Case Definitions

In this multistate outbreak, all sites adopted the Centers for Disease Control and Prevention (CDC) case definitions for anthrax (2). A confirmed case was defined as a clinically compatible illness that was laboratory confirmed either by isolation of B. anthracis from an affected tissue or site, or by two supportive laboratory tests. A suspected case was defined as a clinically compatible illness with no isolation of B. anthracis and no alternative diagnosis, but with one positive supportive laboratory test or a clinically compatible illness epidemiologically linked to a confirmed environmental exposure to B. anthracis.

Supportive laboratory tests included demonstration of B. anthracis in a clinical specimen by immunohistochemical staining; detection of B. anthracis DNA by polymerase chain reaction from specimens collected from an affected tissue or site; or the presence of anti-protective antigen immunoglobulin G (anti-PA IgG) by enzyme-linked immunoadsorbent assay (3).

Case Investigations

Suspected and confirmed case-patients were interviewed about symptoms, employment, and other possible exposures, and their medical records were reviewed. Coworkers and supervisors were also interviewed. For case-patients who were USPS employees, job assignments and time sheets were reviewed, with special attention to dates when letters containing anthrax spores were postmarked. Blood, tissue, and microbiologic samples were obtained and sent for testing. When possible, the incubation period was defined as the time between the date of likely exposure to spore-containing envelopes and the onset of symptoms.

Surveillance

Initial case finding involved investigation of potential cases reported by health-care providers, hospitals, and the public directly to the health department. Subsequently, we initiated stimulated passive hospital-based surveillance to identify additional inhalational anthrax cases (4). Infection control professionals from 61 hospitals in 15 counties in New Jersey, Pennsylvania, and Delaware, serving an area of 6.7 million residents, provided daily totals of emergency department and intensive-care unit admissions and reported all patients meeting broad clinical criteria (such as respiratory failure or febrile illness without a confirmed alternative diagnosis) for possible inhalational anthrax. Passive surveillance for both inhalational and cutaneous anthrax cases was conducted statewide in New Jersey and in parts of Pennsylvania and Delaware that are contiguous to New Jersey. Surveillance was enhanced through electronic communication with local health departments, press releases, and postings on websites of the New Jersey Department of Health and Senior Services (NJDHSS) and two New Jersey medical associations. All persons with possible anthrax identified through surveillance were followed up through telephone calls to the patients, the physicians and nurses treating them, and requests for laboratory specimens.

Exposure Assessment

To identify locations where exposures to letters containing B. anthracis spores might have occurred, we tracked the path of the contaminated letters through the Trenton PDC by obtaining information collected by the USPS for routine tracking and quality control. We also determined how mail flows to and from the PDC as it is brought from and delivered to other postal facilities and to the public.

Attack Rates

We reviewed the time sheets and specific work locations of the PDC employees working on the night of October 9, when the letters destined for Washington, D.C., were sorted. The number of employees working on this shift and the number of employees working on subsequent shifts were determined by review of available records and interviews with the PDC postmaster. Some records remained unavailable for review because the PDC was closed. We calculated attack rates for inhalational anthrax by dividing the number of cases by the total number of employees in the specified area.

Environmental Sampling

Initial sampling, conducted October 18–19, focused on the identified path of the letters in the Trenton PDC and public access areas of the PDC. When samples taken from areas along the path of the letters were found to be positive for B. anthracis on the following day, we developed a sampling strategy to evaluate the extent of contamination in the building and further characterize the risk to postal employees and visitors.

During October 21–November 9, sampling was conducted in a wider horizontal distribution around the areas of the initial positive samples and vertically upward toward the ceiling of the PDC. Sampling was performed on machinery located beyond the original path of the letters, the ventilation system, lookout galleries (enclosed elevated corridors), administrative areas on the mezzanine level, and the roof rafters. Sampling techniques included swab sampling with sterile moist swabs to collect settled dust and vacuum sock sampling with portable HEPA-filtered vacuum to collect surface dust over large areas (5).

Other mail facilities in New Jersey through which the recognized contaminated letters could have passed were identified and sampled. Most samples from these facilities were collected from areas where the initial mail-sorting activities were conducted. Additional samples were collected from customer areas, receiving bins of indoor mailboxes, cleaning equipment, loading docks, ventilation systems, computer work stations, and at least one delivery vehicle from each site. After the identification of cutaneous anthrax in an office worker who was not a PDC employee, sampling was performed at this case-patient’s workplace and home; the focus was on areas where mail might have been placed or opened.

Laboratory

B. anthracis screening identification of human and environmental samples was performed at the NJDHSS Public Health and Environmental Laboratories according to Bioterrorism Laboratory Response Network Level A and B protocols, with modifications to enhance the recovery rate of spores from environmental samples (6,7). Agar slants with isolates identified as B. anthracis by the Public Health and Environmental laboratories were sent to CDC’s Anthrax Laboratory, where identification of B. anthracis was confirmed by standard microbiologic procedures and the Laboratory Response Network testing algorithm (6–8). Antimicrobial susceptibility patterns were determined for B. anthracis isolates by using National Committee for Clinical Laboratory Standards breakpoints for staphylococci (9). Isolates of B. anthracis recovered from clinical specimens and environmental samples were typed to determine their genetic relatedness by using multiple-locus variable-number tandem repeat analysis (MLVA) (10).

Intervention

Postexposure prophylaxis was made available to potentially exposed persons pending results of environmental testing. We recommended continuation of postexposure prophylaxis for a total of 60 days for persons considered to be at risk for inhalational anthrax (11). A series of three postexposure prophylaxis clinic periods, each involving several sessions, were organized at two local hospitals. Seven or 10 days of antibiotics were dispensed at the initial clinic, and 25 days of antibiotics were dispensed at each of the two follow-up clinics. Hospital staff were available for consultation with persons who could not attend the formal clinics. At the initial clinic, ciprofloxacin was provided. The recommended antibiotic for postexposure prophylaxis was changed to doxycycline for the two follow-up clinics, after testing showed the B. anthracis isolates were susceptible to doxycycline (12). Antibiotics were obtained from the National Pharmaceutical Stockpile.

Employees who did not attend the clinics were contacted by telephone and encouraged to come to the clinic. To promote adherence, fact sheets and a newsletter were developed and distributed, reminders for postexposure prophylaxis clinics were posted at work sites, and weekly meetings were held with USPS management and representatives from each of the four postal unions. A health education team conducted focus groups with postal employees and conducted a health education campaign.

Demographic and Clinical Characteristics of Cases

From October 18 to October 24, six persons with anthrax were identified in the New Jersey area, including three with confirmed cutaneous anthrax, one with suspected cutaneous anthrax, and two with confirmed inhalational anthrax (Table 1). Their median age was 44 years (range 35–56 years); four were women. Five were USPS employees; four worked at the Trenton PDC, and one was a mail carrier at the West Trenton post office. The sixth case-patient was a bookkeeper at a Hamilton Township, New Jersey, office.

Figure 1

Figure 1. Timeline of events during bioterrorism-related epidemic, New Jersey, September–October, 2001. Red box = l case-patient with onset of inhalational anthrax; blue box = l case-patient with onset of cutaneous anthrax.

The incubation period was 5–9 days (median 8 days) for the three cutaneous cases whose exposure date could be estimated, and 5 and 6 days for the two inhalational cases. The dates of onset were clustered: two case-patients had onset of symptoms 8 and 9 days after the letters sent to New York City were processed at the Trenton PDC on September 18, and four case-patients had onset of symptoms 5–6 days after the letters sent to Washington, D.C., were processed on October 9 (Figure 1). Five of the patients were hospitalized—both persons with inhalational anthrax and three persons with cutaneous anthrax. No case-patients died. Demographic and clinical descriptions of the New Jersey case-patients are summarized in Tables 1–3 and presented in detail elsewhere (1,12–14).

Surveillance

Surveillance was initiated on October 24, and from October 24 to December 17, 2001, hospital infection control practitioners reviewed 240,160 emergency department visits and 7,109 intensive-care unit admissions. Four hundred sixty-four patients who met initial criteria for possible inhalational anthrax were reported to the NJDHSS; 214 (46%) required additional follow-up to rule out inhalational anthrax. Ninety-eight patients with suspicious cutaneous lesions were reported; 26 (27%) were assessed further to rule out cutaneous anthrax. Anthrax was ruled out in all patients; no additional cases were identified (4).

Exposure Assessment and Mail Flow

Figure 2

Figure 2. Floor map of the Trenton Postal Distribution Center in Hamilton Township with locations of environmental samples taken October–November, 2001, and work stations of New Jersey case-patients on dates when letters containing...

The Trenton PDC occupies 281,387 square feet or approximately 7,000,000 cubic feet and is divided into a mail-processing area and administrative and public access areas. Approximately 2 million pieces of mail are processed through the facility each day. The recognized spore-containing letters destined for New York City and Washington, D.C., took similar paths as they were processed through the facility. The letters received a barcode on one of three advanced facer canceller system machines (AFCS) and were then sorted through one of two delivery barcode sorters (DBCS 70 and 71), high-speed machines that read the barcode and sort approximately 30,000 letters per hour into bins according to destination (Figure 2). The letters destined for New York City were sorted through DBCS 70 or 71 in the late afternoon of September 18. Both letters destined for Washington, D.C., were processed in the late afternoon of October 9 through AFCS 3 within approximately 15 minutes of each other, followed by sorting on DBCS 70 within 2 minutes of each other. After sorting, the letters were packed into trays in the packing area and loaded onto mail trucks (Figure 2).

In general, mail that receives the Trenton postmark at the Trenton PDC comes from one of 50 local post offices in central New Jersey, or it is dropped off in a mailbox at the Trenton PDC. We could not determine the source of the letters containing B. anthracis. After processing at the Trenton PDC, mail with a Trenton postmark follows one of three routes:1) it is returned to one of the 50 local post offices for local delivery; 2) it is transferred to one of 12 other PDCs in New Jersey, Philadelphia, or Delaware; 3) or it is routed through the transfer facility in Carteret, New Jersey or the airmail center in Newark en route to destinations throughout the world. We confirmed that the recognized spore-containing letters were routed through the Carteret transfer facility en route to their destinations in New York City and Washington, D.C. At the Carteret transfer facility, mail is not unwrapped or handled; it remains in tubs that are transferred from one truck to another.

Potential Case Exposures

Case-patient 1 (Table 1) was a mail carrier at the West Trenton post office, which sends and receives its mail through the Trenton PDC. She never worked at the Trenton PDC and did not visit that facility. The mail that this carrier delivered on September 19 had been sorted at the PDC on September 18 on the same machines that had sorted the New York City letters earlier that day. On September 28, 10 days later, cutaneous anthrax developed in this mail carrier.

Case-patient 2 (Table 1) was a machine technician. He worked on September 18, 2001, when the letters to the New York City media outlets were sorted. This technician circulated throughout the letter-sorting area, was responsible for maintenance and repair of the high-speed sorters, and used compressed air to blow out dust and debris from the machines. Cutaneous anthrax developed in this man 8 days later.

Case-patient 3 (Table 1) was working on October 9, when the letters containing B. anthracis bound for Washington, D.C., were processed. Although he began his shift working in a different area of the facility, he later moved to operate the DBCS 70 that had sorted the letters containing B. anthracis earlier that evening (Figure 2). Cutaneous anthrax developed 5 days later.

The two New Jersey inhalational case-patients (Case-patients 4 and 5, Table 1) were also working on the night of October 9. They stood side by side at the input subsystem sorters, machines located next to the AFCS and the DBCS that sorted the B. anthracis–containing letters (Figure 2). Inhalational anthrax developed in these case-patients 5 and 6 days later, respectively.

Case-patient 6 was a bookkeeper at a Hamilton Township accounting firm; she did not visit the Trenton PDC. However, mail delivered to both her home and workplace came directly from the Trenton PDC without passing through an intermediate local post office. The bookkeeper’s onset of cutaneous anthrax was October 17, eight days after the Washington, D.C., destined letters were processed at the Trenton PDC.

Attack Rates for Inhalational Anthrax after Exposure to Washington, D.C.–Destined Letters

The two case-patients with inhalational anthrax (Case-patients 3, 4, Table 1) were identified among 750 Trenton PDC employees who worked in the processing area of the facility during or after the letters addressed to Washington, D.C., were processed on October 9 (overall attack rate 0.25%). The two persons with inhalation anthrax were among 170 who worked in the sorting area on the October 9 shift when the letters transited, for an attack rate of 1.2%.

Environmental Sampling

Of the 137 samples obtained at the Trenton PDC, 57 (42%) were positive for B. anthracis (Figure 2, Table 4). Positive samples were located throughout the facility, including samples taken from rafters as high as 25 feet above the plant floor and samples from the ventilation system (Figure 2). Twenty-five (83%) of 30 samples were positive in the area where the letters containing B. anthracis were sorted. Positive samples were identified from the machines at which Case-patients 4 and 5 worked and from the sorting machine that processed mail destined for the workplace and home of Case-patient 6.

In addition to the samples collected at West Trenton Post Office, we obtained a mean of 18 samples (range 4–27 samples) from each of 49 local post offices. Five of the local post offices had one positive sample each. The positive sample in each facility came from an area where mail from the Trenton PDC is deposited. One of the samples was obtained underneath a sorting machine, three were obtained from mail containers or the place where mail containers are stored, and one was from a bin inside a mailbox outside the post office. All 54 samples collected from the West Trenton post office (where Case-patient 1 worked) were negative. All 14 samples from the Carteret facility were negative.

Of 21 samples collected from the workplace of Case-patient 6, one grew B. anthracis. This sample was obtained from a tray near the receptionist’s desk that held delivered and outgoing mail. None of the samples collected from the home of Case-patient 6 were positive, including samples collected from her mailbox and areas where she stored and opened her mail.

Of the 10 environmental isolates typed by MLVA (4 from locations throughout the Trenton PDC, 5 from the local post offices, and 1 from the workplace of Case-patient 6), all were indistinguishable from clinical isolates.

Interventions

We recommended 60 days of postexposure prophylaxis for 1,069 employees of the Trenton PDC, as well as for persons who visited the facility and spent >1 hour on the plant floor from September 18 (the date the first letter containing B. anthracis was processed in the Trenton PDC) to October 18, 2001 (the date the facility was closed). Beginning October 20, a total of 885 (83%) Trenton PDC postal workers were provided with the full 60-day course of postexposure prophylaxis. Of the 184 (17%) postal workers who did not receive 60 days of antibiotics, 29 (3%) did not receive any antibiotics, 40 (4%) came to the initial clinic only; and 115 (11%) came to the initial and first follow-up clinics. Most postal workers (1,032 [97%]) obtained their antibiotics from Hospital A; 37 (3%) obtained antibiotics from their private physicians.

Three hundred twenty-four visitors to Trenton PDC went to Hospital A (n= 175), Hospital B (n=129), or their private physicians (n=20) for prophylaxis. Of these, 206 (64%) received 60 days of antibiotics, 85 (26%) received <60 days, and 33 (10%) did not receive any antibiotics.