New questions have arisen as public health authorities have investigated and responded to the intentional release of B. anthracis
in the United States. Studies by Canadian investigators with a sophisticated preparation of Bacillus globigii
have shown that a contaminated envelope may, even unopened, cause a substantial primary aerosol event (6
). In light of this new appreciation, we investigated whether, after a remote contamination event and initial decontamination, a Delivery Bar Code Sorter machine could be a continual source of aerosolized B. anthracis
spores and, if so, whether the particle concentration in the air could be estimated.
Initial reports indicated that no specific remediation had yet been undertaken on the contaminated machine. Subsequently, we learned that the surface of the sorter implicated in processing the contaminated envelopes had been cleaned with 0.5% hypochlorite solution. We proceeded with testing because the expectation was that topical cleaning would provide only fractional decontamination of a contaminated machine. By focused sampling, we found that, despite topical cleaning, the sorter remained contaminated with B. anthracis. By either swab technique or Rodac plates, 9 of 10 sites on the machine were positive and 4 sites produced B. anthracis colonies that were too numerous to count.
Air sampling detected B. anthracis before and after the sorter was activated. Before the sorter was turned on, the samplers detected a single B. anthracis-containing particle (0.0010 agent-containing particles per liter of air [ACPL]). Six colonies of B. anthracis (0.0061 ACPL) were identified in the 990 L of air sampled after sorter activation. The difference between the number of B. anthracis-containing particles detected by the samples collected as background and those collected after the sorter was activated was not significant at the 0.05 level (p=0.06); however, analysis suggests a trend toward a significant increase.
Environmental surface sampling done shortly after the Brentwood mail facility was closed found widespread contamination of the facility with B. anthracis
). Both aerosolization of B. anthracis
spores and direct cross-contamination of surfaces were considered likely mechanisms for contamination. Approximately 30 h of air sampling with open-faced 37-mm mixed cellulose ester filters (0.8-µm pore size) was negative. The previous report of negative air sampling despite extensive testing suggests our detection of airborne B. anthracis
while the sorter was inactive may have been spurious and possibly related to investigator activities while the experiment was being set up.
Based on these concentrations and assuming 100% sampler collection efficiency, the estimated number of B. anthracis–
containing particles that a worker might inhale near this activated sorter can be calculated. If we assume a normal ventilation rate (10 L/min), during 8 h working around this partially cleaned, but still contaminated sorter, a worker might be expected to inhale approximately 30 B. anthracis-
containing particles. This finding of very low-level airborne B. anthracis
contamination is supported by the negative testing of the mask filters. If all the airborne particles are assumed to be of optimal size for inhalation, this estimate is approximately 100-fold less than the lower boundary of the 50% lethal dose estimates for inhalational anthrax in nonhuman primate studies (8
). This number is also approximately 20-fold less than estimates of the number of routinely inhaled B. anthracis
–containing particles from a 1960 study of asymptomatic, unvaccinated workers in a goat-hair mill in Pennsylvania (9
). In that study, investigators calculated that, in an 8-h workday, workers inhaled >1,300 viable B. anthracis-
containing particles, 510 of which were <5 μm in size. Thus, although detected in the Brentwood facility, airborne contamination was at a relatively low level.
The comparison of this type of exposure with nonhuman primate anthrax data and historical industrial anthrax data is problematic for several reasons. Our understanding of human infection risk at very low-dose B. anthracis
exposures is limited, as illustrated by the death from inhalational anthrax of an elderly Connecticut woman for whom no exposure could be determined, despite extensive environmental testing of her home and areas she frequented (1
). A well-known contributor to the rate of alveolar deposition of a bioaerosol is the particle size distribution; because the slit sampling method does not measure the aerodynamic particle size distribution, we were unable to measure this attribute. Finally, historical comparisons to goat-hair mill workers are limited by the unknown contributions of prior host immunity, incomplete surveillance, and the lack of additional environmental sampling data other than the study from Pennsylvania.
This study shows that a mail sorter may remain contaminated, as indicated by surface sampling, many days after processing B. anthracis–contaminated letters and despite topical bleach cleaning. In addition, even after processing >1.2 million subsequent letters, as this sorter did, aerosolized B. anthracis–containing particles can still be detected around a contaminated sorter when active, at a level likely increased over background levels. At the time of our study, the level of B. anthracis–containing particles around this contaminated sorter at Brentwood was low, but any level of aerosolized B. anthracis spores is undesirable in this occupational setting.
Further studies are essential to define the risks of inhalational anthrax in the settings of both primary and secondary aerosolization of B. anthracis spores. In anticipation of potential future B. anthracis exposures, re-aerosolization potential should be evaluated in other environments, such as an office setting. In addition, size stratification of the re-aerosolized portion of a primary release should be part of any testing, to give some guidance as to risk stratification for exposed persons. Finally, better understanding of human health risk of low-dose exposure of B. anthracis spores is critical to guide optimal public health response.