Executive Summary

On May 19, 2003, General Accounting Office officials testified before the Subcommittee on National Security, Emerging Threats, and International Relations, Committee on Government Reform. The title of the testimony was “U.S. Postal Service: Issues Associated with Anthrax Testing at the Wallingford Facility.”

The following recommendations were made during the testimony:

The impact of additional anthrax cases could result in illness or loss of life as well as loss of confidence in the nation’s postal system. Further, even though the health risk is probably low, it is uncertain; we therefore recommend that the Postmaster General, in consultation with CDC, EPA, OSHA, as well as any other relevant agencies and postal unions, for those facilities that were deemed to [be] free of anthrax spores based solely on a single negative sampling result, (1) reassess the risk level for postal workers at those facilities and the general public served by those facilities, (2) reconsider the advisability of retesting those facilities and employing the most effective sampling methods and procedures, and (3) communicate to the postal workers and the general public the results of the reassessment of health risk, the advisability of retesting, the rationale for these decisions, and other relevant information that may be helpful regarding the health of the postal workers and the general public (GAO 2003).

This report has been prepared by the U.S. Postal Service with assistance from several federal agencies and labor unions. The following agencies and unions participated in preparing the report: Department of Homeland Security (DHS)^{^[1]}, Centers for Disease Control and Prevention (CDC), Environmental Protection Agency (EPA), Occupational Safety and Health Administration (OSHA), American Postal Workers Union (APWU), National Association of Letter Carriers (NALC), National Postal Mail Handlers Union (NPMHU), and the National Rural Letter Carriers Association (NRLCA).

The participating organizations formed a workgroup to review and share information relating to the anthrax incidents, including sampling processes, epidemiology of the incidents, work practices and engineering controls, and other precautions instituted since October of 2001. The workgroup, with subject matter experts from CDC, EPA, and OSHA, discussed and analyzed the facts and reached conclusions on the current risk that may be posed to postal workers and the public as a result of the anthrax-related events of 2001. The mail handler's, city and rural letter carrier's unions reviewed and concurred with the workgroup's conclusions. The APWU reviewed and provided comments concerning those same conclusions.

The workgroup concluded that the anthrax risk level for postal workers in the facilities tested, and the general public served by those facilities, is negligible. No further sampling is warranted for those facilities that tested negative for anthrax spores. It further concluded that additional testing would not appreciably increase the safety of postal premises for employees and customers. Several factors contributed to this conclusion:

¿ Continuation of illness tracking by the Postal Service and federal, state, and local health agencies reveal that no epidemiological evidence of inhalational or cutaneous anthrax has occurred in postal employees or customers since November 2001.

¿ The Postal Service continues to use anthrax-related engineering controls and work practices that reduce the potential for a re-aerosolization event.

¿ Under no circumstances was a single sample used to assess a facility and clear it for continued operation, and no facilities were “deemed to [be] free of anthrax spores based solely on a single negative sampling result” (GAO 2003).

The Postal Service, in conjunction with the workgroup members, has developed a plan to communicate the findings of this report to employees and the public.

Contents

Chapter 1 Background................................................................ 1-1

Important Terms and Definitions................................................................... 1-1

The Inhalational Anthrax Disease Process................................................. 1-5

The Anthrax Attacks of 2001......................................................................... 1-6

Chapter 2 Postal Service and Governmental Testing of Postal Facilities 2-1

Postal Service Rationale and Procedures for Testing in the Fall of 2001.... 2-1

EPA and CDC Rationale and Procedures for Testing in the Fall of 2001....... 2-3

EPA ..................................................................................................... 2-3

CDC ...................................................................................................... 2-3

Summary of Testing Data............................................................................. 2-4

Remedial Efforts........................................................................................ 2-5

Chapter 3 Review of Existing Information....................................... 3-1

Epidemiology and Surveillance.............................................................. 3-1

Environmental Sampling and Analysis................................................ 3-2

Engineering Controls and Work Practices.................................................... 3-5

Verification Survey for Anthrax Engineering and Administrative Controls... 3-7

Chapter 4 Conclusions........................................................... 4-1

Review of the Basic Risk Questions.................................................... 4-1

Question One................................................................................ 4-1

Question Two............................................................................. 4-2

Question Three............................................................................. 4-3

Question Four.............................................................................. 4-3

Question Five.............................................................................. 4-4

Responses to the GAO Recommendations......................................... 4-4

Recommendation One................................................................. 4-5

Recommendation Two................................................................. 4-5

Recommendation Three............................................................ 4-6

Appendix A. Postal Service Sampling Strategy: A Chronology

Appendix B. Anthrax Pre-Screening Sampling Strategy

Appendix C. Anthrax Testing Data

Appendix D. Anthrax Verification Survey Template/Instructions

Appendix E. Supplemental Data from the Verification Survey

Appendix F. References

Appendix G. Abbreviations

Figures

Figure 1-1. Epidemic Curve for 22 Cases of Bioterrorism-Related Anthrax, United State, 2001 1-6

Figure 1-2. Cases of Anthrax Associated with Mailed Paths of Implicated Envelopes and Intended Target Sites 1-8

Tables

Table 3-1. Summary of Respondents to Verification Questionnaire.................................... 3-8

Table 3-2. Summary of Results of Verification Survey (%).................................................... 3-8

Chapter 1
Background

Important Terms and Definitions

This section defines the terms and definitions that are important for discussing risk in the context of this report. They are presented logically by topic area.

¿ Anthrax: The name of the acute infectious disease caused by the spore-forming bacterium Bacillus anthracis. Although it most commonly occurs in hoofed mammals, anthrax can infect humans. Anthrax spores can remain dormant for many years.

¿ Contamination: Bioterrorism involves the deliberate introduction of viable anthrax spores (or other infectious hazards) into an area or workplace, leading to contamination of surfaces and airspaces in the targeted location.^{^[2]} Naturally occurring anthrax spores tend to clump together, making them difficult to spread into the air. However, “weaponized” anthrax spores are treated to reduce clumping, which makes the spores easier to “aerosolize,” prolongs their ability to stay in the air, and increases the likelihood that they will be breathed in by potential victims and lead to inhalational anthrax. After the initial contamination episode ends, additional activity may resuspend the spores in the air (also known as re-aerosolization), which can lead to additional contamination events, especially when the contamination event is not easily observable. It is not known how common re-aerosolization is or how much of the original material is available for resuspension.

Environmental air samples taken during an anthrax contamination event can identify the presence of B. anthracis spores in a tested airspace, but air samples collected after an event may be negative if the spores have already settled onto surfaces. Environmental surface samples taken after an anthrax contamination event can help identify the presence of B. anthracis spores on a tested surface. A positive result indicates contamination at that location. Surface samples are the primary tool for evaluating contamination, but surface contamination is not the same as exposure. Surface contamination indicates that a given location has a potential reservoir of spores that, if contacted or disturbed, could lead to exposure.

¿ Cross-contamination: The transfer of anthrax spores from one surface to another after contact. Postal examples of cross-contamination include one mail piece contaminating an adjoining mail piece, a mail piece contaminating the sorting machine it passes through, and a contaminated sorting machine contaminating subsequent mail pieces. Cross-contamination is the mechanism that led to contamination in downstream postal facilities.

¿ Decontamination: A variety of physical and chemical treatments can be used to inactivate,^{^[3]} kill, or remove B. anthracis spores. Decontamination can “clean” a surface to greatly reduce any potential for future re-aerosolization and exposure episodes. For example, bleach or other specific wet cleaning methods can be used to reduce surface contamination levels. Also, fumigation (e.g., chlorine dioxide gas) can kill anthrax spores.

¿ Exposure: Human contact with spores in a contaminated location. Surfaces and airspaces may serve as exposure pathways. Touching a contaminated surface may cause skin exposure that may potentially lead to cutaneous anthrax. Mechanical forces applied to contaminated mail during the sorting process may create a plume of airborne spores, leading to possible inhalation exposures for workers operating the equipment or working nearby. The use of compressed air for cleaning has the potential to re-aerosolize settled spores and contaminate the nearby airspace, which may then lead to exposure for the workers in the vicinity and to contamination of new surfaces when the spores settle out.

Several factors and considerations influence the likelihood that a contaminated surface can contribute to exposure. They include worker proximity to such surfaces, worker activities and task patterns that might create aerosols, the ability of machines to aerosolize spores, use of compressed air, and locations of fans and ventilation systems.

¿ Primary aerosolization: Available information, from previously known cases, (Meselson et al.1994 and Inglesby et al. 2002) suggests that the greatest risk for inhalational disease to humans exposed to an aerosol of B. anthracis spores occurs when spores are first made airborne. An initial event may involve multiple aerosolizations. This period is called “primary aerosolization.”

¿ Secondary aerosolization: Following the period of primary aerosolization, B. anthracis spores may settle out on surfaces. The rate of settling is influenced by various factors such as particle size and air movement. Secondary aerosolization results from the disruption and resuspension of the settled particles and spores via physical force or activity (e.g., use of compressed air cleaning). Many variables affect the likelihood of a secondary aerosolization, including powder characteristics, particle charge, how spores are weaponized, the surface involved, climatic factors (e.g., humidity), and the nature of the human or mechanical activity that occurs in the affected area. Studies done in the Hart Senate Office building (Weis et al. 2002) show that some treated spores can be resuspended by routine activities such as walking near contaminated surfaces. The risk to humans associated with this level of resuspension is unclear.

¿ Infectious dose: The minimum dose needed to cause infections in humans. This dose is still uncertain with anthrax exposure. Individuals typically vary in their susceptibility to infection, so a given dose can affect some but not all individuals in a given group. However, the theoretical lower range of infectious dose for inhalational anthrax may be as few as one to three spores based on extrapolation from primate studies (Inglesby et al. 2002).

¿ Incubation period: The time interval between initial exposure to an infectious agent and the appearance of the first sign or symptom of the disease.

¿ Cutaneous anthrax: This is the most common type of naturally acquired anthrax infection (greater than 95 percent) and usually occurs after skin contact with contaminated products from infected animals (e.g., carcasses, meat, wool, hair, hides, or leather). Infection begins as a pruritic papule (itchy bump) or vesicle that enlarges and erodes (1–2 days), leaving a necrotic ulcer with subsequent formation of a central black eschar (scab). The lesion is usually painless, with surrounding edema, hyperemia (increased blood supply), and regional lymphadenopathy (swollen lymph glands). Patients may have associated fever, malaise, and headache. Historically, the case-fatality rate for cutaneous anthrax has been less than 1 percent with antibiotic treatment and 20 percent without antibiotic treatment. Following the bioterrorism attack in fall 2001, there were 11 patients with cutaneous disease with no fatalities.

¿ Inhalational anthrax: The most lethal form of anthrax; it can result from aerosolization of B. anthracis spores through industrial processing or intentional release. Inhaled spores may remain dormant in the lungs or lymphatic system for weeks to months before germinating. After germination in alveolar macrophages, vegetative organisms may replicate and cause symptomatic disease. Reported incubation periods have ranged from 1 to 43 days after initial exposure, depending on the dose of B. anthracis inhaled and the use of antibiotics. Person-to-person spread of inhalational anthrax has not been documented.

Disease may initially involve a prodrome (preliminary phase) of fever, chills, nonproductive cough, chest pain, headache, myalgia, and malaise. However, more distinctive clinical hallmarks include hemorrhagic mediastinal lymphadenitis, hemorrhagic pleural effusions, bacteremia, and

toxemia resulting in severe dyspnea (difficulty breathing), hypoxia, and septic shock.

Case-fatality rates for inhalational anthrax are high, even with appropriate antibiotics and supportive care. Among the 18 reported cases of inhalational anthrax in the United States during the 20th century, the overall case fatality was greater than 75 percent. Following the bioterrorism attack in fall 2001, the case fatality rate among patients with inhalational disease was 45 percent (5 of 11 cases), despite the availability of antibiotics and intensive medical care.

¿ Risk: In this report, risk is defined as the overall likelihood for contracting anthrax and the further likelihood of death. Risk is dependent on all the variables that affect contamination, exposure, disease detection, and medical treatment. The virulence of different anthrax strains can affect risk. The type of exposure (skin vs. inhalation) affects the risk of death since cutaneous anthrax is significantly less lethal than inhalational anthrax. The extent of anthrax spore weaponization can affect the risk of disease because weaponized spores are more likely to form aerosols, stay in the air longer to increase exposure, and are more likely to be inhaled deeply into the lungs. Variation in individual susceptibility further influences risk.

In general, all available information and evidence were considered when evaluating risk. This includes epidemiological findings, environmental sampling results, and engineering and exposure factor information. Environmental contamination results alone cannot be the sole determinant of risk.

¿ Negligible risk: For this report, the term negligible risk means that the presence of some residual anthrax spores from the 2001 attacks is possible, but the unknown amounts are unlikely to cause disease. It is not scientifically accurate to speak of “zero risk” given the limitations of current detection methods and data gaps concerning infectious dose and risk factor issues.

The Inhalational Anthrax Disease Process

The terms and definitions in the previous section help to describe the multi-step process associated with anthrax disease. Given its much higher mortality rate, the primary disease concern is inhalational anthrax. The prerequisites for inhalational anthrax are that contamination must occur, conditions must lead to aerosolization of spores, and individuals must be present during aerosolization for exposure to occur. Each of the steps in this sequence must occur to create the conditions for disease development. A variety of factors (e.g., the physical characteristics of spores and the proximity of workers to the contaminated location) affects the likelihood of each step to influence the resulting risk of disease. In addition, there are opportunities for prevention and control interventions to interrupt the sequence leading to risk and disease. For example, steps can be taken post-exposure, such as provision of prophylactic antibiotic medication, to reduce the likelihood of disease development.

In examining the risk of inhalational anthrax disease, this report addresses fundamental questions related to the prerequisite steps. The following hypothetical questions provide a framework for evaluating the key issues associated with potential risks:

1. Aside from the postal facilities already identified, what is the likelihood that other postal facilities or locations were contaminated via cross-contaminated mail in fall 2001?

2. Given that the Postal Service has taken measures intended to clean locations that were most likely to have been contaminated, what is the likelihood that the spores remain if undetected contamination occurred?

3. If these hypothetical undetected spores are still present in a facility, what is the likelihood that secondary aerosolization can still occur?

4. If secondary aerosolization were to occur, what is the likelihood that it would result in sufficient employee exposure to cause disease?

5. If employee exposure and illness were to occur as a result of the above steps, what is the likelihood that timely treatment would be provided?

This report presents background information for looking at these questions in the sections that follow, beginning with a review of the events of 2001.

The Anthrax Attacks of 2001

The following information is derived from CDC reports.

From October 4 to November 20, 2001, 22 cases of anthrax (11 inhalational, 11 cutaneous) were identified; five of the inhalational cases were fatal. Twenty (91 percent) case-patients either were mail handlers or exposed to worksites where contaminated mail was processed or received. Nine of the 11 inhalational cases occurred with postal and non-postal workers who directly handled mail or serviced mail-processing equipment (MPE). Four letters are known to have been mailed. The following figures, excerpted from a CDC report (Jernigan et al. 2002), summarize the mail path and resultant disease.

Figure 1-1 shows the relationship between the number, location, and timeline of all 22 known cases of anthrax in the United States, both inhalational and cutaneous, resulting from the two bioterrorism attacks. Two distinct case clusters were separated in time, with no cases occurring during the 13-day period between clusters. One case of inhalational anthrax occurred in a Connecticut resident 20 days after the second case cluster.

Figure 1-1. Epidemic Curve for 22 Cases of Bioterrorism-Related Anthrax,

United States, 2001

The first cluster of nine cases began approximately 4 days after the September 18 envelopes were mailed. The Jernigan study summarized the findings in the following manner:

All seven cases from New York City and New Jersey in the first case cluster were cutaneous anthrax; all five New York City cases included media company employees or visitors. Both New Jersey cases were in postal employees. The two cases from Florida were both inhalational anthrax and were in media company employees. Overall, eight of the nine persons in the first case cluster were exposed to worksites (postal facilities or media companies) that had environmental samples positive for B. anthracis [2002].

The second case cluster began about 5 days after the October 9 envelopes were mailed. Jernigan et al. summarized these findings as follows:

All five cases from the D.C. metropolitan area were in the second case cluster, all were inhalational anthrax, and all case-patients worked in postal facilities contaminated by the B. anthracis-containing October 9 envelopes. The last two cutaneous cases from New York City, whose onsets of illness occurred in the second case cluster, were known to have handled the September 18 New York Post envelope when it was moved in mid-October before its identification. Of the four New Jersey cases in the second cluster, two were inhalational anthrax in postal employees, one was cutaneous anthrax in a postal worker, and one was cutaneous anthrax in a bookkeeper who worked at a nearby commercial office building; all four case-patients were exposed to worksites that had environmental samples positive for B. anthracis [2002].

Based on established mail paths, the remaining two cases in this cluster are reported to result from cross-contamination with the B. anthracis-containing envelopes sorted at Trenton Processing and Distribution Center (P&DC) in Hamilton Township, NJ.

P&DCs employ high-speed, automated MPE to mechanically sort mail. The various sorting machines mechanically squeeze the letters during processing, which can create a bellows effect that may result in the formation of aerosols. During the 2001 attacks, there were nine anthrax cases involving Postal Service employees (six inhalational and three cutaneous), eight of which occurred at two of the four processing and distribution facilities—Trenton P&DC in Hamilton Township, NJ, and Brentwood P&DC in Washington, DC. No anthrax cases occurred at the other two P&DCs that processed B. anthracis-containing envelopes (West Palm Beach, FL, and Morgan Station, New York City).

The one postal worker anthrax case that occurred at a non-P&DC facility was a cutaneous anthrax case in a city letter carrier employed at the West Trenton Post Office (PO). Since this facility sends and receives mail through the Trenton P&DC, this case was likely caused by exposure via cross-contaminated mail. The West Trenton PO city letter carrier anthrax case was (1) the first Postal Service anthrax case, (2) the first case linked to unopened mail, and (3) the only Postal Service case that occurred at a facility without MPE. This city letter carrier never worked at nor visited the Trenton P&DC, but the mail that this carrier delivered on September 19 had been sorted at the Trenton P&DC on September 18 using the same machines that had sorted the New York City letters earlier that day (Greene et al. 2002).

Figure 1-2 shows the relationships among the four Postal Service P&DCs known to have processed a B. anthracis-positive, powder-containing envelope. It also shows where the four contaminated envelopes were recovered.

Figure 1-2. Cases of Anthrax Associated with Mailed Paths of Implicated Envelopes and Intended Target Sites

* Unopened envelope addressed to Senator Leahy found in a barrel of unopened mail sent to Capitol Hill, on November 16, 2001.

** Dotted line indicates intended path of the envelope addressed to Senator Leahy.

Chapter 2
Postal Service and Governmental Testing
of Postal Facilities

The initial sampling of potentially contaminated mail-handling facilities focused on following the mail trail and the pathway of disease occurrence from the AMI facility, in Florida, and the Trenton, NJ, and Washington (Brentwood), DC, P&DCs. The actual sampling of these facilities was performed by several organizations:

¿ U.S. Postal Service environmental contractors

¿ U.S. Environmental Protection Agency (EPA)

¿ Centers for Disease Control and Prevention (CDC)

¿ Federal Bureau of Investigation (FBI), generally for evidentiary purposes only.

Postal Service Rationale and Procedures for Testing in Fall 2001

In response to findings of contamination in a number of postal facilities, the Postal Service prepared interim guidance addressing issues such as decontamination, communication, employee notification, and interim cleaning procedures. It also launched a “pre-screening sampling” initiative to further determine whether anthrax spores were present in other postal facilities (USPS 2001). Mail flow data were used as initial criteria to identify facilities that received 1 percent or more of their mail from either the Trenton or Brentwood P&DCs. Several facilities (e.g., the Indianapolis Repair Facility and Kansas City Stamp Fulfillment Center) were included in the sampling initiative on the basis of plausible non-mail pathways such as potentially contaminated machine parts or stamp stock. The initiative included guidance for contractors on what sampling methods to use and which areas to test. This initiative resulted in the targeting of 179 facilities, including 109 P&DCs, 22 POs, and 48 other facility types. The three tables in Appendix C contain relevant information on the facilities tested.

The table in Appendix A shows the evolution of the sampling strategy that developed from October 16 to November 16, 2001, as the anthrax incident information flows were identified and pre-screening was assessed.

Several sampling methods were used at the outset of the initial response before the pre-screening initiative. They included the following:

(1) Wet swab sampling with RAPID™ polymerase chain reaction (PCR) sample analysis and laboratory plate culture analysis backup

(2) Dry swab sampling with analyses performed by PathCon Laboratories, (Pathogen Control Associates, Inc.)

(3) Dry swab, wet swab, wet wipe, and high efficiency particulate air sock (HEPA sock) sampling used by various agencies, with analyses performed at various state or contract laboratories using Laboratory Response Network (LRN) procedures.

The Postal Service’s objective was to provide consistent and uniform sampling procedures. Sampling protocols, using dry swab techniques and analyses of the samples by laboratories belonging to the Association of Public Health Laboratories (APHL), were finalized on November 5, 2001. The need for analysis of a large number of samples (more than 6,600) across 43 states necessitated close coordination with APHL to ensure that reliable analyses could be obtained in a short time.

At the time, a number of sample collection methods had been successfully used for evaluation of anthrax contamination. Of these methods, use of dry swabs was preferred by analytical laboratories because it minimized potential interferences and was the safest procedure for laboratory personnel. Laboratory capability and capacity were important issues, given the large surge in requests for analysis during November 2001. In summary, dry swabs were viewed, at this time, as an acceptable method for environmental evaluation. Dry swab issues are discussed further in the section on Environmental Sampling and Analysis of Chapter 3.

The new sampling procedures were transmitted to all sampling contractors on November 8, 2001. These procedures (see Appendix B) were uniformly implemented by the Postal Service in sampling that occurred on or after November 8, 2001. The Postal Service’s Interim Anthrax Guidelines draft sampling procedures, and supporting rationale for development of a consistent sampling approach and protocol, were submitted for Postal Service Headquarters review on November 16, 2001, and published as an interim guidance document on November 28, 2001 (USPS 2001).

EPA and CDC Rationale and Procedures for Testing in Fall 2001

EPA

In coordination with CDC and state health departments, EPA participated in the initial response activities in the Florida outbreak investigation. Those activities involved collecting wet-swab and wet-wipe samples in locations determined (targeted) to be the most likely to be contaminated. At Postal Service facilities, these areas included critical locations such as processing equipment, sorting boxes, and drop locations. Analysis was performed using CDC-approved culture methods at public or contracted laboratories.

Post-remediation sampling to verify the efficacy of decontamination involved collecting samples in the same locations where contamination was originally discovered. Additional samples also were collected in peripheral locations to determine whether the contaminant had spread.

CDC

In fall 2001, CDC participation was triggered by reports of anthrax cases from local health departments. In coordination with local health departments and the Postal Service, CDC provided technical assistance and performed outbreak investigations to identify the source of the exposure and determine whether additional public health interventions were needed (e.g., antibiotic prophylaxis and vaccines). In some cases, it selected facilities for testing because postal employees at that particular facility had contracted anthrax (e.g., Trenton and Brentwood P&DCs). In other cases, facilities were tested as part of epidemiologic investigations looking for clues on the role that cross-contaminated mail might have played in non-postal cases (such as the news media cases in West Palm Beach and New York City). Lastly, some facilities were tested based on sampling, epidemiologic, or mail-flow patterns that suggested cross-contamination of mail may have resulted in their contamination (e.g., all 50 downstream post offices from Trenton P&DC). Environmental testing targeted locations such as sorting machines and bins considered most likely to be contaminated. Sampling primarily employed wet swabs, along with some wet wipes and HEPA sock samples. CDC’s environmental sampling procedures are described in “Comprehensive Procedures for Collecting Environmental Samples for Culturing Bacillus anthracis” (CDC 2002). Samples were analyzed using CDC-approved methods by state health departments, CDC, and contract laboratories. CDC, the Agency for Toxic Substances and Disease Registry (ATSDR), and EPA collaborated on sample collection at the facilities affected by the Florida anthrax cases. In general, CDC testing efforts utilized wet swab methods along with wet wipes and HEPA sock samples. Dry swabs were used in the New York City outbreak investigations because they were consistent with the sampling analysis protocol used by the New York City
Department of Health Public Health Laboratory (Marfin et al. 2002). The overall sampling effort identified several cross-contaminated locations during October and November 2001.

Summary of Testing Data

The overall sampling process occurred as follows:

¿ There were 298 sampling efforts at 286 facilities.^{^[4]} These totals include efforts by the Postal Service (or its contractors), CDC, EPA, and FBI. More than one agency performed sampling at 12 of the facilities.

¿ Twenty-three facilities were considered to have some degree of anthrax contamination. This was based on positive environmental test results.

¿ The West Trenton PO was considered a special case in that it experienced a cutaneous anthrax case but no positive environmental test results. The FBI and CDC took a total of 42 samples at the West Trenton PO, including 6 HEPA sock samples.

¿ Three facilities were found positive in sampling conducted by more than one agency—Morgan P&DC, Brentwood P&DC, and West Palm Beach P&DC.

¿ Environmental sampling methods included wet swabs, HEPA socks, RAPID™ PCR backed by laboratory culture analysis, dry swabs, and wet wipes.

¿ Analytical methods included RAPID™ PCR and culture methods.^{^[5]}

¿ Sampling highlights for each of the agencies involved with sampling are provided below:

† Postal Service contractors tested 179 facilities and found six that tested positive. Three of the six facilities were found positive only by the Postal Service (Raleigh P&DC, Indianapolis Repair Facility, and Kansas City Stamp Fulfillment Services). However, the remaining three facilities were also found positive by other agencies (Brentwood P&DC, Morgan P&DC, and West Palm Beach P&DC).

† The CDC tested 112 facilities and found 12 that tested positive. Of the facilities testing positive, Morgan P&DC and Brentwood P&DC were also tested and found positive by other agencies. The 10 additional positive facilities were located downstream from the Brentwood and Trenton facilities.

† The EPA, with CDC involvement, tested and found seven facilities that tested positive. Of this group, West Palm Beach P&DC was also found positive by another agency. The remaining six facilities were located downstream from the West Palm Beach facility.

† The FBI tested an unknown number of facilities for forensic purposes. Of the facilities tested by the FBI, only South Jersey P&DC (also known as Bell Mawr P&DC) and the West Windsor/Princeton PO have been included in the total. These were the two facilities where the only positive results were reported by the FBI.

¿ The determination of contamination of a facility was never based on just a single sample; the fewest number of samples collected at any facility was four—at three downstream post office locations, one in the Washington, DC, area and two in Florida.

The specific sampling information and results, supporting these summary statements, are found in Appendix C.

Remedial Efforts

A detailed discussion of remediation efforts is beyond the scope of the GAO recommendations. However, it is appropriate to note that:

¿ All facilities suspected of contamination with B. anthracis were remediated (cleaned).

¿ The method and scope of remediation was based on sampling and epidemiological data.

A summary table of available decontamination and verification data, for facilities initially testing positive for B. anthracis spores, is found in Appendix C.

Chapter 3
Review of Existing Information

The workgroup analyzed available information to determine whether the risk level for postal employees had changed since earlier recommendations regarding additional sampling. This chapter summarizes observations made using existing information on epidemiology and surveillance, environmental sampling and analysis, and engineering controls and work practices put in place to reduce the risk. It also includes a summary of results from a Postal Service survey of downstream facilities that focused on implementation of the nationally required administrative and engineering controls and work practices.

Epidemiology and Surveillance

Observations made from information on epidemiology and surveillance include the following:

¿ Among the six cases of inhalational anthrax affecting postal workers, the mean duration between exposure and onset of symptoms was 4.5 days (range 4-6 days).

¿ Aside from the 2001 attacks, the longest reported incubation period from exposure to onset of clinical disease is estimated at 43 days on the basis of data from an accidental release of B. anthracis spores from a military microbiology facility in the former Soviet Union in 1979 (Meselson et al. 1994).

¿ Longer incubation periods were observed in experimental animal infection studies. The longest animal incubation period seen was 98 days after exposure (Glassman 1966; Henderson et al. 1956; Friedlander et al. 1993).

¿ The risk of developing a clinical case of anthrax is highest right after exposure occurs and declines afterward (Inglesby et al. 2002).

¿ The last known contaminated envelope was postmarked October 9, 2001, while the last case of inhalational anthrax occurred in a 94 year-old Connecticut woman, with the onset of symptoms reported on November 14, 2001.

¿ The FBI reported differences in the consistency of the B. anthracis powders used in the September and October mailings (Broad 2002; Collingwood 2002). Two possible explanations for the absence of inhalational anthrax in the postal workers in New York and Florida are that (1) the
B. anthracis spore preparation in the October 9 envelopes had a higher potential for aerosolization than that in the September 18 envelopes, and (2) the two mailings were made under or exposed to different environmental conditions (e.g., different amounts of moisture) (Perkins et al. 2002).

¿ The Postal Service had a worker health surveillance program underway from November 2001 through April 2002. Continuous surveillance at state health departments is still ongoing. None of these organizations are finding any new cases of inhalational or cutaneous anthrax.

¿ More than 2 years have passed since the original contamination events occurred. The absence of a new case provides strong evidence that continuing exposure to airborne spores, or spores on surfaces, is unlikely and that current conditions present only a negligible risk. This is especially important when consideration is given to the continuous operation of the mail system and ongoing use of automated MPE that would serve as the primary means of re-aerosolization of anthrax spores in sufficient quantity to cause disease.

Environmental Sampling and Analysis

Once the primary anthrax response was completed in fall 2001, steps were immediately taken to improve understanding of the existing environmental sampling tools. Limited information was available to compare the accuracy and consistency of the swab and wipe surface sampling methods or any of the air sampling methods. Furthermore, while the existing methods had been previously evaluated in laboratory settings with other types of spores expected to behave similarly to B. anthracis, the methods had not been validated specifically for B. anthracis or for treated spores. At that time, the dry swab method was believed to be an acceptable method (i.e., positive results were obtained using the method). The need for comparative studies was an important priority. Therefore, in December 2001, CDC and USPS partnered to use the closed Brentwood and Trenton facilities to initiate additional studies.

The research performed to evaluate surface sampling methods involved side-by-side sampling to compare the relative effectiveness of the different methods for detecting anthrax spores. This sampling was performed at the Brentwood P&DC (renamed Curseen-Morris P&DC) and in targeted locations believed to be contaminated. The research found that dry swabs, analyzed by the CDC method, detected B. anthracis spores 14 percent of the time, while other sampling methods were significantly better—wet swabs detected spores 54 percent of the time, HEPA sock samples, 80 percent of the time, and wet wipe samples, 87 percent of the time (Sanderson et al. 2002).^{^[6]} The swab and wipe samples were analyzed using APHL-approved procedures. These data were compiled by CDC and shared with the Postal Service, EPA, OSHA, and others in February 2002. These results led to a shift away from the use of dry swabs to the use of wet swabs, wet wipes, and HEPA sock samples. In summary, studies done after the bulk of USPS testing had been completed found that the dry swab method used in conjunction with the approved laboratory analytical technique was less sensitive than other methods to detect anthrax spores. The Postal Service incorporated this information and ceased reliance on dry swabs for sampling performed at Brentwood, Trenton, Wallingford, and other facilities undergoing remediation.

The studies referenced above were not conducted early enough and other field study research was not available to help inform the Postal Service in the choice of methods for the pre-screening sampling initiative. However, it is important to point out that (1) the need to complete the initial public health response was appropriately viewed as a higher priority than the research effort, (2) the research was performed as soon as possible, and (3) Postal Service support and cooperation were critical to completion of the expedited research.

The following items are relevant to the preceding sampling and analysis discussion:

¿ Approximately 48% of the 179 postal facilities addressed in the Postal Service pre-screening initiative were sampled with dry swabs before November 13, 2001. The swabs were cultured and any bacterial growth was analyzed using the PathCon Laboratories method that is somewhat different from the LRN APHL method.[7] It is possible that this earlier PathCon culture method involved a more efficient recovery of spores from the swab. However, no formal side-by-side studies are available.

¿ Dry swab testing performed at one of the facilities where employee cases occurred (Brentwood P&DC) and two facilities implicated in mail processing associated with non-employee cases (West Palm Beach P&DC and Morgan P&DC) did provide positive results.

¿ Most of the sampling performed at facilities with a high probability for contamination was done by methods other than dry swabs. For example, 13 facilities (4 P&DCs and 9 POs) were known to or may have processed one of the source letters.^{^[8]} Of those 13 facilities, only three were tested solely with dry swabs. This suggests that the facilities, where concerns would be expected to be highest, were evaluated with higher sensitivity sampling methods, thus reducing the likelihood that contamination was overlooked at the facilities where impacts were most likely.^{^[9]} In addition, it is recognized that the less sensitive method was used in those facilities or locales suspected to have less contamination. This issue is also addressed in Question 1 of Chapter 4.

¿ Much of the environmental surface sampling performed as part of the outbreak investigations by CDC (e.g., Trenton) involved wet swab methods and other methods such as wet wipes and HEPA sock samples. This sampling targeted key post offices downstream from P&DCs that processed the original contaminated letters.

¿ In no case was a facility “deemed to be free of B. anthracis spores based solely on a single negative sampling result,” during initial sampling. The lowest number of samples collected at any Postal Service facility was four, but the average was considerably higher.

In general, the Postal Service facilities that were found to have surface contamination fit an understandable pattern because they had a clear relationship to the path of contaminated mail pieces through the system. For example, contamination was most often found at the P&DCs and POs that processed the source letters. Contamination was less frequent at postal facilities directly downstream from those P&DCs and POs. Contamination at the downstream facilities was most likely caused by cross-contamination of mail. However, there were exceptions, such as the Kansas City Stamp Fulfillment Center, which cancels stamps for sale to stamp collectors. Nevertheless, even these situations could be explained. In this particular situation, some of the contaminated stamp materials had been stored in the registry cage at the Brentwood P&DC. Samples collected at Brentwood showed that spores did settle out on surfaces in this area. Thus, there was a clear relationship established and explainable pathway for this finding. These types of associations between positive findings and known exposure pathways increase the level of confidence that the main exposure pathways have been identified.

The main source of uncertainty in understanding the system-wide exposure pathways is the path of the letters sent to the Florida media company resulting in the index inhalational anthrax case. The letters were not recovered, so no postal code information was available to identify the specific P&DCs that may have been involved other than the West Palm Beach P&DC.

The following bullet items address other important environmental sampling points:

¿ Although not the focus of this report, ‘characterization’ is the additional sampling used to help delineate the extent and understand the spread of contamination once anthrax (or other sampling target) is identified. Based on experience from the recent anthrax attacks, many environmental sampling experts believe that characterization that is more extensive should be performed earlier in the investigative process if another attack takes place. This conclusion is a key finding in the May 2002 report of an independent peer review panel that evaluated the adequacy of the characterization performed at the Morgan P&DC in New York City (EPA 2003).

¿ In general, environmental surface samples were the most effective in identifying contamination, whether taken during initial assessment or characterization. While surface samples help to identify the location of contamination, without knowledge of other exposure factors they do not provide results that are directly translatable to the level of worker risk.

¿ No data exist on the limits of detection (i.e., the minimum concentration of anthrax spores that can be detected) for environmental sampling methods. CDC is collaborating with EPA and the U.S. Army to define such limits.

¿ Additional sampling would reduce, but not eliminate, uncertainty because the sensitivity of the methods used is unknown.

Engineering Controls and Work Practices

It is recognized that engineering, work practices, and administrative controls are the primary means of reducing exposure to workplace hazards. Engineering controls minimize employee exposure by either physically reducing or removing the hazard at the source, or isolating employees from the hazard. Work practices and administrative controls minimize employee exposure by altering the process in which tasks are performed or managed.

The Postal Service issued nationwide directives, requirements and guidance that formally established the following engineering controls and work practices for postal facilities:

¿ On October 26, 2001, the Postal Service issued “Interim Custodial Cleaning Procedures,” which eliminated the use of compressed air for all custodial cleaning. In a policy memorandum, November 1, 2001, the interim procedures were made mandatory. Compressed air was identified as a major exposure factor associated with concern for inhalational anthrax. Data collected suggest facilities are complying with the policy.

¿ In these interim procedures and policy memorandum, the Postal Service replaced compressed air cleaning of sorting machines with HEPA vacuum cleaning. The daily use of HEPA vacuuming over 2½ years should further reduce any hypothetical contamination. For example, a delivery bar code sorter that was HEPA-cleaned once daily, 6 days per week from November 1, 2001, through May 1, 2004, has undergone 806 cleanings. This repetitive cleaning lowers the likelihood that resampling of machine mail paths would result in any positive findings.

¿ At a teleconference held on or about October 26, 2001, the U. S. Postal Service Chief Operating Officer directed the Area Vice Presidents to have all facilities with automated MPE engage in a one-time bleach cleaning of that equipment.

¿ The Postal Service developed and implemented engineering controls and new work practices to reduce potential exposure to anthrax and educate employees for improved job safety (original policy start dates are in parentheses):

† Mandating a one-time cleaning of automated MPE with bleach and subsequent cleanings with a HEPA vacuum (October 26, 2001).

† Custodial mopping and cleaning of workroom floors and other surfaces with wet methods using a 10 percent solution of household bleach (November 1, 2001; in effect until June 19, 2002).

† Eliminating dry sweeping and dusting (November 1, 2001). Permission and instruction were given on the use of treated dust mops for cleaning floors and cleaning of treated dust mops with HEPA vacuums (November 13, 2001).

† Using HEPA vacuums for custodial cleaning and cleaning of mail processing equipment (November 1, 2001); heating, ventilation, and air-conditioning (HVAC) systems (March 5, 2002); and vehicles (October 30, 2001).

† Using HEPA vacuums and wet methods to clean high bay areas (February 28, 2002).^{^[10]}

† Banning the use of personnel cooling fans (October 2001).^{^[11]} The ban was modified to allow for cooling fan use in delivery units, manual distribution operations, docks, trailers and non-mail
processing locations. Directing fans at automated MPE continues to be banned (February 28, 2002).

† Providing filtering facepiece respirators and gloves to employees who request them (October 16, 2001) and instructing the washing of hands with soap and water when gloves are removed and before eating (October 26, 2001).

† Establishing “suspicious mail and powder” handling protocols (November 28, 2001).

† Providing continuous training on emergency plans and anthrax-related subjects in mandatory safety talks (November 28, 2001).

¿ During the ensuing 2½ years, numerous routine operations have occurred representing potential sources of aerosol formation. These include sorting machine maintenance, machine removal, high bay cleaning, custodial cleaning, and renovation and alterations.

Verification Survey for Anthrax Engineering and Administrative Controls

Between October 16, 2001 and June 27, 2002, the Postal Service issued nationwide directives and updated guidance and requirements documenting the establishment of the engineering controls and work practices described previously. These actions culminated in the August 2003 release of MMO-047-03, Consolidated Policy on Custodial Cleaning, which formalized current policy on cleaning activities for postal operations.

On the basis of preliminary totals of facilities tested, the Postal Service distributed 284 questionnaires (see Appendix D) to the tested facilities on September 5, 2003. The Postal Service sent seven more surveys to facilities identified after agencies reconciled the total. The purpose of this data call was to verify that the engineering controls, work practices, and administrative measures put in place to protect employees from anthrax exposure had actually been implemented. Of the 291 questionnaires sent out, 274 were returned.^{^[12]} Table 3-1 summarizes the numerical breakdown of respondents by facility type and postal area.

Table 3-2 summarizes the results of responses to each of the verification questions and reflects telephone follow-up information described in the paragraphs following the table.

Table 3‑1. Summary of Respondents to Verification Questionnaire

Postal Area	Large Facilities with MPE^a	Small Facilities with MPE^b	Facilities with no MPE	Total
Capital Metro	16	5	48	69
Eastern	19	0	1	20
Great Lakes	14	1	0	15
USPS Headquarters	0