Anthrax in injecting drug users: the need for increased vigilance in the clinic

Abstract

The emergence of a previously unrecognized route of Bacillus anthracis infection over the last few years has led to concern: sporadic anthrax outbreaks among heroin users in northern Europe have demonstrated the severe pathology associated with the newly described ‘injectional anthrax’. With a high case fatality rate and non-specific early symptoms, this is a novel clinical manifestation of an old disease. Lack of awareness of this syndrome among emergency room clinicians can lead to a delayed diagnosis among heroin users; indeed, for many health workers in developed countries, where infection by B. anthracis is rare, this may be the first time they have encountered anthrax infections. As the putative route of contamination of the heroin supply is potentially ongoing, it is important that clinicians and public health workers remain vigilant for early signs of injectional anthrax.

Keywords:

• edema factor
• heroin
• injectional anthrax
• intravenous drug use
• lethal factor
• protective antigen

Bacillus anthracis, classical disease etiology & ‘skin-popping’

Since the turn of this century, there has been an upsurge in the numbers of soft tissue infections in persons who inject drugs and these have been primarily associated with spore-forming bacteria of the Clostridium and Bacillus species [1]. It was Ringertz et al. who first described a new route of infection for the pathogen Bacillus anthracis, the causative organism of the disease anthrax, in the year 2000 [2]. In contrast to the classical etiology of the disease (inhalational, cutaneous or gastrointestinal), this newly described injectional anthrax (IA) route, involved ‘skin-popping’ or injection of heroin contaminated with anthrax spores directly into the subcutaneous tissue. The Norwegian intravenous drug user, presenting as the first case of IA, progressed from edema of the soft tissues surrounding the injection site to septic shock, and eventually death from hemorrhagic encephalitis [2].

The gram-positive bacterium, B. anthracis, has gained prominence as a bioterrorism agent due to a high case fatality rate and ability to form resilient spores [3,4]. The bacteria are obligate pathogens, and upon entry to host tissues, the spores germinate rapidly. Following exposure to oxygen, the bacteria, which are facultative anaerobes, produce spores, which persist in the environment. This leads to anthrax infection being endemic in many countries, particularly in the developing world where it presents a significant economic burden to rural populations as a zoonotic disease, primarily affecting wild and domestic mammals, with sporadic outbreaks in humans, exposed either through ingestion of contaminated meat or by occupational exposure to animal hides. The three classical forms of the disease are each capable of progressing to lethal systemic infections, but all three are associated with widely varying fatality rates. Cutaneous anthrax represents the most prevalent form of the disease, with skin lesions at the site of infection progressing to an ulcerated vesicle complete with localized edema and a characteristic black eschar, which forms as the surrounding skin becomes necrotic [5]. Although infection is fatal in approximately 20% of untreated cases, appropriate antibiotic therapy reduces the risk of death to <2% [6]. Gastrointestinal anthrax infections are relatively rare, and generally follow the ingestion of spores in contaminated meat leading to ulceration and acute inflammation of the gastrointestinal tract, nausea, vomiting, hematemesis and bloody diarrhea. Without treatment, intestinal anthrax results in death in 25–60% of cases [7]. Inhalation anthrax is characterized by a rapid onset of disease, with the initial symptoms of mild fever, myalgia and non-productive cough progressing to severe respiratory distress, necrotizing pneumonic lesions, septic shock, coagulopathy, widened mediastinum and associated hemorrhagic mediastinitis [4,5,8]. Inhalational disease may also be associated with hemorrhagic meningitis, characterized by the presence of blood in the cerebrospinal fluid, and as the most severe form of the disease, it can result in fatality rates approaching 100% if not treated appropriately [3]. The relatively low mortality rate of 45% in the US during the 2001 attacks, when anthrax spores were distributed through the mail, was attributed to rapid diagnosis and aggressive antibiotic therapy [4].

IA outbreaks in 2009–2010 & 2012–2013

Following the Norwegian case of 2000, IA infection was not seen again in the clinic, until two outbreaks occurred in northern Europe during 2009–2010 and 2012–2013. The first outbreak, which centered on persons who inject drugs in Scotland, with cases in England and Germany, was originally diagnosed in two heroin users from the Greater Glasgow and Clyde area, presenting with blood cultures positive for B. anthracis [9]. Eventually, 54 laboratory-confirmed cases, with a 33% fatality rate, were reported for the initial outbreak [10]. The second, smaller, wave of cases was more widely distributed geographically, with 15 laboratory-confirmed cases diagnosed in England, Scotland, Wales, Germany, Denmark and France; a case fatality rate of 47% was reported for this cluster [10].

The etiology of IA differs from the classical routes of infection in several ways. Notably, while marked edema was observed in a number of patients, the characteristic cutaneous anthrax eschars were reported to be absent in almost all cases published so far [2,10,11]. Necrosis and blistering of the tissues in some cases led to the misdiagnosis of compartment syndrome or necrotizing fasciitis; however, unlike these conditions, erythema, and fever and pain are not pronounced features of IA, and white cell count, C-reactive protein and lactate are not significantly elevated [10,11]. Equally, clinicians should be aware of the possibility for cellulitis and necrotizing fasciitis to develop in IA cases [9,12]. Respiratory symptoms associated with IA were rare and possibly related to smoking or snorting the heroin [13]. Despite fears that this route of administration would lead to a surge in inhalational anthrax cases, an intriguing case–control study found that during the 2009–2010 outbreak, smoking heroin was associated with a lower risk of infection [14]. This finding should be treated with caution, however, as a number of the patients admitted with systemic anthrax in the absence of localized tissue edema reported a history of smoking the drug, rather than injection, suggesting a correlation of this route of administration with a more severe form of the disease.

Due to the complex constellation of symptoms, treatment of cases during the two outbreaks was inconsistent. In general, management consisted of intravenous administration of conventional antimicrobial agents and supportive therapy [10,15]. It is important to note that B. anthracis isolates have been identified which display resistance to frontline antibiotics used in the treatment of anthrax, particularly β-lactams [16]; therefore, combination therapy is recommended in preference to monotherapy. Surgical debridement of tissue was indicated where appropriate, with the aim of removing bacterial toxin reserves and necrotic tissue [17]. Anthrax immune globulin intravenous, donated by the US CDC, was also made available for management of patients with confirmed IA, as an adjunctive therapy directed against the bacterial toxins [13,15]. The pathology of IA is particularly severe and early diagnosis and treatment is critical to improve the prognosis of patients. Enhanced awareness of IA among Scottish clinicians possibly accounted for the lower fatality rate during the 2009–2010 outbreak. Interestingly, we were able to show that despite the well-known capacity of anthrax toxins to induce immune dysfunction, exposure via this route can be associated with the induction of strong immunological memory [18].

Tracing the source of contamination

Early theories that the heroin had become contaminated at its likely source in Afghanistan, or through infected cutting or dissolving agents immediately before use, were challenged by phylogenetic analysis indicating that all patients (including the original Norwegian case) were infected with a single B. anthracis strain type that is closely related to the strains found in Turkey [19,20]. As anthrax is endemic in Turkey, it is possible that animal skins used to smuggle heroin into Europe or an animal-derived cutting agent led to accidental contamination. However, as there has not been extensive genotyping of B. anthracis strains originating along the heroin trafficking route, it is currently not possible to draw definitive conclusions about the source of the contamination.

B. anthracis belongs to Bacillus cereus group of bacteria, which contains seven species, the most closely related of which are B. anthracis, B. cereus and Bacillus thuringiensis [21]. The difficulty in discriminating between these species was tragically illustrated by a fatal case of IA in Germany, where appropriate treatment was delayed due to misidentification of the suspect organism as B. cereus [22]. B. anthracis is highly monomorphic, showing low levels of genetic diversity. This high level of clonality is presumably related to the episodic nature of the B. anthracis reproductive cycle, with limited opportunities for accumulating DNA mutations. B. anthracis appears to have arisen from the B. cereus group relatively recently in the evolutionary past, with the acquisition of two large plasmids, pXO1 and pXO2.

Anthrax toxins & disease course

The virulence factors of B. anthracis are encoded by genes found on the two plasmids. These include the three toxin subunits: Protective Antigen (PA), Lethal Factor (LF) and Edema Factor. PA combines with LF or Edema Factor to form the binary exotoxins, Lethal Toxin or Edema Toxin. PA and LF have been identified as key immunogens, inducing both humoral and cellular immunity following natural infection [18,23,24].

During the IA outbreaks, identification of B. anthracis was carried out by culture of colonies showing a characteristic non-hemolytic ground-glass appearance, and either immunohistochemical staining showing gram-positive chains of rod-shaped bacteria, 1–10 μm in length, or real-time PCR analysis of the bacterial chromosome and plasmids [9]. A number of further diagnostic tests may be employed to confirm IA cases; these fall broadly into assays which detect the presence of the bacteria, such as fluorescence immunoassays and mass spectrometry identifying cell-associated proteins [25–27] and assays which detect the presence of antibodies to PA and LF [28,29]. Several caveats must be borne in mind though when considering these methods, as phenotypic characteristics may vary between strains and hemolytic, motile bacilli, which are resistant to phage lysis, have been identified in B. anthracis isolates, closely related Bacillus species may lead to misdiagnosis, and (perhaps most importantly) antimicrobial therapy rapidly decreases the sensitivity of these assays [30]. Although it has been found that LF-specific IgG can be detected early in infection [29], assays based upon the development of an immune response to anthrax antigens may have limited utility in the clinic [30], and instead may be useful in retrospective epidemiological investigations [19,24].

Indeed, due to the severity of the disease and the impact of prompt therapy, IA should be considered in the initial differential diagnosis when persons who inject drugs present to the clinic with signs of soft tissue inflammation or severe systemic septic disease. The importance of increased vigilance among health care professionals is underlined by the biphasic nature of the outbreaks, which suggests the possibility that multiple batches of contaminated heroin are still circulating in the user population.

Financial & competing interests disclosure

The authors disclose funding from NIH-NIAID (contract number HHSN266200400084C). The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.