The Conundrum of the Gram-Positive Rod: Are We Missing Important Pathogens in Complicated Skin and Soft-Tissue Infections? A Case Report and Review of the Literature

Introduction

Infections of the skin and soft tissue are generally classified as “complicated” when the inflammatory process incorporates deep tissue layers or when management requires surgical intervention [1]. The “necrotizing” subset of complicated infections is a highly morbid group of aggressive and potentially devastating inflammatory and tissue destroying infections that may involve and spread along any of the layers of subcutaneous tissue, deep fascia, and skeletal muscle [2–5]. This group of diseases, including in part necrotizing fasciitis, Fournier gangrene, and myonecrosis, are known collectively as necrotizing soft tissue infections (NSTIs).

Although NSTIs vary according to site, depth, and extent of infection, it is unclear whether they should be grouped based solely on common physiology and management strategies or whether they represent a continuum of infection through progressively deeper layers [1–5]. What is known, however, is that, given the varying depths of origin, NSTIs may be difficult to diagnosis in their early stages and that this delay can affect outcome [2–4,6–8]. In necrotizing fasciitis, for example, the diagnosis is generally made once the disease has made its rapid progression from fasciitis to secondary edema, emphysema, inflammation, and ultimately necrosis of the overlying subcutaneous tissue and skin [6–10]. Sepsis and concomitant multiple organ dysfunction syndrome followed this destructive cutaneous process [11,12]. This is a common scenario seen with all of the NSTIs.

The disease course begins after inoculation of the deep tissues with bacteria, the source of which categorizes the resultant infection as primary or idiopathic. Primary NSTIs occur as a complication of trauma (cuts, scrapes, punctures, surgical wounds), in which case the responsible pathogens are often skin flora, most commonly Streptococcus or Staphylococcus species [4,5,8,13]. Primary infection can also occur as the result of insect or animal bites, thereby including organisms native to a host species as potential pathogens [14–16]. Hematogenous seeding from distant or occult sites, known as idiopathic infection, can come from any source [2,6,17–19]. Anaerobic and gram-positive aerobic bacteria are common culprits.

Any immunocompromise, as seen with diabetes mellitus, advanced age, malignant neoplasm, chronic disease, and human immunodeficiency virus, and with patients taking steroids or immunosuppressive medications, can contribute to the development of and mortality from NSTIs [1–5,20–22]. Loss of host defenses likewise expands the list of potential causative pathogens to include an ever-increasing list of microorganisms, alone or in combination [12,17,18]. A typical polymicrobial scenario is synergistic infection perpetrated by opportunistic anaerobes in the presence of pathogenic gram-negative aerobes [6,23,24]. This piggy-backing is generally believed to occur because anaerobes proliferate well in the hypoxic environment of the wound, where polymorphonuclear leukocyte function is diminished [25].

Still, NSTIs are traditionally caused by gram-positive cocci, one of the most common and virulent being group A β-hemolytic Streptococcus, found in primary and idiopathic cases [1–5,6,19,24], but the virulence of gram-positive bacteria does not end with the cocci.

The family of gram-positive rods, including Corynebacterium, Turicella, Dermabacter, Microbacterium, Actinomyces, and Arcanobacterium, covers a wide range of pathogenicities, but all are morphologically similar. With the advent and use of 16S ribosomal ribonucleic acid (rRNA) gene sequencing, these bacteria are more frequently identified to species, and as a result, the literature has increasingly identified gram-positive bacilli as present in monomicrobial and polymicrobial cultures taken from complicated skin and soft tissue infections (Table 1). Arcano hemolyticum, for example, has been cultured from a myriad of sources, from wound infections to cellulitis to necrotizing fasciitis [26]. One recent study found that A. bernardiae was among the polymicrobial culture isolates recovered from wounds in moderate-to-severe diabetic foot infections [27]. These are bacteria that, without genetic testing, would have been missed or misaligned as another, similar species. Although the pathogenic role of these organisms in polymicrobial cultures is not self-evident, their increasingly common presence in complicated infections bears clinical significance. Tables 1 and 2 list recent publications of serious infections caused by these and other coryneform organisms.

Discussion

In 1987, the U.S. Centers for Disease Control and Prevention (CDC) had an undesignated collection of 11 strains of coryneform bacteria, designated “group 2” [28]. Funke et al. studied this assortment of taxonomic misfits in detail in 1995 [29]. Isolates obtained from the Laboratory of Bacteriology at the CDC were examined for their morphologic, biochemical, molecular, and genetic characteristics; 16s rRNA gene sequencing was the final stage of these investigations. This examination included 1,350 nucleotides that demonstrated that coryneform group 2 was 97.5% genetically similar to Actinomyces pyogenes, with mismatch divergence sufficient to indicate a discrete species. The result was the addition of coryneform group 2 to the genus Actinomyces, as the new species Actinomyces bernardiae, named for Canadian microbiologist Kathryn A. Bernard.

The genus Actinomyces is composed of gram-positive bacilli that Bollinger first identified in 1877 as an isolate from “lumpy jaw” lesions in cows [30]. Harz coined the name, Actinomyces bovis, to identify the pathogen responsible for this disease [30–32]. As the type species for this genus, Actinomyces bovis was aggressively studied for years as the suspected cause of human actinomycosis. James Israel identified and fully described similar coryneform organisms isolated from lesions in two human patients in 1878, earning him the eponymous, Actinomyces israelii [33]. By 1891, Israel and Max Wolfe had studied this organism in detail and were able to describe its phenotypic characteristics and anaerobic growth [34]. Despite this new insight into the species, many considered A. bovis and A. israelii to be the same organism. It was not until 1964, that Georg et al. were able to show that A. israelii was not only the responsible pathogen in human actinomycosis, but was a microbiologically distinct organism from A. bovis [35,36].

Over the next 2 decades, the family of gram-positive, coryneform, or “diphtheroid” organisms continued to expand and, by this time, included several Corynebacterium species and Actinomyces as its primary members [37,38]. Collins et al. added to this familial extension in 1982 with the creation of the genus Arcanobacterium. In his studies of Corynebacterium haemolyticum, Collins found that this rare cause of tonsillar and pharyngeal infections was phenotypically distinct from C. diphtheriae—so much so that C. haemolyticum was reclassified as A. haemolyticum, the type species of this new genus [39].

With the continued expansion of this nebulous phylogeny of genera, the various species have displayed marked phenotypic heterogeneity, and it has proven difficult to reliably distinguish them from one another using standard phenotypic methods [40,41]. Generally, tests for morphology, carbohydrate fermentation, hemolysis, esculin and gelatin hydrolysis, nitrate reduction, and end-product analysis can usually identify the genus of these microorganisms [40–44] Species-level identification is another story. By the late 1990s, the genus Actinomyces had expanded to include 18 validated species, primarily inhabitants of the mucosal, usually oral, surfaces of humans, cows, dogs, and seals [40,41,43]. Santala et al. described a comparison of four commercially available test kits for the identification of Actinomyces and related species [41]. They found that none of the systems was able to reliably identify classical or newly described strains, their sensitivity peaking at 65%. Molecular-based techniques, such as 16S rRNA gene sequencing, PCR, and spectrometry, have, therefore, become a criterion standard to differentiate these bacteria. Unfortunately, these techniques are expensive, labor intensive, and often inaccessible to smaller facilities.

Multiple laboratories have attempted to address this problem. Sarkonen et al. tried to create a biochemical algorithm for strain differentiation [43]. The morphologic and biochemical differences between Actinomyces, Arcanobacterium, and Actinobaculum species were used to create flowcharts to be used as guides for species identification. Unfortunately, the biochemical overlap and metabolic diversity of these organisms were a persistent problem. Genetic confirmation of strains by reference or research laboratories was still recommended.

The taxonomic composition of the family has remained under substantial scrutiny. In 1997, Ramos et al. described the phylogenic analysis of 13 known Actinomyces and one Arcanobacterium species, as well as an unnamed Actinomyces strain [40]. Using 16S rRNA gene sequencing, 800 nucleotide pairs, including segments of variable regions V1-V4, were sequenced and aligned to create a phylogenetic tree. The ultimate goal was to determine whether the genotypic relatedness organized therein supported the continued grouping of these organisms into one genus. Bootstrap re-sampling analysis was used to test the reliability of the data set, and confirmed that the results were “highly significant.” The result was creation of the new species A. phocae. Phylogenic divergence observed here also led to the reassignment of Actinomyces bernardiae and Actinomyces pyogenes to the genus Arcanobacterium.

Actinomyces bernardiae is a facultatively anaerobic, gram-positive, coryneform bacillus, classically recovered from urine, blood, bone, or wounds [40,43,45]. Arcanobacterium is probably a member of the normal flora in the oral cavity and gastrointestinal tract and, like many other species of Actinomyces and Arcanobacterium, acts as an opportunistic pathogen [40–43]. To our knowledge, no animal studies have been done to demonstrate its virulence, although the strains in the CDC collection came from infections of sufficient magnitude to be submitted to the CDC by the local laboratories. The isolate described in our case was from skin, soft tissue, and deep fascia. Some unusual sites have been presented as case reports in the literature, including diabetic foot infection, as described earlier [27]. The first case of A. bernardiae was reported in septic arthritis in 1998 [46]. A more-recent report described an episode of gluteal necrotizing myofascitis caused by infected mesh after an abdominal sacropexy. There, cultures of the mesh grew Gardnerella vaginalis and Actinomyces (species not reported) [47]. Gardnerella vaginalis and A. bernardiae have overlapping profile numbers in the API Coryne identification kit (API-bioMérieux, Inc., La Balme les Grottes, France) and thus could easily be confused. To our knowledge, this case presentation is the first description of A. bernardiae as a pathogen in necrotizing fasciitis.

In the clinical setting, infections with A. bernardiae or M. morganii are seen in hospitalized, immunocompromised patients. This subgroup of patients is likewise more prone to idiopathic NSTIs than the immunocompetent. As such, the episode of community-acquired, idiopathic necrotizing fasciitis presented here appears unique. Unfortunately, because coryneforms or “diphtheroids” are generally considered to be nonpathogenic skin contaminants when they appear on culture results for skin and soft tissue infections, they are often neither identified to species nor tested for antibiotic sensitivities. Furthermore, diphtheroids, coryneform species, coccobacilli, and small bacilli are all microscopically similar and difficult to differentiate without genetic testing. We suggest, therefore, that gram-positive rods may be more common in NSTIs than the literature would suggest but are simply missed and go unreported. These organisms have enzymatic activities that make them well suited to cause deep-tissue infections, but further study is needed to determine the frequency of pathogenicity and extent of their virulence in complicated infections of the skin and adnexa.

Although it is not realistic or cost-effective to genetically sequence all colonies showing pleomorphic, gram-positive rods if they are suspected to be contaminants, in cases involving complicated infections or pathogens recovered from surgical specimens, it may be beneficial to pursue identification to the genus and species level [42]. To overlook a potentially causative pathogen could have had serious consequences for our patient. As these genera expand their breadth to include an ever-increasing number of organisms, gene sequencing may be required more frequently for species identification. These highly sensitive but expensive and time-consuming techniques may benefit from process simplification and commercial availability, making them more inexpensive and accessible to small labs.