Outbreak, 2001

The index case of inhalational anthrax was associated with meningitis and occurred on 2 October 2001 in a patient who worked for America Media in Boca Raton, Florida. The case was detected by an alert infectious diseases physician, Dr. Larry Bush, who raised the diagnostic possibility of inhalational anthrax when examination of CSF specimens revealed typical gram-positive rods [1]. The second case was cutaneous anthrax, which was also detected by an alert infectious diseases physician in an NBC employee in New York City; it was reported to the US Centers for Disease Control and Prevention (CDC) on 9 October 2001 [2]. By the end of the year, there were 11 confirmed cases of inhalational anthrax and 7 confirmed cases of cutaneous anthrax, with an additional 4 suspected cases of cutaneous anthrax. These cases were identified in 4 states and in Washington, DC (table 1). Of the 22 cases, ⩾20 were definitively linked to mail contaminated with a single strain of Bacillus anthracis. The source of the anthrax in these cases appeared to be 5 letters sent through the US Postal Service to recipients in Florida, New York City, and Washington, DC. A letter sent to Senator Tom Daschle contained 2 g of powder containing B. anthracis spores that measured ∼1.5 × 3 µm, with an estimated concentration of B. anthracis spores of 10¹¹–10¹²cfu/g [3]. When extrapolated from a model of experimentally infected monkeys, the quantity of spores present in that letter is >10 million LD₅₀ doses [4, 5]. With regard to the putative agent, partial DNA analysis of isolates obtained from diverse sources (including environmental samples, clinical samples, and envelope contents) implicates a single strain. This strain is the Ames strain, which was originally isolated from an infected cow in Texas in 1981 [5, 6].

The CDC has provided leadership in dealing with the medical and epidemiological aspects of these cases, including guidelines for medical management [2, 7, 8] and prophylaxis [8–11]. These were based in part on the 1999 consensus statement from the Johns Hopkins Working Group on Civilian Biodefense [12]. Although some of these decisions have been controversial, the mortality rate for inhalational anthrax has been substantially lower than that previously reported, there have been no relapses after treatment, and there have been no relapses among the ∼10,000 persons who received prophylaxis after possible exposure to B. anthracis spores [13, 14].

Microbiology

B. anthracis is an aerobic, gram-positive, sporulating bacillus measuring 1–1.5 × 3–5 µm that grows readily on conventional microbiology media, including blood agar. Spores are not generally seen on direct Gram stain of specimens, but they appear in culture when nutrients are depleted. The usual time for growth is 6–24 h. The organism is easily recovered if samples are obtained before the administration of antibiotics. Thus, each of 8 patients with inhalational anthrax in the recent outbreak provided samples for culture before they received antibiotic treatment, the results of which were positive [14, 15]. In the Sverdlovsk anthrax outbreak in the former Soviet Union, which occurred after the accidental release of anthrax spores from a bioweapons facility, blood cultures were positive in all cases in which there was no antecedent receipt of antibiotics and for patients who had received antibiotics for <21 h [16]. In many cases, the microbial concentration in blood is extraordinary, as indicated by a positive Gram stain of the buffy coat in some cases, detection of positive growth within 6 h of incubation in some patients, and quantitative cultures that indicate concentrations of 10⁸ cfu/mL [14–18]. Nevertheless, antibiotic administration for >24 h virtually precluded recovery of pathogens from cultures of samples obtained from any site in this recent outbreak.

A tentative microbial diagnosis can be made in most clinical laboratories on the basis of recovery of gram-positive bacilli that are spore forming, nonmotile, nonhemolytic, penicillin sensitive, and encapsulated. Definitive diagnosis requires specialized testing, such as phage γ lysis, detection of capsule and cell-wall antigens by direct fluorescent antibody, B. anthracis—specific PCR, or serologic tests, by means of techniques and reagents that are available only in public health laboratories in the Laboratory Response Network [13]. The serologic test is an ELISA to detect IgG to B. anthracis protective antigen; it has sensitivity of 98.6% and specificity of 80%. Specificity is improved with use of a protective antigen—competitive inhibition ELISA as an additional confirmatory step [13]. Preliminary studies indicate that positive serologic results can be achieved as early as 10 days after the onset of symptoms, with a peak titer at 40 days after the onset of symptoms [13]. As noted, this test is not commercially available.

Inhalational Anthrax

Pathophysiological characteristics. Historically, in the United States, inhalational anthrax is a rare occupational disease. There have previously been only 18 reported cases in the United States in the past 100 years, the most recent of which was in 1976 [17]. Nevertheless, the pathophysiological characteristics, pathological changes, and clinical events have been well studied in nonhuman primate models [2, 19], the Sverdlovsk incident [16], and the 11 bioterrorism-related cases in the United States in 2001 [1, 14, 15, 17].

Particle size is a critical factor for aerosolization and effective inhalation and deposition of spores in the alveoli, where alveolar macrophages ingest the organisms and initiate the pathological process. With regard to aerosolization, it has been observed that opening an envelope with 0.1 g of Bacillus globigii, a nonpathogenic Bacillus species and a surrogate marker of B. anthracis with a particle size of 5 µm, results in extensive contamination throughout a 10 × 10 × 18—foot room within seconds [20, 21]. The organism has high affinity for alveolar macrophages, which endocytose the bacilli and transport them to mediastinal lymph nodes, where they may persist as spores or germinate to vegetative forms with the production of the 2 toxins (lethal toxin and edema toxin). The inoculum size necessary to initiate this process (the minimum infecting dose) is not known; the highly quoted figure of 8000–40,000 spores is the LD₅₀ in experimental infection of nonhuman primates, but the relevance of these observations to human disease is unclear, and extrapolation to define the minimum infecting dose is not possible [2, 19].

Pathological findings are available from the autopsies of 42 persons who died in the Sverdlovsk incident from 1979 (table 2) [16, 22]. These 42 patients were previously healthy and died within 1–4 days after the onset of symptoms. The most characteristic features were hemorrhagic thoracic lymphadenitis and hemorrhagic mediastinitis, which were found in all 42 patients. There was also spread of B. anthracis to peribronchial lymph nodes. The pulmonary portal of entry showed a focal hemorrhagic, necrotizing pneumonia in 11 patients, and 21 patients had hemorrhagic meningitis. Most of the patients had large hemorrhagic pleural effusions averaging 1700 mL per patient. Submucosal intestinal hemorrhages were noted in 93% of patients. Remarkable features of the disease were gelatinous edema of the mediastinum, large bloody pleural effusions, and arterial necrosis with hemorrhage. B. anthracis was recovered from blood cultures for all patients treated with antibiotics for <21 h; quantitative cultures showed concentrations of up to 10⁸ cfu/mL in blood, and Gram stains revealed large numbers of organisms at all infected sites. Gastrointestinal submucosal hemorrhagic lesions were noted in 39 (93%) of 42 patients.

Clinical features. The incubation period in the Sverdlovsk outbreak had a range of 2–43 days from the time of release (2 April 1979) to the onset of symptoms. For 9 of the 11 US patients for whom the time of exposure is known, the range was 4–6 days [14, 20]. In both the Sverdlovsk and US epidemics, most hosts were previously healthy adults with no comorbidities. In the United States, the mean age was 56 years (range, 42–93 years), which is somewhat older than the demographic profile of the exposed population [20]. The disease tends to be fulminant, and it tends to occur in 2 stages that may merge imperceptibly. The first stage is characterized by nonspecific flulike symptoms with fever, chills, drenching sweats, gastrointestinal complaints (nausea, vomiting, diarrhea, and abdominal pain), headache, cough, and chest pain. The second stage is characterized by dyspnea, fever, and shock. In the Sverdlovsk experience, the mean duration of the first stage was 3–4 days, and the median duration of the second stage before death was 1 day [16, 22].

Laboratory tests showed leukocytosis with a left shift; the median admission peripheral leukocyte count in 10 US patients was 9800 leukocytes/mm³, and the median peak WBC in these cases was 26,400 cells/mm³ [14]. The chest radiograph revealed the characteristic wide mediastinum, but this may be subtle and is considered nonspecific, because similar changes may be noted in tuberculosis, tularemia, sarcoidosis, histoplasmosis, lymphoma, silicosis, tumor, aneurysm, and alveolar proteinosis. Pleural effusions are particularly characteristic, and they are often large and usually contain bloody fluid. Although pneumonia is not a prominent feature of disease at autopsy, pulmonary infiltrates mimicking pneumonia may be seen radiographically; these infiltrates were seen in 7 of 10 patients in the 2001 outbreak [14, 20]. Particularly helpful have been findings of CT scans of the chest, which show highly characteristic features of hyperdense (hemorrhagic) mediastinal and hilar lymph nodes, mediastinal edema, and pleural effusions. As noted, blood cultures are positive for nearly all patients when blood samples are obtained before antibiotic therapy is initiated.

Clinical features in the US cases are summarized in table 3. Factors that appear to be particularly useful in distinguishing inhalational anthrax from other forms of pulmonary infections include the following: (1) the link to an epidemiological source —in this case, mail (for 9 of 11 patients) —with a relatively short incubation period (4–6 days); (2) selected clinical features, which include lack of coryza, presence of drenching sweats, and prominent gastrointestinal symptoms; (3) chest radiograph and chest CT scans that show typical mediastinal changes and pleural effusions; (4) bloody pleural effusions; (5) fulminant progression to shock in a previously healthy person; and (6) established or suspected isolation of B. anthracis in cultures for patients with clinically compatible illness.

The mortality rate for inhalational anthrax is high, but it is also variable (table 4). Historically, of the 18 naturally occurring cases in the United States in the 20th century, 19 were fatal, despite the use of penicillin to treat some patients. The Sverdlovsk incident was associated with a high mortality rate (68 deaths), but the actual figure is confused by uncertainty about the number of infected persons, which is the result of the nonavailability of clinical records [16].

Treatment. Essential elements of care include rapid initiation of antibiotic therapy; supportive care (i.e., intravascular volume repletion, with vasopressor and ventilatory support as necessary) and drainage of the large pleural effusions, usually with chest tubes, was required in most of the recent cases. Therapy with corticosteroids should be considered for patients with inhalational anthrax associated with meningitis or for patients who have severe mediastinal edema (table 5) [8]. Recommendations for antibiotic treatment of inhalational anthrax include use of either ciprofloxacin or doxycycline and are based on the observation that these agents are active in vitro against this outbreak strain. These agents have been approved by the US Food and Drug Administration (FDA) for the treatment of anthrax, and they have been shown to be effective in preventing infection in monkey models when given after aerosol challenge with a dose of 8 LD₅₀ [4]. The CDC recommendation is for combination therapy, which includes 1 of these 2 antibiotics plus 1–2 additional antibiotics that are known to be active in vitro against B. anthracis (table 6) [5, 7]. Because there was only a single strain responsible for the 2001 outbreak, decisions regarding initial treatment could be made empirically. Nevertheless, there have been prior reports of a genetically engineered strain with resistance to tetracycline and another to fluoroquinolones [23]. The implication is that antibiotic selection for this and for any future, similar episodes will be dictated by the results of in vitro susceptibility testing of the implicated strain.

The selection of the second or third antibiotic is arbitrary. Some favor the use of clindamycin on the basis of its property of inhibiting toxin production in static culture. Agents that penetrate the blood-brain barrier and that have established clinical efficacy in the treatment of bacterial meningitis, such as high-dose penicillin, are preferred for patients with inhalational anthrax associated with meningitis; these agents may be considered for all patients in view of the relatively high rates of this complication in older reports [18, 22]. Penicillin has been approved by the FDA for treatment of inhalational anthrax, and it performed well in the experimental nonhuman primate model [4], and penicillin is recommended as a second or third antibiotic (added to either ciprofloxacin or doxycycline). The CDC guidelines [2, 7] state that penicillin should not be used as monotherapy, because the strain implicated in this outbreak, as well as multiple historical strains, produces an inducible β-lactamase. The outbreak strain appeared highly susceptible to penicillin in vitro, but in patients with a high microbial load, as with inhalational anthrax, penicillin therapy might induce the β-lactamase and result in penicillin resistance in vivo. With regard to other fluoroquinolones, those commonly used for pneumonia (levofloxacin, gatifloxacin, moxifloxacin) show in vitro activity against B. anthracis comparable to that of ciprofloxacin, but these agents lack FDA approval for this indication and have not been tested in studies involving nonhuman primates; many authorities think that these agents would be comparable to ciprofloxacin.

The duration of therapy has created substantial controversy. Although there is consensus that the duration needs to be long, there is debate about how long is long enough. Prolonged treatment is supported by the observation that 5 of 29 monkeys treated for 30 days after inhalation challenge relapsed after antimicrobial therapy was discontinued [4], and, after aerosol challenge, spores persisted in trace amounts in lung homogenates of monkeys for up to 100 days [19]. The implication is that the organism persists as a spore (presumably in mediastinal nodes) during antibiotic exposure, and that it germinates to cause lethal disease after the withdrawal of antibiotics. It has been suggested that the probability of this late relapse may be directly correlated with the inoculum size [20], because a study involving nonhuman primates has shown that the retained number of spores at any given time is directly correlated with the initial inoculum used at the time of challenge [19]. The monkey model also shows that, when antibiotics alone are administered in the postexposure setting, animals do not develop protective antibodies and do not survive when rechallenged with B. anthracis; animals given vaccine plus antibiotics develop protective antibodies and are protected from rechallenge [4].

The original recommendation from the CDC for postexposure prophylaxis was a 60-day course of antimicrobials. However, on the basis of the aforementioned observations, recommendations were subsequently modified to permit 1 of 3 options: (1) a 60-day course of antibiotics followed by careful clinical observation, (2) extension of the course of antibiotics to 100 days, or (3) extension of antibiotic therapy to 100 days combined with administration of anthrax vaccine in 3 doses at 2-week intervals [24].

Cutaneous Anthrax

Epidemiological characteristics. Cutaneous anthrax occurs as a result of direct contact with B. anthracis. This has accounted for >95% of reported cases in the world, which are primarily due to occupational exposure. These include ∼200 cases per year in the United States in the early 1900s, 203 cases reported during 1955–1992, and 1 case of occupational acquisition since 1992 [25]. During the 2001 outbreak, there were 7 confirmed and 4 suspected cases of cutaneous anthrax. All 11 were associated with exposure to contaminated mail.

Pathophysiological characteristics. Cutaneous anthrax results from skin exposure to B. anthracis, presumably from surface contamination. It was previously thought that there needed to be a cutaneous breech, but this has not been a documented feature of the cutaneous cases from 2001. The inoculum size is unknown.

Clinical features. The incubation period generally has a range of 1–12 days [25], and, in the 2001 outbreak, it had a range of 1–10 days. The initial symptom is pruritus at the infected site, which is followed by a painless papule that progresses to a vesicle in 1–2 days, then erodes to a highly characteristic necrotic ulcer with a black center. The disease may remain localized, but some patients experience systemic symptoms, including fever, malaise, and headache, and there may be extensive edema surrounding the lesion that presumably reflects the activity of “edema toxin.” Major considerations in the differential diagnosis include receipt of a brown recluse spider bite, erysipelas, cellulitis, cat scratch disease, ecthyma gangrenosum, and ulceroglandular tularemia. Recommendations to establish the diagnosis are Gram stain and culture of a lesion specimen, the results of which should be positive in the absence of antecedent antibiotic treatment. Alternative tests, which are suggested in the event of negative culture results despite clinical suspicion, include biopsy for silver stain and immunohistochemical testing and serologic tests [5].

Treatment. The standard treatment advocated before the current outbreak was penicillin G (4–6 million U iv q.i.d.) or doxycycline (100 mg iv or po b.i.d.) for 7–10 days; for localized or uncomplicated cases, the recommendation is for penicillin V (500 mg q.i.d.) for 5–7 days [25]. The current recommendation for treatment of severe cutaneous disease (i.e., disease with signs of systemic involvement, extensive edema, lesions on the head and neck) is the same as summarized above for inhalational anthrax. For patients with less serious cutaneous disease, including those treated as outpatients, the recommendation is for ciprofloxacin or doxycycline provided orally (table 7) [5]. The initial recommendation was for an abbreviated treatment course, but the current recommendation is that antibiotic therapy should be continued for at least 60 days because of the assumption that cutaneous anthrax is likely to be associated with an inhalation exposure as well. Corticosteroid therapy is sometimes advocated for patients who have extensive edema.

Guidelines for children are difficult to formulate because of the infrequency of anthrax in this population. However, a 7-month-old child with cutaneous anthrax in the recent outbreak had serious, life-threatening disease with microangiopathic hemolytic anemia, which suggests that this population may also need aggressive therapy [26]. One adult had a similar complication with inhalational anthrax [14].

Anthrax Prophylaxis

Recommendations for prophylaxis were made by public health officials on the basis of epidemiological investigations. The primary goal is to prevent the rapidly progressing inhalational anthrax [5, 8–11]. Possible inhalation exposure to spores is determined by environmental cultures, review of clinical cases, tests of potentially contaminated source materials (such as envelope contents), and culture of nasal specimens. Highly variable results have been obtained with culture of nasal specimens and may reflect the lag time between inhalation of spores and the obtaining of nasal swabs [11]. This is consistent with an experimental model in which the yield of cultures of nasal samples was virtually zero when swab samples were obtained >24 h after exposure to spores [27].

Administration of prophylactic antibiotics is recommended for persons exposed to aerosolized B. anthracis in the space of contaminated areas (table 8). The clinical experience for ∼10,000 persons with a recommendation for prophylactic antibiotics given for 60 days showed adherence rates with a range of 45%–85%, depending on the rate of adverse reactions, method of study, and perception of risk [11 20]. With regard to adverse advents, surveys of patients who received anthrax prophylaxis indicated rates of adverse reactions of ⩾19% [20]. This high rate was determined on the basis of a self-administered or nurse-administered questionnaire, but there were no clearly identified cases in which there was a reaction considered serious in terms of hospitalization or death [20].

The anthrax vaccine (anthrax vaccine absorbed [AVA]) has been licensed by the FDA since 1970 for the preexposure prophylaxis for inhalational anthrax in persons at risk of acquiring the disease occupationally, such as woolen mill workers, laboratory workers, and veterinarians. The vaccine provides almost 100% protection to nonhuman primates against an aerosol challenge with B. anthracis (Ames strain) [24, 28]. The vaccine has not been approved by the FDA for use as postexposure prophylaxis. However, when administered in conjunction with antimicrobial prophylaxis, it is 1 of 3 options recommended by the CDC for treatment of those individuals thought to have been exposed to anthrax spores. In these patients, the vaccine should be administered as an investigational new drug. In an open-label safety study in which ∼16,000 doses of this vaccine were administered to ∼7000 at-risk individuals, local reactions at the vaccination site were the most commonly reported adverse reaction. Mild reactions (e.g., erythema and tenderness) occurred in up to 20% of recipients, but more-severe reactions (e.g., forearm edema) and systemic reactions (e.g., fever and chills) occurred in <1% of recipients [20]. Additional information on AVA (BioThrax; BioPort), such as product approval information and licensing action, is available at http://www.fda.gov/cber/products/biopava0131022.htm.

Conclusion

Bioterrorism must now be considered in the differential diagnosis of many enigmatic or common conditions, particularly those that are severe or have unusual epidemiological features. It is critical for physicians, public health officials, and law enforcement officials to work together to minimize the potential morbidity associated with these events. It has become clear with the recent outbreak that, in 2002, inhalational anthrax is virtually diagnostic of a bioterrorism event, and anthrax must be considered in the differential diagnosis of patients who have an epidemiologically defined risk and the clinical features summarized above. The role of the infectious diseases physician will be to remain alert for this possibility, to notify public health and law enforcement officials when appropriate, to provide appropriate consultation with regard to disease management, and to facilitate the development of local and regional plans for an appropriate response.