitting respirator.

Correlation Between Quantitative Fit Factors and Workplace Protection Factors Measured in Actual Workplace Environments at a Steel Foundry

Past studies have found little or no correlation between workplace protection factors (WPFs) and quantitative fit factors (FFs). This study investigated the effect of good- and poor-fitting half-facepiece, air-purifying respirators on protection in actual workplace environments at a steel foundry and the correlation between WPFs and FFs. Fifteen burners and welders, who wore respirators voluntarily, and chippers participated in this study. Each subject was fit-tested with two respirator models each with three sizes, for a total of six fit-tests. Models and sizes were assigned this way to provide a wide range of FFs among study participants. Each worker donned the respirator twice per day (at the beginning of the shift and following the lunch break) for 2 days. Quantitative FFs were first obtained for each donning using the PortaCount Plus trade mark in a separate room. Without redonning the respirators, workers performed normal work for 1 to 2 hours, and WPFs were measured by collecting ambient and in-facepiece samples simultaneously. A second fit-test was conducted without disturbing the respirator. FFs were obtained by averaging the results from the first and second fit-tests. The resulting FFs had a geometric mean (GM) of 400 (range=10-6010) and a geometric standard deviation (GSD) of 6.1. Of the 55 valid donnings, 43 were good fitting (FFs> or =100) and 12 were poor fitting (FFs<100). The WPFs had a GM of 920 (range=13-230,000) and a GSD of 17.8. The WPFs were found to be significantly correlated with the FFs (R(2)=.55 and p-value=.0001). Therefore, FF was shown to be a meaningful indicator of respirator performance in actual workplace environments.

Comparison of Six Respirator Fit-Test Methods with an Actual Measurement of Exposure in a Simulated Health Care Environment: Part II — Method Comparison Testing

This article, the second in a series of three, describes the method comparison testing portion of a study conducted to compare the fit factors from six quantitative fit-tests (QNFT) with a measure of a respirator wearers actual exposure assessed by end-exhaled air analysis for 1,1,2-trichloro-1,2,2-trifluoroethane (Freon-113) under the same conditions. The six QNFT methods were (1) continuous low flow, flush probe; (2) continuous high flow, deep probe (CHD); (3) exhalation valve discharge (EVD); (4) controlled negative pressure; (5) 10-minute Ambient Aerosol 1 (AA1); and (6) 30-minute Ambient Aerosol 2. The first three methods utilized corn oil and a forward light scattering photometer. The last two methods used the TSI Portacount. Respirators used in the study were both disposable and elastomeric organic vapor/high efficiency half-masks. The characterization equations from the preliminary research (described previously) were used to determine the actual exposure to Freon-113 during the method comparison testing. The fit factors resulting from the QNFT methods were then individually correlated with the Freon-113 exposures using the coefficient of determination, R2. The lowest R2 value, 0.20, was found with the EVD method. The highest R2 values, 0.81 and 0.78, were associated, respectively, with the CHD and AA1 methods. This study suggests that some QNFT methods may be used to estimate actual respirator performance under laboratory conditions.

Simulated Workplace Protection Factors for Half-Facepiece Respiratory Protective Devices

This study investigates two different methods (random effects model and 5th percentile) for determining the performance of three types of respiratory protective devices (elastomeric N95 respirators, N95 filtering-facepiece respirators, and surgical masks) during a simulated workplace test. This study recalculated the protection level of three types of respiratory protective devices using the random effects model, compared the two methods with each other and the APF of 10 for half-facepiece respirators, and determined the value of each of the fit test protocols in attaining the desired level of simulated workplace protection factor (SWPF). Twenty-five test subjects with varying face sizes tested 15 models of elastomeric N95 respirators, 15 models of N95 filtering-facepiece respirators, and 6 models of surgical masks. Simulated workplace testing was conducted using a TSI PORTACOUNT Plus model 8020 and consisted of a series of seven exercises. Six simulated workplace tests were performed with redonning of the respirator/mask occurring between each test. Each of the six tests produced an SWPF. To determine the level of protection provided by the respiratory protective devices, a 90% lower confidence limit for the simulated workplace protection factor (SWPF LCL90% ) and the 5th percentile of simulated workplace protection factor were computed. The 5th percentile method values could be up to seven times higher than the SWPF LCL90% values. Without fit testing, all half-facepiece N95 respirators had a 5th percentile of 4.6 and an SWPF LCL90% value of 2.7. N95 filtering-facepiece respirators as a class had values of 3.3 and 2.0, respectively, whereas N95 elastomeric respirators had values of 7.3 and 4.6, respectively. Surgical masks did not provide any protection, with values of 1.2 and 1.4, respectively. Passing either the Bitrex, saccharin, or Companion fit test resulted in the respirators providing the expected level of protection with 5th percentiles greater than or equal to 10 except when passing the Bitrex test with N95 filtering-facepiece respirators, which resulted in a 5th percentile of only 7.9. No substantial difference was seen between the three fit tests. All of the SWPF LCL90% values after passing a fit test were less than 10. The random model method provides a more conservative estimate of the protection provided by a respirator because it takes into account both between- and within-wearer variability.

Respiratory protection as a function of respirator fitting characteristics and fit-test accuracy.

The fitting characteristics of particulate respirators are no longer assessed in the National Institute for Occupational Safety and Health respirator certification program. It is important for respirator program administrators to understand the implications of that change and the additional burden it may impose. To address that issue, a typical respirator fit-testing program is analyzed using a mathematical model that describes the effectiveness of a fit-testing program as a function of the fitting characteristics of the respirator and the accuracy of the fittesting method. The model is used to estimate (1) the respirator assignment error, the percentage of respirator wearers mistakenly assigned an ill-fitting respirator; (2) the number of fit-test trials necessary to qualify a group of workers for respirator use; and (3) the number of workers who will fail the fit-test with any candidate respirator model and thereby fail to qualify for respirator use. Using data from previous studies, the model predicts respirator assignment errors ranging from 0 to 20%, depending on the fitting characteristics of the respirator models selected and the fit-testing method used. This analysis indicates that when respirators do not necessarily have good fitting characteristics, respirator program administrators should exercise increased care in the selection of respirator models and increased care in fit-testing. Also presented are ways to assess the fitting characteristics of candidate respirator models by monitoring the first-time fit-testing results. The model demonstrates that significant public health and economic benefits can result when only respirators having good fitting characteristics are purchased and respirators are assigned to workers using highly accurate fit-testing methods.

Comparison of Six Respirator Fit-Test Methods with an Actual Measurement of Exposure in a Simulated Health Care Environment: Part I — Protocol Development

Quantitative fit tests (QNFT) have been assumed to be predictive of the protection respirators would provide to a wearer in the workplace. Workplace studies have consistently found no correlation between quantitative fit factors and workplace protection factors. This article is the first in a series of three describing a study designed to compare the fit factors from six QNFT methods against the actual dose of 1,1,2 trichloro-1,2,2 trifluoroethane (Freon-113) received under the same laboratory conditions. Five preliminary studies conducted to develop the protocol to assess the respirator wearers dose through end-exhaled air analysis are described in this article: (1) chamber characterization, (2) end-exhaled air sampling, (3) skin absorption testing, (4) pharmacokinetic modeling, and (5) subject characterization. It was established that the concentration of corn oil aerosol and Freon-113 could be generated simultaneously in the chamber. It was ascertained that the optimum time to sample the exhaled breath was 30 minutes after the subject exited the chamber. It was also found that in a chamber concentration of 500 ppm, without any respiratory exposure, Freon-113 was still present in the end-exhaled air. This was attributed to skin absorption. The end-exhaled air of subjects exposed to 0.5, 3, 5, 25, 50, and 100 ppm (30 minute time-weighted average) of Freon-113 was evaluated at 30 minutes postexposure. This characterization was then used to predict the actual dose of Freon-113 received during the method comparison and validation testing to be described in subsequent articles.