Comparison of spiral gradient endpoint and agar dilution methods for susceptibility testing of anaerobic bacteria: a multilaboratory collaborative evaluation

A multilaboratory collaborative study was carried out to assess the utility of the spiral gradient endpoint (SGE) method for the determination of the antimicrobial susceptibilities of anaerobes and to evaluate the equivalence of the MICs obtained by the SGE method with those obtained by the reference agar dilution method of the National Committee for Clinical Laboratory Standards. The standard deviation of the MIC obtained by the SGE method for the five participating laboratories was +/- 0.26 of a twofold dilution, whereas it was +/- 1 twofold dilution by the reference method. The interlaboratory reproducibility of the results for two control strains tested with imipenem, chloramphenicol, and metronidazole indicated that 96% of the measurements fell within +/- 1 twofold dilution of the mode. The equivalence of the SGE method with the agar dilution method was assessed with a wide variety of anaerobic organisms. The MICs by both methods were within 1 doubling dilution in 93% of the measurements (n = 1,074). Discrepancies generally occurred with those organism-drug combinations that resulted in tailing endpoints (Fusobacterium nucleatum, 86% agreement) or in cases of light growth (Peptostreptococcus spp., 86% agreement).

In vitro activities of two new glycylcyclines, N,N-dimethylglycylamido derivatives of minocycline and 6-demethyl-6-deoxytetracycline, against 339 strains of anaerobic bacteria

The in vitro activities of the N,N-dimethylglycylamido derivatives of minocycline (DMG-MINO) and 6-demethyl-6-deoxytetracycline (DMG-DMDOT) were compared with those of minocycline, tetracycline, clindamycin, and metronidazole by using the National Committee for Clinical Laboratory Standards-approved Wadsworth agar dilution method. The MICs of DMG-MINO, DMG-DMDOT, and metronidazole at which 90% of the strains were susceptible (0.5, 1, and 1 micrograms/ml, respectively) were lower than those for clindamycin, minocycline, and tetracycline (4, 8, and 32 micrograms/ml, respectively). All of the strains of anaerobes tested, except one strain of Bacteroides ovatus (MIC, 16 micrograms/ml), were susceptible to DMG-MINO and DMG-DMDOT at 8 micrograms/ml.

Preliminary study using species-specific oligonucleotide probe for rRNA of Bilophila wadsworthia

Portions of the 16S RNA from a urease-positive Bilophila wadsworthia strain were sequenced, and a probe was constructed. The probe was end labeled with [32P]ATP and polynucleotide kinase and hybridized on a nylon filter (by dot blot hybridization) to the immobilized rRNA of 12 B. wadsworthia strains and eight other anaerobic isolates. The probe efficiently hybridized only to the Bilophila strains. Cross-reactivity at high RNA levels (2,000 ng) was observed with one strain of Bacteroides thetaiotamicron and one strain of Bacteroides fragilis (with 10x SET buffer [20x SET buffer is 0.5 M NaCl, 0.03 M Tris, and 2 mM EDTA]) but was not seen at lower RNA levels or with 5x SET buffer. When tested against mixed cultures of aerobic and anaerobic isolates representative of appendiceal abscess flora, the probe did not react with mixed cultures containing no Bilophila cells and could detect > or = 10(5) Bilophila CFU/ml when the mixture was seeded with Bilophila cells. This probe is of potential use in the rapid identification of pure isolates and in the direct identification of B. wadsworthia in clinical specimens.

In-vitro activity of clinafloxacin (CI-960) and PD 131628-2 against anaerobic bacteria

The antimicrobial activities of two new quinolones, CI-960 and PD 131628-2 were determined against 339 strains of anaerobic bacteria and compared to cefoxitin, imipenem and metronidazole. The NCCLS-approved Wadsworth agar dilution technique with Brucella-lysed blood agar was used throughout the study. Breakpoints of the new quinolones are 2 mg/L, and breakpoints for cefoxitin, imipenem and metronidazole are 32, 8 and 16 mg/L, respectively. CI-960 displayed excellent activity, inhibiting all strains tested at 1 mg/L. PD 131628-2 inhibited 94% of Bacteroides fragilis, 75% of other B. fragilis group isolates, 87% of Prevotella spp, 79% of the Fusobacterium mortiferum-varium group, 74% of non-sporing gram-positive bacilli, and 89-100% of Clostridium spp other than Clostridium difficile at 2 mg/L. None of the eight strains of C. difficile was inhibited at 2 mg/L although they were inhibited at 4 mg/L. PD 131628-2 inhibited all strains of other Bacteroides spp, Porphyromonas spp, and Fusobacterium nucleatum at < or = 1 mg/L.

In vitro activity of DU-6859a against anaerobic bacteria

The activity of a new quinolone agent, DU-6859a, against 330 strains of anaerobic bacteria was determined by using the National Committee for Clinical Laboratory Standards-approved Wadsworth brucella laked blood agar method; the activity of DU-6859a was compared with those of amoxicillin-clavulanate (2:1), chloramphenicol, ciprofloxacin, clindamycin, fleroxacin, imipenem, lomefloxacin, metronidazole, sparfloxacin, and temafloxacin. DU-6859a and chloramphenicol inhibited all of the isolates at concentrations of 1 and 16 micrograms/ml, respectively; amoxicillin-clavulanate, imipenem, and metronidazole inhibited > or = 94% of the isolates at their respective breakpoints (8, 8, and 16 micrograms/ml). MICs of DU-6859a at which 90% of the strains were susceptible were 1 to 5 twofold dilutions lower than those of the other quinolones for every group of organisms. MICs of DU-6859a at which 90% of the strains were susceptible (total numbers of strains tested are in parentheses) were < or = 0.25 micrograms/ml for Bacteroides fragilis (57), other B. fragilis group species (84), Bilophila wadsworthia (15), Clostridium species (27) (including C. difficile, C. perfringens, and C. ramosum), Fusobacterium nucleatum (16), Fusobacterium mortiferum-F. varium group species (10), Peptostreptococcus species (20), non-spore-forming gram-positive rods (20), and Prevotella species (25).

Inactivation of anaerobic bacteria by various photosensitized porphyrins or by hemin

The photodynamic effects of deuteroporphyrin (DP), hematoporphyrin derivative (HPD), hematoporphyrin (HP), or protoporphyrin (PP) on a variety of anaerobic microorganisms were examined in this study. The majority of the species, among the 350 strains tested, were inhibited by concentrations of < or = 2.5 micrograms/ml of light-activated DP. Species found to be resistant to this treatment included Bilophila wadsworthia, Fusobacterium mortiferum, Fusobacterium varium, and Bacteroides gracilis. These species were inhibited by concentrations of > 60 micrograms/ml of DP. The porphyrin-producing species, Porphyromonas and Prevotella spp, were all inhibited by < or = 2.5 micrograms/ml DP and light. Comparing the photodynamic activity of the porphyrins used on Porphyromonas strains resulted in the following pattern: DP > HPD > HP > PP. Porphyromonas spp., Gram-positive cocci, and many Gram-positive rods (excluding clostridia) were inactivated by hemin (a metal-containing porphyrin) at 10-20 micrograms/ml. Hemin inhibitory action was not affected by light. Binding and insertion of DP into bacteria (both inactivated and non-inactivated strains by DP and light) were monitored by the characteristic fluorescence band of bound DP at 622 nm. Porphyromonas spp. bound DP tightly, whereas only low binding was seen with B. wadsworthia and other DP-resistant species. High binding of DP to B. wadsworthia can be achieved by pretreatment of the bacteria with imipenem or cefoxitin, beta-lactam agents known to interfere with the integrity of the cell wall. If cell wall integrity is disturbed (e.g., by these agents), inactivation of B. wadsworthia by DP can occur.

In vitro activity of grepafloxacin (OPC-17116) against anaerobic bacteria

In vitro activity of the quinolone grepafloxacin (OPC-17116) was compared with that of ciprofloxacin, fleroxacin, clindamycin, imipenem, and metronidazole by using the NCCLS-approved Brucella-base-laked blood agar dilution method and breakpoints, when available. Clindamycin, metronidazole, and imipenem inhibited > or = 98% of Bacteroides fragilis at the breakpoint; grepafloxacin, ciprofloxacin, and fleroxacin inhibited 83%, 6%, and 0, respectively, at 2 micrograms/ml. Grepafloxacin inhibited 39% of other B. fragilis group species isolated (80) at breakpoint (< or = 2 micrograms/ml) compared with 100% for metronidazole and imipenem, 83% for clindamycin, 6% for ciprofloxacin, and 1% for fleroxacin. Grepafloxacin demonstrated substantially better activity against B. fragilis than did ciprofloxacin or fleroxacin; overall activity against anaerobes was marginally better than that of ciprofloxacin or fleroxacin.

Morphological response of Bilophila wadsworthia to imipenem: correlation with properties of penicillin-binding proteins

The penicillin-binding protein (PBP) patterns of six strains of Bilophila wadsworthia were investigated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis and subsequent fluorography of membrane preparations labelled with [3H]benzylpenicillin. The PBP profiles among the strains were similar; generally, seven to nine PBP-reactive bands could be visualized; their molecular weights ranged from 31 to 137 kDa. The relative affinities of the PBPs of four strains of B. wadsworthia for imipenem were examined and correlated with the morphological responses of the cells to imipenem. Morphological changes were examined by light and electron microscopies. Light microscopy revealed that at low concentrations (less than the MIC), imipenem induced the formation of rounded and bulging cells; rarely, elongation without filamentation was observed. In the presence of imipenem at the MIC, spheroplast formation was observed. Scanning and transmission electron microscopies revealed round forms together with larger, multilobate cells in the presence of subinhibitory concentrations of imipenem, suggesting that new growth sites were initiated while cell division was inhibited. Peeling of the outer membrane was also seen. Spheroplasts were very large (up to 30 microns in diameter) and stable in aqueous solution. Inhibition of the PBPs could be seen in the presence of low imipenem concentrations.

In vitro activity of Bay Y3118 against anaerobic bacteria

The antimicrobial activity of a new quinolone, Bay Y3118, was determined against 326 strains of anaerobic bacteria and compared with the activities of ampicillin-sulbactam, cefotetan, clindamycin, imipenem, metronidazole, and sparfloxacin. The National Committee for Clinical Laboratory Standards-approved Wadsworth agar dilution technique with Brucella-laked blood agar was used throughout the study. Breakpoints used to determine the percent susceptible were 2 micrograms/ml for Bay Y3118 and sparfloxacin, 4 micrograms/ml for clindamycin, 8 micrograms/ml for imipenem, 16 micrograms/ml for metronidazole and ampicillin-sulbactam, and 32 micrograms/ml for cefotetan. Species tested included Bacteroides fragilis (57 strains), other B. fragilis group species (79 strains), Bacteroides gracilis (10 strains), other Bacteroides spp. (9 strains), Prevotella spp. (30 strains), Porphyromonas spp. (9 strains), Fusobacterium spp. (36 strains), Bilophila wadsworthia (14 strains), Clostridium spp. (36 strains), Peptostreptococcus spp. (20 strains), and gram-positive non-spore-forming rods (26 strains). Bay Y3118 inhibited all but 1 of 326 anaerobic bacteria tested at the breakpoint level or lower.

Susceptibility testing of anaerobic bacteria--the state of the art

Demand for susceptibility testing of anaerobes has increased, but no consensus on procedure and interpretation has been achieved. The need for reliable methods for testing anaerobic bacteria extends from small hospital laboratories to large research centers. Agar dilution testing is too costly and labor intensive for many clinical laboratories. Microbroth dilution is more convenient; however, some fastidious anaerobes do not grow well enough in this system, and the choice of antimicrobial agents may not reflect the hospital formulary. Disapproval of the broth disk elution system leaves fewer options open to clinical laboratories and emphasizes the need for more convenient and reliable techniques. Some newer methods are undergoing evaluation. Variables in susceptibility testing of anaerobes include the media and methods used, organisms chosen, breakpoints chosen, and endpoint determination. This latter variable is probably the most problematic since no endpoint based on interaction of organism and antimicrobial agent rather than on subjective observation has been defined. Also, clustering of MICs around the breakpoint may lead to significant variability in reported methods. A more accurate MIC measurement is needed. Adherence of laboratories to approved methods and careful reporting of methods and the interpretive breakpoints would facilitate interlaboratory comparisons and the identification of emerging resistance.

Current perspectives on anaerobic infections: diagnostic approaches

This article discusses when to look for anaerobes, anaerobic infections as clues to other problems in patients, and underlying clinical conditions as clues to the nature of anaerobic infections. Diagnostic approaches, identification methods, and susceptibility testing are reviewed.

Methods for testing the susceptibility of anaerobic bacteria to two fluoroquinolone compounds, PD 131628 and clinafloxacin

The susceptibility of anaerobic bacteria to two new fluoroquinolones, PD 131628 (the bioactive form of PD 131112 or CI-990) and clinafloxacin (CI-960 or PD 127391), was determined with the agar dilution procedures and two media, and one broth microdilution procedure. Sparfloxacin and ciprofloxacin were also tested by the broth microdilution method. One hundred anaerobic isolates and four control strains were tested by the three methods which gave minimum inhibitory concentrations (MICs) that were essentially comparable, but not identical. With the broth microdilution method, the relative potency of the four fluoroquinolones was: clinafloxacin > PD 131628 > sparfloxacin > ciprofloxacin. For the latter three drugs but not clinafloxacin, the MIC values were often near the proposed interpretive breakpoint concentrations, and thus minor methodological differences frequently influenced the interpretive categories. Replicate agar dilution tests in five laboratories established MIC control limits for each of four control strains: those MIC limits could also be used to define the expected performance of the two alternative methods.

The isolation and characterisation of a major outer-membrane protein from Bacteroides distasonis

An outer-membrane protein (OMP) was isolated from a clinical strain of Bacteroides distasonis. Changes in growth media did not appreciably affect the appearance of this protein in crude outer-membrane preparations examined by SDS-PAGE. However, the proportion of the protein relative to other OMPs was greater in 24-h cultures than in 48-h cultures. The protein could not be readily solubilised by various conventional detergent extraction techniques but treatment of the insoluble material at 100 degrees C with SDS released the protein, as did overnight extraction at 37 degrees C with SDS. This OMP was heat-modifiable, and thus was similar to the OmpA protein of Escherichia coli, with a faster mobility on SDS-PAGE when solubilised at 25 degrees C than at 100 degrees C. The critical temperature for conversion was between 80 degrees C and 90 degrees C. Because of the characteristic heat-modifiability, the protein was called B. distasonis HMP-1 (heat modifiable protein-1). Overnight exposure to EDTA or NaCl at 37 degrees C favoured conversion of the 25 degrees C form to the 100 degrees C form. In intact cells, the protein was labelled by a cell-surface radio-iodination procedure, and thus is at least partially exposed at the cell surface. No reactions between the B. distasonis HMP-1 and antibodies to either E. coli OmpA or E. coli porin were found by Western blot analysis. A B. distasonis OM preparation containing predominantly HMP-1 had pore-forming ability in a liposome assay. This study is the first report of the isolation and characterisation of a heat-modifiable OMP in Bacteroides, and it is the first description of pore-forming activity in a Bacteroides OM fraction.

Antimicrobial susceptibility testing of Bilophila wadsworthia by using triphenyltetrazolium chloride to facilitate endpoint determination

Initial susceptibility studies of Bilophila wadsworthia indicated significant resistance to several beta-lactam antibiotics, including imipenem and cefoxitin. NO beta-lactamase production was detected. However, some B. wadsworthia strains may grow as a heavy "haze" at up to the highest concentration of an antibiotic on standard antimicrobial agent-containing plates, and it is often difficult to determine the point at which conventional growth stops and haze growth begins. We investigated the nature of the haze growth of B. wadsworthia by using triphenyltetrazolium chloride (TTC) as an indicator of viability during antimicrobial susceptibility testing, by determining viability counts on antimicrobial agent-containing plates at various times, and by microscopically inspecting stained preparations of the growth on the control plate and the haze area. TTC MICs were determined by applying a TTC solution over the growth on plates inside the anaerobic chamber or within 5 min after exposure to air (aerobic TTC MICs). The haze growth reduced TTC in the chamber but not under aerobic conditions, whereas TTC was reduced by the conventional growth in both atmospheres. The aerobic TTC MICs correlated with the viability counts. Separated proteins resolved by polyacrylamide gel electrophoresis and isoelectric focusing showed TTC-reactive bands only when stained under anaerobic conditions, further demonstrating the sensitivity of TTC reduction to aerobic conditions. Microscopic examination of the haze growth indicated spheroplast formation. A new antibiogram for B. wadsworthia has been established by use of aerobic TTC endpoints; we believe that the lower MICs obtained with the TTC method are likely the ones that are clinically relevant and should be used in tests of B. wadsworthia. Also, we found that when the organisms were grown on pyruvate-containing medium, 87% of 56 strains tested were Beta-lactamase positive.

In vitro activities of three of the newer quinolones against anaerobic bacteria

The antimicrobial activities of three new quinolone compounds, sparfloxacin, temafloxacin, and WIN 57273, against anaerobic bacteria were determined in three separate studies. The Wadsworth agar dilution technique using brucella-laked blood agar was used throughout. The activities of other antimicrobial agents, including ciprofloxacin, imipenem, chloramphenicol, metronidazole, cefotetan, cefoxitin, and amoxicillin-clavulanic acid, were also determined. The breakpoints of the new quinolones were 2 micrograms/ml for sparfloxacin and WIN 57273 and 4 micrograms/ml for temafloxacin. WIN 57273 displayed very good activity against anaerobes, inhibiting all strains of Bacteroides fragilis group species at 2 micrograms/ml. Only two strains of Fusobacterium species were resistant (MIC, 4 micrograms/ml). Sparfloxacin inhibited 78% of B. fragilis strains and 44% of other B. fragilis group isolates at 2 micrograms/ml. At 2 micrograms/ml, the percentages of other anaerobic species susceptible were as follows: B. gracilis, 70%; other Bacteroides species, 61%; Clostridium species, 50%; Fusobacterium species, 70%; Peptostreptococcus species, 91%; non-spore-forming gram-positive rods, 71%. Temafloxacin inhibited 91% of B. fragilis strains and 87% of other B. fragilis group species at 4 micrograms/ml. All strains of other Bacteroides species, 78% of Fusobacterium species, 80% of Clostridium species, and 90% of Peptostreptococcus species were inhibited at 4 micrograms of temafloxacin per ml.

In-vitro activity of temafloxacin against anaerobic bacteria: a comparative study

The antimicrobial activity of temafloxacin for 328 anaerobic bacteria was determined and compared to that of cefotetan, cefoxitin, ciprofloxacin and metronidazole. The Wadsworth agar dilution technique using Brucella-lysed sheep blood agar was used throughout. At the recommended breakpoint concentration 4 mg/L, temafloxacin inhibited 60/62 (97%) of the isolates of Bacteroides fragilis and 82/87 (94%) of the isolates of other species of the B. fragilis group. Ninety-six percent of the 24 isolates of Peptostreptococcus, 97% of the 31 isolates of other Bacteroides species and 88% of 66 isolates of Fusobacterium species were also inhibited by 4 mg/L temafloxacin. Metronidazole (breakpoint 16 mg/L) had a broader spectrum of activity than temafloxacin (judged by the percentage of strains tested susceptible at the breakpoints employed) with the exception of non-sporing Gram-positive bacilli. The cephalosporins tested (breakpoint 32 mg/L) had a narrower spectrum of activity. Ciprofloxacin (breakpoint 2 mg/L) was the least effective agent against the majority of the anaerobes tested.

Susceptibility testing of anaerobic bacteria: myth, magic, or method?

The demand for susceptibility testing of anaerobes has increased, yet consensus as to procedure and interpretation in this area has not been achieved. While routine testing of anaerobic isolates is not needed, certain isolates in specific clinical settings should be tested. Also, laboratories may monitor their local antibiograms by doing periodic surveillance batch testing. The National Committee for Clinical Laboratory Standards has published a protocol of methods approved for susceptibility testing of anaerobic bacteria. Both agar and broth microdilution are included; however, the broth disk elution method is no longer approved by the National Committee for Clinical Laboratory Standards because of method-related interpretive errors. A number of newer methods are undergoing evaluation and seem promising. Clinicians and microbiologists reviewing susceptibility reports should be aware of sources of variability in the test results. Variables in susceptibility testing of anaerobes include the media and methods used, organisms chosen for testing, breakpoints chosen for interpretation, antibiotic, and determination of endpoint. Clustering of MICs around the breakpoint may lead to significant variability in test results. Adherence of testing laboratories to approved methods and careful descriptions of the method and the breakpoints used for interpretation would facilitate interlaboratory comparisons and allow problems of emerging resistance to be noted. A variety of resistance mechanisms occurs in anaerobic bacteria, including the production of beta-lactamase and other drug-inactivating enzymes, alteration of target proteins, and inability of the drug to penetrate the bacterial wall. Antimicrobial resistance patterns in the United States and abroad are described.

Comparison of spiral gradient and conventional agar dilution for susceptibility testing of anaerobic bacteria

Antimicrobial susceptibility tests were performed on brucella laked blood agar with 340 isolates and 14 antimicrobial agents by the standard agar dilution technique and the spiral gradient technique in which antibiotic concentrations were established by diffusion from the agar surface. For comparison, spiral gradient MICs were determined by calculating antimicrobial concentrations at growth endpoints and rounding up to the next twofold incremental concentration. The cumulative percentage of strains susceptible at the breakpoint determined from spiral gradient data was within 10%, generally, of the percentage of strains susceptible at the breakpoint determined from agar dilution data. The overall agreement between the two techniques (within one doubling dilution) was 90.6%. The spiral gradient agar dilution technique is a reasonable alternative to the conventional agar dilution technique for susceptibility testing of anaerobic bacteria.

Effect of beta-lactamase inhibitors on the activities of various beta-lactam agents against anaerobic bacteria

The in vitro activities of several new beta-lactam-beta-lactamase inhibitor combinations (piperacillin plus tazobactam, ceftizoxime and cefonicid with sulbactam and clavulanic acid, and ampicillin plus 8 micrograms of sulbactam per ml) were tested with anaerobic bacteria and compared with known beta-lactam-beta-lactamase inhibitor combinations and other potent antianaerobe agents. All the combinations tested (except for the cefonicid-inhibitor combinations) were active against almost all strains of the Bacteroides fragilis group. This report indicates that beta-lactamase inhibitors may improve the activity of beta-lactam agents with marginal activity against the B. fragilis group.

Media- and method-dependent variation in MIC values of ceftizoxime for clinical isolates of the Bacteroides fragilis group

Although ceftizoxime has been used effectively in several clinical trials for infections involving anaerobic bacteria, reports of its in-vitro activity against anaerobes are contradictory and confusing. In an effort to clarify the discrepant reports, we tested 90 strains of Bacteroides fragilis group organisms from patients with perforated or gangrenous appendicitis using eight different susceptibility testing procedures. The minimal inhibitory concentration values were dependent on the technique used; agar dilution values were often four twofold dilutions higher than microbroth dilution values. Agar techniques (including spiral gradient end point) gave values of 36% to 61% susceptible at breakpoint (depending on the technique), while the microbroth dilution techniques gave values of 84% to 92% susceptible. When a 64 micrograms/ml breakpoint for agar dilution testing was used, the methods were more comparable, with the agar methods giving values of 64% to 83% susceptible. The results of the broth disk elution procedure were difficult to read and often did not agree with other values.

Statistical analysis of the effects of trial, reader, and replicates on MIC determination for cefoxitin

A pilot study was designed to estimate the variance components in the determination of the MIC of cefoxitin for isolates of the Bacteroides fragilis group. Twenty different organisms were tested, and replicate, trial, and reader variabilities were examined. When the total-variance component was used, if the true MIC was 16 micrograms/ml, then the chance that the observed MIC was between 8 and 32 micrograms/ml, inclusive, was 95%. For all analyses, the isolate (P = 0.0001) and reader (P less than 0.03) effects were significant. The probability of specific MIC observations for various true MICs (over the range of 16 to 32 micrograms/ml at 4-micrograms/ml increments) was calculated. For true MICs of 20, 24, and 28 micrograms/ml, the probabilities of observing an MIC of 16 or 32 micrograms/ml (inclusive) were 86, 75, and 62%, respectively. An upward bias was shown to exist in addition to sources of sizeable variation. The recommendation stemming from recognition of this inherent variability is that ranges of percent susceptibility at various concentrations be included in reports of in vitro susceptibility studies.

Gangrenous and/or perforated appendix: clinical outcome and in vitro susceptibility testing

The data from this study indicate that cefoxitin was effective and generally well tolerated in the management of gangrenous and/or perforated appendicitis. No strong correlation was identified between in vitro susceptibility testing results and clinical outcome.

Effects of differences in media and methods on MIC values for anaerobes

In vitro susceptibility of Bacteroides fragilis isolates to cefoxitin at the breakpoint of 32 micrograms/ml varied little when tested by eight different methods. We found that broth disk elution results were difficult to read and interpret and did not correlate with other methods.

Alterations to the penicillin-binding proteins in the Bacteroides fragilis group: a mechanism for non-beta-lactamase mediated cefoxitin resistance

The penicillin-binding proteins (PBPs) of ATTC Type Strains of nine species of the Bacteroides fragilis group were visualized by gel electrophoresis and subsequent fluorography. Each species had a distinctive PBP pattern, although variation within species was seen. Generally, five PBPs could be visualized, ranging in molecular weight from approximately 40,000 to approximately 90,000. A laboratory-derived cefoxitin-resistant mutant of B. distasonis was compared with its wild type parent and cefoxitin-sensitive revertant. The fluorograph of the resistant mutant indicated a marked reduction of labelling to the PBP-1 complex as compared with the wild type and revertant. Cefoxitin-resistant clinical isolates of B. thetaiotaomicron and B. uniformis also showed changes to the PBP-1 complex, in comparison with sensitive strains.

Comparison of in vitro antibiograms of Bacteroides fragilis group isolates: differences in resistance rates in two institutions because of differences in susceptibility testing methodology

With 120 clinical isolates of the Bacteroides fragilis group, a comparison of rates of resistance to selected antimicrobial agents by using two susceptibility tests was performed in two medical institutions. The broth microdilution method produced MICs significantly lower than those determined by the agar dilution method. With ceftizoxime and cefoxitin, 88 and 18%, respectively, of the MICs were greater than or equal to 2 twofold dilutions apart. These differences in MIC results produced major interpretive discrepancies for ceftizoxime and cefoxitin, whereas no significant differences in resistance rates were noted for clindamycin and metronidazole.

Effect of inoculum size and medium on activity of seven antimicrobial agents against Bacteroides fragilis strains

Antimicrobial activity of cefoxitin, cefotetan, cefotaxime, ceftizoxime, ceftazidime, imipenem, and clindamycin against four inocula of Bacteroides fragilis strains was determined on three different media. The inoculum sizes were 10(4), 10(5), 10(6), and 10(7) colony forming units (CFU) per spot. On all three media, substantial effects of inoculum size on minimal inhibitory concentrations (MIC) of cefotaxime and ceftizoxime were found: the doubled dilution differences in MICs between inocula of 10(7) and 10(4) CFU/spot were 2.2, 2.3, and 2.1 micrograms/ml of cefotaxime and 1.8, 4.4, and 4.0 micrograms/ml of ceftizoxime on Brucella base-laked blood agar, Wilkins-Chalgren agar, and a brain-heart infusion medium, respectively. An inoculum difference found on all three media with ceftazidime may also be of practical significance. There was evidence of larger differences between inocula on the Wilkins-Chalgren agar and brain-heart infusion than on the Brucella agar.

Cell-wall-defective variants of Fusobacterium

The activity of antimicrobial agents against Fusobacterium species has been reported as variable in the literature. For some strains, the inconsistency arises from difficulty in determining the endpoint of growth in agar dilution susceptibility tests. Certain strains persist as a subtle haze beyond the levels of antibiotic that permit conventional colonial growth. We have determined by light and electron microscopy that this haze represents the colonial growth of cell-wall-defective (CWD) variants of the parent Fusobacterium. The CWD forms could be propagated indefinitely in hypertonic medium containing the antibiotic inducing agent. However, when the antibiotic was eliminated, the organisms would revert to their native morphology. Formation of CWD variants was observed in the presence of cell-wall-active drugs (e.g., beta-lactam agents) but not with drugs that work by a different mechanism (e.g., clindamycin or chloramphenicol). Fourteen of 22 F. varium strains, 8 of 11 F. mortiferum strains, 2 of 10 F. gonidiaformans strains, and 1 of 4 of F. necrophorum strains could be induced to a CWD form in vitro in the usual agar dilution susceptibility test. Although the clinical significance of CWD variants of Fusobacterium is unknown, they may be a source of confusion in interpreting agar dilution susceptibility tests.

In vitro activity of cefotetan compared with that of other antimicrobial agents against anaerobic bacteria

The activity of cefotetan against 430 strains of anaerobic bacteria was compared with that of cefoxitin, ceftizoxime, clindamycin, metronidazole, and chloramphenicol. Percent susceptible values for the Bacteroides fragilis group were 60, 80, 29, 86, 100, and 100%, respectively. Percent susceptible values for the B. fragilis species were 91, 92, 46, 98, 100, and 100%, respectively. Non-B. fragilis-group Bacteroides species were inhibited very well (90 to 100%) by all drugs except ceftizoxime (80%). Cefotetan and metronidazole were the most active agents against Clostridium difficile. Percent susceptible values for all strains were 72, 79, 44, 82, 93, and 98%, respectively.

In vitro activity of cefoperazone plus sulbactam compared with that of other antimicrobial agents against anaerobic bacteria

The activity of two cefoperazone-sulbactam combinations against anaerobic bacteria was tested and compared both with that of cefoperazone alone and with that of other commonly used antimicrobial agents. Imipenem was the most active of the tested agents, followed by chloramphenicol, metronidazole, and cefoperazone-sulbactam (90 to 100% of bacterial growth inhibited). Clindamycin and cefoxitin inhibited approximately 80%, cefoperazone inhibited 63%, and penicillin G inhibited 47% of the strains tested. The agents were variable in activity against the Bacteroides fragilis group, with percents susceptible as follows: cefoperazone-sulbactam, imipenem, metronidazole, and chloramphenicol, 99 to 100%; cefoxitin and clindamycin, approximately 80%; cefoperazone, 49%; and penicillin G, 15.5%.

Six-year retrospective survey of the resistance of Bacteroides fragilis group species to clindamycin and cefoxitin

Two hundred and forty-six strains of the Bacteroides fragilis group, all clinical isolates, collected at the Wadsworth Veterans Administration Medical Center from 1977 to 1982, were tested for susceptibility to clindamycin and cefoxitin. There was no significant change in resistance to either clindamycin or cefoxitin over the time period tested for any individual species, nor for the B. fragilis group in toto. Striking differences in susceptibility to the two drugs were seen among species of the B. fragilis group. B. fragilis displayed resistance to cefoxitin (32 micrograms/ml) and clindamycin (8 micrograms/ml) of 0.0% and 0.8%, respectively, whereas B. thetaiotaomicron showed resistances of 12.7% to cefoxitin (32 micrograms/ml) and 9% to clindamycin (8 micrograms/ml). B. thetaiotaomicron, B. distasonis, and B. ovatus are distinctly more resistant to cefoxitin than B. fragilis and B. vulgatus. Similarly B. thetaiotaomicron and B. distasonis are much more resistant to clindamycin than are the other B. fragilis group species. It is apparent that determination of species within the B. fragilis group is important in evaluating a potential therapeutic regimen.

In vitro activity of imipenem against anaerobic bacteria

The in vitro activity of imipenem, metronidazole, clindamycin, moxalactam, and cefoxitin against 203 strains of anaerobic bacteria isolated from patients at the Veterans Administration Wadsworth Medical Center in Los Angeles was studied. Imipenem and metronidazole were the most active agents overall, inhibiting 98% and 99%, respectively, of all anaerobes tested. At breakpoint levels all of the agents tested were very active against anaerobic cocci. Clostridium perfringens, and Bacteroides species other than those of the Bacteroides fragilis group. Imipenem, metronidazole, and clindamycin were the most active agents against the B. fragilis group in this study, although more recent experience with clindamycin indicates less potency. Marked variation among the susceptibility results obtained at various centers may be due to differences in technique, including inoculum size, media, and incubation time. In all instances, however, imipenem has clearly been the most active of the beta-lactam agents.

Impact of imipenem/cilastatin therapy on normal fecal flora

The impact of parenteral imipenem/cilastatin therapy on the bowel flora of six patients was evaluated. Stool samples were collected before and during therapy and qualitative and quantitative bacteriologic studies were performed. Imipenem had no effect on total microorganism counts. Two patients acquired Candida albicans during therapy, and three patients acquired Proteus species. Pseudomonas species in one patient acquired resistance. Imipenem appears to have a relatively modest effect on the bowel flora and apparently does not readily induce resistance in the resident flora as compared with other agents.