Antibiotics in Laboratory Medicine, 6 Ed.

Chapter 4. Antimicrobial Susceptibility Testing of Anaerobic Bacteria

Anilrudh A. Venugopal and David W. Hecht

Anaerobic bacteria have been well established in the literature as causing significant infection in humans (1). In certain situations, a single anaerobic organism can be the cause of a specific infection or sequelae, such as tetanus (Clostridium tetani), botulism (Clostridium botulinum), or food poisoning (Clostridium perfringens). However, the majority of infections due to anaerobes occur most often as mixed infections involving intraabdominal, skin and soft tissue, pulmonary, or central nervous system sites (1). Anaerobes have been recognized as either a causative agent or contributor to infection, and appropriate treatment is required for a good clinical outcome (2,3).

Recently, much attention has been given to antibiotic resistance among numerous aerobic and facultative anaerobic bacteria, with clear evidence of clinical failure when an ineffective antibiotic is used (4–7). Much less attention has been given to the role of antibiotic resistance among anaerobes and adverse clinical outcome. In fact, clinical trials that assess the efficacy of new antibiotics with good in vitro activity against anaerobic bacterial pathogens generally show favorable outcomes when compared with less active agents (8). However, until recently, there have been few studies demonstrating a correlation of antibiotic-resistant anaerobes with poor clinical outcome. Factors that have limited the ability to draw such conclusions from any study include the nature of the infection (mixed aerobes and anaerobes), lack of identification of anaerobic bacteria from specimens, absence of clinical data, effects of surgical drainage or debridement (a major factor that obscures the importance of a resistant organism), and previous inaccurate or modified susceptibility testing methods (9,10).

Several retrospective clinical studies were published in the 1980s and early 1990s that supported the association of antibiotic resistance among Bacteroides sp and clinical failure (11–13). However, a sentinel prospective, observational, Bacteroides fragilis group bacteremia study published in 2000 confirmed these previous suppositions and conclusions (14). In that study, mortality rate for patients receiving ineffective therapy (for resistant B. fragilis) was significantly higher than those receiving therapy that was active in vitro against the organism. Similarly, clinical failure and microbiologic persistence were greater for patients receiving ineffective therapy. These findings, along with numerous reports of increased antibiotic resistance (15), have prompted the Clinical and Laboratory Standards Institute (CLSI) (formerly the National Committee for Clinical Laboratory Standards [NCCLS]) to make recommendations for susceptibility testing of anaerobic bacteria in certain situations in their most recent standards publication (10). These recommendations have also been echoed in other publications (3,16).

To accomplish the most recent consensus document, the CLSI working group on antimicrobial susceptibility testing of anaerobic bacteria had previously conducted several multicenter collaborative studies to establish a highly reproducible agar dilution reference standard and comparable broth microdilution method (17,18). Coupled with the commercially available Etest (AB Biodisk, Solna, Sweden), there are now three reproducible and reliable methods for susceptibility testing of anaerobic bacteria (19). These different but comparable methods each have their pros and cons, depending on the testing needs.

Susceptibility testing has been rarely employed at most hospitals and medical centers. Despite the increasing prevalence of resistance among anaerobes, the frequency of testing was reported to be declining in the last survey conducted in the early 1990s partly because of budgetary reductions, a concomitant loss of expertise at these institutions, a lack of automated testing for anaerobes, and a failure to consider resistance as important to clinicians (20,21).

The recent and varied trends in antibiotic resistance, spread of resistance genes, and poor clinical outcomes resulting from ineffective antibiotic therapy argue strongly for more susceptibility testing of anaerobes. The most recent CLSI recommendations suggest that clinical laboratories should strongly consider susceptibility testing to assist in the management of individual cases of serious or life-threatening infections, to test for surveillance purposes for local and regional resistance rates, and to determine the rates of susceptibility of anaerobes when newer antimicrobial agents are introduced (10). The current major indications for testing individual isolates should be based on persistent infection despite an adequate treatment regimen, or difficulty in making empiric decisions based on precedent and known resistance of an organism or species (10). Organisms recognized as highly pathogenic and for which antimicrobial resistance cannot be predicted include members of the B. fragilis group, Prevotella sp, Fusobacterium sp, Clostridium sp, Bilophila wadsworthia, and Sutterella wadsworthensis (10).

Surveillance Testing

Annual surveillance testing is now recommended for clinical laboratories that routinely identify anaerobic bacteria to establish local patterns of resistance for commonly encountered anaerobes (3,10,16). Proper identification of anaerobes is, of course, an important first step in this process. Laboratories should be aware of recent taxonomic changes that have occurred within gram-negative anaerobic bacteria (based on 16S recombinant DNA techniques) that have resulted in regrouping of some organisms and the identification of new species. The reader is referred to excellent descriptions of these changes, as well as a strategy for identification of anaerobic species (22–25).

Strains to be tested for surveillance purposes should be collected over several months and stored until at least 50 to 100 are available for batch testing, allowing for the most efficient use of time, training, and materials. In general, anaerobic isolates to be tested should reflect the distribution of bacteria isolated in the laboratory. Because of the high frequency of resistance among the B. fragilis group, it is recommended that at least 20 isolates be selected from the various species, along with the testing of 10 isolates from other frequently isolated anaerobic genera. When choosing antibiotics to test, laboratories should consider at least one agent from each antibiotic class, and that should also reflect their respective hospital’s formulary (10,16).

Routine Testing

Individual isolate testing may also be appropriate in certain clinical situations. Isolates obtained from severe infections including brain abscess, bacteremia, endovascular infections, and bone and joint infections should be strongly considered for testing (10). Consultation with the physician about the clinical situation will be important in deciding on the need for susceptibility testing of these isolates (10).

Strategy for Testing

The anaerobe working group suggests a strategy for testing gram-positive and gram-negative groups of anaerobes. This includes a listing of primary and supplemental choices for testing B. fragilis group and other gram-negative anaerobes and gram-positive anaerobes (10). Examples of primary choices include one each of the β-lactam–β-lactamase (BLA) inhibitor combinations, carbapenems, clindamycin, and metronidazole for B. fragilis group and gram-negative rods. For the gram-positive anaerobes, penicillin (or ampicillin), β-lactam/BLA inhibitors, clindamycin, carbapenems, and metronidazole are recommended for routine primary testing.

The mechanism of action for most antianaerobic antimicrobial agents is either proven or presumed the same as that demonstrated for other nonanaerobic organisms. Although many of the pathogenic anaerobic bacteria are gram-negative, the either high or low degree of activity of some agents does not always closely match that of aerobic or facultative anaerobic bacteria. Cases where differences in activity are known are included in the discussion. The mechanism of action of clindamycin (and lincomycin), metronidazole, quinolones, aminoglycosides, and tetracyclines are discussed here. The mechanism of action of β-lactam antibiotics is not presented in this section, as this is thoroughly discussed in Chapter 11.

Clindamycin and lincomycin both work by binding to the 50S ribosomal subunit of bacteria, resulting in a disruption of protein synthesis by interfering with the transpeptidation reaction, preventing peptide chain elongation (26). Of note, chloramphenicol and macrolides compete for binding at the same site and are thought to be potentially antagonistic when used together. In aerobic bacteria, clindamycin may potentiate opsonization and phagocytosis of bacteria, presumably by the resulting changes in the cell wall surface decreasing adherence of bacteria to host cells and increasing intracellular killing (27,28). This phenomenon has not been demonstrated for anaerobes but could also occur. Clindamycin is considered bactericidal against B. fragilis, although its activity can be inconsistent. Resistance to this agent is discussed in the next section.

Metronidazole activity against anaerobes is mediated through a four-step process. In the first step, metronidazole must enter the cell, which it does efficiently as a low molecular weight compound that diffuses easily across cell membranes (29). The second step includes reductive activation by intracellular transport proteins. Metronidazole is reduced by the pyruvate:ferredoxin oxidoreductase system in the mitochondria of obligate anaerobes altering its chemical structure. Metronidazole is reduced when its nitro group acts as an electron sink, capturing electrons and reducing the compound, which also results in a concentration gradient driving its own uptake as well as forming intermediate compounds and free radicals toxic to the cell (30,31). In the third step, reduced intermediate particles interact with host cell DNA, resulting in fatal DNA strand breakage (32,33). Lastly, breakdown of cytotoxic intermediates occurs, resulting in inactive end products (34). Metronidazole is rapidly bactericidal in a concentration-dependent manner, killing B. fragilis and C. perfringens more rapidly than does clindamycin (35,36).

Fluoroquinolones directly inhibit bacterial DNA synthesis by binding to the complex of both DNA gyrase and topoisomerase IV, which are required for bacterial replication (37). The key event in quinolone action is reversible trapping of gyrase-DNA and topoisomerase IV–DNA complexes. Complex formation with gyrase is followed by a rapid, reversible inhibition of DNA synthesis and growth, resulting in damage to bacterial DNA and cell death. Thus, quinolones are also bactericidal agents.

Aminoglycosides bind 30S ribosomal subunits of aerobic bacteria but have no activity against anaerobes. Uptake of aminoglycosides by bacteria requires an energy-dependent phase normally provided by an oxygen- or nitrogen-dependent electron transport system that is absent in strictly anaerobic bacteria (38). Thus, anaerobes do no import aminoglycosides. However, aminoglycosides do bind to the ribosomes of B. fragilis and C. perfringens from cell-free extracts, indicating likely activity if they could gain entry into cells (39). Tetracyclines, on the other hand, are able to enter bacteria passively, including anaerobes, and also bind the 30S ribosomal subunit, preventing protein synthesis (40). However, resistance to this agent is widespread among anaerobes and, therefore, not frequently used. Specific resistance mechanisms are discussed in the next section.

As noted previously, antibiotic resistance among some anaerobes has increased significantly over the last few decades and parallels that of nonanaerobic bacteria (41). Organisms for which the most significant change has occurred are members of the B. fragilis group. Three major surveillance studies have reported significant changes in resistance among these bacteria since the 1980s (42–46). All three surveys confirm that resistance is increasing with hospital-to-hospital variation, even within the same geographic area. The most recent anaerobe survey conducted at eight medical centers throughout the United States reporting data from 2006 to 2009 confirms the presence of increasing resistance among anaerobes in geographically diverse areas (47).

Clindamycin Resistance

Clindamycin resistance among Bacteroides sp has increased the most significantly in the last two decades. Starting at only 3% in 1987, resistance to clindamycin in 2000 ranged from 16% to 44% resistance among the members of the B. fragilis group (42,46,48). In the 2006 to 2009 survey, the rates of clindamycin resistance remained stable at 19% to 50%, suggesting that this is not a good empiric initial choice without specific susceptibilities for this group of organisms (47). For many non-Bacteroides anaerobes, resistance has also increased, albeit not as significantly as in the B. fragilis group (48). Other reports have found up to 10% clindamycin resistance for Prevotella sp, Fusobacterium sp, Porphyromonas sp, and Peptostreptococcus sp, with higher rates for some Clostridium sp (especially Clostridium difficile) (44,49).

β-Lactam Antibiotic Resistance

Resistance to β-lactam agents among anaerobes is fairly widespread for the penicillins, cephamycins, and third-generation cephalosporins. Around 97% of the B. fragilis group is resistant to penicillin G by virtue of BLA production. In contrast, cefoxitin retains activity against most B. fragilis group members, although resistance has ranged between 8% and 22% over the period of 1987 to 2000 and similarly it ranged between 6% and 20% from 2006 to 2009 (47). Cefotetan activity is very similar to that of cefoxitin against B. fragilis but is much less potent against other members of the B. fragilis group (30% to 87% resistant) (43). Resistance to piperacillin, the most active ureidopenicillin against anaerobes, is also now widespread among members of the B. fragilis group (average 25% resistant). Resistance is also found among non-Bacteroides anaerobes (42,46,48).

Fortunately, activity of other more potent β-lactams, the β-lactam–BLA inhibitor combinations and carbapenems, remains excellent. The three combination agents of ampicillin/sulbactam, ticarcillin/clavulanate, and piperacillin/tazobactam are highly active against members of the B. fragilis group, with less than 4% resistance reported in the most recent survey (47). The carbapenem class of antibiotics including doripenem, imipenem/cilastatin, meropenem, and ertapenem remain potent agents against members of the B. fragilis group, with rates of resistance being less than 3% in recently tested isolates (47).

Resistance to β-lactam agents among non-Bacteroides anaerobes is generally much lower than that of Bacteroides. However, similar to that of Bacteroides organisms, Prevotella spp are also potent BLA producers, with more than 50% resistant to penicillin and ampicillin (50). Aldridge et al. (44) have reported penicillin resistance for Fusobacterium sp, Porphyromonas sp, and Peptostreptococcus sp. at 9%, 21%, and 6%, respectively. In the same survey, resistance to cefoxitin, cefotetan, β-lactam–BLA inhibitor combinations, and carbapenems was 0%, except for resistance to ampicillin/sulbactam in Peptostreptococcus sp and Porphyromonas sp, which were 4% and 5%, respectively (44).

5-Nitro-Imidazole Resistance

Although metronidazole resistance among gram-negative anaerobes had been reported in a single case in the United States, from a patient returning from Europe, and occasionally but rarely in European countries (51–53), however, more recent reports from United States and Europe have shown resistance rates of less than 1% for metronidazole among tested B. fragilis group isolates (47,54). Metronidazole resistance among gram-positive anaerobes is far more common, especially for most isolates of Propionibacterium acnes and Actinomyces sp (21).

Tetracycline and Glycylcycline Resistance

Among other antibiotic classes, tetracycline resistance is now nearly universal among Bacteroides sp and many other anaerobes, limiting its use in therapy. Tigecycline is the first agent in a newer glycylcycline class that was created by adding on a side chain to minocycline (55). It was designed to overcome the tetracycline-specific efflux pump and increase its activity over both aerobic and anaerobic bacteria (55). Although tigecycline has anaerobic activity against C. difficile, Fusobacterium sp, Prevotella sp, Porphyromonas sp, and the B. fragilis group, resistance in the latter has been reported at 0% to 8% (47,55).

Resistance to Other Antibiotics

Fluoroquinolone resistance among anaerobes has increased the most significantly and rapidly. Moxifloxacin is approved by the U.S. Food and Drug Administration (FDA) for treatment of complicated intraabdominal infections. Currently, use of this agent is very limited due to increasing resistance among the B. fragilis group. In the most recent anaerobe survey, the rates of resistance in the B. fragilis group ranged from 30% to 80% (47).

Resistance to aminoglycosides is universal among anaerobes, with this antibiotic restricted to combination therapy for mixed infections. Chloramphenicol resistance is very rare, but this agent is also rarely used in the clinical setting (56).

Table 4.1 summarizes the current known mechanisms of resistance and resistance genes for anaerobic bacteria. Not surprisingly, antibiotic resistance mechanisms are quite different for each class of antibiotics. Clindamycin resistance is mediated by a macrolide-lincosamide-streptogramin (MLS) type 23S methylase similar to that of staphylococci (48) and is typically encoded by one of several erythromycin ribosome methylation (erm) genes that are typically regulated and expressed at high levels. However, some isolates that contain erm genes demonstrate only moderately elevated minimum inhibitory concentrations (MICs) that would not otherwise be designated as resistant. Some of these latter strains can be detected by testing for erythromycin resistance. It is possible that these isolates could be induced to higher levels of resistance under selective pressure. At the current time, however, there is no specific recommendation to screen for resistance using erythromycin. Transfer of erm genes by conjugation in B. fragilis group organisms is easily demonstrated in the laboratory and likely explains the rapid emergence of this resistance phenotype (57–59). Of note, clindamycin is no longer recommended as empiric therapy for intraabdominal infections in the latest published guidelines, presumably because of the high rate of resistance (3).

Resistance to β-lactam antibiotics can occur by one of three major mechanisms: inactivating enzymes (BLAs), low-affinity penicillin-binding proteins (PBPs), or decreased permeability. BLA is by far the most common mechanism associated with resistance to β-lactam antibiotics. The most common BLAs found among Bacteroides sp and Prevotella sp are cephalosporinases of the type 2e class. These BLAs are inhibited by sulbactam, clavulanic acid, or tazobactam, thus the increased potency of the β-lactam–BLA inhibitor combinations. Cefoxitin- and cefotaxime-inactivating enzymes and other BLAs have also been reported in many B. fragilis group species (60). The most potent BLAs are the zinc metalloenzymes encoded by either ccrA or cfiA genes of the B. fragilis group (61). These enzymes are responsible for the rare resistance to carbapenems, are active against all β-lactam antibiotics with known activity against anaerobes, and are not inactivated by current BLA inhibitors. Although resistance to carbapenems is currently quite rare in the United States, up to 3% of Bacteroides strains have been found to carry one of the genes expressed at a very low level. These strains can be induced to a higher level of resistance in the laboratory under selective pressure caused by a promoter (contained in an insertion sequence) that has inserted upstream of the ccrA or cfiA genes (29,62).

Production of BLAs by other anaerobic bacteria has been generally less well studied, but strains of Clostridium, Porphyromonas, and Fusobacterium organisms express resistance by one or more of these enzymes. Penicillin-resistant Fusobacterium and Clostridium organisms express penicillinases that are typically inhibited by clavulanic acid, although exceptions among some Clostridium sp have been reported (56,63).

Other mechanisms of resistance to β-lactam antibiotics are far less frequent in occurrence and less well studied. Decreased binding to PBP2 or PBP1 complex has been reported in rare clinical isolates in cefoxitin resistance among B. fragilis strains (64). Alterations in pore-forming proteins of gram-negative anaerobes are a third type of resistance, with the absence of one or more outer membrane proteins associated with high MICs to ampicillin/sulbactam in some strains of B. fragilis (65,66).

Metronidazole resistance occurs by the lack of reduction to its active form in anaerobic bacteria. Metronidazole-resistant B. fragilis group organisms, although rare, carry one of six known nim genes that appear to encode a nitroimidazole reductase that converts 4- or 5-nitroimidazole to 4- or 5-aminoimidazole, preventing the formation of the toxic drug form necessary for the agents’ activity (67,68). These genes have been identified on both the chromosome and on transferable plasmids. High-level expression of the nim genes requires an insertion sequence with a promoter, similar to that of carbapenem resistance (69). Differential gene expression affecting cell metabolism in Bacteroides has also been reported as an alternative mechanism for resistance (70). In contrast to Bacteroides, the mechanism of resistance to metronidazole for non-Bacteroides anaerobes is currently not known. Interestingly, metronidazole resistance in the microaerophilic organism Helicobacter pylori has been partially solved for some strains that contain mutations in the rdxA gene, an oxygen-insensitive nitroreductase that converts metronidazole to its active form in this organism (71). Other candidate genes for resistance include the flavin oxidoreductase (frxA), ferredoxin-like proteins (fdxA, fdxB), and pyruvate oxidoreductase (porA, porB) (72). Metronidazole also has activity against Mycobacterium tuberculosis, although apparently only in dormant cells, when reduction of the drug can occur. Resistance among actively growing Mycobacterium organisms is presumed secondary to a lack of sufficient reducing potential (73).

Fluoroquinolone resistance among Bacteroides sp has been attributed to either a mutation in the quinolone resistance–determining region of the gyrase A gene (gyrA) from single or multiple mutations, or an alteration in efflux of the antibiotic (74–77). High-level resistance may be secondary to both mechanisms in the same cell, although only a few strains have been tested to date. Both of these mechanisms appear to be responsible for the cross-class resistance to newer quinolones.

The lack of activity of aminoglycosides against anaerobes is related to the lack of uptake by the bacteria under anaerobic conditions and a failure to reach their ribosome targets (39). Tetracycline resistance is widespread, especially among B. fragilis group and Prevotella sp (15). Several genes encoding resistance have been identified among various anaerobes, which encode protective proteins, resulting in protection of the ribosomes. More importantly, however, is the association of tetracycline resistance and the inducible transfer of this resistance determinant upon exposure to low levels of the antibiotic. Chloramphenicol resistance is extremely rare but, when found, is associated with inactivation of the drug by nitroreduction or acetyltransferase (78).

Several different methods, spanning five decades, have been utilized in antimicrobial susceptibility testing (AST) of anaerobic bacteria. During that time, more than 16 methods, 16 different media, and a host of other variables have been described to test susceptibility of anaerobes. NCCLS took the lead in developing a consensus for AST of anaerobes starting with the first approved standard in 1985, along with alternative methods published in a second document the same year. Subsequently, several revisions have been published that modified, added, or eliminated some methods (10). Following extensive multilaboratory collaborative studies sanctioned by the NCCLS in the late 1990s, a consensus was reached culminating in a single agar dilution standard and one broth microdilution method, both using the same medium (10,17,18). In addition to NCCLS methods, a very useful and highly correlated user-friendly method, Etest (AB Biodisk), has been FDA approved and available for several years (19,79). As a proprietary product, the Etest is not included in the CLSI documents. Of note, all methods can be performed in ambient air but require incubation in an anaerobic jar or glove box.

Choosing a method from among these three recognized and approved methods may depend on a number of factors (Table 4.2). The agar dilution standard has a very high degree of reproducibility but is fairly labor-intensive. A laboratory can test up to 30 isolates plus two controls, making it useful for batch testing. However, individual sets of dilution plates must be pored for each antibiotic, increasing material and labor costs. As the reference standard, agar dilution is most often used for evaluation of new antimicrobial agents. The broth microdilution method is more user-friendly than agar dilution and has the flexibility to test multiple antibiotics using the same microtiter plate, albeit only one isolate at a time. Based on available comparative studies of broth microdilution to the standard, results are considered equivalent when testing members of the B. fragilis group. However, results are not as comparable for non-Bacteroides anaerobes because of poor growth, and broth microdilution is not used for this group at this time. It should be noted that previously published studies testing non-Bacteroides anaerobes using a different broth (anaerobe MIC broth) were able to grow non-Bacteroides anaerobes with MIC results within twofold of those for agar dilution. Thus, broth microdilution testing for non-Bacteroides anaerobes can be considered but only if correlated with the current agar dilution standard. The third method, Etest, is relatively easy to perform and is well suited for testing individual isolates of any anaerobe. For surveillance testing, costs could be prohibitive.

For the agar dilution reference standard, each test concentration of an antibiotic is mixed into molten agar and poured into separate Petri dishes to which an inoculum of an organism is applied, incubated anaerobically, and examined to determine the MIC for the antimicrobial agent tested. Methods described here are adapted from those of CLSI standards (10). For laboratories inexperienced with this method and anaerobe AST, a useful time table for setup and testing is provided as an appendix in the CLSI standards document.

Testing Medium

Brucella blood agar supplemented with 5 µg hemin, 1% vitamin K₁, and 5% laked sheep blood is the recommended testing medium (10). Brucella agar blanks (17 mL) can be prepared in advance containing hemin and vitamin K₁, and flash-autoclaved or microwaved and placed in an approximately 50°C water bath on the day of use (10). One milliliter of laked sheep blood is then added to each melted blank while still in the water bath. Two milliliters of each twofold diluted test antibiotic is then added to the agar, mixed by inverting the tubes, and poured into sterile Petri dishes (10). Following hardening of the plates, and drying briefly in an inverted position in a 37°C incubator, the plates are ready for use. Preferably, plates should be made on the day of testing but can be sealed in plastic bags and stored at 2°C to 8°C for periods of up to 72 hours if necessary (10). Exceptions to storage include plates containing clavulanic acid or imipenem/cilastatin, which must be made on the day of use.

Inoculum Preparation

The inoculum can be prepared by either a direct colony suspension or growth method. Direct colony suspension requires 24- to 48-hour growth on a Brucella blood agar plate (10). Several colonies are touched lightly with an inoculating needle or cotton swab and suspended in reduced Brucella broth to achieve a turbidity equivalent to a McFarland standard of 0.5 (10). The alternative growth method involves the inoculation of enriched thioglycollate medium (without indicator) with portions of five or more colonies from a Brucella blood agar plate, incubating for 6 to 24 hours at 37°C, and adjusting the turbidity to a McFarland standard of 0.5 by addition of reduced Brucellabroth (10).

Inoculation and Incubation of Plates

Once the inoculum is prepared, it is most often applied using an inoculum-replicating apparatus, such as a Steers-Foltz replicator, to deliver 1 to 2 µL on the agar surface, corresponding to 1 × 10⁵ colony-forming units (CFUs) per spot. Depending on the device, either 32 or 36 wells can be filled with different test organisms and controls using a Pasteur or other pipette (10). Application of inoculum to plates includes repeated stamping of plates, starting with lowest to highest dilution of each antibiotic set. One plate of supplemented Brucella blood agar without antibiotic should be stamped prior to and after each set of antibiotics for growth control. Contamination by aerobic bacteria during the inoculation procedure can be detected by inoculating a drug-free plate and incubated aerobically. Once plates are inoculated, they should sit until liquid is absorbed into the medium and then incubated in an anaerobic environment at 35°C to 37°C for 42 to 48 hours (10).

Interpretation of Results

End points are determined by reading each plate against a dark, nonreflecting background and comparing it with the control growth plate. Any growth on the aerobic control should eliminate further interpretation of that test organism. The end point for a given test organism is where a marked reduction occurs in the appearance of growth compared with control. A marked change includes a haze, multiple tiny colonies, or one to several normal-sized colonies. These descriptions have been problematic for those inexperienced with using this method. To that end, CLSI recommends the use of two figures containing 28 full-dilution color photographic examples of end point readings to illustrate the written descriptions (10).

Interpretative categories approved by CLSI for MICs derived for anaerobic bacteria are shown in Table 4.3. This table includes agents that are the most frequently used in the clinical setting and were updated through 2012. Interpretative categories for any organism have been determined based on the population distribution of the bacteria, the pharmacokinetics, and pharmacodynamic properties of the antibiotic with verification of efficacy by clinical studies (see an in-depth description in Chapter 1). This works particularly well for single-organism infections. However, this is rarely the case for anaerobic bacteria, which are typically isolated from mixed infections. Many of the published anaerobic breakpoints were determined on the basis of animal models or the result of clinical trials involving patients with polymicrobial infections as well as pharmacokinetic data. Despite these potential limitations, the use of maximum dosages of antibiotics along with appropriate ancillary therapy (debridement or drainage) should be effective for organisms with susceptible breakpoints, although those with intermediate susceptibilities should be monitored closely (10).

The broth microdilution method has been validated by CLSI for susceptibility testing of the B. fragilis group.

Media

Brucella broth supplemented with hemin (5 µg/mL), vitamin K₁, and lysed horse blood is the recommended medium, which is essentially equivalent to that of agar dilution (10). Trays can be prepared fresh and then frozen or purchased commercially. Trays should be kept at −70°C. Antibiotics are diluted according to an algorithm recommended by CLSI in volumes of 15 to 100 mL (10) and delivered using a device that can simultaneously dispense aliquots of 0.1 mL per well (or 0.05 mL per well if a pipette will be used to deliver an equal volume of inoculum) (10). When a pipette is used for inoculation, antibiotic concentrations should be prepared at 2× the final desired concentration. Volumes of less than 0.1 mL are not recommended (10).

Inoculum Preparation, Inoculation Procedure, and Incubation

Inoculum preparation is the same as for agar dilution to achieve a turbidity of a McFarland standard of 0.5 (10). Commercially available inoculating devices can be used that deliver 10 µL of a 1:15 dilution of a 0.5 McFarland inoculum (10). For commercially prepared trays, follow the manufacturer’s recommendations. The final concentration of inoculum should be 1 × 10⁷ CFU/mL.

Before inoculation, frozen trays should be brought to room temperature and inoculated within 15 minutes of inoculum preparation. It is advisable to perform a colony count and purity check of the inoculum by removing 10 µL from the growth control well and diluting it into 10 mL of saline, streaking 0.1 mL onto the surface of an anaerobic blood agar plate, and incubating anaerobically (10). One hundred colonies on the plate correspond to 1 × 10⁶CFU/mL. Trays are then incubated for 46 to 48 hours at 35°C in an anaerobic atmosphere (see previous discussion), ensuring sufficient humidity to prevent drying (10).

Interpretation of Results

The MIC values are read by viewing the plates from the bottom using a stand and a mirror. A sufficient growth control is required to interpret results. The MIC end point is read as the concentration where no growth, or the most significant reduction of growth, is observed. A trailing effect may be observed for some drug–organism combinations. Again, two figures containing 28 examples of broth microdilution end points are provided by CLSI (10). Breakpoints for broth microdilution interpretation are the same as those for agar dilution (see Table 4.3).

A quality control program to monitor accuracy of testing, reagents, equipment, and persons conducting tests is essential. The quality control strains chosen for anaerobic bacteria are limited to two Bacteroidessp, a nontoxigenic C. difficile strain, and an Eggerthella strain. B. fragilis ATCC 25285 and Bacteroides thetaiotaomicron ATCC 29741 are appropriate for testing using any of the methods listed. C. difficileATCC 700057 is preferred over Eggerthella lenta ATCC 43055 when testing gram-positive anaerobes (10). Two of the four quality control strains should be used for each assessment when agar dilution is used. When an individual strain is being tested by broth microdilution or Etest, one strain should be included. Expected values for end points for both agar and broth microdilution methods are published by CLSI.

The Etest is an excellent and convenient choice for testing individual anaerobic organisms. Several studies have validated this method, demonstrating good correlation with the agar dilution method (19,79). However, Rosenblatt and Gustafson (79) have noted that some Prevotella and Bacteroides strains show false susceptibility when testing penicillin and ceftriaxone that is minimized if BLA-producing strains are eliminated. A more significant warning is potential false resistance to metronidazole as a result of test conditions and medium quality (80). This aberrant result can be avoided by prereducing test plates in an anaerobic chamber the night before testing.

Procedure

The Etest is a familiar technique to most clinical laboratories and does not differ significantly in its application to anaerobic bacteria. The Etest strips are coated with a gradient of antimicrobial on one side with an MIC interpretative scale on the other. The organism to be tested is prepared to a McFarland standard of 1 and applied to a 150-mm diameter Petri dish of supplemented Brucella blood agar, with the strips applied in a radial fashion. Smaller plates can be used with fewer strips. Incubation is recommended for 48 hours at 35°C and read where an elliptical zone of inhibition intersects the strip on the scale of MIC values.

The BLA testing deserves mention, although it is not a true AST test. Testing for BLA activity can be performed on anaerobic organisms, although it is not recommended for the B. fragilis group because of the high prevalence of positivity. This test can be used as a first step to drive additional testing choices. Any BLA-producing anaerobe should be considered resistant to penicillin and ampicillin. However, as noted in the section on antimicrobial resistance, alternative mechanisms of resistance to β-lactams are known, and a negative test does not assure susceptibility to penicillin.

The recommended method for testing is chromogenic and cephalosporin-based, either by a nitrocefin disk assay (Cefinase; BBL Microbiology Systems, Cockeysville, MD) or the S1 chromogenic disk (International BioClinical, Inc., Portland, OR). Tests are performed according to manufacturers’ directions. A positive reaction is denoted by a change in color from yellow to red that typically occurs within 5 to 10 minutes. However, some Bacteroides strains may react more slowly (up to 30 minutes) (10).

 1.  Finegold SM, George WL. Anaerobic infections in humans. San Diego, CA: Academic Press, 1989.

 2.  Proceedings of the 1st North American Congress on Anaerobic Bacteria and Anaerobic Infections. Marina del Rey, California, 24–26, 1992. Clin Infect Dis 1993;16(Suppl 4):S159–S457.

 3.  Solomkin JS, Mazuski JE, Bradley JS, et al. Diagnosis and management of complicated intra-abdominal infection in adults and children: guidelines by the Surgical Infection Society and the Infectious Diseases Society of America. Clin Infect Dis 2010;50:133–164.

 4.  Martinez E, Miro JM, Almirante B, et al. Effect of penicillin resistance of Streptococcus pneumoniae on the presentation, prognosis, and treatment of pneumococcal endocarditis in adults. Clin Infect Dis 2002;35:130–139.

 5.  Paterson DL, Ko WC, Von Gottberg A, et al. International prospective study of Klebsiella pneumoniae bacteremia: implications of extended-spectrum beta-lactamase production in nosocomial infections. Ann Intern Med 2004;140:26–32.

 6.  Peres-Bota D, Rodriguez H, Dimopoulos G, et al. Are infections due to resistant pathogens associated with a worse outcome in critically ill patients? J Infect 2003;47:307–316.

 7.  Yu Y, Zhou W, Chen Y, et al. Epidemiological and antibiotic resistant study on extended-spectrum beta-lactamase-producing Escherichia coli and Klebsiella pneumoniae in Zhejiang Province. Chin Med J (Engl) 2002;115:1479–1482.

 8.  Rosenblatt JE, Brook I. Clinical relevance of susceptibility testing of anaerobic bacteria. Clin Infect Dis 1993;16(Suppl 4):S446–S448.

 9.  National Committee for Clinical Laboratory Standards. Methods for antimicrobial susceptibility testing of anaerobic bacteria. 4th ed. M11-A4. Villanova, PA: National Committee for Clinical Laboratory Standards, 1997.

10.  Clinical and Laboratory Standards Institute. Methods for antimicrobial susceptibility testing of anaerobic bacteria. 8th ed. M11-A8. Wayne, PA: Clinical and Laboratory Standards Institute, 2012.

11.  Bieluch VM, Cuchural GJ, Snydman DR, et al. Clinical importance of cefoxitin-resistant Bacteroides fragilis isolates. Diagn Microbiol Infect Dis 1987;7:119–126.

12.  Dalmau D, Cayouette M, Lamothe F, et al. Clindamycin resistance in the Bacteroides fragilis group: association with hospital-acquired infections. Clin Infect Dis 1997;24:874–877.

13.  Snydman DR, Cuchural GJ Jr, McDermott L, et al. 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.

14.  Nguyen MH, Yu VL, Morris AJ, et al. Antimicrobial resistance and clinical outcome of Bacteroides bacteremia: findings of a multicenter prospective observational trial. Clin Infect Dis 2000;30:870–876.

15.  Nikolich MP, Shoemaker NB, Salyers AA. A Bacteroides tetracycline resistance gene represents a new class of ribosome protection tetracycline resistance. Antimicrob Agents Chemother 1992;36:1005–1012.

16.  Citron DM, Hecht DW. Susceptibility test methods: anaerobic bacteria. In: Murray PR, Baron EJ, Jorgensen JH, et al, eds. Manual of clinical microbiology. 8th ed. Washington, DC: ASM Press, 2003:1141–1148.

17.  Roe DE, Finegold SM, Citron DM, et al. Multilaboratory comparison of anaerobe susceptibility results using 3 different agar media. Clin Infect Dis 2002;35:S40–S46.

18.  Roe DE, Finegold SM, Citron DM, et al. Multilaboratory comparison of growth characteristics for anaerobes, using 5 different agar media. Clin Infect Dis 2002;35:S36–S39.

19.  Citron DM, Ostovari MI, Karlsson A, et al. Evaluation of the E test for susceptibility testing of anaerobic bacteria. J Clin Microbiol 1991;29:2197–2203.

20.  Goldstein EJ, Citron DM, Goldman RJ. National hospital survey of anaerobic culture and susceptibility testing methods: results and recommendations for improvement. J Clin Microbiol 1992;30:1529–1534.

21.  Goldstein EJC. United States national hospital survey of anaerobic culture and susceptibility methods. Anaerobe 1995;1:309–314.

22.  Citron DM, Hecht DW. Susceptibility test methods: anaerobic bacteria. In: Murray PR, Baron EJ, Jorgensen JH, et al, eds. Manual of Clinical Microbiology. 8th ed. Washington, DC: ASM Press, 2003:343.

23.  Jousimies-Somer H, Summanen P. Recent taxonomic changes and terminology update of clinically significant anaerobic gram-negative bacteria (excluding spirochetes). Clin Infect Dis 2002;35:S17–S21.

24.  Song Y, Liu C, McTeague M, et al. Rapid identification of Gram-positive anaerobic coccal species originally classified in the genus Peptostreptococcus by multiplex PCR assays using genus- and species-specific primers. Microbiology 2003;149:1719–1727.

25.  Song Y, Liu C, McTeague M, et al. 16S ribosomal DNA sequence-based analysis of clinically significant gram-positive anaerobic cocci. J Clin Microbiol 2003;41:1363–1369.

26.  Chang FN, Weisblum B. The specificity of lincomycin binding to ribosomes. Biochemistry 1967;6:836–843.

27.  Veringa EM, Ferguson DA Jr, Lambe DW Jr, et al. The role of glycocalyx in surface phagocytosis of Bacteroides spp. in the presence and absence of clindamycin. J Antimicrob Chemother 1989;23:711–720.

28.  Veringa EM, Verhoef J. Influence of subinhibitory concentrations of clindamycin on opsonophagocytosis of Staphylococcus aureus, a protein-A-dependent process. Antimicrob Agents Chemother 1986;30:796–797.

29.  Edwards R, Read PN. Expression of the carbapenemase gene (cfiA) in Bacteroides fragilis. J Antimicrob Chemother 2000;46:1009–1012.

30.  Edwards DI. Reduction of nitroimidazoles in vitro and DNA damage. Biochem Pharmacol 1986;35:53–58.

31.  Muller M. Reductive activation of nitroimidazoles in anaerobic microorganisms. Biochem Pharmacol 1986;35:37–41.

32.  Tocher JH, Edwards DI. The interaction of reduced metronidazole with DNA bases and nucleosides. Int J Radiat Oncol Biol Phys 1992;22:661–663.

33.  Tocher JH, Edwards DI. Evidence for the direct interaction of reduced metronidazole derivatives with DNA bases. Biochem Pharmacol 1994;48:1089–1094.

34.  Goldman P, Koch RL, Yeung TC, et al. Comparing the reduction of nitroimidazoles in bacteria and mammalian tissues and relating it to biological activity. Biochem Pharmacol 1986;35:43–51.

35.  Ralph ED, Kirby WM. Unique bactericidal action of metronidazole against Bacteroides fragilis and Clostridium perfringens. Antimicrob Agents Chemother 1975;8:409–414.

36.  Stratton CW, Weeks LS, Aldridge KE. Comparison of the bactericidal activity of clindamycin and metronidazole against cefoxitin-susceptible and cefoxitin-resistant isolates of the Bacteroides fragilis group. Diagn Microbiol Infect Dis 1991;14:377–382.

37.  Drlica K, Zhao X. DNA gyrase, topoisomerase IV, and the 4-quinolones. Microbiol Mol Biol Rev 1997;61:377–392.

38.  Bryan LE, Van Den Elzen HM. Streptomycin accumulation in susceptible and resistant strains of Escherichia coli and Pseudomonas aeruginosa. Antimicrob Agents Chemother 1976;9:928–938.

39.  Bryan LE, Kowand SK, Van Den Elzen HM. Mechanism of aminoglycoside antibiotic resistance in anaerobic bacteria: Clostridium perfringens and Bacteroides fragilis. Antimicrob Agents Chemother 1979;15:7–13.

40.  Craven GR, Gavin R, Fanning T. The transfer RNA binding site of the 30 S ribosome and the site of tetracycline inhibition. Cold Spring Harb Symp Quant Biol 1969;34:129–137.

41.  Hecht DW. Prevalence of antibiotic resistance in anaerobic bacteria: worrisome developments. Clin Infect Dis 2004;39(1):92–97.

42.  Snydman DR, Jacobus NV, McDermott LA, et al. National survey on the susceptibility of Bacteroides fragilis group: report and analysis of trends for 1997–2000. Clin Infect Dis 2002;35:S126–S134.

43.  Snydman DR, Jacobus NV, McDermott LA, et al. Multicenter study of in vitro susceptibility of the Bacteroides fragilis group, 1995 to 1996, with comparison of resistance trends from 1990 to 1996. Antimicrob Agents Chemother 1999;43:2417–2422.

44.  Aldridge KE, Ashcraft D, Cambre K, et al. Multicenter survey of the changing in vitro antimicrobial susceptibilities of clinical isolates of Bacteroides fragilis group, Prevotella, Fusobacterium, Porphyromonas, and Peptostreptococcus species. Antimicrob Agents Chemother2001;45:1238–1243.

45.  Hecht DW, Osmolski JR, O’Keefe JP. Variation in the susceptibility of Bacteroides fragilis group isolates from six Chicago hospitals. Clin Infect Dis 1993;16(Suppl 4):S357–S360.

46.  Hecht DW, Vedantam G. Anaerobe resistance among anaerobes: what now? Anaerobe 1999;5:421–429.

47.  Snydman DR, Jacobus NV, McDermott LA, et al. Update on resistance of Bacteroides fragilis group and related species with special attention to carbapenems 2006–2009. Anaerobe 2011;17:147–151.

48.  Jimenez-Diaz A, Reig M, Baquero F, et al. Antibiotic sensitivity of ribosomes from wild-type and clindamycin resistant Bacteroides vulgatus strains. J Antimicrob Chemother 1992;30:295–301.

49.  Ackermann G, Degner A, Cohen SH, et al. Prevalence and association of macrolide-lincosamide-streptogramin B (MLS(B)) resistance with resistance to moxifloxacin in Clostridium difficile. J Antimicrob Chemother 2003;51:599–603.

50.  Hecht DW. Anaerobes: antibiotic resistance, clinical significance, and the role of susceptibility testing. Anaerobe 2006;12:115–121.

51.  Breuil J, Dublanchet A, Truffaut N, et al. Transferable 5-nitroimidazole resistance in the Bacteroides fragilis group. Plasmid 1989;21:151–154.

52.  Urban E, Soki J, Brazier JS, et al. Prevalence and characterization of nim genes of Bacteroides sp. isolated in Hungary. Anaerobe 2002;8:175–179.

53.  Ricci V, Peterson ML, Rotschafer JC, et al. Role of topoisomerase mutations and efflux in fluoroquinolone resistance of Bacteroides fragilis clinical isolates and laboratory mutants. Antimicrob Agents Chemother 2004;48(4):1344–1346.

54.  Nagy E, Urban E, Nord CE, et al. Antimicrobial susceptibility of Bacteroides fragilis group isolates in Europe: 20 years of experience. Clin Microbiol Infect 2011;17(3):371–379.

55.  Stein GE, Craig WA. Tigecycline: a critical analysis. Clin Infect Dis 2006;43:518–524.

56.  Rasmussen BA, Bush K, Tally FP. Antimicrobial resistance in anaerobes. Clin Infect Dis 1997;24(Suppl 1):S110–S120.

57.  Bachoual R, Dubreuil L, Soussy CJ, et al. Roles of gyrA mutations in resistance of clinical isolates and in vitro mutants of Bacteroides fragilis to the new fluoroquinolone trovafloxacin. Antimicrob Agents Chemother 2000;44:1842–1845.

58.  Privitera G, Dublanchet A, Sebald M. Transfer of multiple antibiotic resistance between subspecies of Bacteroides fragilis. J Infect Dis 1979;139:97–101.

59.  Welch RA, Jones KR, Macrina FL. Transferable lincosa-mide-macrolide resistance in Bacteroides. Plasmid 1979;2:261–268.

60.  Rogers MB, Parker AC, Smith CJ. Cloning and characterization of the endogenous cephalosporinase gene, cepA, from Bacteroides fragilis reveals a new subgroup of Ambler class A beta-lactamases. Antimicrob Agents Chemother 1993;37:2391–2400.

61.  Yang Y, Rasmussen BA, Bush K. Biochemical characterization of the metallo-beta-lactamase CcrA from Bacteroides fragilis TAL3636. Antimicrob Agents Chemother 1992;36:1155–1157.

62.  Podglajen I, Breuil J, Collatz E. Insertion of a novel DNA sequence, 1S1186, upstream of the silent carbapenemase gene cfiA, promotes expression of carbapenem resistance in clinical isolates of Bacteroides fragilis. Mol Microbiol 1994;12:105–114.

63.  Appelbaum PC, Spangler SK, Pankuch GA, et al. Characterization of a beta-lactamase from Clostridium clostridioforme. J Antimicrob Chemother 1994;33:33–40.

64.  Fang H, Edlund C, Nord CE, et al. Selection of cefoxitin-resistant Bacteroides thetaiotaomicron mutants and mechanisms involved in beta-lactam resistance. Clin Infect Dis 2002;35:S47–S53.

65.  Wexler HM. Outer-membrane pore-forming proteins in gram-negative anaerobic bacteria. Clin Infect Dis 2002;35:S65–S71.

66.  Wexler HM, Halebian S. Alterations to the penicillin-binding proteins in the Bacteroides fragilis group: a mechanism for non-beta-lactamase mediated cefoxitin resistance. J Antimicrob Chemother 1990;26:7–20.

67.  Carlier JP, Sellier N, Rager MN, et al. Metabolism of a 5-nitroimidazole in susceptible and resistant isogenic strains of Bacteroides fragilis. Antimicrob Agents Chemother 1997;41:1495–1499.

68.  Haggoud A, Reysset G, Azeddoug H, et al. Nucleotide sequence analysis of two 5-nitroimidazole resistance determinants from Bacteroides strains and of a new insertion sequence upstream of the two genes. Antimicrob Agents Chemother 1994;38:1047–1051.

69.  Trinh S, Haggoud A, Reysset G, et al. Plasmids pIP419 and pIP421 from Bacteroides: 5-nitroimidazole resistance genes and their upstream insertion sequence elements. Microbiology 1995;141(Pt 4):927–935.

70.  Diniz CG, Farias LM, Carvalho MA, et al. Differential gene expression in a Bacteroides fragilis metronidazole-resistant mutant. J Antimicrob Chemother 2004;54(1):100–108.

71.  van der Wouden EJ, Thijs JC, Kusters JG, et al. Mechanism and clinical significance of metronidazole resistance in Helicobacter pylori. Scand J Gastroenterol Suppl 2001;10–14.

72.  Chisholm SA, Owen RJ. Mutations in Helicobacter pylori rdxA gene sequences may not contribute to metronidazole resistance. J Antimicrob Chemother 2003;51:995–999.

73.  Wayne LG, Sramek HA. Metronidazole is bactericidal to dormant cells of Mycobacterium tuberculosis. Antimicrob Agents Chemother 1994;38:2054–2058.

74.  Oh H, Hedberg M, Edlund C. Efflux-mediated fluoroquinolone resistance in the Bacteroides fragilis group. Anaerobe 2002;8:277–282.

75.  Onodera Y, Sato K. Molecular cloning of the gyrA and gyrB genes of Bacteroides fragilis encoding DNA gyrase. Antimicrob Agents Chemother 1999;43:2423–2429.

76.  Ricci V, Piddock L. Accumulation of garenoxacin by Bacteroides fragilis compared with that of five fluoroquinolones. J Antimicrob Chemother 2003;52:605–609.

77.  Schapiro JM, Gupta R, Stefansson E, et al. Isolation of metronidazole-resistant Bacteroides fragilis carrying the nimA nitroreductase gene from a patient in Washington State. J Clin Microbiol 2004;42(9):4127–4129.

78.  Britz ML, Wilkinson RG. Chloramphenicol acetyltransferase of Bacteroides fragilis. Antimicrob Agents Chemother 1978;14:105–111.

79.  Rosenblatt JE, Gustafson DR. Evaluation of the Etest for susceptibility testing of anaerobic bacteria. Diagn Microbiol Infect Dis 1995;22:279–284.

80.  Cormican MG, Erwin ME, Jones RN. False resistance to metronidazole by E-test among anaerobic bacteria investigations of contributing test conditions and medium quality. Diagn Microbiol Infect Dis 1996;24:117–119.