www.epa.gov/or
United States
Environmental Protection
Agency
Assessment and Evaluation Report
Evaluation of Surface Sampling
for Bacillus Spores Using
Commercially-available
Cleaning Robots
Office of Research and Development
National Homeland Security Research Center
-------
EPA600/R-13/100
June 2013
Evaluation of Surface Sampling for Bacillus Spores
Using Commercially-available Cleaning Robots
Assessment and Evaluation Report
National Homeland Security Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
-------
Disclaimer
The United States Environmental Protection Agency, through its Office of Research and Development's
National Homeland Security Research Center, funded and managed this investigation through EP-C-09-
027 WA 2-29 and 3-29 with ARCADIS U.S., Inc. This report has been peer and administratively reviewed
and has been approved for publication as an Environmental Protection Agency document. It does not
necessarily reflect the views of the Environmental Protection Agency. No official endorsement should be
inferred. This report includes photographs of commercially available products. The photographs are
included for purposes of illustration only and are not intended to imply that EPA approves or endorses the
product or its manufacturer. Environmental Protection Agency does not endorse the purchase or sale of
any commercial products or services.
Questions concerning this document or its application should be addressed to:
Sang Don Lee, Ph.D.
Decontamination and Consequence Management Division
National Homeland Security Research Center
U.S. Environmental Protection Agency (MD-E343-06)
Office of Research and Development
109. T.W. Alexander Drive
Research Triangle Park, NC 27711
Phone:919-541-4531
Fax:919-541-0496
E-mail: Lee.Sangdon@epamail.epa.gov
-------
Acknowledgments
This effort was managed by the principal investigator from ORD's National Homeland Research Center
(NHSRC), and funded through the Wide Area Recovery and Resiliency Program by the Department of
Homeland Security Science and Technology Directorate (DHS S&T) under interagency agreement (#
RW-70-95812401). The authors acknowledge Chris Russell (formerly, DHS S&T) and Lori Miller (DHS
S&T) for their support of this work.
Project Team:
Sang Don Lee, Ph.D. (Principal Investigator)
National Homeland Security Research Center, Office of Research and Development, US Environmental
Protection Agency
Research Triangle Park, NC 27711
M. Worth Calfee, Ph.D.
National Homeland Security Research Center, Office of Research and Development, US Environmental
Protection Agency
Research Triangle Park, NC 27711
Leroy Mickelsen United States Environmental Protection Agency, Office of Emergency Management,
Research Triangle Park, North Carolina, United States
Stephen Wolfe United States Environmental Protection Agency, Region 5, West Lake, Ohio, United
States
Jayson Griffin United States Environmental Protection Agency, Office of Emergency Management,
Research Triangle Park, North Carolina, United States
This effort was completed under U.S. EPA contract #EP-C-09-027 with ARCADIS-US, Inc. The support
and efforts provided by ARCADIS-US, Inc. are gratefully acknowledged. The support provided by Tanya
Medley (NHSRC) in acquiring the vast quantities of supplies required for the completion of this project is
also acknowledged.
Additionally, the authors would like to thank the peer reviewers for their significant contributions.
Specifically, the efforts of Frank Schaefer (NHSRC), Rebecca Connell (US EPA), and Don Bansleben
(DHS S&T) are recognized.
IV
-------
Table of Contents
Disclaimer iii
Acknowledgments iv
Table of Contents v
List of Figures vii
List of Tables vii
List of Appendices viii
List of Acronyms and Abbreviations ix
Executive Summary xi
1 Introduction 1
1.1 Process 1
1.2 Project Objectives 2
1.3 Experimental Approach 2
1.3.1 Testing Approach 2
2 Materials and Methods 3
2.1 AFSD Testing Chamber 3
2.2 AFSD 3
2.3 Test Materials and Deposition 4
2.3.1 Test Coupons Preparation 4
2.3.2 Bacillus Spore Preparation 5
2.3.3 Coupon Inoculation 5
2.3.4 AFSD Testing Procedure 6
2.3.5 Comparative Surface Sampling Methods 6
2.3.6 Aerosol Sampling 7
2.3.7 Sample Extraction and Spore Recovery 7
2.4 Test Matrix 8
2.4.1 Test Facility Sampling Procedures 11
2.4.1.1 Sampling/Monitoring Points 11
2.5 Sampling Handling and Custody 12
2.5.1 Prevention of Cross-contamination of Sampling/Monitoring Equipment 12
2.5.2 Sample Identification 13
2.5.3 Sample Custody 13
2.5.4 Sample Preservation 15
-------
2.5.5 Sample Holding Times 15
2.5.6 Sample Archiving 15
3 Results and Discussion 16
3.1 Inoculation and Recovery 16
3.2 Extraction Efficiency 16
3.3 Sampling Efficiency 17
3.3.1 Scoping Studies 17
3.4 Scenario-based Evaluation 22
4 Quality Assurance 25
4.1 Sampling, Monitoring, and Analysis Equipment Calibration 25
4.2 Data Quality Objectives 26
4.3 QA/QC Checks 27
4.4 Acceptance Criteria for Critical Measurements 28
4.5 Data Quality Audits 30
4.6 QA/QC Reporting 31
5 Summary and Recommendations 32
VI
-------
List of Figures
Figure 2-1: Isolation chamber for coupon and AFSD 3
Figure 3-2: Recovery from Laminate Floor by AFSD Type 18
Figure 3-3: Recovery from tile surfaces by AFSD type 20
Figure 3-4: Recovery from carpet surfaces by AFSD type 21
List of Tables
Table 2-1. List of AFSD evaluated in the current study 4
Table 2-2. Components from AFSD extracted for analysis 6
Table 2-3. Test matrix for small coupon tests 9
Table 2-4. Test matrix for large-scale testing 10
Table 2-5. Frequency of sampling monitoring events 11
Table 2-6. Critical and non-critical measurements 12
Table 2-7. Coupon sample coding 14
Table 3-1. AFSD extraction efficiency test results 16
Table 3-2. Summary of laminate surface sampling comparative recovery using AFSD 18
Table 3-3. Aerosol recoveries during laminate tests 19
Table 3-4. Summary of tile surface sampling comparative recovery using AFSD 19
Table 3-5. Recoveries from aerosol samples collected during tile tests 20
Table 3-6. Summary of carpet surface sampling comparative recovery using AFSD 21
Table 3-7. Recoveries from aerosol samples during carpet tests 22
Table 3-8. Partitioning of recovered spores in AFSD 22
Table 3-9. Results from "Hot Spot" testing 23
Table 3-10. Results from wide area "release" testing 23
Table 3-11. Spread of spores by AFSD 24
Table 4-1. Sampling and monitoring equipment calibration frequency 25
Table 4-2. Analysis equipment calibration frequency 25
Table 4-3. QA/QC sample acceptance criteria 29
Table 4-4. Critical measurement acceptance criteria 30
VII
-------
List of Appendices
Appendix A Miscellaneous Operating Procedures (MOPs)
VIM
-------
List of Acronyms and Abbreviations
ADA
AFSD
ATCC
B.
BSC
CBRN
CPU
CM
CMAT
COC
COMMANDER
COTS
CR
CT
DCMD
DHS
Dl
DF
DPG
DQI
DQO
ECBC
EPA
FIFRA
GMP
HEPA
Ipm
LR
m
MDI
MOP
NOT
NHSRC
NIST
ORLS
OPP
ORD
OSWER
Aerosol Deposition Apparatus
Automatic Floor Sampling Device
American Type Culture Collection
Bacillus
Biological Safety Cabinet
Chemical, Biological, Radiological, and Nuclear
Colony Forming Units(s)
Critical Measurements
Consequence Management Advisory Team
Chain of custody
Consequence Management and Decontamination Evaluation Room
Consumer Off-the-Shelf
Comparable Recovery
Concentration xTime
Decontamination and Consequence Management Division
Department of Homeland Security
Deionized
Decimal Factor
Dugway Proving Ground
Data Quality Indicator
Data Quality Objective
Edgewood Chemical Biological Center
U. S. Environmental Protection Agency
Federal Insecticide, Fungicide, and Rodenticide Act
A product name, rather than an acronym
High-Efficiency Particulate Air
Liter per minute
Log reduction
Meter
Metered Dose Inhaler
Miscellaneous Operating Procedure
National Decontamination Team
National Homeland Security Research Center
National Institute of Standards and Technology
On-Site Research Laboratory Support
Office of Pesticides Programs
Office of Research and Development
Office of Solid Waste and Emergency Response
IX
-------
PARTNER Program to Align Research and Technology with the Needs of Environmental
Response
PBST Phosphate Buffered Saline with Tween20
PPE Person Protective Equipment
ppm parts per million
ppmv parts per million by volume
QA Quality Assurance
QAPP Quality Assurance Project Plan
QC Quality Control
RH Relative Humidity
AFSD Automated Floor Sampling Device
SOP Standard Operating Procedure
TBD To Be Determined
TSA Tryptic Soy Agar
VHP Vaporized Hydrogen Peroxide
WAM Work Assignment Manager
-------
Executive Summary
The existing surface sampling strategy for a post-terror incident involving the release of Bacillus anthracis
spores requires the use of various methods depending on the surface type. The established surface
sampling methods for 8. anthracis spores include wet wipes (for smooth nonporous surfaces) or wet
sponge wipes, vacuuming (for rough and porous surfaces), and wet swabs (for small and/or hard to
sample areas such as keyboards). These methods can be labor intensive and expensive to deploy since
they require sampling personnel to wear appropriate personal protective equipment (PPE) to reduce the
risk of exposure to pathogenic agents. The general process being investigated in this project is to assess
an alternative cost-effective, reliable sampling technique for various surfaces contaminated with Bacillus
spores (i.e., surrogates of 8. anthracis) using commercially available, off-the shelf Automated Floor
Sampling Devices (AFSD).
Three commercially available autonomous (robotic) vacuum-based cleaning robots (R1, R2 and R3), one
wipe-based robot (R4) and one wet vacuum-based robot (R5) were evaluated as AFSD for their sampling
efficiency on non-porous surfaces (laminate and tile). The first three vacuum-based AFSD were also
evaluated on a porous surface (carpet). The two wipe and wet vacuum-based AFSD were not tested on
carpet because of their recommended usage only on hard surfaces, according to the instruction manuals.
The evaluation criteria for the robotic cleaners included vacuum efficiency, availability, and cost. The top
two performers were then further evaluated to investigate their sampling capabilities at multiple levels of
contamination. In addition, these two AFSD were each challenged with two contamination scenarios, a
low level, large spatial extent contamination (wide contamination) scenario in which ~40% of the total
area sampled was experimentally inoculated with spores (~0.1 and 10 colony forming units (CFUs) per
cm2), and a high level, small spatial extent contamination (hot spot) scenario in which ~2% of the test
area was experimentally loaded with spores (~104 CFUs per cm2).
The sampling efficiencies of these AFSD were assessed by comparing their recoveries (CFUs) to
recoveries obtained using currently-used surface sampling methods. The overall results show that
sampling via AFSD is a viable option, when compared to traditional sampling methods. Some AFSD for
porous and non-porous materials were as efficacious as the respective surface sampling methods that
are currently recommended.
The AFSD sampling comparative recovery (CR) results for a laminate surface were higher for the wet
wipe and wet vacuum-based AFSD (up to 62% and 32%, respectively) than for the vacuum-based AFSD
(CR less than 10%) that were tested. The sampling process of the wet wipe-based AFSD is similar to the
well established wet wipe surface sampling method since both methods use a wetted cloth in conjunction
with a rubbing action on the surface. Low CRs from vacuum units were expected since previous
sampling studies have shown that the surface sampling using the wet wipe or the sponge wipe method on
nonporous surfaces has higher recovery efficiency than vacuum-based sampling methods.
The CRs for porous material (carpet) sampling were determined by comparison of the number of spores
(CFUs) recovered using three vacuum-based AFSD to that of the vacuum sock sampling method. The
test results showed CR values on the same order or greater (in some cases up to 161%) than the
vacuum sock sampling method. The differences in CRs among the three vacuum-based AFSD may be
related to the unique design and operating conditions of each device.
XI
-------
Two AFSD types from the scoping test were selected for further evaluation in a more complicated
environment, such as a larger spatial scale. The test results demonstrated the capability of AFSD
sampling of spores from carpet and laminate surfaces under two test scenarios (hot spot and wide
contamination). Further, only minimal contamination of the non-inoculated adjacent surfaces was
observed. The same AFSD were tested on larger floor areas inoculated at lower concentrations and
showed comparable results to the comparative surface sampling methods. This information may help
design targeted decontamination strategies, and possibly assist in the determination of the spatial
distribution of the spore attack.
Aerosol recoveries of spores observed during sampling for all five types of AFSD and all type of materials
tested showed small, but detectable, spore re-aerosolization. The observed relative differences in the
level of spore re-aerosolization for each AFSD/material combination are presumably due to the presence
of surface agitation devices (brush or a beater bar) on these units, and the type of AFSD sampling
scheme (vacuum-based versus wet-wipe sampling).
The current test method focused only on the sampling mechanism of the individual AFSD by limiting
sampling surface area. Varying the area cleaning logics or algorithms of individual AFSD was not part of
this study. However, varying the area cleaning logics or algorithms of individual AFSD could be a way to
increase collection efficiency when sampling a wide area.
Currently available AFSD have various convenient functions such as self-recharging, mapping,
navigation, etc. These functions will allow large contaminated areas to be sampled systematically. Two
obvious benefits of using AFSD for wide area sampling rather than the currently used sampling methods
include (1) fewer samples, because one composite sample is generated per deployment, and (2) less risk
of personnel exposure to 8. anthracis spores. In addition to wide area sampling, these AFSD could be
deployed to areas where human sampling is difficult, such as inside HVAC ductwork and in highly
contaminated areas (hot zones). However, for real world application, these AFSD need further evaluation
with various surfaces, deposition types, surface loadings, and environmental conditions (relative humidity
variation, exposure duration, etc.).
XII
-------
1 Introduction
After the 2001 intentional Bacillus anthracis spore contamination incidents in the U.S., many studies have
been conducted to develop and improve the remediation process of contaminated buildings [1-6]. Since
2001, surface sampling studies have been especially emphasized because of their direct impact on
decision making during on-site remediation activities.[7, 8] Accordingly, sound and defensible protocols
and implementation plans for surface sampling are needed but not yet adequately developed.[9]
Numerous studies have tested surface sampling methods to evaluate and/or validate their efficacy on
various surface types under numerous environmental conditions.[10-18] As a result, surface sampling
methods have been improved and optimized for real world application. However, there are still large
gaps surrounding sampling and analysis following a large urban area biological terror attack. [19] A 8.
anthracis release over a wide and highly populated area would tremendously increase the time, cost and
complexity to return the contaminated area to normalcy. Currently-used sampling methods are limited to
small areas {10 cm2 to 1 m2 (0.01 ft2 to 10.8 ft2)} and would require the collection of a large number of
samples in order to be representative if deployed over a large spatial scale. Such a sample burden would
strain sample processing laboratories during characterization and remediation and delay the overall
recovery. Although efforts have been made to increase the number of laboratories capable of processing
biological agent samples, the current capacity may be a limiting resource during recovery operations
following a wide-area attack.[9, 20]
The currently-used surface sampling methods include wet wipes, vacuuming, and wet swabs. The
existing spore sampling strategy requires the use of varied methods depending on the surface types, e.g.,
wet wipes for smooth nonporous surfaces, vacuuming for rough and porous surfaces, and wet swabs for
small and/or hard to sample areas such as keyboards. These methods can be labor intensive and
expensive to deploy since they require sampling personnel to wear appropriate personal protective
equipment (PPE) to reduce the risk of exposure to pathogenic agents. Commercially-available domestic
cleaning robots could be an alternative for 8. anthracis spore surface sampling. These cleaning robots
were introduced and commercialized since the early 1980s for home and industrial use.[21] The cleaning
mechanisms of these robots are similar to the current surface sampling methods such as vacuuming,
sweeping, and scrubbing. Robots have been developed with various convenient functions and sensors to
improve cleaning perform a nee [22], and can clean approximately 2 to 4 rooms {100 to 400 m2 (1076 ft2 to
4305 ft2)} per charge according to the manufacturers' claims. Using such cleaning robots as "automated
floor sampling devices" (AFSD) for B. anthracis spore surface sampling would reduce the number of
required samples (consequently reducing the burden on laboratories) and personnel compared to the
current sampling methods. AFSD can collect composite samples, thereby sampling numerous buildings
and large surface areas efficiently and economically. This study investigates the collection efficiency of
AFSD for 8. anthracis spore sampling compared to the current, conventional, surface sampling methods.
1.1 Process
The general process being investigated in this project is sampling of surfaces contaminated with Bacillus
spores (i.e., surrogates of 8. anthracis). Commercially available cleaning robots (AFSD) were evaluated
for their suitability and robustness for such surface sampling. The cleaning robots for this study were
selected based on their availability in the United States and price ($50 - $500). Three vacuum-based
AFSD and two mopping-based AFSD were chosen for laboratory evaluation of their efficacy at sampling
Bacillus spores. The top two performers based upon comparative sampling efficiencies were then further
-------
evaluated at multiple levels of contamination and on several spatial scales. In all cases, the AFSD were
compared to traditional surface sampling techniques.
1.2 Project Objectives
This work was designed to evaluate AFSD by generating data on the effectiveness of the device for
collection of Bacillus spores on different materials and under varied environmental conditions compared
to currently-used surface sampling methods.
1.3 Experimental Approach
In this study, AFSD were evaluated for their ability to collect Bacillus spores from environmental surfaces.
The current study determined the sampling efficiency of each AFSD, without modifying the sensors,
algorithms, or logics set by the manufacturers. Aerosol deposited 8. atrophaeus spores were used as a
surrogate of 8. anthracis spores. Test results were compared to currently-used surface sampling methods
(vacuum sock and sponge wipe). Air was sampled using a bio filter sampler to evaluate the potential for
re-aerosolization of spores during the sampling process using AFSD.
1.3.1 Testing Approach
Coupons representing three types of flooring materials were fabricated and sterilized before use. Floor
types included laminate flooring, carpet and tile. Coupons were inoculated with Bacillus atrophaeus
spores by aerosol inoculation using custom designed dose chambers. After inoculation, coupons were
transported into the Consequence Management and Decontamination Evaluation Room (COMMANDER),
a specially constructed enclosed, single-access-point chamber (henceforth, chamber) within the current
Homeland Security Enclosure located within High-Bay Room 130 (H130) at EPA's Research Triangle
Park, NC campus. The dosing chambers were removed from the coupons and then each coupon was
placed in a small secondary isolation AFSD testing chamber to prevent cross-contamination and to help
quantify re-aerosolized spores. AFSD were then used to sample the inoculated material, during a
predetermined or robot-determined amount of time. Collected spores were then recovered from the
collection bins and filters of each AFSD using liquid extraction-based techniques. Culture-based methods
were subsequently used to quantify the number of spores recovered by enumeration of CFUs on
microbiological growth media after plating serially-diluted aliquots of sample extracts. Therefore, in this
report "recovery" is defined by the number of CFUs observed following sample collection, extraction, and
analysis (dilution-plating or filter-plating).
-------
2 Materials and Methods
2.1 AFSD Testing Chamber
Each AFSD testing chamber (91 cm x 91 cm x 46 cm) was constructed of clear acrylic material (5 mm
thickness) and the inside surface was coated with antistatic film (chemical-resistant PVC (Type I)
antistatic Film, McMaster-Carr, Princeton, NJ). The diagram of the chamber is shown in Figure 2.1. The
chamber had one port located on the lid that was used for air sampling. Another port was located on the
front of the chamber and was outfitted with a HEPA filter so that sterile make-up air could be supplied
during sampling. Chamber air was sampled (15 LPM for 20 min) using a bio filter sampler (Via-Cell®
Bioaerosol Sampling Cassette, p/n VIA010, Zefon International, Inc., Ocala, FL) to determine the potential
for re-aerosolization of spores during AFSD sampling.
Port for air sampling
Inlet for
filtered make-up
45 cm
Figure 2-1: Isolation chamber for coupon and AFSD
2.2 AFSD
The five commercially-available cleaning robots purchased from an internet retail store and evaluated as
AFSD are summarized in Table 2-1.
-------
Table 2-1. List of AFSD evaluated in the current study
AFSD
Rl
R2
R3
R4
R5
Model
Roomba 760
XV-11
P3 P4920
Mint 4200
Scooba 390
Manufacturer
i Robot
Neato
P3 International
Evolution Robotics
[Robot
Cleaning type
Vacuum with bristle brush
Vacuum with silicone flat beater
Vacuum (no surface agitation tool)
Sweep and mop
Wet vacuum
Applicable
Surfaces
All surfaces
All surfaces
All surfaces
Hard floor
Hard floor
R1, R2, and R3 are vacuum-based cleaning AFSD and were tested on both carpet and laminate
surfaces. R4 and R5 are wet wipe- and wet vacuum- based AFSD, respectively, and were tested only on
laminate surfaces (not on carpet) as instructed by the factory manuals. All AFSD were removed from the
shipping box and sterilized inside COMMANDER by exposure to 250 ppmv of vaporized hydrogen
peroxide (VHP®, 1000ED, Steris, Mentor, OH) for 4 hours before testing. The sterilized AFSD were
degassed one to three days to remove residual fumigants. All AFSD retained their factory settings during
testing, and each AFSD was used only once before being discarded. All AFSD, except R3, possess
internal logic that allows the device to sample a discrete or predefined space and subsequently deactivate
itself. For these AFSD, the sampling duration was, therefore, determined by the AFSD itself. R3 units
were manually operated for an amount of time equivalent to those AFSD with the longest sampling
duration (R1 for carpet and R5 for laminate). When operating manuals required liquid inputs, sterile
Phosphate Buffered Saline with Tween-20 (PBST) was used rather than water or soapy water. For
example, the R4 wipe material was soaked with the PBST and the R5 "clean tank" was filled with PBST
before testing. Spore recovery efficiencies from the collection components (i.e., filters or collection bins)
of individual AFSD types were separately evaluated prior to conducting coupon-based testing. For these
preliminary recovery tests, a predetermined amount of B. atrophaeus spores (in PBST) were spiked onto
the filters and collection bins of each AFSD and allowed to dry. Extraction procedures were conducted
according to the procedures outlined in Section 2.3.5. Extraction efficiencies were determined by
comparing recoveries from AFSD to recoveries where extraction buffer was directly spiked with the same
liquid inoculum.
2.3 Test Materials and Deposition
2.3.1 Test Coupons Preparation
AFSD sampling tests were conducted with three floor surface types: laminate (Pergo Estate Oak , PE-
191113), carpet (Beaulieu Laredo Sagebrush loop carpet, Model 6666-01-1200-AB), and tile (Marazzi
Island Sand Glazed Ceramic Tile, Model UG4W). These materials were purchased from a local retail
store (Home Depot, Durham, NC). Coupons were fabricated from all three surface types into 107 cm x
107 cm and 71 cm x 71 cm size pieces for AFSD sampling tests and 36 cm x 36 cm for vacuum or
sponge wipe sampling tests. Both coupon types were backed with an equal-sized piece of 1.1 cm thick
Oriented Strand Board (OSB) plywood. Prior to use in tests, carpet coupons were vacuumed to remove
-------
the detachable foreign debris and particles, while laminate and tile coupons were cleaned with a dry wipe
(SIMWyPE tack cloth). After surface cleaning, coupons were sterilized by exposure to VHP® (250 ppmv
hydrogen peroxide vapor for 4 hours). The sterilized coupons were stored in sterilization bags (General
Econopak Inc., Philadelphia, PA, P/N 63636TW) until tested. After sterilization, coupons were degassed
for a minimum of three days before testing. The sterility of the coupons and other equipment needed for
the inoculation were confirmed by sampling at least one coupon per sterilization batch and one
representative piece of each inoculation equipment by using a Bactiswab (Remel Products, Lenexa KS,
P/N R12100) for sampling their respective surfaces. The swabs were subsequently streaked onto tryptic
soy agar (TSA) (BD, Franklin Lakes, NJ) plates, and the plates were incubated at 35 ± 2°C for at least 18
hours before being visually inspected to determine if bacterial growth (i.e., contamination) was present.
2.3.2 Bacillus Spore Preparation
The B. anthracis surrogate used for this study was a powdered spore preparation of B. atrophaeus
(ATCC 9372, Manassas, VA) and silicon dioxide particles. This bacterial species was formerly known as
B. subtilis var niger and subsequently B. globigii. The preparation was obtained from the U.S. Army
Dugway Proving Grounds (DPG) Life Science Division. The preparation procedure is reported in Brown
et al. [23] Briefly, after 80 - 90 percent sporulation, the suspension was centrifuged to generate a
preparation of about 20 percent solids. A preparation resulting in a powdered matrix containing
approximately 1011 viable spores per gram was prepared by dry blending and jet milling the dried spores
with fumed silica particles (Deguss, Frankfurt am Main, Germany). The powdered preparation was loaded
into metered dose inhalers (MDIs) by the U.S. Army Edgewood Chemical Biological Center (ECBC) or by
ARCADIS-US, Inc., according to a proprietary protocol. The MDI preparation and characteristics can be
found in Lee et al. and the references therein.[24]
2.3.3 Coupon Inoculation
The sterilized coupons were inoculated using the method described in the study by Calfee et al. [25, 26]
Coupons were inoculated with between ~5 x 102 and ~1 x 109 spores, depending on the desired target
inoculum levels. The consistency and loading levels of inoculums were verified using four stainless steel
control coupons during each inoculation event. These inoculation control stainless steel coupons were
sterilized via steam autoclave (Steris, Mentor, OH), inoculated, and then sampled with the sponge wipe
method as described in Miscellaneous Operating Procedure (MOP) 3165. Stainless steel surfaces were
inoculated concurrently with test samples and used to verify the magnitude and repeatability of the
inoculation procedures, since stainless steel surfaces have been shown to provide highly repeatable
recoveries.[10, 12, 27] Test coupons used for AFSD sampling tests were inoculated in the centermost 30
cm x 30 cm area of the coupon. The same size area was inoculated on the comparative surface sampling
method coupons. All test coupons underwent the same inoculation procedures, and were stored together
until used in testing. Each coupon was inoculated independently using separate dosing chambers,
originally designed for inoculation of the centermost 30 cm x 30 cm area of 36 cm x 36 cm coupons.
Following a metered dissemination, spores were allowed to settle onto the coupons for a minimum period
of 18 hours.
-------
2.3.4 AFSD Testing Procedure
AFSD sampling tests were conducted inside COMMANDER, a controlled chamber. COMMANDER was
controlled for the temperature (22 ± 0.7°C) and relative humidity (57 ± 5%). More detailed information
about COMMANDER can be found in prior publications.[28, 29] AFSD sampling tests were conducted
with carpet, laminate, and tile surfaces inside AFSD testing chambers. Each test (one AFSD and one
surface type) was conducted with four AFSD testing chambers (one blank and three test replicates) inside
COMMANDER. First, the test started with a blank AFSD sampling. A sterilized coupon was placed in the
middle of an AFSD testing chamber and a sterilized AFSD was placed on the bottom left corner of the
coupon. The cleaning start button was pressed and the lid to the AFSD testing chamber was closed. Air
sampling was initiated simultaneously with the onset of AFSD sampling process and the sampling
duration was monitored. After the blank sampling, B. atrophaeus spore-inoculated test coupons (one
coupon per an AFSD testing chamber) were sampled with AFSD. The inoculated coupons were placed in
an individual AFSD testing chamber, taking care not to touch the inoculated surface. After the completion
of sampling, the AFSD were powered off, removed from the testing chamber one at a time, and
disassembled for retrieval of the sample per Table 2-2. The AFSD sampling duration was recorded and
included the total time that each AFSD was sampling actively.
Table 2-2. Components from AFSD extracted for analysis
AFSD
R1-R3
R4
R5
Components extracted for sample
Collection Bin and Filter
Mop cloth
Retrieved liquid from "Dirty" tank
Filter
Treatment
Inlet sealed with Sterilized
Parafilm during transport
Aseptically placed in sterile
sample bag.
Tank triple rinsed with sterile
PBST, packed in leak-proof jar
Aseptically placed in sterile
sample bag.
The AFSD components were placed in a sterilized plastic bag. Each bag was then secondarily contained
in another bag and transported to the on-site Microbiology Laboratory for processing. Filters from air
sampling units were aseptically removed and placed in a sterile plastic bag or specimen cup for analysis
in the Microbiology Laboratory.
2.3.5 Comparative Surface Sampling Methods
To evaluate the collection performance of the AFSD, AFSD recoveries were compared to recoveries
obtained by currently-used surface sampling methods. Control coupons of carpet, tile, and laminate were
sampled using currently-used surface sampling methods.[9, 11, 15, 16, 20] Laminate and tile surfaces
were sampled with a sponge wipe sampling method and carpet surfaces with a vacuum sock method. An
area of 34 cm x 34 cm was sampled with the sponge wipe, delineated with a sterile stainless steel
template placed over the target area. Sponge wipe samples, described in MOP 3165, were collected
-------
using the following 5 steps: (1) using one flat side of the sponge wipe, the surface was sampled using
horizontal S-strokes, covering the entire template area; (2) the sponge wipe was then flipped over to the
opposite flat side to sample the surface in a vertical S-stroke pattern, covering the entire template area;
(3) using the narrow edges of the sponge wipe, the surface was sampled using the same S-strokes but
applied diagonally across the template, (4) rotating the sponge to use the opposite side at the midway
point of the coupon; and (5) the tip of the sponge wipe was then used to sample the perimeter of the
sampling area. The sampling method is described in detail in the study by Rose et al.[16]
Vacuum socks are the currently-used method for sampling porous surfaces, and were therefore used as
the comparative method for collection of spores from carpet surfaces. During vacuum sampling, a 34 cm
x 34 cm sterile stainless steel template and a sterile sock/nozzle attachment were used to collect the
sample. The nozzle was lightly pressed against the coupon surface while holding the nozzle at a 45
degree angle to the sampled surface. The samples were collected using horizontal and vertical S-
strokes. This method, described in MOP 3145, is a modified version of the method detailed in the study
by Brown et al.[11].
Wipe sampling was done on non-porous surfaces that had undergone AFSD sampling. These samples
were collected as a quality control check to validate inoculation had occurred. Gauze wipe sampling,
described in MOP 6567 was performed on the entire 71 cm x 71 cm surface. This method is better
adaptable to large surface area sampling than sponge wipe samplers; the sponges are pre-moistened
with a fixed volume of liquid and can become dry if sampling a large area. Checks were done to verify
sterility using swab samples collected according to MOP 3135. All relevant MOPs are included in the
appendix.
2.3.6 A erosol Sampling
ViaCell® bio-aerosol cassette samples were collected according to MOP 3155 for 20 minutes beginning
with the start of AFSD sampling to evaluate the potential for re-aerosolization. A flow rate of 15 liters per
minute (Ipm) was used and measured by a calibrated dry gas meter.
2.3.7 Sample Extraction and Spore Recovery
Sponge wipe (PN SSL10NB, 3M Inc., St. Paul, MN) samples were extracted by stomaching (1 minute,
260 RPM) in 90 ml of PBST using a Seward® Model 400 circulator (Seward® Laboratory Systems, Inc,
Port Saint Lucie, FL). Vacuum sock samples were extracted by first wetting the collection (white) portion
of the filter by dipping in PBST, then cutting it with sterile scissors (vertically then horizontally) into small
pieces (approx. 1 cm x 4 cm). As the filter was fractioned, the resulting pieces were allowed to fall into a
120 ml sterile specimen cup (Starplex Scientific LeakBuster Specimen Containers - Fisher Scientific #14-
375-459) containing 20 ml sterile PBST. The cups were then agitated (30 minutes, 300 RPM, ambient
temperature) using an orbital platform shaker incubator (Lab-Line, Model 3625). Spores collected by R4
were recovered from the mopping cloth by stomaching the cloth (2 minutes, 230 RPM) in 133 ml PBST
using a Seward® Model 400 circulator.
Two extraction procedures were required for R1, R2, and R3, as collected spores could have partitioned
to either the collection bin or the filter. Recovery from the filters of R1 proceeded by placing both filters
(each AFSD is equipped with two filters) into a 120 mL specimen cup (Starplex Scientific Inc., Cleveland,
TN, PN 3008-3TN), adding 90 mL PBST, and then agitating (30 minutes, 300 RPM) on an orbital platform
-------
shaker (Lab-Line, Model 3625). Recovery from the filters of R2, and R3 proceeded by placing each filter
into two 14 cm x 23 cm sterile sample bags (Fisher Scientific, P/N 01-002-53), one inside the other for
double containment. 180 ml of sterile PBST was then added to the innermost bag, and the samples were
agitated (30 minutes, 300 RPM) on an orbital platform shaker (Lab-Line, Model 3625). Spore recovery
from the particle bins of R1, R2, and R3 was accomplished by placing the bins into double 25 cm x 38 cm
sterile sample bags, aseptically adding 180 ml of PBST to each bag containing the bin, and then agitating
(30 minutes, 300 RPM, ambient temperature) on an orbital platform shaker incubator (Lab-Line, Model
3625).
The wet vacuum AFSD (R5) also required two extraction procedures, one procedure for the liquid fraction
and liquid collection reservoir and one procedure for the filter. First, the original 60 mL sterile PBST R5-
collected liquid was retrieved from the "dirty tank" using a 100 ml sterile serological pipette. The reservoir
was then rinsed twice with 60 ml PBST, and the three fractions were combined (for a total of 180 ml).
The filters from R5 were extracted with the same procedures used for filters from R2 and R3. The
resulting liquid extracts from all AFSD and all fractions were then each concentrated by centrifugation
where briefly, each sample was retrieved from its respective extraction bag or cup, and dispensed equally
into four 50 ml conical tubes (~45 ml for each tube). The samples were then centrifuged (3500 x g, 15
min, 4°C) to sediment the collected spores. All but 5 ml of the supernatant was carefully removed using a
50 ml sterile serological pipette. Each spore pellet was resuspended in the remaining 5 ml by three
cycles of alternating vortex mixing (30 seconds) and sonication (30 seconds, 40 kHz, Branson Model
8510).
Following resuspension, the four fractions per sample were recombined into one ~20 ml sample extract.
All sample extracts (AFSD, vacuum sock, and sponge wipe) were then subjected to a series of ten-fold
dilutions, as necessary, by adding 0.1 ml of the sample to 0.9 ml of PBST using a micropipette.
Appropriate dilutions were spread in triplicate (0.1 ml each) onto trypticase soy agar (BD™; Becton,
Dickinson, and Company; Franklin Lakes, NJ) plates and incubated at 35 ± 2°C. Plates were visually
examined and CFUs were enumerated after approximately 18 hours. The sampling results were
determined by averaging the observed CFUs from triplicate plates (subsamples), multiplying by the
inverse of the dilution factor, dividing by the volume plated (typically 0.1 ml), and multiplying by the total
volume of the sample extract. Mean recovery (CFUs) for each device and material type was determined.
2.4 Test Matrix
The test plan consisted of two tasks that were completed sequentially. The first task (tests 1-8 listed in
Table 2-3) consisted of evaluating and characterizing the five AFSD types on two different surfaces.
Three AFSD types were used for carpet, (R1, R2, and R3), and all five for laminate flooring. The target
spore surface loadings for this series of tests ranged between 5 x 105 and 5 x 106 CFU per ft2. Based on
the results of Tests 1 through 8 in Table 2-3, two AFSD (R2 and R4) were chosen for further evaluation at
additional inoculum levels and materials (tests 11-O1 through 17-O2). Each test in Table 2-3 included one
blank AFSD surface sample coupon, three inoculated AFSD sample coupons, four positive control wipe
sample coupons, four ViaCell® samples collected during AFSD operation, as well as field blank samples
for vacuum socks, aerosol samples, and sponge and wipe sample kits.
-------
Table 2-3. Test matrix for small coupon tests
Test
Number
1
2
3
4
5
6
7
8
11-01
13-01
13-02
14-O2
15-01
15-02
16-O2
17-O2
AFSD
Rl
R2
R3
Rl
R2
R3
R4
R5
R2
R4
R2
R4
Number of AFSD
4
Material
Type
Carpet
Laminate
Carpet
Laminate
Tile
Target Spore Loading
(CPU per cm2 )
5 x 102 to 5 x 103
2xl05
2X101
5 x 10"1
IxlO6
2X101
5 x 10"1
The second task was designed to further evaluate two AFSD types (R2 and R4) on a larger spatial scale.
The testing was conducted inside the COMMANDER, and utilized a total test area equal to 1.8 m x 2.5 m
(Figure 2-2). Subsections of the floor were inoculated under aerosol deposition apparatus (ADAs). This
task consisted of four tests for each material (carpet and laminate) that were conducted in duplicate, as
listed in Table 2-4.
COMMANDER was fitted with pre-sterilized coupons as shown in Figure 1. The set-up consisted of two
16.5 cm x 16.5 cm coupons, four 11 cm x 11 cm coupons, and a single 5.5 cm x 5.5 cm coupon in the
center. The spore inoculation was performed at the center of the room (for Hot Spot inoculation) or
throughout the whole area (for Wide Area inoculation). For the wide area release, the two 16.5 cm x 16.5
cm coupons and the center coupon were inoculated.
-------
Figure 2-2. COMMANDER floor testing setup
Table 2-4.
Test matrix for large-scale testing
Test
Number
18
19
20
21
22
23
24
25
Inoculation
Area
Hot Spot
Hot Spot
Wide Area
Wide Area
Hot Spot
Hot Spot
Wide Area
Wide Area
Number of
AFSD
1
1
1
1
1
1
1
1
Material
Type
Carpet
Laminate
AFSD
type
R2
R4
Target Spore
Loading
(CPU per cm2)
2xl04CFU
2xl04CFU
2xl01CFU
2xl01CFU
2xl04CFU
2xl04CFU
2 x 10"1 CFU
2 x 10"1 CFU
Comparative
Surface
Sampling
Method
Vacuum sock
Vacuum sock
Vacuum sock
Vacuum sock
Sponge wipe
Sponge wipe
Sponge wipe
Sponge wipe
10
-------
2.4.1 Test Facility Sampling Procedures
2.4.1.1 Sampling/Monitoring Points
Table 2-5 lists the samples collected for each test.
Table 2-5. Frequency of sampling monitoring events
Sample Type
Test AFSD
Negative control AFSD
Positive control coupon -
vacuum sock sample
Positive control coupon -
sponge wipe sample
Reference coupon -
sponge wipe sample
Laboratory blank
coupons
Biocontaminant
Laboratory material
blanks
Aerosol Samples
Aerosol sample during
blank
RH/Temp
Sample Number
3 per test condition
1 per test
3 per test condition
3 per test condition
a set of 4 stainless steel
coupons inoculated at the
beginning, middle, and
end of test coupon
inoculations
3 sterile coupons
3 per material
1 per AFSD
1 per test
Logged every 10
seconds
Purpose
To determine the number of viable
spores recovered from the AFSD
To determine extent of cross-
contamination
To determine the number of viable
spores recovered by conventional
methods
To determine the number of viable
spores recovered by conventional
methods
To provide the best estimate of the
number of viable spores deposited
onto the material test coupons
To demonstrate sterility of coupons
and extraction materials.
To demonstrate sterility of
extraction and plating materials
To determine the extent of
resuspended spores during
operation
To determine extent of cross-
contamination
To determine environmental
conditions during AFSD operation
Table 2-6 lists the critical and non-critical measurements for each sample.
11
-------
Table 2-6. Critical and non-critical measurements
Sample Type
Test AFSD
Negative control AFSD
Positive control coupon -vacuum sock
sample
Positive control coupon - sponge wipe
sample
Reference coupon - sponge wipe
sample
Laboratory blank coupons
Biocontaminant Laboratory material
blanks
Aerosol Samples
Aerosol sample during blank
RH/Temp
Critical Measurements
Plated volume, incubation temperature, extracted
volume, CPU
Plated volume, incubation temperature, extracted
volume, CPU
Plated volume, incubation temperature, extracted
volume, CPU
Plated volume, incubation temperature, extracted
volume, CPU
Plated volume, incubation temperature, extracted
volume, CPU
Plated volume, incubation temperature, extracted
volume, CPU
Plated volume, incubation temperature, extracted
volume, CPU
Plated volume, incubation temperature, extracted
volume, volume of air sampled, CPU
Plated volume, incubation temperature, extracted
volume, volume of air sampled, CPU
Non-critical
Measurement
Storage time, storage
temperature, operation
time
Storage time, storage
temperature, operation
time
Storage time, storage
temperature
Storage time, storage
temperature
Storage time, storage
temperature
Storage time, storage
temperature
Storage time, storage
temperature
Ambient temperature
during air sampling,
duration of air sampling
Ambient temperature
during air sampling,
duration of air sampling
RH and temperature
during AFSD operation
2.5 Sampling Handling and Custody
2.5.1 Prevention of Cross-contamination of Sampling/Monitoring Equipment
Several management controls were instituted to prevent cross-contamination. This project was labor
intensive and required that many activities be performed on coupons that were intentionally contaminated
(test coupons and positive controls). Specific procedures were put in place in the effort to prevent cross-
contamination among the samples. Adequate cleaning of all common materials and equipment was
critical in preventing cross-contamination.
There were three primary activities for each test in the experimental matrix. These activities were
preparation of the coupons, sampling, and analysis. The AFSD were sterilized prior to use with VHP®
(250 ppmv vaporous hydrogen peroxide for four hours). Specific management controls for each of the
three following activities are described below.
• Negative control coupons were present for each test. Growth on these coupons would indicate
contamination during inoculation or sample collection activities.
• Swabs were used to sample coupon surfaces prior to inoculation. Growth of these swab samples
would indicate the failure of the sterilization methods. While some swabs did indicate contamination,
12
-------
no systematic changes to sterilization protocols were required since the level of contamination was
insignificant compared to the test inoculum amounts.
• Only one AFSD was handled at a time.
• The AFSD was sealed with sterilized parafilm and placed in sample bags immediately following use.
General aseptic laboratory technique was followed and is embedded in the standard operating
procedures (SOPs) and MOPs used by the on-site Biocontaminant Laboratory to recover and analyze
samples. The SOPs and MOPs document the aseptic technique employed to prevent cross-
contamination. Additionally, the order of analysis was always as follows: (1) all blank coupons; (2) all test
coupons; and (3) all positive control coupons.
2.5.2 Sample Identification
Each coupon was identified by a unique sample number. The sampling team maintained an explicit
laboratory log which included records of each unique sample number and its associated test number,
inoculant amount, any preconditioning and treatment specifics, and the date treated. The sample codes
eased written identification. Once the coupons were transferred to the on-site Biocontaminant Laboratory
for microbiological analysis, each sample was additionally identified by replicate plate (Petri dish) number
and dilution. Table 2-7 specifies the sample identification (e.g., 28-4-C1-O1). The Biocontaminant
Laboratory also included on each plate the date it was placed in the incubator.
Swabs collected as sterility checks were identified by the code 29-[Test Number]-SW-[unique area code].
The swabs were collected according to MOP3135 from coupons and inoculation materials prior to
inoculation.
2.5.3 Sample Custody
Careful coordination with the on-site Biocontaminant Laboratory was required to achieve successful
transfer of uncompromised samples in a timely manner for analysis. Test schedules were confirmed with
the Biocontaminant Laboratory prior to the start of each test. To ensure the integrity of samples and to
maintain a timely and traceable transfer of samples, established and well-documented chain of custody
(COC) procedures are mandatory. Accurate records were maintained whenever samples were created,
transferred, stored, analyzed, or destroyed. The primary objective of these procedures was to create an
accurate written record that can be used to trace the possession of the sample from the moment of its
creation through the reporting of the results. A sample was in custody in any one of the following states:
• In actual physical possession
• In view, after being in physical possession
• In physical possession and locked up so that no one can tamper with it
• In a secured area, restricted except to authorized personnel
• In transit, secure and sealed so any tampering is evident
Laboratory test team members received copies of the test plans prior to each test. Pre-study briefings
were held to apprise all participants of the objectives, test protocols, and COC procedures to be followed.
These protocols were required to be consistent with any protocols established by EPA.
13
-------
Table 2-7. Coupon sample coding
Coupon Identification: 28-T-(X)MM-SS
Category
T
(X)MM
(Material)
SS
(Sample Type)
Example
Code
1
X
C#
L#
S#
T#
FB
R#
O#
P
HS
A#
Test Number
Procedural Blank
Carpet, where # is replicate (1-3)
Laminate Flooring, where # is replicate (1-3)
Stainless Steel (for QC purposes), where # is replicate
(1-4)
T, where number is replicate (1-3)
Field Blank
As purchased AFSD, where # is for Type 1, type 2, type
3, type 4 or type 5
Optimized AFSD, # for selected AFSD and optimization
Sponge wipe Sample
Vacuum sock sample
Aerosol sample, where # is replicate (1-3)
Microbiology Lab Plate Identification 28-T-(X)MM-SS -Rd
25-T-(X)M-SS
R (Replicate)
d (Dilution)
As above
R A-°
Oto4, for 10EO to 10E4
In the transfer of custody, each custodian signed, recorded, and dated the transfer on the COC. Sample
transfer could be on a sample-by-sample basis or on a bulk basis. The following protocol was followed for
all samples as they were collected and prepared for distribution:
• A COC record accompanied the samples. When turning over possession of samples, the transferor
and recipient signed, dated, and noted the time on the record sheet. This record sheet allowed
transfer of custody of a group of samples from the testing laboratories to the on-site Biocontaminant
Laboratory.
• If the custodian had not been assigned, the laboratory operator had the responsibility of packaging
the samples for transport. Samples were carefully packed and hand-carried between on-site
laboratories. The COC record showing the identity of the contents accompanied all packages.
14
-------
2.5.4 Sample Preservation
Following transfer to the on-site Biocontaminant Laboratory, all samples were stored at 4 ± 2 °C until
analyzed. All samples were allowed to equilibrate at room temperature for one hour prior to analysis.
2.5.5 Sample Holding Times
After sample collection for a single test was complete, all biological samples were immediately
transported to the on-site Biocontaminant Laboratory, with appropriate COC form(s). Samples were
stored no longer than five days before the primary analysis. Typical hold times, prior to analyses, for most
biological samples was < 2 days.
2.5.6 Sample A re hiving
All samples and diluted samples were archived for two weeks following completion of analysis. This time
allowed for review of the data to be performed to determine if any re-plating of selected samples was
required. Samples were archived by maintaining the primary extract at 4 ± 2 °C in sealed 50 mL conical
tubes.
15
-------
3 Results and Discussion
3.1 Inoculation and Recovery
Stainless steel coupons were used to verify the magnitude and repeatability of spore loadings for every
inoculation event. A total of 32 stainless steel coupons were inoculated and sampled using the sponge
wipe method. The recovery results showed the average loading level of each inoculation event ranged
from 9.12x105through 1.61 x 107 CPUs between tests. The coefficients of variation of spore inoculations
were between 8 and 63%. The sampling efficiency from stainless steel using the sponge wipe method is
approximately 48% according to the study by Krauter et al.[15] Therefore, using this crude assessment, it
can be assumed that the inoculated spores were between 106 and 107 CPUs per coupon.
The recoveries from blank coupon sampling using AFSD were all negative (0 CPUs) except R1 and R5
from laminate. For these samples, viable spores were recovered; however, the recoveries were below
the quantification limit (less than 30 CPUs per plate). The air sampling results from blank coupons were
negative for all blank tests. The recoveries from blank tests were minimal compared to the test spore
loadings (106- 107 CPUs) and no further action was required.
3.2 Extraction Efficiency
As described in the previous section, the extraction efficiency was determined for each AFSD prior to
conducting surface recovery tests since extraction efficiency may affect the overall sampling efficacy.
Recovery from the AFSD depends on two distinct properties: the efficiency of the AFSD to collect the
sample (spores) and the efficiency of the method to then remove and quantify the sample from the
device. In order to focus on the AFSD, rather than to optimize recovery efficiency from each AFSD,
extraction techniques were the same for all AFSD as much as possible, given that each had unique filter
and bin designs. These data are summarized in Table 3-1.
Table 3-1. AFSD extraction efficiency test results
AFSD
Rl
R2
R3
R4
R5
Extracted Parts
Filter, dust bin
Filter, dust bin
Filter, dust bin
Wipe cloth
Tank
Extraction
Method
Orbital shaking
Orbital shaking
Orbital shaking
Stomaching
Rinsing
Average Extraction
Efficiency (%)
65 ±14
57 ±8
67 ±4
49 ±7
90 ±7
Test sample
size
3
3
3
3
3
The efficiency was calculated by normalizing the recovery (CFUs) from AFSD parts by the number of
spores spiked onto the parts. The results showed the extraction efficiency was approximately 50 - 90 %.
The extraction efficiencies for vacuum units were within 10% difference. The maximum difference
between R4 and R5 was approximately 40%. The recovery efficiency test was conducted with a liquid
16
-------
spore inoculum on clean, unused AFSD parts. While the extraction efficiency during surface sampling
tests may be different due to the complexity of the collected sample matrix, the similarities of the
extraction recovery suggests that results from the AFSD may be directly compared.
3.3 Sampling Efficiency
The sampling efficiency of the AFSD is a measure of the spores recovered from a contaminated material
surface by the AFSD as compared to the spores recovered by traditional methods. The sampling
efficiencies of AFSD were compared to the sponge wipe method for laminate coupons and the vacuum
sock method for carpet coupons. The AFSD sampling comparative recovery (CR) was calculated using
Equation 1.
CR (o/o) = average recovery fromrobot (CPUs) x WQ
^ J average recovery from comparative surface sampling method (CPUs)
Results are based on the total recovery (CFUs) from an AFSD, in some cases determined as the sum of
the recoveries (CFUs) from two or more parts.
3.3.1 Scoping Studies
Laminate Surfaces: Initially, laminate coupons were tested with five AFSD to compare surface sampling
recoveries among all AFSD types. The test results from laminate surfaces are summarized in Table 3-2
and Figure 3-2. The highest CR was achieved by R4 (62%) among the five AFSD types tested. The
second highest CR was achieved by R5 (32%). The vacuum units (R1, R2, and R3) demonstrated CRs
equal to or less than 10%. Low CRs from vacuum units were expected since previous sampling studies
have shown that the wet wipe or the sponge wipe method on nonporous surfaces has higher recovery
efficiency than vacuum-based methods. The sampling process by R4 was similar to the wet wipe or the
sponge wipe method because R4 used a PBST-wetted cloth in conjunction with a rubbing action on the
surface. The size of cloth used by R4 for cleaning was approximately 25 cm x 10 cm and, therefore,
resulted in a 10 times larger contact area than that of the sponge wipe. The sampling efficiency of R4
was found to be lower than that of the sponge wipe method; however, the sampling efficiency of the
sponge wipe may depend not only on the wipe size, but also on applied pressure, actual surface contact
area, residence, and other parameters pertinent to the given sampled surface.
The relatively lower CR of the R5 units compared to the R4 units is due to the fact that the latter unit
collects the spores directly by the wetted wipe while the former unit sample surfaces by first releasing
clean PBST onto the surface followed by scrubbing and wiping the surface. The dispensed PBST liquid,
along with any particles contained in the liquid, is recollected by the unit from the surface and stored in
the "dirty tank". The spore recovery by R5 is mainly dependent on recollection of the dispensed liquid. It
is expected that the recollection of liquid may vary significantly depending on the surface morphology
sampled. Lower sampling efficiencies would be expected when sampling coarse or irregular surfaces.
17
-------
Table 3-2. Summary of laminate surface sampling comparative recovery using AFSD
AFSD
Rl
R2
R3
R4
R5
Average Sampling
Duration
(sec)
422
76
540
130
545
Mean Recovery
from AFSD
(CFUs x!05)a
1.7 ±1.5
0.16 ±0.07
1.3b±0.7
19 ±4.6
21 ±17
Mean Recovery from
Sponge Wipe
(CFUs x!05)a
21 ±11
1.4 ±1.1
53 ±30
31 ±6
64 ±2
CR (%)
8.1
11
2.5
62
32
a. Mean recovery ± standard deviation
b. One R3 unit stopped after approximately 10 seconds of operation. Data are calculated with
duplicate sample results.
Recovery from Laminate Floor
surfaces
l.OE+07
l.OE+06
l.OE+05
l.OE+04
<-> l.OE+03
l.OE+02
l.OE+01
l.OE+00
I I I
70%
60%
50%
40% ^
30% g
20%
10%
0%
Mean recovery from AFSD
Mean recovery from
sponge wipe
Rl R2 R3 R4 R5
Figure 3-2: Recovery from Laminate Floor by AFSD Type
The sampling duration (the total time that each AFSD was sampling actively) was shortest for R2 (mean
of 76 seconds) and longest for R5 (mean of 545 seconds). R3 was operated for approximately the same
time as R5. R4 was also the most efficient unit in terms of the number of spores collected per unit area
and time.
18
-------
Table 3-3. Aerosol recoveries during laminate tests
AFSD
Rl
R2
R3
R4
R5*
Recovery from Air Filters (CPUs)
67
81
4
127
607
% CPU from sponge wipe
0.0032%
0.0566%
0.0001%
0.0041%
0.0094%
Beater Bar
Yes
Yes
No
No
Yes
"These data were collected from ViaCell cassettes which were past the expiration date.
Aerosol sample results observed during sampling are listed in Table 3-3. The CPU/sample values have
been reported because all samples were operated for the same length of time (20 minutes) and the same
approximate flow rate (15 Ipm). It is unknown when the CPUs are resuspended, but it is much more likely
that the concentration in the chamber air is much higher at the beginning of the AFSD operation than at
the end. The results are further confounded by the total number of CPU present on the coupon, which
varied by nearly two orders of magnitude. The results showed a small but detectable spore re-
aerosolization, slightly higher from R4 and R5 sampling tests, and relatively high (in comparison to
coupon CPU) for R2. The re-aerosolization from R1 and R2 was presumably due to the presence of
surface agitation devices (brush or a beater bar) on these units. Spore re-aerosolization from R3 was
minimal compared to R1 and R2, likely due to its lack of a surface agitation tool. Though R5 uses a wet
method for collection, R5 also has a surface agitation tool, and a filter that was designed for collection of
large liquid particles rather than aerosolized spores.
Tile surfaces: Two AFSD types, R4 and R2, were tested for sampling efficiency on tile surfaces. These
were designed as follow-up tests to the laminate series to determine if different hard material surfaces
affected efficiency. Only AFSD with demonstrated efficiency were chosen for this extended study. Table 3-4
summarizes the results. Limits of detection were investigated for R4 by lowering the inoculation.
Table 3-4. Summary of tile surface sampling comparative recovery using AFSD
AFSD
R2
R4
R4
R4
Average sampling
duration (sec)
65
209
170
206
Mean recovery from
AFSD
(CFUs)a
1.68 x 10s + 2.05 x 10s
1.97 x 10s + 4.04 x 10s
1.44 x!04 + 9.76 xlO3
9.12xl02 + 6.10xl02
Mean recovery from sponge
wipe (CFUs)
7. 25 x 10s + 1.53 x 10s
1.31 x!07 + 1.46 xlO7
1.62 x!04 + 2.60 xlO3
6.41xl02 + 5.10xl02
CR (%)
2%
15%
89%
142%
aMean CFUs + one standard deviation
19
-------
Recovery from Tile surfaces
1 np+na icno/
1 nF-i-nv
1 nF-i-nfi
r? 1 nF-i-n^
u
1 f)Fj-f)A
1 fiF-ufi^
1 nF-i.n9
•
•
:
i
R2 R4 R4 R4
- 140%
- 120%
Mean recovery from AFSD
- 100%
- 80/o • Mean recovery from
60% sponge wipe
40% BCR(%)
20%
fW
Figure 3-3: Recovery from tile surfaces by AFSD type
There was an inverse relationship of CR to initial surface spore loading. This suggests that R4 can be
used for detection of relatively low spore loadings on non-porous surfaces.
Recoveries from aerosol samples are listed in Table 3-5.
Table 3-5. Recoveries from aerosol samples collected during tile tests
AFSD
R2
R4
R4
R4
CPU Present
(Stainless Steel)
5. 83 x 10s
9.96 x 10s
1.57 xlO4
2.80x 102
Mean Air filter
(CPUs)
6239
76
<1
<1
% CPU from sponge
wipe
0.086%
0.001%
0.003%
0.087%
Beater Bar
Yes
No
No
No
Re-aerosolization from R2 operation was much higher from tile than from laminate surfaces. This
suggests that the condition and type of the floor surface may be a larger factor in re-aerosolization than
the type of AFSD used within the current test conditions.
Carpet surfaces: Three AFSD types (R1, R2 and R3) were tested for sampling efficiency on carpet
coupon surfaces. R4 and R5 were not tested on carpet surfaces because of their recommended usage
only on hard surfaces, according to the instruction manuals. Similar to the laminate surface tests,
comparative recovery efficiencies for carpet sampling were determined by comparison of AFSD
recoveries to that of the vacuum sock sampling method. The test results are summarized in Table 3-6
20
-------
and Figure 3-4. The highest average CR was achieved with R2 (161%). This unit was also the most
effective AFSD type per surface area and time. R3, a vacuum only unit, demonstrated AFSD sampling
results similar to that of the vacuum sock method. R1 showed the lowest CR among three AFSD. It is
not clear why R1, which was equipped with an agitating brush bar, showed a lower CR than R2 or R3.
One reason might be that a rotating brush bar, such as that on R1, may not be effective for resuspending
sparsely distributed micron-size spores on surfaces. The brush is likely more effective for collection of
fibrous materials such as animal hair. However, a rotating flexible beater, such as that on R2, may come
in contact with a larger portion of the surface and therefore more effectively dislodge spores. There may
be other reasons to explain the low CR from R1, such as vacuum power, sampling speed, sampling
coverage area, etc. However, determining the effect of these variables was not part of the study
objective. One thing to note is that two R3 units failed during testing. It is questionable whether R3 is
reliable enough for incident field sampling following an actual incident.
Table 3-6. Summary of carpet surface sampling comparative recovery using AFSD
AFSD
Rl
R2
R3
Average sampling
duration (sec)
423
81
422
Mean recovery from AFSD
(CFUs xlO5)
1.3 ±0.4
1.0 ±0.2
2.4 ±0.4
Mean recovery from vacuum
sock (CFUs xlO5)
5. 2 ±2.7
0.6 ±0.2
2.6 ±0.6
CR (%)
26
161
92
Recovery from Carpet surfaces
l.OOE+06
l.OOE+05
l.OOE+04
l.OOE+03
Mean recovery from AFSD
I Mean recovery from
Vacuum sock
R2
R3
Figure 3-4: Recovery from carpet surfaces by AFSD type
The data from the air filter analyses during carpet sampling tests are shown in Table 3-7, and suggest
that re-aerosolization is likely for any AFSD.
21
-------
Table 3-7. Recoveries from aerosol samples during carpet tests
AFSD
R1
R2
R3
Recovery from Air Filters (CPUs)
507
273
760
% CPU from sponge wipe
0.098%
0.429%
0.288%
Beater Bar
Yes
Yes
No
Table 3-8 shows an indication of where within AFSD the spores were partitioned. The reservoir was
difficult to seal, and would be problematic to ship because of this. In nearly all cases, analysis of only the
filter would allow detection of the spores, while lowering demand on laboratory and shipping. Only R1
came with a HEPA filter, which should improve collection efficiency, while an aftermarket HEPA filter is
available for R2. The filter for R5 was not designed for dry particulates, but rather for liquid, so it is not
surprising that R5 demonstrated such low recovered from the filter.
Table 3-8. Partitioning of recovered spores in AFSD
AFSD
Rl
R2
R3
R5
Material Type
Carpet
Laminate
Carpet
Laminate
Carpet
Laminate
Tile
Carpet
Laminate
Laminate
Filter recovery
8.66 xlO4
4.20 xlO4
5.28xl04
5.36 xlO3
9.42 xlO1
1.77 xlO1
1.15xl04
2. 16 x 10s
9.51 xlO4
4.76 xlO3
Bin recovery
4.03 x 104
1.27 x 10s
S.OOxlO4
1.02 xlO4
8.19 xlO1
3.44 xlO1
1.56 x 10s
2.63 x 104
3.20xl04
2.05 x 10s
% on filter
65%
25%
51%
34%
54%
34%
7%
89%
75%
0%
3.4 Scenario-based Evaluation
Following the tests outlined in Table 2-3 and discussed in Section 3.3.1; two AFSD, R2 and R4, were
chosen for further evaluation with a larger sample area; R2 for evaluation on carpet coupons, and R4 for
evaluation on laminate coupons. Duplicate tests were performed with each of these AFSD. Results of the
hot spot tests are summarized in Table 3-9.
22
-------
Table 3-9. Results from "Hot Spot" testing
AFSD
R4
R4
R2
R2
AFSD Sampling
area (ft2)
47.6
47.6
47.6
47.6
Mean Recovery
from AFSD
(CFU/sample)
2. 09 x 10s
1.23xl07
5. 82 x 10s
1.41 x 10s
Surface Type and
(Sampling
Method)
Laminate
(sponge wipe)
Laminate
(sponge wipe)
Carpet (vacuum
sock)
Carpet (vacuum
sock)
Mean recovery from
surface CFU/cm2
1.85 xlO3 +1.52xl03
1.71 x!04 + 4.20 xlO3
2.00 x!03 + 1.81 xlO3
1.75 x!03 + 6.52 xlO2
CR (%)
121
78
31
9
A single inoculated coupon was placed in the middle of the larger sample area for these "hot spot" tests.
For all of these tests, the AFSD successfully sampled the inoculated section within the large floor and
recovered spores from that section. The recoveries from AFSD were higher from laminate floors than
from carpeted floors, as was the case for the scoping tests. When the R2 AFSD samples spores, it
sequesters them in the filter and bin, while the R4 AFSD keeps the spores on the mop in contact with the
floor. In theory then, the efficacy of an R4 AFSD could be inversely proportional to the area it samples.
The results from wide contamination tests are shown in Table 3-10. For these tests, lower inoculums were
used on a wider area of floor surface. Total CFU present was estimated by multiplying the total number of
inoculated coupons by the recovery (CFUs) from control coupons.
Table 3-10. Results from wide area "release" testing
AFSD
R4
R4
R2
R2
R2
R2
Surface
Type
Laminate
Laminate
Carpet
Carpet
Carpet
Carpet
Mean Recovery from
Materials (CFU/929 cm2)
4.51 x!02 + 4.53 xlO2
6.6xl01±3.8xl01
9.55xl025.18xl02
3.95 x!02 + 2.70 xlO2
2.27 x!02 + 4.2 xlO1
1.5xl02 + 9xl01
Inoculated
area
(929 cm2)
19
19
19
19
19
19
Estimated
Total CFU
present
8.57 xlO3
1.26 xlO3
1.81 X 104
7.50 XlO3
4.31 XlO3
2.84 xlO3
AFSD
Recovery
(CFU)
2.79 xlO3
4.90 xlO2
5.00 XlO3
1.60 XlO3
3. 16 XlO2
5.59 xlO3
CR (%)
33%
39%
28%
21%
7%
197%
23
-------
The test plan required collecting samples after AFSD operation from areas of test coupons not originally
inoculated. These samples were used to determine if the AFSD transferred spores from the contaminated
(inoculated) areas to areas not previously contaminated (i.e., cross-contamination). The percent of spores
transferred to sterile areas by the AFSD is the ratio of the estimated recovery from the non-inoculated
areas to the initial spore loading recovered from the inoculated areas using the respective comparative
sampling methods. These results are shown in Table 3-11.
Table 3-11. Spread of spores by AFSD
Dispersal
Hot Spot
Hot Spot
Wide Area
Wide Area
Hot Spot
Hot Spot
Wide Area
Wide Area
AFSD
R2
R2
R2
R2
R4
R4
R4
R4
Surface
Type
Carpet
Carpet
Carpet
Carpet
Laminate
Laminate
Laminate
Laminate
Pre-Test
Recovery on
Hot Spot
Post-Test
Recovery
on Hot
Spot
Avg. Post-
Test
Recovery on
non-
inoculated
areas
CFU/929 cm2
1.86 X 10s
1.63 X 10s
955
227
1.72 X 10s
1.59 X107
451
66
5.83 X 10s
4.84 X 10s
7
29
4.71 X 104
3. 17 X 10s
31
31
2.03 X102
2.09 X102
4
4
3.33 X103
6.68 X104
13
3
Avg. Post-
Test
Recovery on
non-
inoculated
areas
Total CPU
9.69 X 103
9.96 X 103
208
175
1.59 X 10s
3. 18 X 10s
606
152
Spores
transferred
to sterile
areas by
AFSD
%
0.5%
0.0%
21.8%
77.2%
9.2%
20.0%
7.1%
12.0%
These results are completely anticipated and can be used to great advantage in a field response. Any
area from which an AFSD proved positive for bacterial spores could then be characterized with traditional
surface sampling methods, yielding a greater chance of detection using methods accepted by the
response community. Collection of point surface samples within the hot zone by personnel could then
provide additional information, such as concentration gradients, which may help characterize original
distribution history. It is expected that an AFSD which encountered a hot spot towards the end of its
operation would cross-contaminate a smaller area, thereby introducing a bit of randomness in
interpretation. Regardless, further research is needed to determine the consequences of contamination
redistribution by the samplers.
24
-------
4 Quality Assurance
This project was performed under an approved Category III Quality Assurance Project Plan titled
Development of Automated Floor Sampling Device for Bacillus anthracis Spores (May 2012)
(available upon request).
4.1 Sampling, Monitoring, and Analysis Equipment Calibration
There were standard operating procedures for the maintenance and calibration of all laboratory
equipment. All equipment was verified as being certified calibrated or having the calibration validated by
EPA's on-site (RTP, NC) Metrology Laboratory at the time of use. Standard laboratory equipment such as
balances, pH meters, biological safety cabinets and incubators were routinely monitored for proper
performance. Calibration of instruments was done at the frequency shown in Tables 4-1 and 4-2. Any
deficiencies were noted. The instrument was adjusted to meet calibration tolerances and recalibrated
within 24 hours. If tolerances were not met after recalibration, additional corrective action was taken,
possibly including, recalibration or/and replacement of the equipment.
Table 4-1. Sampling and monitoring equipment calibration frequency
Equipment
Meter box
RH sensor
Stopwatch
Clock
Calibration/Certification
Volume of gas is compared to NIST-traceable dry gas
meter annually
Compare to 3 calibration salts once a week.
Compare against NIST Official U.S. time at
http://nist.time.gOV/timezone.cgi7Eastern/d/-5/java
once every 30 days.
Compare to office U.S. Time @ time.gov every 30
days.
Expected Tolerance
±2%
±5%
±1 min/30 days
±1 min/30 days
Table 4-2. Analysis equipment calibration frequency
Equipment
Pipettes
Incubator
thermometers
Scale
Calibration
Frequency
Annually
Annually
Before each
use
Calibration Method
Gravimetric
Compared to NIST-
traceable thermometer
Compared to Class S
weights
Responsible
Party
External
Contractor
ARCADIS
Metrology
Laboratory
ARCADIS
Acceptance
Criteria
±1% target
value
± 0.2 °C
±0.01% target
25
-------
4.2 Data Quality Objectives
The primary objective of this project was to determine the efficacy of various AFSD to collect biological
samples from contaminated floor surfaces. This section discusses the Quality Assurance/Quality Control
(QA/QC) checks (Section 4.3) and Acceptance Criteria for Critical Measurements (Section 4.4)
considered critical to accomplishing the project objectives.
The Quality Assurance Project Plan (QAPP) in place for this testing was followed with several deviations,
many of which were documented in the relevant sections of this report. Deviations included:
• Some samples from Test 01 ruptured primary containment during analysis. Samples were
recoverable due to secondary containment.
• Stainless steel control samples from Test 03 were inadvertently combined.
• Some Test 6 AFSD did not operate.
• The Test 7 blank AFSD recovery (CFUs) was high due to contamination
• Test 08B ViaCell Cassettes were used past the expiration date.
• Test 13 O2 AFSD became wedged on the coupon. This was likely due to a splinter on the edge of
the coupon which snagged the mop cloth.
• Test 14 O2 had only two positive control coupons. The third, which showed no detect, was
considered an outlier and removed from the data set.
• Two MDIs were used for Test 15 O2, and seem to be different, however not all stainless steel
controls were collected for both MDIs.
• Some samples required heat shock to enumerate due to contamination from another organism.
• One tile coupon for Test 16 O2 broke in half during installation. This did not seem to affect AFSD
operation or recovery.
• The Test 20 AFSD required manipulation from a stuck brush error.
• Test 20b was included despite a high lab blank value.
26
-------
4.3 QA/QC Checks
Uniformity of the test materials was a critical attribute to assuring reliable test results. Uniformity was
maintained by obtaining a large enough quantity of material that multiple material sections and coupons
could be constructed with presumably uniform characteristics. Samples and test chemicals were
maintained to ensure their integrity. Samples were stored away from standards or other samples which
could cross-contaminate them.
Supplies and consumables were acquired from reputable sources and were NIST-traceable when
possible. Supplies and consumables were examined for evidence of tampering or damage upon receipt
and prior to use, as appropriate. Supplies and consumables showing evidence of tampering or damage
were not used. All examinations were documented and supplies were appropriately labeled. Project
personnel checked supplies and consumables prior to use to verify that they met specified task quality
objectives and did not exceed expiration dates.
Quantitative standards do not exist for biological agents. Quantitative determinations of organisms in this
investigation did not involve the use of analytical measurement devices. Rather, the CPU were
enumerated manually and recorded. Critical QC checks are shown in Table 4-3. The acceptance criteria
were set at the most stringent level that could be routinely achieved and are consistent with the data
quality objectives described in Section 4.4. Positive controls and procedural blanks were included along
with the test samples in the experiments so that well-controlled quantitative values were obtained.
Background checks were also included as part of the standard protocol. Replicate coupons were included
for each set of test conditions when possible. Qualified, trained and experienced personnel using
SOPs/MOPs ensure data collection consistency. When necessary, training sessions were conducted by
knowledgeable parties, and in-house practice runs were used to gain expertise and proficiency prior to
initiating the research.
27
-------
4.4 Acceptance Criteria for Critical Measurements
Critical measurements (CM) are used to determine and assess the stated objectives and specify tolerable
levels of potential errors associated with simulating the prescribed decontamination environments. The
following measurements were deemed to be critical to accomplish part or all of the project objectives:
• enumeration of spores from traditionally surface sampling methods
• enumeration of spores from AFSD samples
The Data Quality Indicators (DQIs) listed in Table 4-4 are specific criteria used to quantify how well the
collected data met the DQOs. Failure to provide a measurement method or device that meets these goals
results in the rejection of results derived from the CM. For instance, if the plated volume of a sample is not
known (i.e., is not 100% complete), then that sample is invalid
28
-------
Table 4-3. QA/QC sample acceptance criteria
Sample Type
Negative Aerosol
Background
Samples
Negative coupon
control sample
Field Blank
Laboratory
Materials
Blank Tryptic Soy
Agar Sterility
Control
(plate incubated,
but not inoculated)
Reference
Control Coupons
(also puffing
control)
Puffing Control
Coupons (also
positive control)
Biological
Samples
Purpose
Determine extent of
cross-
contamination in
COMMANDER and
from each sampling
technique
Determine extent of
cross-
contamination in
COMMANDER
Verify the process
of moving coupons
does not introduce
contamination
Verify the sterility of
materials used to
analyze viable
spore count
Controls for sterility
of plates
Used to determine
the extent of
inoculation on the
coupon.
Used to determine
drift in the MDI
Controls for outliers
in colony growth
Acceptance Criteria
None
None
No detectable spores
No detectable spores or
some if Dahman says it
is okay.
No observed growth
following incubation
Target CPU, ±0.5 log
Target varies per test.
First set must be within
0.5 log of second set
The recovered (CPUs)
from the first set of
positive controls must
be within 0.5 log of the
second set of positive
controls
CPU counts between
30-300
Corrective Actions
If CPU detected, discuss
potential impact on
results with EPAWAM.
Repeat test if necessary
after identifying and
removing source of
contamination
Values on test coupons
of the same order of
magnitude will be
considered to have
resulted from cross-
contamination
Determine source of
contamination and
remove
Determine source of
contamination and
remove
All plates are incubated
prior to use, so any
contaminated ones will
be discarded
Outside target range:
discuss potential impact
on results with EPA
WAM; correct loading
procedure for next test
and repeat depending on
decided impact
Reject results and repeat
test
Replate or filter plate if
CPU outside criteria
Frequency
1 per sample
per sampling
technique per
test
3 per test
1 per sampling
type
1-3 per material
per test
Each plate
4 per test
Each sample
29
-------
Table 4-4. Critical measurement acceptance criteria
Critical
Measurement
Plated Volume
CPU/Plate
ViaCell® Total
Volume
Measurement
Device
Pipette
Visual
Inspection
Dry gas meter
Accuracy
±2%
± 10% (between
2 counters)
±10%
Precision
±1%
±5
±5%
Detection Limit
NA
1CFU
0.002 ft3
Completeness
100%
100%
100%
Plated volume critical measurement goals were met. All pipettes are calibrated yearly by an outside
contractor (Calibrate, Inc.).
Plates were quantitatively analyzed (CFU/plate) using a visual inspection method. For each set of results
(per test), a second count was performed on 25 percent of the plates with significant data (data found to
be between 30-300 CPU). All second counts were found to be within 10 percent of the original count.
There are many QA/QC checks used to validate microbiological measurements. These checks include
samples which demonstrate the ability of the NHSRC Biocontaminant Laboratory to culture the test
organism, as well as to demonstrate that materials used in this effort do not themselves contain spores.
The checks include:
• Negative control coupons: sterile coupons sampled alongside inoculated coupons
• Laboratory blank coupons: sterile coupons not removed from NHSRC Biocontaminant Laboratory
• Laboratory material coupons: includes all materials, individually, used by the NHSRC Biocontaminant
Laboratory in sample analysis
• Positive control coupons: coupons inoculated but not subjected to AFSD operation
• Inoculation control coupons: stainless steel coupons puffed at beginning, and end of each inoculation
campaign, not subjected to AFSD operation, to assess the stability of the puffer during the inoculation
operation.
The Vaisala RH meters were calibrated weekly and were within the factory specifications during each
AFSD operation.
4.5 Data Quality Audits
This project was assigned QA Category III and did not require technical systems or performance
evaluation audits.
30
-------
4.6 QA/QC Reporting
QA/QC procedures were performed in accordance with the QAPP for this investigation.
31
-------
5 Summary and Recommendations
The initial scoping tests consisted of testing three vacuum-based AFSD (R1, R2 and R3) and one wipe-
and one wet vacuum-based AFSD (R4 and R5, respectively) for sampling efficiency on a non-porous
surface (laminate). These tests showed that CRs for laminate surfaces were higher for the wet wipe- and
wet vacuum-based AFSD than the vacuum-based AFSD that were tested. The sampling process used by
the wet wipe-based AFSD is similar to the currently established wet wipe surface sampling method since
both methods use a PBST-wetted cloth in conjunction with a rubbing action on the surface. Low CRs
from vacuum units were expected since previous sampling studies have shown that the surface sampling
using the wet wipe or sponge wipe method on nonporous surfaces has higher recovery efficiency than
vacuum-based methods.
Similar to the laminate surface tests, CRs for porous material (carpet) sampling were determined by
comparison of the recoveries of three vacuum-based AFSD to that of the vacuum sock sampling method.
The test results showed that the recoveries from AFSD were on the same order or greater than the
vacuum sock sampling method. The differences in the test results among the three vacuum-based AFSD
may be related to the unique design of each AFSD and operating conditions.
Aerosol recoveries of spores observed during sampling for all five types of AFSD and surface material
types showed that small, but detectable, spore re-aerosolization can occur. The observed relative
differences in the level of spore re-aerosolization for each AFSD/material combination may be due to the
presence of surface agitation devices (brush or a beater bar) on these units, and the type of AFSD
sampling scheme (vacuum-based versus wet-wipe sampling).
Two top performers (R2 and R4) from the scoping tests were evaluated further in a more complicated
environment. The results from this test demonstrated that the AFSD were capable of sampling a hot spot
placed in the middle of the large area. For all of these tests, the AFSD successfully sampled the
inoculated section within the large floor and recovered spores from that section. Further, minimal
contamination of the non-inoculated adjacent surfaces was observed. The same type of AFSD used on
lower inoculated wide areas, showed comparable results to the more established comparative surface
sampling methods. These results are of a great importance in a field response to localize "hot spots" and
"secondary contamination" that may help design targeted decontamination strategies. Moreover, this
AFSD approach may possibly enable assessment of the contamination spatial distribution.
In addition to wide area sampling, these AFSD could be deployed to areas where human sampling is
difficult, such as inside HVAC ductwork and in highly contaminated areas (hot zones). Extending the use
of these devices for sampling of other biological, chemical, or radiological agents may also be pursued.
However, for real world application, these AFSD need further evaluation on larger spatial scales, with an
extended set of surface types, dissemination types (contamination scenario), surface loadings
(contamination surface concentration), and environmental conditions (relative humidity variation,
exposure duration, etc.).
32
-------
References
1. Rogers, J.V., C.L. Sabourin, Y.W. Choi, W.R. Richter, D.C. Rudnicki, K.B. Riggs, M.L. Taylor, and
J. Chang, Decontamination Assessment of Bacillus anthracis, Bacillus subtilis, and Geobacillus
stearothermophilus Spores on Indoor Surfaces Using a Hydrogen Peroxide Gas Generator.
Journal of Applied Microbiology, 2005. 99(4): p. 739-48.
2. Beecher, D.J., Forensic application of microbiological culture analysis to identify mail intentionally
contaminated with Bacillus anthracis spores. Appl Environ Microbiol, 2006. 72(8): p. 5304-10.
3. Weis, C.P., A.J. Intrepido, A.K. Miller, P.G. Cowin, M.A. Durno, J.S. Gebhardt, and R. Bull,
Secondary Aerosolization of Viable Bacillus anthracis Spores in a Contaminated US Senate
Office. JAMA, 2002. 288(22): p. 2853-8.
4. Teshale, E.H., J. Painter, G.A. Burr, P. Mead, S.V. Wright, L.F. Cseh, R. Zabrocki, R. Collins,
K.A. Kelley, J.L. Hadler, D.L. Swerdlow, and T. Connecticut Anthrax Response, Environmental
Sampling for Spores of Bacillus anthracis. Emerging Infectious Diseases, 2002. 8(10): p. 1083-
1087.
5. Raber, E., A. Jin, K. Noonan, R. McGuire, and R.D. Kirvel, Decontamination Issues for Chemical
and Biological Warfare Agents: How Clean is Clean Enough? International Journal of
Environmental Health Research, 2001.11(2): p. 128-148.
6. Schmitt, K. and N.A. Zacchia, Total Decontamination Cost of the anthrax Letter Attacks.
Biosecurity and bioterrorism : biodefense strategy, practice, and science, 2012. 10(1): p. 98-107.
7. Price, P.M., M.D. Sohn, K.S. Lacommare, and J.A. McWilliams, Framework for Evaluating
Anthrax Risk in Buildings. Environ Sci Technol, 2009. 43(6): p. 1783-7.
8. Bresnitz, E.A., Indoor anthrax Decontamination: How Clean is Clean? Journal of public health
management and practice : JPHMP, 2010.16(3): p. 185-8.
9. GAO, Anthrax Detection: Agencies Need to Validate Sampling Activities in Order to Increase
Confidence in Negative Results, 2005, U.S. Government Accountability Office: Washington, DC.
10. Brown, G.S., R.G. Betty, J.E. Brockmann, D.A. Lucero, C.A. Souza, K.S. Walsh, R.M. Boucher,
M.S. Tezak, M.C. Wilson, T. Rudolph, H.D.A. Lindquist, and K.F. Martinez, Evaluation of Rayon
Swab Surface Sample Collection Method for Bacillus Spores from Nonporous Surfaces. Journal
of Applied Microbiology, 2007. 103(4): p. 1074-1080.
11. Brown, G.S., R.G. Betty, J.E. Brockmann, D.A. Lucero, C.A. Souza, K.S. Walsh, R.M. Boucher,
M.S. Tezak, and M.C. Wilson, Evaluation of Vacuum Filter Sock Surface Sample Collection
Method for Bacillus Spores from Porous and Non-porous Surfaces. Journal of Environmental
Monitoring, 2007. 9(7): p. 666-671.
12. Brown, G.S., R.G. Betty, J.E. Brockmann, D.A. Lucero, C.A. Souza, K.S. Walsh, R.M. Boucher,
M. Tezak, M.C. Wilson, and T. Rudolph, Evaluation of a Wipe Surface Sample Method for
Collection of Bacillus Spores from Nonporous Surfaces. Applied and Environmental Microbiology,
2007. 73(3): p. 706-710.
13. Buttner, M.P., P. Cruz, L.D. Stetzenbach, A.K. Klima-Comba, V.L. Stevens, and P.A. Emanuel,
Evaluation of the Biological Sampling Kit (BiSKit) for Large-area Surface Sampling. Applied and
Environmental Microbiology, 2004. 70(12): p. 7040-7045.
14. Edmonds, J.M., P.J. Collett, E.R. Valdes, E.W. Skowronski, G.J. Pellar, and P.A. Emanuel,
Surface Sampling of Spores in Dry-Deposition Aerosols. Applied and Environmental
Microbiology, 2009. 75(1): p. 39-44.
15. Krauter, P.A., G.F. Piepel, R. Boucher, M. Tezak, B.C. Amidan, and W. Einfeld, False-Negative
Rate and Recovery Efficiency Performance of a Validated Sponge Wipe Sampling Method.
Applied and Environmental Microbiology, 2012. 78(3): p. 846-854.
16. Rose, L.J., L. Hodges, H. O'Connell, and J. Noble-Wang, National Validation Study of a Cellulose
Sponge Wipe-Processing Method for Use after Sampling Bacillus anthracis Spores from
Surfaces. Applied and Environmental Microbiology, 2011. 77(23): p. 8355-8359.
17. Estill, C.F., P.A. Baron, J.K. Beard, M.J. Hein, L.D. Larsen, L. Rose, F.W. Schaefer, J. Noble-
Wang, L. Hodges, H.D.A. Lindquist, G.J. Deye, and M.J. Arduino, Recovery Efficiency and Limit
33
-------
of Detection of Aerosolized Bacillus anthracis Sterne from Environmental Surface Samples.
Applied and Environmental Microbiology, 2009. 75(13): p. 4297-4306.
18. Hodges, L.R., L.J. Rose, H. O'Connell, and M.J. Arduino, National Validation Study of a Swab
Protocol for the Recovery of Bacillus anthracis Spores from Surfaces. J Microbiol Methods 2010.
81(2): p. 141-146.
19. Raber, E., W.J. Hibbard, and R. Greenwalt, The National Framework and Consequence
Management Guidance Following a Biological Attack. Biosecurity and bioterrorism : biodefense
strategy, practice, and science, 2011. 9(3): p. 271-9.
20. Valiante, D.J., DP. Schill, E.A. Bresnitz, G.A. Burr, and K.R. Mead, Responding to a Bioterrorist
Attack: Environmental Investigation of anthrax in New Jersey. Applied occupational and
environmental hygiene, 2003.18(10): p. 780-5.
21. Prassler, E., A. Ritter, C. Schaeffer, and P. Fiorini, A short history of cleaning robots. Autonomous
Robots, 2000. 9(3): p. 211-226.
22. Ulrich, I., F. Mondada, and J.D. Nicoud, Autonomous Vacuum Cleaner. Robotics and
Autonomous Systems, 1997. 19(3-4): p. 233-245.
23. Brown, J.S., J.A. Graham, L.C. Chen, E.M. Postlethwait, A.J. Ghio, W.M. Foster, and T. Gordon,
Panel discussion review: session four - assessing biological plausibility of epidemiological
findings in air pollution research. Journal of Exposure Science and Environmental Epidemiology,
2007. 17: p. S97-S105.
24. Lee, S.D., S.P. Ryan, and E.G. Snyder, Development of an Aerosol Surface Inoculation Method
for Bacillus Spores. Applied and Environmental Microbiology, 2011. 77(5): p. 1638-1645.
25. Worth Calfee, M., S.P. Ryan, J.P. Wood, L. Mickelsen, C. Kempter, L. Miller, M. Colby, A. Touati,
M. Clayton, N. Griffin-Gatchalian, S. McDonald, and R. Delafield, Laboratory Evaluation ofLarge-
Scale Decontamination Approaches. 2012.
26. Calfee, M.W., S.D. Lee, and S.P. Ryan, A rapid and repeatable method to deposit bioaerosols on
material surfaces. J Microbiol Methods, 2013. 92: 375-3967
27. Probst, A., R. Facius, R. Wirth, M. Wolf, and C. Moissl-Eichinger, Recovery of bacillus spore
contaminants from rough surfaces: a challenge to space mission cleanliness control. Appl
Environ Microbiol, 2011. 77(5): p. 1628-37.
28. EPA, Assessment of Liquid and Physical Decontamination Methods for Environmental Surfaces
Contaminated with Bacterial Spores: Development and Evaluation of the Decontamination
Procedural Steps 2012: Washington, DC.
29. EPA, Effectiveness of Physical and Chemical Cleaning and Disinfection Methods for Removing,
Reducing or Inactivating Agricultural Biological Threat Agents 2011: Washington, DC.
34
-------
Appendix A: Miscellaneous Operating Procedures (MOPs)
MOP 3135 Procedure for Sample Collection using BactiSwab™ Collection and Transport Systems
MOP 3144 Procedure for Wipe Sampling of Coupons
MOP 3145 Procedure for HEPA Vacuum Sampling of Large and Small Coupons
MOP 3150-AII Procedure for Fabrication of 14" x 14", 28" x 28", and 42" x 42" Material Coupons
MOP 3155 Procedure for Via-Cell Air Sampling
MOP 3165 Sponge Sample Collection Protocol
MOP 6535a: Serial Dilution: Spread Plate Procedure to Quantify Viable Bacterial Spores
MOP 6587: Retrieval and Processing of Biological Samples Collected by the iRobot Scooba Robot
Vacuum
MOP 6588: Retrieval and Processing of Biological Samples Collected by the Mint Automatic Floor
Cleaner
MOP 6589: Retrieval and Processing of Biological Samples Collected by the iRobot Roomba
MOP 6590 Retrieval and Processing of Biological Samples Collected by the P3 International P4920
MOP 6592 Retrieval and Processing of Biological Samples Collected by the Neato
-------
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
------- |