If You Can't Measure Respirator Leakage,Can You Measure Respirator Fit?

(Reprinted with permission from: Respiratory Protection Update 5(4):5-7. ISSN 1048-6658,Respirator Support Services, 2028 Virts Lane, Jefferson, MD 21755, 1995.)

Clifton D. Crutchfield, Ph.D., CIH
cdcrutch@u.arizona.edu
Environmental and Occupational Health
University of Arizona
1435 N. Fremont Avenue
Tucson, AZ 85719

Determination of respirator fit has always been an important element of respiratory protection. Once you have decided what type of respirator is needed for a job, the two most important questions that remain regarding respirator selection are 1) How does it feel? and 2) How does it fit? Only the respirator wearer can provide the answer to the first question. Our job is to help him or her make a good decision by making sure that he knows how to don the respirator correctly or that she wears it long enough to develop a feel for how comfortable or uncomfortable it will be during actual use. We all have a pretty good idea about the kind of correlation that exists between respirator comfort and the degree of actual respirator use in the workplace.

While the worker plays the major role in the decision about respirator comfort, the job of providing an answer to the second question about respirator fit falls squarely on our shoulders. Most of us also have a pretty good idea about the kind of relationship that surely exists between respirator fit and respirator performance in the workplace, even though researchers have not been able to demonstrate a correlation between those two factors with any degree of certainty.

What are we really asking when we ask how a respirator fits? What we really want to know is how much does it leak, and is that level of leakage OK for the intended use. These are the bottom line questions we need to answer before sending a worker out into a contaminated workplace. If we are using qualitative fit testing to provide answers to these questions, we figure the respirator leaks if the worker tells us it leaks, and if he or she tells us it leaks then it leaks too much. We have answers to our questions about fit provided

we did a good job of screening the subject for adequate sensory detection,

we did a good job of generating an appropriate challenge concentration of the sensory test agent, and

the worker provided true sensory response information. The quality in qualitative fit testing is inextricably bound to such questions, and it is often difficult for either us or the worker to know the level of quality that accompanies the answers we generate with qualitative fit testing.

If we opt to use quantitative fit testing to provide answers about respirator fit, then we have hard numbers to back up our decisions. Before we get too comfortable with those numbers and decisions, however, we may need to take a closer look at what the researchers have really been telling us about those numbers.

Quantitative respirator fit test (QNFT) systems based on generated aerosol were developed more than 30 years ago, and have since been adopted as the "gold standard" QNFT method by both ANSI ⁽¹⁾ and OSHA.⁽²⁾ Aerosol fit test methods are based on the assumption that the aerosol concentration measured inside a respirator divided by the aerosol concentration measured outside the respirator provides a good representation of actual respirator penetration or leakage. That sounds good in theory, and since we have a long history of being able to measure a wide range of aerosol concentrations with photometers, the development of an aerosol-based QNFT system was certainly predictable. The inability to validate the basic assumptions underlying such systems when they were developed should have raised at least a small red flag of concern about their acceptability as a "gold standard".

The leakage measurement capabilities of aerosol-based fit test systems have never been validated. That fact has not generated any visible evidence of concern on the part of OSHA during its promulgation of numerous substance-specific standards that address quantitative respirator fit testing. In those standards, OSHA has consistently referred to generated aerosol fit test systems as if they are in fact a validated "gold standard", against which every other fit test method must be tested.

Several researchers have raised a number of red flags about aerosol QNFT methods. Myers ⁽³⁾ added the terms measurement bias and streamlining to the QNFT vocabulary. His seminal research found that in-mask aerosol measurements are significantly affected by factors such as sampling probe location, probe depth, nose versus mouth breathing, and inhalation versus exhalation cycles. Hinds ⁽⁴⁾ described differential aerosol losses during penetration of respirator leak sites, as well as the impact of aerosol lung deposition on fit test measurements. Willeke ⁽⁵⁾ found that aerosol-based measurements of respirator leakage were affected by the location of the leak, as well as by changes in inspiratory flow rates. Oestenstad ⁽⁶⁾ provided graphic evidence of aerosol streamlining inside respirators worn by human subjects. He used a flourescent challenge agent and ultraviolet illumination to show evidence of aerosol particles streamlining into the subject’s nostrils or mouth instead of migrating to the sampling probe where they could be detected as leakage by the aerosol system.

Although a number of researchers were reporting strong evidence that aerosol-based fit test systems do not measure respirator leakage very well, such information was evidently fairly easy to overlook or dismiss since no method was available for quantifying just how bad aerosol measurements of respirator leakage were. That situation changed with the development of the controlled negative pressure (CNP) fit test method. ⁽⁷⁾ Since the development of CNP, a number of comparison studies have reported that CNP fit test systems consistently detect far more respirator leakage than aerosol-based systems used to sequentially measure the same respirators on the same subjects. ^(8-10) A validation study ⁽¹¹⁾ of CNP performance versus a generated aerosol standard included the introduction of fixed leakage of > 530 mL/sec into both half-mask and full-face respirators worn by 50 human subjects. The CNP system detected unacceptable leakage in all 50 respirators, while the "gold standard" generated aerosol system passed 30 of the 50 respirators.

Despite the consistency with which CNP fit test systems detected up to 10 times more respirator leakage than aerosol-based systems during sequential fit test studies, the question of which system actually measured respirator leakage more accurately remained open to debate. As a consequence, a different research design was employed in an attempt to answer that question. Most of the prior research involving comparisons of different fit test methods had simply measured respirator fit sequentially with each system, without any reference to a known true value. During most of that research, the aerosol-based system was assumed to represent truth, even though it had never been validated.

A method of introducing a known amount of leakage into a respirator worn by a human subject was developed by inserting fixed leaks (hypodermic needles) into the respirator facepiece before it was donned. ⁽¹²⁾ The needle was then capped while a complete fit test series was completed with a CNP fit test system (Dynatech Nevada FitTester 3000). The leak needle was then uncapped, and an identical fit test series was repeated using the CNP system. The difference between the two fit tests represents the leakage introduced into the respirator when the leak needle was uncapped. Needle leakage was a direct function of the pressure gradient across the needle during the fit test, which was measured for each subject. A primary standard flow rate measurement device (Gilian Gilibrator) was used to determine flow rate through the needle at the measured pressure gradient.

Fit tests of the same respirators and leak needles worn by the same subjects were then completed in the manner described above using an ambient aerosol fit test system (TSI Portacount Plus). During 75 separate fit test pairs conducted with half-mask respirators on five separate subjects, the CNP system measured an average of 105.2% of the known leakage introduced into the subjects' respirators, while the Portacount Plus measured an average of only 20.8% of the known leakage. CNP fit tests were completed in approximately one-fifth of the time required to complete the ambient aerosol tests.

The same study design was employed in a follow-on study of the FitTester 3000 and Portacount Plus systems using a breathing machine-headform system. ⁽¹³⁾ A series of fixed leaks were introduced via matched hypodermic needles into three separate locations (bridge of nose, cheek, chin) in both half-mask and full-face respirators mounted on the headform. The FitTester system detected an average of 98.4% of the known leakage introduced into the respirators during 96 separate fit tests. The coefficient of variation (COV) was 4.3%. An analysis of variance revealed that CNP measurements were not affected by leak location.

Portacount Plus measurements of the same leaks averaged 40.3% of the known leak rates, with a COV of 46.9%. An analysis of variance detected significant differences in the ambient aerosol system's measurements of leakage as a function of leak location. An inverse relationship was observed in Portacount leak detection as a function of leak location between half-mask and full-face respirators. The pattern of leak detection as a function of leak location reported by the Portacount supports the theory of in-mask sampling bias related to particle streamlining that has reported by other investigators. The Portacount's overall increase in percent of the known leak detected, relative to the study done with human subjects (40.3% vs. 20.8%), was attributed to the lack of aerosol lung deposition in the breathing machine.

The CNP fit test method was designed to measure respirator leakage directly. To do so, it employs a well-established scientific principle that can be calibrated and validated using primary standards. During a number of studies conducted with both fixed systems and human subjects, the FitTester CNP system has consistently been able to measure respirator leakage with exceptional accuracy, precision, and speed.

The leak measurement capabilities of aerosol-based fit test systems cannot be calibrated against primary standards, and have never been validated. They measure respirator leakage indirectly as a function of measured aerosol penetration. An array of factors have been identified that significantly limit the ability of aerosol systems to detect respirator leakage. As a consequence, aerosol systems tend to over-state respirator fit by a wide margin.

 

References

1. ANSI Z88.10 (draft): "Respirator Fit Methods". American National Standard Institute, Inc., 1430 Broadway, New York, March 24, 1988.

2. U.S. Department of Labor: "OSHA Lead Standard: Quantitative Fit Test Procedures," Code of Federal Regulations Title 29 Part 1910.1025. Washington D.C.:U.S. Government Printing Office, 1983.

3. Myers, W.R., J. Allender, R. Plummer and T. Stobbe: "Parameters that Bias the Measurement of Airborne Concentration Within a Respirator". Am. Ind. Hyg. Assoc. J. 47(2):106-114, 1986.

4. Hinds, W.C. and G. Kraske: "Performance of Dust Respirators with Facial Seal Leaks". I. Experimental. Am. Ind. Hyg. Assoc. J. 48(10):836-841, 1987.

5. Willeke, K., and U. Krishnan: "Present Procedures in Quantitative Respirator Fit Testing: Problems and Potential Solutions." Appl. Occup. Environ. Hyg. 5(11): 762-768, 1990.

6. Oestenstad, R.K., J.L. Perkins, and V.E. Rose: "Identification of Faceseal Leak Sites on a Half-Mask Respirator". Am. Ind. Hyg. Assoc. J. 51(5):280-284, 1990.

7. Crutchfield, C.D., M.P. Eroh and M.D. Van Ert: "A Feasibility Study of Quantitative Respirator Fit Test by Controlled Negative Pressure." Am. Ind. Hyg. Assoc. J. 52(4):172-176, 1991.

8. Crutchfield, C.D., R.M. Murphy, and M.D. Van Ert: "A Comparison of Controlled Negative Pressure and Aerosol Quantitative Respirator Fit Test Systems Using Fixed Leaks." Am. Ind. Hyg. Assoc. J.52(6):249-251, 1991.

9. Graffeo, J.B.: "The Classification of Respirator Fit as Determined by an Aerosol Method and a Negative Pressure Method." Master's Research Project Report, Department of Environmental Health Sciences, University of Alabama at Birmingham, 1992.

10. Burnis, R.: Comparison of Fit Factors Determined by an Aerosol Test Method and a Dynamic Pressure Test Method. Masters Thesis. Dept. of Environmental Health Sciences, University of Alabama, Birmingham, 1991.

11. Crutchfield, C., Ruiz, A., and Van Ert, M.: A Validation Study of Respirator Fit Testing by Controlled Negative Pressure. Appl. Occ. Environ. Hyg. 9(5):362-366, 1994.

12. Crutchfield, C.D., Park, D.L. Henshel, J.L., et al.: "Determinations of Known Respirator Leakage Using Controlled Negative Pressure and Ambient Aerosol QNFT Systems." Am. Ind. Hyg. Assoc. J. 56(1):16-23, 1995.

13. Crutchfield, C.D., and D.L. Park: "Effect of Leak Location on Measured Respirator Fit." Am. Ind. Hyg. Assoc. J. 58(6):413-417, 1997.