We identified 22 cases of anthrax that occurred after envelopes containing B. anthracis–positive powder were mailed to persons in news media and government. Inhalational and cutaneous disease followed exposure to B. anthracis spores; five people died. These cases represent the first reported bioterrorism-related outbreak of anthrax. The investigation of these cases reveals important findings for detecting and preventing infections from bioterrorist attacks.
We tested B. anthracis isolates from patients, powder-containing envelopes, and environmental samples from news media, government, and postal processing worksites and found all tested isolates to be indistinguishable by molecular typing methods. Similar U.S. postal service-issue envelopes containing powder preparations of these B. anthracis spores were mailed from the Trenton, New Jersey, area on at least two dates. Although isolates, envelopes, and originating postal paths were similar, characteristics of cases differed by date of mailing and geographic region.
Patients in the cluster that occurred after the September 18 mailing were more likely to have cutaneous disease and to have been exposed at news media facilities rather than at postal facilities. Case-patients in the cluster that occurred after the October 9 mailing were more likely to have inhalational disease and to have been exposed at postal facilities along the path of envelopes sent to U.S. senators. Postal workers exposed to B. anthracis from the October mailings had predominantly inhalational disease. The case-fatality ratio for all cases of inhalational anthrax was 45%, a ratio lower than previously reported (33
); the estimated incubation period of 4.5 days for inhalational cases was consistent with previously reported findings (1
The fulminant systemic illness associated with the October mailing to U.S. senators differed greatly from the less severe cutaneous cases in media company employees in New York City, suggesting that substantial illness and death likely might have occurred among senate office staff after implicated envelopes were opened. Exposure to B. anthracis spores from processing unopened envelopes at the Hamilton and Brentwood postal facilities went unrecognized until after the implicated envelope was opened at the Hart Senate Office Building. Administration of postexposure chemoprophylaxis likely prevented further cases in postal workers and almost certainly averted disease in senate staff. Estimates derived from mathematical models support this conclusion (34
). Our findings suggest that prompt use of antimicrobial prophylaxis following suspected bioterrorist attacks can prevent disease.
Differences in the consistency of B. anthracis powders between the September and October mailings have been reported by the Federal Bureau of Investigation and may account for the preponderance of inhalational cases in the second cluster (35
). The later mailings may have intentionally contained a smaller particle-sized powder to produce greater harm. Media company employees had less severe disease than did the postal workers along the path of envelopes sent to senators.
Our findings indicate that the clinical and epidemiologic presentations of a bioterrorist attack depend on the population targeted, the characteristics of the agent, and the mode of transmission. With naturally occurring outbreaks of infection, early cases identified often provide clues to the mode and source of exposure. For bioterrorism-related disease, characteristics of initial cases may be misleading if terrorists vary the mode and source of exposure. Further understanding is needed of the role of different B. anthracis powder formulations in the mode of exposure and illness characteristics of persons exposed.
Cases of anthrax occurred in persons near those targeted for infection and also in those along the mail path of spore-containing envelopes. After the mailing of the September 18 envelopes, cases of cutaneous anthrax occurred, but were initially unrecognized, in workers at the postal processing center in New Jersey where the implicated envelopes originated. After the mailing of the October 9 envelopes, inhalational disease was identified in workers at postal facilities in the District of Columbia and New Jersey. Investigators did not anticipate the exposures and fulminant disease in those exposed to aerosols of B. anthracis spores from unopened envelopes along the path of the mail. No prior experience with mailed B. anthracis–positive, powder-containing envelopes is described in published reports; previous descriptions of aerosolized B. anthracis spores indicated that risk for re-aerosolization or resuspension of spores was low (37
). Previous preventive strategies for presumed B. anthracis exposures now appear inadequate in light of recent findings. Before this incident, antimicrobial prophylaxis was recommended only for direct exposures to the envelopes, and limited decontamination was suggested only for the immediate site of envelope opening (38
). Cutaneous and inhalational disease in postal workers in our investigation clearly shows that sealed, B. anthracis–positive, powder-containing envelopes can be a source of infection, presumably via the airborne route, for persons processing contaminated mail in postal facilities. Airborne transmission at the Brentwood and Hamilton facilities may have been facilitated by the use of high-speed sorters, as well as air-blowers used for routine cleaning (12
). Any future investigations of bioterrorism-related anthrax should evaluate persons potentially exposed along the path of the delivery vehicle as well as those targeted by the attack.
We found most cases of anthrax to be epidemiologically linked to sites contaminated by implicated envelopes; however, not all cases had direct exposures to targeted worksites, implicated envelopes, or mail-processing facilities along the mail path. Two cutaneous anthrax patients, a mail carrier and a bookkeeper in New Jersey, were not exposed to contaminated postal facilities or media companies. Only one of many environmental samples of surfaces at the bookkeeper’s office, where mail was received, was positive for B. anthracis. Cross-contaminated mail may be a likely exposure source for anthrax for both these cases.
The possibility of B. anthracis exposure from envelopes secondarily contaminated from implicated postal facilities greatly extended the group of potentially exposed persons in our investigation. Experience with anthrax related to agricultural or industrial sources indicated that direct exposure to animals, animal products, and wool-processing facilities accounted for most reported cases (1
). Contamination of the environment in animal and wool-processing facilities has been shown, and occasional cases due to secondarily contaminated items have been reported as a possible source of anthrax (1
For our investigation, contamination found at postal processing facilities off the direct mail path of implicated envelopes indicates that cross-contamination of mail occurred; however, enhanced surveillance for anthrax cases in multiple regions has not identified additional cases. Two patients with inhalational anthrax, a hospital worker in New York City and a retired woman in Connecticut, had no exposure to media or government worksites, implicated postal facilities, or possible sources of naturally occurring anthrax (40
). Neither patient had evidence of B. anthracis contamination at her home (or workplace for the New York City case), yet both were infected with B. anthracis isolates indistinguishable from the outbreak strain. Postal processing facilities in New York City and Wallingford, Connecticut, were contaminated with B. anthracis, suggesting cross-contaminated mail as a possible source of B. anthracis exposure for both cases.
From our investigation, B. anthracis–positive powder appears capable of contaminating other mail during processing, leading to exposure and subsequent development of cutaneous and possibly inhalational anthrax. The risk from cross-contaminated mail appears to be extremely low; 85 million pieces of mail were processed at facilities in New Jersey and District of Columbia after the October 9 envelopes, and no additional anthrax cases were detected through stimulated enhanced hospital-based surveillance of 10.5 million people in metropolitan areas around those postal facilities (41
). Although the risk for B. anthracis infection from cross-contaminated mail may be low, investigations of future bioterrorist attacks with B. anthracis–positive powders should consider the potential role of secondarily contaminated items in transmission of disease. An attack using a greater number of spore-containing envelopes would likely lead to many more cases due to cross-contaminated mail (42
Throughout the investigation, various reporting mechanisms were used to enhance detection of cases, including prospective syndromic surveillance in emergency departments and intensive-care units, laboratory-based surveillance, networks of clinicians such as dermatologists, and worksite absenteeism monitoring. In general, most cases of anthrax were detected through reports from clinical laboratorians and clinicians and from patient self-reporting. The role of the news media in increasing patient, clinician, and laboratorian awareness of anthrax was likely an important factor in stimulating case detection and reporting. Health departments sent alerts to health-care providers and provided training seminars for clinicians to improve case detection. Before the bioterrorism-related anthrax cases in 2001, clinician recognition of clinical findings suggestive of cutaneous or inhalational anthrax is presumed to have been very low (43
). For our investigation, cases in the first cluster associated with the September 18 mailing went unrecognized until B. anthracis was identified in a culture of cerebrospinal fluid from the index case in Florida, underscoring the critical role of the laboratory in initiating the investigation.
These first unrecognized cutaneous cases demonstrate the potential difficulties in detecting cases from a covert bioterrorism agent release. Once the possibility of anthrax exposures at media companies was recognized, along with subsequent environmental work site samples positive for B. anthracis, cases of cutaneous anthrax were more readily detected and reported. During the investigation, rapid dissemination of clinical findings through broadcast e-mail and fax alerts to hospitals and providers, public health reports, and networks of clinical, laboratory, and public health officials provided important tools to frontline clinicians to improve recognition of anthrax. Enhancing the knowledge and skills of clinicians and laboratorians for diagnosing bioterrorism-related infections and improving collaborations between clinicians and public health practitioners will set the stage for better detection of cases associated with any future acts of bioterrorism.
Our investigation had several limitations. The detection of anthrax cases involved numerous local, state, and federal public health and law enforcement officials. Because of the widely distributed activities of various investigators and the need to act quickly in identifying potential exposure sources, data collection instruments were not uniform. Collation of information across sites was limited to a select set of demographic, exposure, and risk factor data elements. The wide use of postexposure prophylaxis, along with difficulty in obtaining detailed information about potentially exposed persons, prevented general estimates of anthrax attack rates for many sites. Surveillance case definitions required laboratory confirmation of disease or of environmental exposure and thus may have missed cases of disease that were treated empirically without appropriate cultures (e.g., illness empirically treated as infected spider bites, which was actually cutaneous anthrax). Environmental sampling of potentially contaminated facilities used different testing methods; because less sensitive testing methods were used, certain sites may have underrepresented the degree of contamination. Throughout the investigation, there was a continuing need to refine study methods and redetermine intervention recommendations, since prior experience with bioterrorism-related anthrax was lacking. Finally, because the public health investigation was also a criminal investigation, information that may have contributed epidemiologic information may not have been available to many public health investigators because it was protected for use in prosecution.
The attacks initiated response activities in all states across the United States and in other countries and required considerable resources to support investigative efforts at the local, state, and federal levels. The perpetrator has not been apprehended, and new cases can still occur. Continued collaboration with law enforcement officials is required, and clinicians, laboratorians, public health officials, and the general public should remain alert for patient symptoms or findings that might indicate additional cases of bioterrorism-related anthrax.