7.3.1. Developing countries, naturally acquired anthrax
7.3.1.1. Mild uncomplicated cases
Penicillin G is still the drug of choice in the therapy of naturally-occurring anthrax in most parts of the world. In mild uncomplicated cases of cutaneous anthrax, the treatment usually recommended is intramuscular procaine penicillin, 500 to 600 mg (800 000 to 1 million units) every 12–24 hours for 3–7 days (procaine penicillin is frequently marketed in vials containing 800 000 units in developing countries). Intravenous antibiotic therapy is not recommended in mild cases. If the patient rejects intramuscular injection, penicillin V (500 mg taken orally every 6 hours) or amoxicillin (500 mg orally every 8 hours for 3–7 days) are acceptable alternatives. Cutaneous lesions usually become sterile within the first 24 hours of such regimens and the accompanying oedema usually subsides within 24 to 48 hours but, although early treatment will limit the size of the lesion, it will not alter the evolutionary stages it must go through (Gold, 1955; Kobuch et al., 1990).
7.3.1.2. Severe or potentially life-threatening cases
In patients exhibiting signs of systemic involvement, such as with inhalational or gastrointestinal anthrax, meningoencephalitis, sepsis, or cutaneous anthrax with extensive oedema, antibiotics should be given intravenously. Penicillin G is generally the first choice, 2400 mg (4 million units) every 4–6 hours by infusion (total daily dose should be 20–24 million units) until the patient’s symptoms resolve with the temperature returning to normal. At this point consideration may be given to switching to the intramuscular procaine penicillin regimen described in section 7.3.1.1. Also in severe cases, penicillin or another chosen antibiotic, e.g., ciprofloxacin, may be combined with another appropriate agent, preferably one that penetrates well to the central nervous system (CNS) (see also section 7.3.1.7). Penicillin G may be combined with clindamycin or clarithromycin in treating inhalational anthrax, or with an aminoglycoside (streptomycin is suggested) in gastrointestinal anthrax. The synergistic action of streptomycin when combined with penicillin has already been noted in relation to treatment of animals (section 7.2.2.1). In the USA, the most up-to-date recommendation for anthrax meningoencephalitis (Sejvar et al., 2005) is a fluoroquinolone (levofloxacin has also been shown to be effective in an animal model – Deziel et al., 2005) with two additional agents that penetrate well to the CSF (see section 7.3.1.9).
Use of more than one antibiotic is discussed further in section 7.3.1.6. Suggested doses are given in . As a general principle, in life-threatening infection with B. anthracis, low doses of antibiotics should not be used.
General measures for treatment of shock may be life-saving since death is caused, at least in part, by toxin-induced shock. Intubation, tracheotomy or ventilatory support may be required in the event of respiratory problems, for example in cases where compression on the neck from oedema is leading to danger of tracheal obstruction; once oedema is extensive it can be difficult to find the trachea at operation. Vasomotor support may be called for when there is haemodynamic instability. If administration of the appropriate volume of intravenous fluid fails to raise the systolic blood pressure above 90 mm Hg, a vasoactive drug such as dopamine or dobutamine can be given. Primary haematological, renal or liver function disorders are not generally seen.
7.3.1.3. Therapy in children
Penicillin is again the antibiotic of choice for treatment of anthrax in children. Generally the approach taken at least in Africa has been to give half-adult doses to children of less than 10 years of age (Martin, 1975; Turnbull, personal observations).
In mild uncomplicated paediatric cases of cutaneous anthrax, penicillin V should be given by the oral route at a dosage of 25–50 mg/kg/day in 3–4 doses, or amoxicillin according to the following formulae:
weight > 20 kg, 500 mg orally every 8 hours for 3–7 days;
weight < 20 kg, 40 mg/kg orally every 8 hours or intramuscular procaine penicillin (25 000 to 50 000 units/kg/day divided into 1–2 doses) for 3–7 days.
In severe or life-threatening cases, penicillin G should be administered intravenously in a dose of 300 000 to 400 000 units/kg/day for systemic anthrax and for cutaneous anthrax if (i) there are signs of systemic toxicity, (ii) the lesion(s) is/are located on the head and neck, and (iii) extensive oedema is present. As with adults (section 7.3.1.2), consideration should be given to combining the penicillin with rifampicin or vancomycin (Sejvar et al., 2005). Penicillin combined with streptomycin may be given in gastrointestinal anthrax, and penicillin combined with clindamycin or clarithromycin may be given in inhalation anthrax. Suggested doses are given in . As with severe cases in adults, once the patient’s temperature returns to normal, consideration may be given to switching to the regimen described for mild cases.
According to Sejvar et al. (2005), in view of the recognized potential of ciprofloxacin to cause arthropathy in children, the United States Working Group on Civilian Biodefense has recommended the use of ciprofloxacin in anthrax-related emergencies in children owing to the severity of the disease.
See section 7.3.1.7 for a discussion of alternatives to penicillin and section 7.3.2.1 for a discussion of postexposure prophylaxis.
7.3.1.4. Adequate doses of penicillin
The possibility that subinhibitory concentrations of penicillins may induce β-lactamase and the importance of using adequate doses of penicillin when this is chosen for treatment is covered in section 7.1.2. This needs to be appreciated in developing country situations, where there may be a temptation to economize.
7.3.1.5. Duration of treatment
The appropriate duration of treatment is a subject for debate. B. anthracis cannot be isolated from cutaneous lesions 24–48 hours after commencement of antibiotic therapy (Ellingson et al., 1946) and Kobuch et al. (1990) could see no advantage to continuing treatment of cutaneous anthrax beyond 4 days. A report from Ethiopia (Martin, 1975) records that 100 patients with cutaneous anthrax were treated with a single intramuscular dose of procaine penicillin, 600 000 units, and 99 of these were sent home with the invitation to return if complications occurred. Only 5 returned on account of further developments. Continuation of treatment for 7–14 days or longer has become standard, but frequently it is not appreciated that the lesions, or other toxin-related systemic damage, will continue to progress through their cycles of development and resolution regardless of the elimination of the infecting B. anthracis. Excessive antibiotic treatment may be wasteful and counterproductive, possibly giving rise to adverse side-effects. Supportive therapy becomes more important after the first few days. For guidance, it is suggested that the duration of antibiotic therapy in uncomplicated cutaneous anthrax be 3–7 days but, in the absence of clinical experience with short-course antibiotic therapy in systemic anthrax, therapy in cases of systemic anthrax should be continued for 10–14 days.
The same reasoning applies to treatment of alimentary tract cases; after the first 24 to 72 hours of antibiotic treatment, which can be expected to have eliminated viable anthrax bacilli, patient survival becomes dependent on supportive therapy while systemic damage caused by the toxin is overcome.
The topic of postexposure prophylaxis following substantial aerosol exposure in deliberate release events is covered in section 7.3.2.1.
7.3.1.6. Use of more than one antibiotic
The concept of using two antibiotics is not a new one. Forty years ago, Lincoln et al. (1964), following animal studies showing that penicillin together with streptomycin was more effective than either antibiotic alone, recommended that streptomycin should be administered together with penicillin in the treatment of septicaemic anthrax.
Recently, in the aftermath of the anthrax letter events in the USA in 2001, it was felt that the use of more than one antibiotic may have been life-saving in the 6 of 11 individuals who survived inhalational anthrax infection (Inglesby et al., 2002; Jernigan, 2001) and, while acknowledging that controlled studies to support a multidrug approach are not available, multidrug regimens were recommended for cases with signs of systemic involvement. As stated in section 7.1.2, it is certainly reasonable to add a second drug in cases showing signs of systemic involvement, utilizing two antibiotics to which B. anthracis is typically sensitive initially, with at least one of these able to penetrate well to the CNS. It may be appropriate to revert to one drug when progression of symptoms ceases, temperature returns to normal and the sensitivity profiles of any isolate have been established. Suggested antibiotic combinations for severe or life-threatening anthrax infections are given in section 7.3.1.2.
7.3.1.7. Alternatives to penicillin
In the event of allergy to penicillin, effective alternatives that are likely to be available in developing countries are tetracyclines and erythromycin (see also section 7.1.1). If ciprofloxacin or doxycycline are available, these are now accepted as best alternatives, although it needs to be remembered that doxycycline penetrates poorly to the CNS (see section 7.3.1.6). For uncomplicated cases, suggested doses are given in .
In severe, potentially life-threatening cases, ciprofloxacin or doxycycline (also see section 7.3.2.1) should be given intravenously, preferably together with another antibiotic having good CNS penetrability (see section 7.3.1.2) until the patient’s clinical condition permits switching to oral therapy. For adults, the recommended dose of intravenous ciprofloxacin is 400 mg every 12 hours and the recommended intravenous dose for doxycycline is 100 mg every 12 hours. The intravenous form of doxycycline is not available in some developing countries.
Ciprofloxacin and doxycycline are normally considered not ideal for children owing to their potential side-effects. Doxycycline is generally not recommended for use in children < 8 years old, owing to staining of teeth and the inhibition of bone growth associated with tetracyclines, and ciprofloxacin has been shown to cause cartilage toxicity in immature animals. However, as discussed in section 7.3.1.3, ciprofloxacin is recommended in anthrax-related emergencies in children owing to the severity of the disease.
The topic of prolonged postexposure prophylaxis for children following substantial aerosol exposure in a deliberate release event is covered in section 7.3.2.1.
7.3.1.8. Pregnancy
There are few reports of anthrax during pregnancy both in humans and animals. Kadanali et al. (2003) found only four human cases reported in the literature; three of these occurred before the availability of antibiotics and all three patients died. The fourth patient, in the Islamic Republic of Iran, went into labour within 48 hours after the appearance of the lesion; she died shortly after admission to hospital from massive oedema of head and neck, which interfered with her breathing. The neonate had no evidence of congenital infection and lived.
All the reports and Kadanali et al.’s own two (cutaneous anthrax) cases were in the last trimester of pregnancy. Kadanali et al. considered it advisable to use penicillin instead of more fetotoxic drugs, such as ciprofloxacin. Treatment was successful in both mothers, though both delivered preterm babies; neither of these showed any evidence of congenital infection. The authors concluded that anthrax during pregnancy could be successfully managed with penicillin, but preterm delivery could be a complication.
Sejvar et al. (2005) support the use of penicillin G (in combination with rifampicin or vancomycin) to treat naturally-occurring anthrax meningitis in pregnant women in endemic countries. As with children (section 7.3.1.3), while recognizing that ciprofloxacin has the potential to cause arthropathy in children, owing to the severity of anthrax meningitis, they state that ciprofloxacin should still be considered for pregnant women under such circumstances.
Published recommendations from the United States Centers for Disease Control and Prevention (CDC) following the anthrax letter events were generally careful to specify that the recommendations were for postexposure prophylaxis for prevention of inhalational anthrax after exposure to intentionally released anthrax spores, and specified the same regimen of ciprofloxacin or doxycycline for pregnant women under such circumstances as for non-pregnant ones (Anon., 2001c). For cutaneous anthrax, Carucci et al. (2002) similarly recommend the same doses of ciprofloxacin or doxycycline for pregnant and non-pregnant women, adding that ciprofloxacin is favoured over doxycycline. Carucci et al. do not state that this is related to bioterrorism, but their recommendation that duration of treatment should be 60 days indicates that it was.
Further discussion of prolonged postexposure prophylaxis for pregnant women and nursing mothers following substantial aerosol exposure in a deliberate release event is given in section 7.3.2.1.
7.3.1.9. Treatment of anthrax meningoencephalitis
Anthrax meningoencephalitis is a life-threatening infection, and mortality is very high. Currently, the first-choice treatment is penicillin G or a fluoroquinolone combined with rifampicin. These have the dual benefits of effective activity against B. anthracis and rapid penetration into the CSF. The intravenous form of rifampicin is not yet available in some developing countries. Another choice would be penicillin G or a fluoroquinolone in combination with vancomycin, but vancomycin is costly. For patients allergic to penicillin, a combination of a fluoroquinolone with two additional agents that penetrate well to the CSF, such as a β-lactam (penicillin, ampicillin or meropenem), rifampicin or vancomycin is recommended (Sejvar et al., 2005). A regime of vancomycin, 1 g intravenously every 12 hours, combined with rifampicin (600–1200 mg/day) for 10–14 days is another alternative.
Lessons from the few recorded instances of survival in cases of anthrax meningoencephalitis suggest that the doses of penicillin G should be 20–24 million units/day divided for intravenous administration every 2–4 hours, and a daily dose of rifampicin, 600–1200 mg/day (intravenous administration is suggested but it may also be given via an enteral tube).
The suggestion given in the third edition of these guidelines (1998) that penicillin may be replaced by chloramphenicol for patients who are hypersensitive to penicillin has now been withdrawn. Chloramphenicol has been stated to be a satisfactory alternative but, in early experimental studies, Lincoln et al. (1964) found that chloramphenicol was not effective in mice or monkeys. It is bacteriostatic rather than bactericidal against B. anthracis, and some strains may be resistant to it (Athamna et al., 2004). Furthermore, chloramphenicol has the potential for serious side-effects on bone marrow. In vivo data regarding its effectiveness in the treatment of severe anthrax infections is lacking, and there is a wide availability of more effective alternatives (Sejvar et al., 2005).
Doxycycline should not be used if meningitis is suspected because of its lack of adequate central nervous system penetration (Bell et al., 2002; Sejvar et al., 2005).
It should be stressed that the therapy failure rate is very high in cases of anthrax meningoencephalitis; in 47 cases covered by four reports (Levy et al., 1981; George et al., 1994; Lalitha et al., 1996; Kanungo et al., 2002), there were only two survivors. Delayed suspicion of the true cause of illness, with initial diagnoses ranging from cerebral malaria to subarachnoid haemorrhage, was considered by Kanungo et al. to be responsible in part for the high treatment failure rate.
Supportive therapy is very important in anthrax meningoencephalitis; respiratory support and anti-oedema therapy for the brain may be required. Essential supportive therapy includes the early institution of assisted respiration, fluid and electrolyte supplement and anticerebral oedema measures, such as 100 ml of 20% mannitol intravenously every 8 hours and hydrocortisone, 100 mg every 6 hours. Dexamethasone is generally suggested.
Useful references concerning anthrax meningoencephalitis are given in section 4.4.5.
7.3.1.10. Immuno- or otherwise compromised individuals
Although written with the deliberate release event in mind, recommended therapy for the immunocompromised individual was the same as that for immunocompetent persons in Anon. (2001c) and Carucci et al. (2002). Special consideration may be needed for patients with renal or hepatic insufficiency.
7.3.1.11. Corticosteroids
The swelling seen in an anthrax infection is caused by the action of oedema toxin, and inflammation is fairly minimal. As discussed in section 5.5.3, the EF component of the toxin is anti-inflammatory in nature. Theoretically, therefore, steroids should be of little value. In practice, some reports indicate that these have been administered with evidence of benefit, but others (Kobuch et al., 1990) have concluded that they were ineffective, discontinuing their use. In experimental studies on therapy in rhesus monkeys, Lincoln et al. (1964) concluded that hydrocortisone did not have a statistically significant effect on survival. In section 7.3.1.9, hydrocortisone is suggested as part of supportive treatment in the event of the very dangerous meningoencephalitic complication. In such situations, it is reasonable to give corticosteroids the benefit of the doubt on the basis that any help they may offer is welcome.
7.3.1.12. Surgery in gastrointestinal anthrax
It should be pointed out that some value a surgical approach to management of advanced intestinal anthrax (Binkley et al., 2002; Kanafani et al., 2003). Kanafani et al. propose that the management of cases of gastrointestinal anthrax should consist of:
initiation of intensive intravenous antibiotic therapy as soon as diagnosis is made;
in patients not improving with medical therapy, wide resection into seemingly healthy tissue with primary anastomosis;
continuous drainage of the ascites, as fluid will continue to accumulate for several days after surgery;
aggressive replacement of protein and electrolyte losses.
7.3.4. Immunotherapy (hyperimmune serum therapy)
The use of serum therapy for treatment of anthrax predates by several decades the realization that death was caused by toxin action. In 1930, Eurich & Hewlett wrote that “as regards malignant pustule, serum treatment is now the method of choice, taking the place of excision”.
A couple of anecdotes concerning Professor Eurich that appeared in the Sunday Telegraph (United Kingdom) in June 1992 may be of interest. The first states: “Professor F.W. Eurich, a German immigrant, settled in Bradford and devoted much of his life to seeking a cure for anthrax, or woolsorter’s disease, so prevalent in the city at the turn of the century. ... He liked to recall how at Heidelberg he had watched one student challenge another to a duel at dinner with the accusation: you have been staring at my sausage”. In another, a letter, Mrs E.G. Croisdale, Hereford, United Kingdom, wrote: “My late father, J.W. Hillas, was one of Professor Eurich’s patients to be cured of anthrax in the early years of this century. … My father owed his life to Professor Eurich ... thus I owe my life to this great man”.
At the time Eurich was working, antisera were developed in asses, sheep and oxen (horses are not named) and standardized by a protection test in rats. Doses and routes varied from 30–40 ml daily by the subcutaneous route, or occasionally intravenously, to single intravenous doses of 100 ml. Efficacies were widely acclaimed. Gold (1955) refers to treatment (apparently in the 1930s) of 21 cases with “an optimum dose [of anti-anthrax serum] that varied from 250 to > 1000 ml given intravenously. This therapy, although effective, resulted in prolonged morbidity from the severe horse sensitivity reactions that ensued”. According to Lincoln et al. (1964), “the cures effected by antiserum alone in the 1920s and 1930s appear to be more numerous than those following use of modern antibiotics in the 1940s and 1950s”. In the former Soviet Union, recommendations on treatment of any type of anthrax infection put primary emphasis on use of antiserum (Shlyakov, 1957, cited by Klein et al., 1962). Klein et al. (1962) appreciated that the role of antiserum was to neutralize the toxin. Lincoln et al. (1964) were demonstrating the greater efficacy of a combination of antibiotics, antiserum and active immunization over any one of these approaches alone in the treatment of late-stage inhalational anthrax in rhesus monkeys.
Despite the history, the use of anti-anthrax serum for treatment of human anthrax appears to have been abandoned by about 1950 in most western countries. However, purified immunoglobulin fractions from hyperimmune horse serum are still prepared for human treatment in China (Dong, 1990; Dong, personal communication, 2002) and the Russian Federation (Anon., 1996; Cherkasskiy, personal communication, 2002) (see Annex 5, section 4). According to Cherkasskiy (personal communication, 2002), this treatment is administered to approximately 2–3 persons each year in the Russian Federation.
The topic was revisited scientifically by Little et al. (1997), who found in protection tests in guinea-pigs that only antibodies to the PA component of the anthrax toxin (section 5.5.3) provided passive protection against live spore challenge. Antibodies to the lethal and oedema factor components did not. One monoclonal antibody to PA of several monoclonal antibodies tested provided a delay in time to death.
Following the anthrax letter events in the USA in 2001, a call was made for a biological defence initiative based on developing, producing and stockpiling specific antibody reagents, for use in protection against biological warfare threats including anthrax (Casadevall, 2002). Certainly one response to this has been the development of clinical-grade hyperimmunoglobulin from donor persons immunized with the United States anthrax vaccine.1 At the research level, reports have appeared on toxin-neutralizing Fabs from human donors vaccinated against anthrax (Wild et al., 2003), human monoclonal anti-PA antibodies (Maynard et al., 2002; Sawada-Hirai et al., 2004; Cui et al., 2005; Mohamed et al., 2005; Peterson et al., 2006) and murine anti-LF monoclonal antibodies (Lim et al., 2005). The protective effects in all but the tests of Mohamed et al. (2005) and Peterson et al. (2006) were based on antitoxic activities. In live spore challenge tests, Mohamed et al. (2005) and Peterson et al. (2006) obtained good protection in rabbits when they were combined with low doses of ciprofloxacin, and also in mice and guinea-pigs. Mohamed et al. (2005) found their antibody still protected 50% of animals when administered 36 hours after challenge, In Peterson et al.’s case, the antibody had to be administered within 12 hours of challenge to be highly effective. Kozel et al. (2004) targeted their monoclonal antibodies to capsular antigen, obtaining a high degree of protection in mice against virulent B. anthracis infection in a pulmonary model.
The value of these types of antibodies for amelioration of infection once symptoms have commenced has yet to be demonstrated. Nevertheless, they remain important for potential situations in which the infecting agent has been engineered to be multiresistant to antibiotics.
7.3.4.1. Novel immune and other therapies
Antibiotics are only effective against anthrax if administered early enough in the course of the infection (section 7.1.1). After a certain point, enough toxin has been formed to cause death of the host even if the antibiotic treatment has killed all the B. anthracis. Among other aftermaths of the 2001 anthrax letter events in the USA have been some intense efforts to develop therapies that will continue to be effective after antibiotics can offer no hope. Among the approaches being taken are therapies targeting:
Probably among the most advanced in their development are humanized monoclonal antibodies targeting toxin-component interactions. For example, the toxin-neutralizing antibodies mentioned by Maynard et al. (2002), referred to in 7.3.4, operated by competing with the host-cell receptor for PA binding. Similarly, the Fab fragments developed by Wild et al. (2003) functioned by binding in a manner that inhibited the interaction of LF with PA. EluSys Therapeutics2 reports on the Web the development of a heteropolymer system using two monoclonal antibodies chemically joined together; one binds to anthrax toxin and the other to red blood cells which then carry the bound toxin to the liver for destruction. Human Genome Sciences, Inc.3 also has a product well into its development stage based on human or humanized neutralizing anti-PA monoclonal antibodies. A non-immunologically based approach is illustrated in the report of Sarac et al. (2004) that furin inhibitor, hexa-D-arginine, might be an effective therapeutic agent by preventing cleavage of PA to its active form.
An example of therapy targeting toxin-induced events within the host cell is the observation by Shen et al. (2004) that adefovir diphosphate, a drug approved to treat chronic hepatitis B virus infection, inhibits the adenylyl cyclase activity of EF and thereby EF-induced cAMP accumulation (see section 5.5.3).
Illustrating other virulence factor entities are the reports of Gold et al. (2004) suggesting that interferon may have a benefit as an immunoadjuvant therapy and of Kozel et al. (2004) on their monoclonal antibodies to capsular antigen (sections 7.3.4). An intriguing suggestion for novel therapy was the proposed use of gamma phage lysin by Schuch et al. (2002) that lysed B. anthracis cells from without in in vitro tests, and protected mice against infection with a phage-sensitive B. cereus used as a simulant for B. anthracis. Success in this case for real anthrax treatment would depend on the lysin’s ability to penetrate the B. anthracis capsule in genuine anthrax infections. Direct phage therapy, an idea that has been considered from time to time in the past for anthrax, has to overcome the fact that the phage does not attack capsulated B. anthracis cells (section 6.3.1.5).4