Secondary Logo

Journal Logo

Bacterial Infections

Infections due to anaerobic bacteria and the role of antimicrobial susceptibility testing of anaerobes

Jenkins, Stephen G.

Author Information
Reviews in Medical Microbiology: January 2001 - Volume 12 - Issue 1 - p 1-12
  • Free



In terms of sheer numbers, anaerobic bacteria are the predominant indigenous microflora of humans and, as a result, play an important role in infections, some of which are serious and carry a high mortality rate. These opportunists are recovered only infrequently from cultures of infected materials because of common shortcomings in collection and transport procedures. In this era of cost-containment, many clinical laboratories do not routinely identify anaerobic isolates to the species level nor do they perform antimicrobial susceptibility studies on such organisms. Furthermore, anaerobes are relatively slow growing in culture and delays in the laboratory reporting of isolates from the infections that they cause are frequent. Because surgical intervention, either through debridement or drainage, is usually required for successful medical management of anaerobic infections, cultures are frequently not collected and submitted to the Microbiology Laboratory.

Antimicrobial resistance among anaerobes, particularly among Bacteroides spp., is an important factor in the selection of agents for both treatment of clinical infections and for surgical prophylaxis in anatomic sites where anaerobes represent the predominant indigenous flora. An appreciation of the various types of infections in which anaerobic bacteria are involved and the mechanisms of antimicrobial resistance among this important group of pathogens will lead to a more directed approach to the collaborative management of such infections with surgical colleagues.


An anaerobe is a bacterium that will not grow on the surface of routine laboratory media incubated in 5–10 % CO2 in room air, whereas facultative organisms grow comparably in either the presence or absence of oxygen. By comparison, an organism is considered to be microaerophilic when its growth is enhanced by 5–10 % CO2 or when it grows better under reduced oxygen tension than in room air. Some anaerobes have been described as strict (failing to grow on an agar surface in the presence of > 0.5 % O2) and others as moderate (tolerant of 2–8 % O2) [1]. Clostridium haemolyticum and Treponema spp. are examples of strict anaerobes whereas Prevotella melaninogenica, Bacteroides fragilis and Fusobacterium nucleatum, all significant clinical pathogens, are classified as moderate anaerobes. However, strain-to-strain differences in aerotolerance within a species are common. Most anaerobes do not produce catalase, but clinically important anaerobes often produce superoxide dismutase. There is generally a correlation between the amount of superoxide dismutase produced and the degree of aerotolerance of the isolate [2,3]. Certain organisms, such as Cl. tertium, that are generally considered to be anaerobes, grow quite well under aerobic conditions [4].


Despite the fact that bacteriological and clinical reports of anaerobic infections date back only to the late 1890s, ancient scholars accurately described syndromes that are now recognized as being related to anaerobic bacteria. In the fourth century BC, Hippocrates rendered an acceptable clinical description of tetanus and in the fourth century AD, Xenophon accurately described acute necrotizing ulcerative gingivitis in Greek soldiers. The clinical manifestations of actinomycosis were first described accurately by Von Langenbeck in 1845.

Antonie van Leeuwenhoek was the first to present evidence that certain microorganisms might be able to live under anaerobic conditions, noting that some of his ‘animalcules’ were capable of living and moving about in the absence of air. While microscopically examining a wet preparation, Louis Pasteur more accurately described anaerobiosis when he discovered that Vibrion butyrique (Cl. butyricum) became non-motile as it approached the edge of a coverglass where it was exposed to air. Pasteur used the word ‘anaérobies’ when describing this phenomenon, forming the basis for the currently used term ‘anaerobes’ [5].

In 1891 Levy published the first report of a patient with an anaerobic infection, a postpartum parametrial mass extending to produce a ‘gas abscess’ in the upper thigh [6]. Anaerobic Gram-positive bacilli in chains were recovered from the foul-smelling pus along with a small amount of Streptococcus pyogenes. In 1893 Veillon recovered and maintained in culture an anaerobic coccus from four patients with various infections. One was found in pure culture; the other three were isolated along with S. pyogenes. The foul odour of the pus was shown to be solely attributable to the anaerobic coccus [7]. Veillon and Zuber published a classic paper in 1897 characterizing a large number of diverse morphologic types of anaerobes cultured from 25 cases of foetid or gangrenous suppurative infections including pulmonary gangrene, brain abscess, appendicitis, septic arthritis and pelvic abscess [8].

Subsequent studies characterized further the role of anaerobes in human infections and facilitated the manipulation of these frequently fastidious organisms in culture. McIntosh and Fildes [9] introduced anaerobic jars in 1916. More recently, the introduction of commercial packets with cold catalysts that generate hydrogen and CO2 upon the addition of water have enhanced the abilities of hospital clinical microbiology laboratories to recover these organisms in culture. Prereduced, anaerobically sterilized media (PRAS) have further improved the clinical microbiology laboratory's ability to recover these sometimes difficult to grow, generally opportunistic pathogens [10]. Other techniques that have been developed in an effort to promote recovery of these organisms further include the use of holding jar systems designed to maintain some level of reduced oxygen tension pending placement of inoculated media into anaerobiosis, use of anaerobe chambers including glove boxes, and inclusion of reducing substances, such as Oxyrase, into media in an effort to ‘prereduce’ the plates. Caution must be exercised with this latter approach, however, as some of these agents (such as palladium chloride) may be toxic to certain anaerobes [11].

During a time span of 55 years from 1926 to 1981, Prévot published extensively on anaerobic bacteriology, taxonomy and clinical infections. In the 1940s and 1950s, antimicrobial agents with potent activity against anaerobes (e.g. penicillin G, chloramphenicol and tetracycline) first became available. In the 1960s and 1970s additional antibiotics useful for treatment of infections caused by anaerobes helped to define further the role of these bacteria in intra-abdominal, pleuropulmonary and female genital tract infections. Over time, as with aerobic pathogens, anaerobes have developed resistance to a lesser or greater degree to essentially all agents available for treatment of such infections.


Anaerobic bacteria predominate as indigenous flora in and on the human body, including certain areas exposed to air (e.g. the skin, nose, throat and mouth). Well over 99 % of all intestinal organisms are commensal anaerobes. Hentges published an abbreviated but useful overview of this subject [12]. Table 1 outlines specific anaerobes as indigenous flora at various sites of the body. Two explanations have been put forth for the survival and proliferation of anaerobes in locations exposed to air: oxygen consumption by co-habitating aerobic bacteria and anaerobic microenvironmental conditions (e.g. in the tonsillar crypts, gingival crevices and hair follicles) [13].

Table 1
Table 1:
Incidence of various anaerobes as human indigenous flora

The indigenous anaerobic microflora are, however, often profoundly affected by pathophysiological states and by exposure to antibiotics and other pharmacological compounds. The stomach characteristically harbours few organisms of a transient nature, but in certain disease states (e.g. peptic ulcer disease with or without bleeding, obstruction or gastric carcinoma), significant increases in number and types of bacteria (e.g. Escherichia coli and B. fragilis) may result. In some cases, the flora may even approximate that of the colon [14]. Following resection of significant amounts of large intestine with accompanying occlusion of the superior mesenteric artery, Gram-positive anaerobes may sometimes replace Gram-negatives as the predominating bowel flora [15]. Similarly, marked overgrowth with anaerobes occurs in the bypassed segment of bowel in patients undergoing ileal bypass for treatment of morbid obesity and in other conditions resulting in bowel stasis (e.g. diverticula of the small intestine) and blind loop syndrome [16,17].


Essentially all anaerobic infections emanate from the indigenous flora of the body. For example, long-term use of a contraceptive intra-uterine device increases the number of anaerobic bacteria among the cervical flora and may predispose to localized disease (e.g. actinomycosis) as well as to bacterial vaginosis and pelvic inflammatory disease [18]. Rarely, as in certain clostridial infections, they may be acquired exogenously. Contrary to popular belief, only a minority of all cases of post-traumatic gas gangrene falls into this category. The majority of cases of clostridial myonecrosis secondary to battle wounds are, in fact, of endogenous origin as a result of diminished hygiene under wartime conditions. Likewise, post-surgical clostridial myonecrosis is generally attributable to the patient's own colonic flora. Exogenously acquired cases also sometimes occur via the gastrointestinal tract and may present as intoxication or as overt infection. Examples include type A Cl. perfringens food poisoning and antibiotic-related diarrhoea, type C Cl. perfringens-induced enteritis necroticans, infant botulism, and infections caused by Cl. difficile such as nosocomially-acquired antibiotic-associated pseudomembranous colitis. Wound botulism may also represent an exogenously acquired syndrome and tetanus, caused by Cl. tetani, is the paramount example of an exogenously acquired anaerobic intoxication. Occasionally anaerobes causing infection are transmitted from animals to humans following a bite injury and a zoonotic role for Anaerobiospirillum species has been postulated in patients with diarrhoea [19]. Human bites also frequently result in anaerobic infection.

Anaerobes have been incriminated in essentially all types of infections. A partial list of important syndromes caused by anaerobes is given in Table 2. Factors that influence whether infection with anaerobes occurs are the inoculum size and the innate virulence of the infecting organisms as well as their ability to interact synergistically with other bacterial species. Three of the more important virulence factors in anaerobes are: the ability to adhere to or invade epithelial surfaces; the production of enzymes, toxins or other pathogenicity factors; and the presence of cell surface constituents such as lipopolysaccharide (endotoxin) or capsular polysaccharides. Host factors including breaks in mucosal barriers and conditions that lower the oxidation–reduction potential also play critical roles pathogenetically. Some important predisposing host conditions to anaerobic infection are listed in Table 3.

Table 2
Table 2:
Partial list of infections in which anaerobic bacteria are involved
Table 3
Table 3:
Conditions and events predisposing to anaerobic infection

Characterized by suppuration and frank pus formation, tissue destruction during anaerobic infections may be extensive and gangrenous. Gas may form in infected tissues and septic thrombophlebitis may become manifest. Members of the B. fragilis (Table 4) group are the most commonly recovered and the most resistant of all anaerobic pathogens to antimicrobial therapy. B. fragilis is the most frequently encountered and most pathogenic of the group, but others including B. thetaiotaomicron, are also commonly seen and are typically more resistant to the antibiotics generally used for the treatment of anaerobic infections than is B. fragilis. The pigmented Porphyromonas and Prevotella spp. are recovered infrequently in pure culture and are typically more fastidious in their in-vitro growth requirements than are members of the B. fragilis group. The Prevotella spp. are important contributors to infection, particularly following insult to the oropharyngeal or female genital tract mucosal surfaces. One recently described organism, Sutterella wadsworthensis, recovered from mixed intra-abdominal infections, has demonstrated significant resistance to a number of antibiotics with recognized activity against anaerobes including metronidazole, clindamycin, cefotetan, ceftizoxime, piperacillin, piperacillin/tazobactam, and trova- floxacin [21]. This organism is also apparently more virulent than the related species B. ureolyticus and Campylobacter gracilis.

Table 4
Table 4:
Members of theBacteroides fragilis group

Of the non-spore-forming anaerobes, Fusobacterium necrophorum is without question the most virulent. Despite typical susceptibility to any of a number of antimicrobial agents, the organism often produces overwhelming bacteraemia with metastatic infections. Lemierre's syndrome (necrobacillosis or postanginal anaerobic septicemia) is an acute oropharyngeal infection with secondary septic thrombophlebitis of the internal jugular vein, frequently leading to metastatic infections that are caused by this pathogen. The disease characteristically occurs in young healthy adults. One recent report suggests an association between the syndrome and Epstein–Barr virus-induced acute mononucleosis [22].

Another recently described Gram-negative anaerobic bacillus, Bilophila wadsworthia, is seen in approximately half of all cases of gangrenous or perforated appendicitis. This virulent species is nutritionally fastidious and may require up to 1 week to grow in culture. Almost always β-lactamase-positive, the organism has also been reported in cases of liver abscess, pericarditis, purulent arthritis, empyema, bacteraemia and soft tissue infection [23,24].

Extremely virulent, Cl. perfringens is the most commonly isolated of the clostridial species. Cl. tertium and Cl. septicum are of importance because of their association with lower gastrointestinal malignancy, particularly the cecum, and their role in the enterocolitis that sometimes strikes neutropenic individuals. Among the anaerobic cocci , Peptostreptococcus magnus appears to be the most pathogenic [25].


To manage individual patients, to determine the activity of new compounds, and to monitor resistance trends locally, nationally, or internationally, antibiotic susceptibility testing of anaerobic bacteria is of significant value. When used for guiding therapy in individual patients, these tests are usually performed on isolates recovered in the following situations: serious infections, in pure culture or from usually sterile body sites, infections requiring extended therapy (e.g. osteomyelitis and endocarditis), patients failing empiric treatment and particularly virulent species. Examples of serious infections from which isolates should be tested for antimicrobial susceptibility include brain abscess, joint infection, infection of a vascular graft or prosthetic device, and recurrent bacteraemia. Organisms recognized as being commonly resistant to antibiotics and/or especially virulent include members of the B. fragilis group, Prevotella and Porphyromonas spp., certain fusobacteria, Cl. ramosum, Cl. septicum, Cl. perfringens, Bilophila wadsworthia, and Campylobacter gracilis. Appropriate methods for susceptibility testing include agar dilution on supplemented Brucella blood agar, broth (macro) dilution, and broth microdilution as described by the National Committee for Clinical Laboratory Standards (NCCLS) [26]. The E-test represents another commonly used option [11].

Historically it has been argued that antimicrobial susceptibility testing of anaerobes is not necessary because: (i) they remain highly susceptible to antibiotics with recognized anaerobic activity; (ii) their slow-growing nature results in delays that limit the test's clinical utility; (iii) standardized methods for such testing are not available. These arguments are now invalid. Unfortunately, the antimicrobial susceptibility patterns of anaerobes are no longer predictable and growing resistance has been clearly documented [27,28]. In addition, standardized techniques are now commonly used in clinical laboratories following the guidelines of the NCCLS. Furthermore, Snydman et al. have shown that results of in-vitro susceptibility testing for cefoxitin are predictive of patients’ response to therapy [29]. Although surgical intervention and drainage, when appropriate, remain the mainstays of therapy for most anaerobic infections, antibiotics clearly play a major role in treating many of these patients successfully. Patients’ overall health and the microenvironments in which infections occur are also contributing factors in the determination of how patients will respond to therapeutic interventions. Furthermore, most anaerobic infections are polymicrobial in nature. Therefore, to expect precise correlation between susceptibility test results and clinical outcome with anaerobic infections is unrealistic. The fact, however, that most such testing is conducted on isolates from chronic or refractory infections underscores its clinical relevance, even if 48 h of incubation are required for accurate testing. In the most serious infections caused by anaerobic bacteria, a consensus group of infectious disease physicians have concluded that patient clinical response does, in fact, correlate with in-vitro antimicrobial susceptibility testing results [30], clearly supporting the use of these tests in selected cases.


Among anaerobes, the mechanisms resulting in antimicrobial resistance to various classes of antibiotics are described below and are outlined in Table 5. Multiple-resistant organisms are sometimes isolated.

Table 5
Table 5:
Mechanisms of antimicrobial resistance of anaerobic bacteria

β-lactam class

Production of β-lactamases by anaerobes is common. Up to 97–100 % of B. fragilis group isolates in the USA [31,32] and 76 % of such isolates in the UK [33] have been shown to produce these enzymes. Among Bacteroides spp., other than those in the B. fragilis group, approximately 65 % produce β-lactamases [34]. Most of the β-lactamases produced by Bacteroides spp. are chromosomally-mediated and produced constitutively. Some contain serine at their active site whereas others, known as metallo-enzymes, require a Zn2+ at their active site for effective β-lactam hydrolysis [35]. Organisms producing these enzymes are particularly disturbing because, with the exception of the monobactams, metallo-β-lactamases hydrolyse all β-lactam agents, including carbapenems such as imipenem and meropenem, and are not inhibited by available β-lactamase inhibitors. Although usually chromosomal, plasmid-mediated metallo-β-lactamases also occur.

β-lactamase production has also been described among a number of other species of anaerobes including Cl. clostridioforme [36], Cl. butyricum [37], Cl. ramosum [38], Fusobacterium nucleatum [34,39], Prevotella spp. and Porphyromonas spp. [39]. Although β-lactamase production by Clostridium spp. remains uncommon, when produced such enzymes are generally inducible, with the exception of TEM-1 β-lactamase in Cl. ramosum.

Members of the B. fragilis group produce group 2e cephalosporinases, which are inhibited by the β-lactamase inhibitors sulbactam, clavulanic acid and tazobactam, thereby explaining their general susceptibility to the commercially available β-lactam/β-lactamase inhibitor antibiotic combinations. Many of the enzymes produced by Bacteroides spp. have been characterized molecularly as class A serine cephalosporinases, which are smaller in size than the inducible group 1 (class C) common to many aerobic Gram-negative species. However, a strain of B. intermedius has been reported that produces a β-lactamase refractory to inhibition by clavulanic acid. Strains of B. vulgatus have been described that produce a class A cefoxitin-hydrolysing enzyme capable of degrading the compound slowly in overnight culture. This degradation may be accompanied by decreased permeability, thus explaining their clinical resistance to cefoxitin therapy.

Rather than cephalosporinases, clostridia and fusobacteria produce at least three types of penicillinase. Enzymes from Cl. butyricum and fusobacteria are inhibited by clavulanic acid whereas the penicillinase from Cl. clostridioforme is refractory to inhibition by all three commercially available β-lactamase inhibitors. This penicillinase is similar in nature to the class D, group 2d cloxacillin-hydrolysing enzymes.

In addition to enzyme production, permeability changes lead to β-lactam resistance among anaerobes. Pore-forming molecules associated with decreased permeability and β-lactam resistance have been described in B. fragilis, Fusobacterium nucleatum and Porphyromonas endodontalis [40]. Decreased permeability has also been associated with increased β-lactamase production, resulting in even higher levels of resistance to β-lactam antibiotics [41]. Cefoxitin resistance has been correlated specifically with a decrease in outer membrane permeability and the loss of a 49–50 kDa outer membrane protein [42].

Binding to penicillin binding proteins (PBP) is the critical factor in determining whether a β-lactam antibiotic will effectively inhibit cell wall synthesis in a bacterium. Essential for bacterial growth, PBP function is the terminal stage of cell wall synthesis. When a β-lactam effectively competes for the active site of an essential PBP, bacterial cell death results. Three to five PBP exist in strains of Bacteroides spp.: a PBP1 complex comprised of from one to three different enzymes, PBP2 and PBP3. Most β-lactams bind well to PBP2 as well as the PBP1 complex of Bacteroides spp. By comparison, monobactams such as aztreonam have very poor affinity for these Bacteroides PBP, accounting for their lack of activity against them [43]. Although resistance to cephalosporins is usually attributable to β-lactamase production, resistance in B. fragilis due to decreased affinity for PBP3 has also been incriminated and decreased susceptibility to cefoxitin has been attributed to decreased affinity for PBP2 as well as the PBP1 complex.


Rates of resistance to clindamycin vary geographically and, therefore, surveillance for resistance is critical to assess the utility of clindamycin as a therapeutic option at any given institution [44]. Although all three major mechanisms of antimicrobial resistance (enzyme inactivation, target site modification, and permeability alterations) result in resistance to clindamycin among various species of anaerobes, macrolide/lincosamide/streptogramin B (MLS) is the mechanism most frequently encountered in Bacteroides spp. Based on DNA and protein sequence analysis, it apparently occurs by a mechanism similar to that for clindamycin resistance in staphylococci.

Among Bacteroides spp., resistance is apparently mediated by methylation of 23S rRNA at one of two adenine residues which prevents effective binding of clindamycin to the ribosomes, rendering them refractory to the drug's inhibitory properties. This mechanism may be either inducible or expressed constitutively. Three closely related MLS genes, cloned from different Bacteroides spp., lie either on a conjugal element or on a transposon:ermFS is encoded on Tn 4551, ermFU is encoded on a B. vulgatus conjugal element, and ermF is encoded on Tn 4351. The resulting proteins have sequence identities highly similar to those encoded by the MLS resistance genes of Gram-positive organisms. However, not all Bacteroides clindamycin resistance DNA sequences cross-hybridize with the ermF gene, indicating that another mechanism of resistance is also present in some Bacteroides spp.

Conjugal transfer of clindamycin resistance has been shown to be plasmid-mediated. Many of these plasmids are self-transmissible and range in size from 14.6 kb to approximately 82 kb. Chromosomally-encoded clindamycin resistance is linked to tetracycline resistance and the gene has been shown to lie within the tetracycline resistance transfer element [45].


Resistance to tetracycline is almost universal among Bacteroides spp.: the rate exceeds 80–90 % at most medical centres. Due to this high rate of resistance, tetracycline should not be used therapeutically for anaerobic infections, as it was through the 1960s, without evidence generated from susceptibility testing that a patient's infecting pathogens are susceptible. Protection or modification of the target site is the only documented mechanism of resistance to tetracycline in Bacteroides spp. The chromosomal tetQ gene encodes a protein that renders the ribosomal protein synthesis apparatus refractory to the inhibitory effects of tetracyclines. DNA cross-hybridization studies indicate that a tetQ or tetQ-like gene is present in most tetracycline-resistant Bacteroides isolates. Identification of tetracycline-resistant isolates that do not contain tetQ DNA sequences indicates that another mechanism (e.g. efflux of tetracycline) or another class of ribosomal protection proteins also contributes to tetracycline resistance.

Tetracycline resistance genes have also been identified in Clostridium spp. The tetA (P) and tetB (P) genes together form an operon encoding two unrelated proteins that result in tetracycline resistance mediated by two separate mechanisms [46]. Two additional genes that cause tetracycline resistance have been identified in Bacteroides species. The tetX gene product inactivates tetracycline by oxidation of the molecule, but is active only under aerobic conditions and has, therefore, not been shown to be operational in Bacteroides spp. A gene encoding a protein that is able to actively efflux tetracycline has also been identified, but has not been shown to confer resistance to tetracycline in these species.

The tetQ resistance gene is both inducible and transferable. The frequency of transfer of resistance to tetracycline is, however, very low unless an organism is exposed to tetracycline. Gene transfer occurs by conjugation mediated by the tetracycline resistance transfer element itself. Transfer is controlled by a two-component regulatory system. The two regulatory genes, rteA and rteB, are located in the tetQ operon downstream from the tetQ gene. Their expression is enhanced greatly by the presence of tetracycline, explaining the more efficient transfer of genetic material following tetracycline exposure. The rteA gene encodes the cytoplasmic membrane component of the system, RteA. Encoded by the rteB gene, the regulatory response protein, RteB, appears to elicit its effect through an interaction with a σ54-like protein [47]. RteB plays an essential role in the mobilization and transfer of the tetracycline transfer element. A third gene, rteC, is also involved in the self-transfer of tetracycline resistance. Although its specific role remains undefined, RteC, its gene product, follows RteB in the regulation cascade.

In addition to regulating the movement of the tetracycline resistance transfer element, RteA and RteB are also involved in the regulation of transfer of unlinked chromosomal elements called non-replicating Bacteroides units (NBU). Although most NBU do not result in any type of phenotypic expression, a cefoxitin-hydrolysing β-lactamase gene, cfxA, has been shown to reside on a NBU. Transfer of the cefoxitin-hydrolysing β-lactamase is facilitated by tetracycline pretreatment. Similar to the conjugal transposon Tn 916 in Enterococcus faecalis, these tetracycline resistance transfer elements, with estimated sizes ranging from 70 to 80 kbp, reside in the chromosome and frequently contain other resistance genes (e.g. ermF).

5-Nitroimidazoles (metronidazole)

Metronidazole, the first 5-nitroimidazole to be used clinically, was introduced in 1960. It was not until 1978, however, that a metronidazole-resistant B. fragilis isolate was recovered from a patient following long-term therapy [48]. Metronidazole-resistant isolates from patients not treated with the compound have also been reported. One theory put forward for clinical failure with metronidazole is that the drug is inactivated by other organisms present in polymicrobial infections. Enterococcus faecalis has been reported to inactivate metronidazole and protect B. fragilis from its antimicrobial effects in mixed cultures. Despite these reports, metronidazole resistance among anaerobic Gram-negative bacilli remains extraordinarily low, typically < 1 %, and the compound continues to be viewed as a primary drug for treatment of infections caused by Bacteroides and other anaerobic Gram-negative species.

The 5-nitroimidazoles, including metronidazole, tinidazole and ornidazole, must undergo reduction to form an active antibacterial agent. This reduction is rapidly reversible in the presence of oxygen and is stable only under anaerobic conditions. Resistance to metronidazole is attributed to a combination of reduced nitroreductase and decreased drug uptake by the organism. Both of these occur simultaneously. Because entry of 5-nitroimidazoles into the cell depends on the rate of reduction of the nitro group, any decrease in the reducing environment within the bacterium will result in reduced nitroreductase activity and a concomitant reduction in drug uptake. Decreased pyruvate : ferredoxin oxireductase activity in combination with a compensatory increase in the lactate dehydrogenase activity results in a decrease in the reducing power of the bacterium. These changes in enzyme activity are uncommon and difficult for the organism to accomplish. Metronidazole-resistant phenotypes have, however, been reported, albeit rarely. Although the exact mechanism of this metronidazole resistance has not been identified, efflux and reduced permeability to the drug have been excluded.

Two genes, designated nimA and nimB, capable of conferring moderate- to high-level resistance have been described. DNA sequencing of these two genes shows approximately 73 % similarity, and they presumably represent two unique genes that confer resistance by the same mechanism. These genes have been localized to the chromosome as well as to a variety of different non-self-transmissible plasmids. These plasmids can, however, be mobilized by other conjugal elements or acquired by transformation. Of note, the transcriptional start genome for both nimA and nimB has been sequenced and is located on insertion sequence (IS) elements integrated 12–14 bp upstream from the protein-coding regions. IS 1118, the element for the nimA gene, is very closely related to the IS element providing the transcriptional start information for the nimB gene and is almost identical to IS 1186, an IS element that provides the transcriptional initiation signals for the metallo-β-lactamase gene described in Bacteroides spp.


Resistance to chloramphenicol among anaerobes, as reported in most antimicrobial susceptibility surveys, remains uncommon probably due to its relatively infrequent use in areas where such surveillance studies have been conducted. Two different classes of chloramphenicol resistance genes have, however, been reported in Bacteroides spp. Both result in drug inactivation, either through acetylation or by nitro-reduction at the p-nitro group of the benzene ring. The chloramphenicol acetyltransferase gene is transferable and plasmid-mediated.


Most currently available quinolones and fluoroquinolones have limited activity against many anaerobic bacterial species. Recently-developed agents including trovafloxacin, however, demonstrate far superior activity than previously released drugs [49]. Whether the typically low rate of susceptibility to most quinolones is due to poor drug penetration, low affinity for the target topoisomerases II and IV, or some other mechanism has not yet been clearly ascertained. Concomitant reduced cell permeability has, however, been described in clinical isolates of B. fragilis to cefoxitin and norfloxacin.

Another reservation to the use of quinolone class antibiotics for treatment of anaerobic infections is the finding that they are generally bacteriostatic rather than bactericidal under anaerobic conditions. As a group, fluoroquinolones have significantly more activity against oral anaerobes than members of the B. fragilis group. As a result, they may eventually prove useful for treatment of mixed aerobic and anaerobic pulmonary infections such as aspiration pneumonia.


Resistance to aminoglycosides is uniform among anaerobic bacteria. This universal resistance is not, however, due to decreased target sensitivity. Both gentamicin and streptomycin bind to and inhibit protein synthesis occurring on Cl. perfringens and B. fragilis ribosomes in a cell-free system [50]. Drug inactivation was not detectable in cell extracts of these same two species [50] and, therefore, does not appear to account for this resistance. Instead, it appears as if aminoglycoside resistance is the result of the drugs’ failure to reach their target sites. Aminoglycoside uptake is a two-step process involving both energy-independent and energy-dependent phases. Either an oxygen- or a nitrogen-dependent electron transport system has the capability to provide the energy necessary for the energy-driven phase of aminoglycoside uptake. Strict anaerobes lack these electron transport systems and are therefore impermeable to aminoglycosides.


Anaerobic bacteria are common causes of clinical infection. Their recovery from patient specimens is compromised, however, by their innate sensitivity to molecular oxygen. As a result, these important pathogens are often underappreciated as significant contributors to infectious processes. Because infections caused by anaerobes are frequently rapidly destructive, resulting in significant morbidity and on occasion mortality, hospitals should take measures to assure that optimal specimen transport systems (e.g. PRAS transport media) are in place so that these organisms can be recovered efficiently in culture.

In addition, clinical microbiology laboratories should assess their methods of anaerobic culture to optimize recovery of these pathogens further and to assure that their presence in patient specimens is reported rapidly and accurately. Although surgical intervention remains critical to the effective management of many types of infectious processes involving anaerobes, antimicrobial therapy is often an equally important therapeutic modality.

Because of evolving resistance to all major classes of antimicrobial agents used for treatment of clinical infection, antibiotic susceptibility testing of anaerobes is important to direct therapy. Many clinical microbiology laboratories have to date either been unable or unwilling to conduct such testing. All hospital laboratories should develop a strategy to offer susceptibility testing to clinicians when clinically relevant anaerobes are recovered from infections. In addition, a mechanism should be developed whereby cumulative antimicrobial susceptibility results are collected such that changing patterns of resistance can be monitored and direction can be afforded to physicians in terms of drugs useful for empirical treatment of anaerobic and mixed aerobic/anaerobic infections. Only by instituting such measures will optimal patient care be guaranteed in cases of serious anaerobic infection.


1. Loesche WJ. Oxygen sensitivity of various anaerobic bacteria. Appl Microbiol 1969, 18: 723–727.
2. Tally FP, Jacobus NV, Goldin BR. et al.: Superoxide dismutase in anaerobic bacteria [abstract]. Clin Res 1975, 23: 418.418.
3. Tally FP, Stewart PR, Sutter VL. et al.: Oxygen tolerance of fresh clinical anaerobic bacteria. J Clin Microbiol 1975, 1: 161–164.
4. Summanen P, Baron EJ, Citron DM. et al. Wadsworth Anaerobic Bacteriology Manual. 5th edn. Belmont, CA: Star Publishing Co.; 1993.
5. Finegold SM. A century of anaerobes: A look backward and a call to arms. Clin Infect Dis 1993, 16 (Suppl 4): S453–SS457.
6. Levy E. Ueber einen Fall von Gasabscess (VI). Dtsch Z Chir 1891, 32: 248–251.
7. Veillon A. Sur un microcoque anaérobie trouvé dans des suppurations fetides. C R Soc Biol (Paris) 1893, 45: 807–809.
8. Veillon A, Zuber A. Sur quelques microbes strictement anaérobies et leur role dans la pathologie humaine. C R Soc Biol (Paris) 1897, 49: 253.253.
9. McIntosh J, Fildes P. A new apparatus for the isolation of anaerobic microorganisms. Lancet 1916, 1: 768–770.
10. Holdeman LV, Cato EP, Moore WEC (eds):Anaerobe Laboratory Manual. 4th edn. Blacksburg: Virginia Polytechnic Institute and State University; 1977.
11. Engelkirk PG, Duben-Engelkirk J, Dowell VR (eds):Clinical Anaerobic Bacteriology. Belmont, CA: Star Publishing Co.; 1992.
12. Hentges DJ. The anaerobic microflora of the human body. Clin Infect Dis 1993, 16 (Suppl 4): S175–S180.
13. Finegold SM. Anaerobic bacteria: general concepts. In Principles and Practice of Infectious Diseases, vol. 2. Edited by Mandell GL, Bennett JE, Dolin R. Philadelphia: Churchill Livingstone; 2000: 2519–2537.
14. Nichols RN, Smith JW. Intragastric microbial colonization in common disease states of the stomach and duodenum. Ann Surg 1975, 182: 557–561.
15. Stolberg L, Rolfe R, Gitlin N. et al.: D-lactic acidosis due to abnormal gut flora: Diagnosis and treatment of two cases. N Engl J Med 1982, 306: 1344–1348.
16. Corrodi P, Wideman PA, Sutter VL. et al.: Bacterial flora of the small bowel before and after bypass procedure for morbid obesity. J Infect Dis 1978, 137: 1–6.
17. Polter DE, Boyle JD, Miller LG. et al.: Anaerobic bacteria as cause of the blind loop syndrome. A case with observations on response to antibacterial agents. Gastroenterology 1968, 54: 1148–1154.
18. Haukkamaa M, Stranden P, Jousimies-Somer H. et al.: Bacterial flora of the cervix in women using different methods of contraception. Am J Obstet Gynecol 1986, 154: 520–524.
19. Malnick H, Jones A, Vickers JC. Anaerobiospirillum: cause of a new zoonosis? Lancet 1989, 1: 1145–1146.
20. Takahara S. Progressive oral gangrene probably due to lack of catalase in the blood (acatalasaemia). Lancet 1952, 2: 1101–1104.
21. Molitoris E, Wexler HM, Finegold SM. Sources and antimicrobial susceptibilities of Campylobacter gracilis and Sutterella wadsworthensis. Clin Infect Dis 1997, 25 (Suppl 2): S264–S265.
22. Moller K, Dreijer B. Post-anginal sepsis (Lemierre's disease): a persistent challenge. Presentation of 4 cases. Scand J Infect Dis 1997, 29: 191–194.
23. Finegold S, Summanen P, Gerardo SH. et al.: Clinical importance of Bilophila wadsworthia. Eur J Clin Microbiol Infect Dis 1992, 11: 1058–1063.
24. Kasten MJ, Rosenblatt JE, Gustafson DR. Bilophila wadsworthia bacteremia in two patients with hepatic abscess. J Clin Microbiol 1992, 30: 2502–2503.
25. Bourgault A-M, Rosenblatt JE, Fitzgerald RH. Peptostreptococcus magnus: A significant human pathogen. Ann Intern Med 1980, 93: 244–248.
26. Hecht DW, NCCLS Working Group on Susceptibility testing of Anaerobic Bacteria:Methods for Antimicrobial Susceptibility Testing of Anaerobic Bacteria. 4th edn. Approved Standard. Wayne, PA: NCCLS Document M11-A4, v.13, no. 26:1997.
27. Snydman DR, McDermott L, Cuchural GL. et al.: Analysis of trends in antimicrobial resistance patterns among clinical isolates of Bacteroides fragilis group species from 1990 to 1994. Clin Infect Dis 1996, 23 (Suppl 1): S54–S65.
28. Snydman DR, Jacobus NV, McDermott LA. et al.: Multicenter study of in vitro susceptibility of the Bacteroides fragilis group, 1995 to with comparison of resistance trends from 1990 to 1996. Antimicrob Agents Chemother 1999, 1996, 43: 2417–2422.
29. Snydman DR, Cuchural GJ, McDermott L, Gill M. Correlation of various in vitro testing methods with clinical outcomes in patients with Bacteroides fragilis group infections treated with cefoxitin: a retrospective analysis. Antimicrob Agents Chemother 1992, 36: 540–544.
30. Rosenblatt JE, Brook I. Clinical relevance of susceptibility testing of anaerobic bacteria. Clin Infect Dis 1993, 16 (Suppl 4): S446–S448.
31. Jacobs MR, Spangler SK, Appelbaum PC. Beta-lactamase production and susceptibility of US and European anaerobic gram-negative bacilli to beta-lactams and other agents. Eur J Clin Microbiol Infect Dis 1992, 11: 1081–1093.
32. Aldridge KE, Henderberg A, Schiro DD, Sanders CV. Susceptibility of Bacteroides fragilis group isolates to broad spectrum β-lactams, clindamycin, and metronidazole: rates of resistance, cross-resistance, and importance of β-lactamase production. Adv Ther 1988, 5: 273–282.
33. Edwards R, Greenwood D. An investigation of β-lactamases from clinical isolates of Bacteroides species. J Med Microbiol 1992, 36: 89–95.
34. Appelbaum PC, Spangler SK, Jacobs MR. β-Lactamase production and susceptibilities to amoxicillin, amoxicillin-clavulanate, ticarcillin, ticarcillin-clavulanate, cefoxitin, imipenem, and metronidazole of 320 non-Bacteroides fragilis Bacteroides isolates and 129 fusobacteria from 28 U.S. centers. Antimicrob Agents Chemother 1990, 34: 1546–1550.
35. Rasmussen BA, Bush K, Tally FP. Antimicrobial resistance in anaerobes. Clin Infect Dis 1997, 24 (Suppl 1): S110–S120.
36. Appelbaum PC, Spangler SK, Pankuch GA. et al.: Characterization of a β-lactamase from Clostridium clostridioforme. J Antimicrob Chemother 1994, 33: 33–40.
37. Hart CA, Barr K, Makin T, Brown P, Cooke RWI. Characteristics of a β-lactamase produced by Clostridium butyricum. J Antimicrob Chemother 1982, 10: 31–35.
38. Matthew M. Plasmid-mediated β-lactamases of Gram-negative bacteria; properties and distribution. J Antimicrob Chemother 1979, 5: 349–358.
39. Appelbaum PC, Spangler SK, Jacobs MR. Evaluation of two methods for rapid testing for beta-lactamase production in Bacteroides and Fusobacterium. Eur J Clin Microbiol Infect Dis 1990, 9: 47–50.
40. Wexler HM. Pore-forming molecules in Gram-negative anaerobic bacteria. Clin Infect Dis 1997, 25 (Suppl 2): S284–S286.
41. Rasmussen BA, Yang Y, Jacobus N, Bush K. Contribution of enzymatic properties, cell permeability, and enzyme expression to microbiological activities of β-lactams in three Bacteroides fragilis isolates that harbor a metallo-β-lactamase gene. Antimicrob Agents Chemother 1994, 38: 2116–2120.
42. Piddock LJV, Wise R. Properties of the penicillin-binding proteins of four species of the genus Bacteroides. Antimicrob Agents Chemother 1986, 29: 825–832.
43. Georgopadakou NH, Smith SA, Sykes RB. Mode of action of azthreonam. Antimicrob Agents Chemother 1982, 21: 950–956.
44. Jenkins SG, Lewis JW: Changing patterns of antimicrobial susceptibility in Bacteroides fragilis over twelve years. In Abstracts of the 94th General Meeting of the American Society for Microbiology; 1994.
45. Shoemaker NB, Guthrie EP, Salyers AA, Gardner JF. Evidence that the clindamycin-erythromycin resistance gene of Bacteroides plasmid pBF4 is on a transposable element. J Bacteriol 1985, 162: 626–632.
46. Sloan J, McMurry LM, Lyras D, Levy SB, Rood JI. The Clostridium perfringens TetP determinant comprises two overlapping genes; tetA (P), which mediates active tetracycline efflux, and tetB (P), which is related to the ribosomal protection family of tetracycline-resistance determinants. Mol Microbiol 1994, 11: 403–415.
47. Stevens AM, Sanders JM, Shoemaker NB, Salyers AA. Genes involved in production of plasmidlike forms by a Bacteroides conjugal chromosomal element share amino acid homology with two-component regulatory systems. J Bacteriol 1992, 174: 2935–2942.
48. Ingham HR, Eaton S, Venables CW, Adams PC. Bacteroides fragilis resistant to metronidazole after long-term therapy [letter]. Lancet 1978, 1: 214.214.
49. Aldridge KE, Ashcraft D, Bowman KA. Comparative in vitro activities of trovafloxacin (CP 99,219) and other antimicrobials against clinically significant anaerobes. Antimicrob Agents Chemother 1997, 41: 484–487.
50. Bryan LE, Kowand SK, Van Den Elzen HM. Mechanism of aminoglycoside antibiotic resistance in anaerobic bacteria:Clostridium perfringens and Bacteroides fragilis. Antimicrob Agents Chemother 1976, 9: 928–938.

Anaerobes; gangrene; susceptibility; antibiotics; Bacteroides; resistance

© 2001 Lippincott Williams & Wilkins, Inc.