In Vitro Activity of Levofloxacin against a Selected Group of Anaerobic Bacteria Isolated from Skin and Soft Tissue Infections

Abstract

The in vitro activity of levofloxacin was compared to the activities of ofloxacin, ciprofloxacin, ampicillin-sulbactam (2:1), cefoxitin, and metronidazole for a selected group of anaerobes (n = 175) isolated from skin and soft tissue infections by using the National Committee for Clinical Laboratory Standards-approved Wadsworth method. Ampicillin-sulbactam and cefoxitin inhibited 99% of the strains of this select group, levofloxacin and ofloxacin inhibited 73 and 50%, respectively, at 2 μg/ml, and ciprofloxacin inhibited 51% at 1 μg/ml. The geometric mean MIC of levofloxacin was lower than those of ofloxacin and ciprofloxacin for every group except Veillonella.

Fluoroquinolones have become important agents in the therapeutic arsenal for a wide range of infections. Their utility for anaerobic infections, however, is thought to be limited. Most of the quinolone agents introduced over the past several years, including ciprofloxacin, lomefloxacin, norfloxacin, pefloxacin, enoxacin, cinoxacin, and ofloxacin, have had only limited activity against anaerobes (3, 5, 8, 10, 11, 16, 17) and have had particularly poor activity against the Bacteroides fragilis group organisms. Variable activity against Prevotella species, such as the Prevotella melaninogenica and Prevotella oralis group, and against Bacteroides ureolyticus has been reported (10). Levofloxacin, the optically active l isomer of ofloxacin, is a tricyclic fluoroquinolone with a methyl group at the C-3 position in the oxazine ring. Levofloxacin is two- to fourfold more active than ofloxacin in vitro and possesses similar pharmacokinetic properties (13).

It can be useful to describe the antibacterial efficacy of agents against organisms found in a specific clinical setting. Documentation of the association of particular anaerobic species with specific infections, along with information on their antimicrobial susceptibility patterns, is helpful both to the laboratory scientist in identification protocols and to the clinician prescribing antimicrobial therapy. Rather than aiming at the entire anaerobic population, empiric therapy can be directed against the organism(s) most likely to be present. Our laboratory performed such a study on organisms isolated from skin and soft tissue infections. The purpose of this study was to evaluate the in vitro activity of levofloxacin and other agents against clinical isolates from these infections.

(Results of this study were presented at the 36th Interscience Conference on Antimicrobial Agents and Chemotherapy, 15 to 18 September 1996.)

The bacteriologic data from this study have been published in detail (14); the susceptibility data are presented here. Briefly, cutaneous or subcutaneous abscesses (86 specimens) of intravenous-drug users (IVDUs) were cultured and results were compared with those for abscesses (74 specimens) of patients with no history of intravenous drug use (non-IVDUs). The IVDU abscesses yielded 173 aerobes and 131 anaerobes. Specimens (n = 160) were primarily from upper extremity (n = 83, 52%), lower extremity (n = 36, 22%), and buttock and perirectal (n = 18, 11%) locations. Staphylococcus aureus was the most common aerobe isolated (50% of the specimens yielded this isolate), followed by “Streptococcus milleri” (46%). The commonly encountered anaerobes were Fusobacterium nucleatum (17% of the specimens yielded this isolate), pigmented Prevotella species (22%), Peptostreptococcus micros (17%), Actinomyces odontolyticus (15%), and Veillonella species (13%). The non-IVDU isolates included 116 aerobes and 106 anaerobes. S. aureus was isolated from 53% of these specimens, followed by coagulase-negative staphylococci (19%), “S. milleri” (19%), and Streptococcus pyogenes (16%). The main groups of anaerobes were Peptostreptococcus species (35% of the specimens yielded this isolate), Bacteroides species (19%), and gram-positive bacilli (31%). Overall, 67% of the IVDU isolates were of oral origin, compared with 25% of the non-IVDU isolates. Of the specimens from IVDUs and non-IVDUs, 48 and 67%, respectively, yielded only aerobes, and 2 and 4%, respectively, yielded only anaerobes. Sixty-four percent of the patients had one or more β-lactamase-producing organisms (14).

Anaerobic bacteria used in this study were isolated from the skin and soft tissue infections described above and were identified according to established procedures (9, 15). MICs were determined by an agar dilution technique described previously (15) by using an inoculum of 10⁵ CFU and brucella base-laked blood agar. Plates were incubated in an anaerobic chamber (Anaerobe Systems, San Jose, Calif.) for 48 h at 37°C. The MIC was defined as the lowest concentration of antimicrobial resulting in no growth, a haze, one discrete colony or multiple tiny colonies, or a marked change in the appearance of growth compared to the control plate (in the case of persistent light [slight] growth) (12). Reference strains of B. fragilis (ATCC 25285) and Bacteroides thetaiotaomicron (ATCC 29741) were used as controls in each test. Antimicrobial agents were obtained as powders from the indicated companies: ampicillin-sulbactam from Pfizer Pharmaceuticals, New York, N.Y.; cefoxitin from Merck Sharp and Dohme, Rahway, N.J.; ciprofloxacin from Miles Pharmaceuticals, West Haven, Conn.; ofloxacin and levofloxacin from R. W. Johnson, Raritan, N.J.; and metronidazole from Sigma, St. Louis, Mo.

National Committee for Clinical Laboratory Standards-approved breakpoints for anaerobic bacteria are established for ampicillin-sulbactam (8 μg/ml for the susceptible category, 16 μg/ml for the intermediate category), cefoxitin (16 μg/ml for the susceptible category, 32 μg/ml for the intermediate category), and metronidazole (8 μg/ml for the susceptible category, 16 μg/ml for the intermediate category).

Results are shown in Tables 1 and 2. Levofloxacin was the most active quinolone against B. fragilis, inhibiting 83% of the strains at 2 μg/ml (compared to 0% inhibition for both ciprofloxacin and ofloxacin). However, at 4 μg/ml, both levofloxacin and ofloxacin inhibited 100% of the strains. (The difference between the results at 2 μg/ml may be due to the 1-twofold-dilution error inherent in this technique.) Ciprofloxacin inhibited only 33% of the strains at 2 μg/ml. Levofloxacin was much more active than the other two quinolones against F. nucleatum (18 strains), inhibiting all of the tested strains at breakpoint compared to 22 and 17% of strains inhibited by ofloxacin and ciprofloxacin, respectively. Ofloxacin inhibited 100% of the strains at 4 μg/ml and ciprofloxacin inhibited 94% at 2 μg/ml. For all the anaerobes tested, levofloxacin was more active than ofloxacin at 2 μg/ml (73 versus 50% of strains inhibited) and at 4 μg/ml (81 versus 74%). Ciprofloxacin, in comparison, inhibited 51% of all anaerobes at 1 μg/ml. The MIC at which 90% of the isolates are inhibited (MIC₉₀) of levofloxacin was 1 twofold dilution lower than those of ciprofloxacin and ofloxacin (and the levofloxacin breakpoint for aerobes is 2 μg/ml while the ciprofloxacin breakpoint is 1 μg/ml). The geometric mean MIC of levofloxacin was lower than those of ofloxacin and ciprofloxacin for every group of anaerobes except Veillonella species (10) and ranged from 0.4 to 2.2 μg/ml. In general, the MICs of levofloxacin were 1 to 2 doubling dilutions lower than those of ofloxacin, a result that has been reported by other investigators as well (1).

TABLE 1.

Activities of antimicrobial agents against anaerobic organisms

Organism (no. of isolates) and antimicrobial agent	MIC (μg/ml)
	Range	50%	90%
B. fragilis (6)
Ampicillin-sulbactam	0.5–4	1
Cefoxitin	8–16	8
Ciprofloxacin	2–4	4
Levofloxacin	1–2	1
Metronidazole	0.5–1	0.5
Ofloxacin	2–2	2
Other B. fragilis group species (6)a
Ampicillin-sulbactam	0.5–16	1
Cefoxitin	2–16	8
Ciprofloxacin	4–16	8
Levofloxacin	1–8	1
Metronidazole	0.5–1	0.5
Ofloxacin	2–16	2
Other Bacteroides species (10)b
Ampicillin-sulbactam	0.25–0.25	0.25	0.25
Cefoxitin	1–1	1	1
Ciprofloxacin	0.12–8	2	4
Levofloxacin	0.12–4	2	2
Metronidazole	0.12–1	0.25	1
Ofloxacin	0.12–8	4	8
Clostridium species (10)c
Ampicillin-sulbactam	0.25–0.25	0.25	0.25
Cefoxitin	1–64	1	8
Ciprofloxacin	0.12–32	2	32
Levofloxacin	0.12–16	1	8
Metronidazole	0.12–1	0.12	0.5
Ofloxacin	0.25–>32	2	16
Peptostreptococcus species (37)d
Ampicillin-sulbactam	0.12–8	0.25	0.25
Cefoxitin	1–8	1	1
Ciprofloxacin	0.12–32	0.5	8
Levofloxacin	0.12–16	0.5	16
Metronidazole	0.12–2	0.25	1
Ofloxacin	0.25–32	0.5	16
Gram-positive (non-spore-forming) rods (35)e
Ampicillin-sulbactam	0.25–0.5	0.25	0.25
Cefoxitin	1–2	1	2
Ciprofloxacin	0.12–16	1	16
Levofloxacin	0.12–4	0.5	4
Metronidazole	0.12–>64	8	>64
Ofloxacin	0.12–8	1	8
F. nucleatum (18)
Ampicillin-sulbactam	0.25–4	0.25	0.25
Cefoxitin	1–1	1	1
Ciprofloxacin	0.5–4	2	2
Levofloxacin	0.5–1	1	1
Metronidazole	0.12–0.5	0.12	0.25
Ofloxacin	1–2	2	2
Prevotella species (43)f
Ampicillin-sulbactam	0.12–4	0.25	2
Cefoxitin	1–8	1	4
Ciprofloxacin	0.5–16	2	8
Levofloxacin	0.5–16	1	4
Metronidazole	0.12–2	0.5	1
Ofloxacin	1–32	1	8
Veillonella species (10)
Ampicillin-sulbactam	0.25–0.5	0.25	0.25
Cefoxitin	1–4	1	4
Ciprofloxacin	0.12–8	0.12	0.12
Levofloxacin	0.12–8	0.25	0.5
Metronidazole	0.25–2	1	2
Ofloxacin	0.25–16	0.5	1
Total (175)
Ampicillin-sulbactam	0.12–16	0.25	1
Cefoxitin	1–64	1	4
Ciprofloxacin	0.12–32	1	8
Levofloxacin	0.12–16	1	4
Metronidazole	0.12–>64	0.5	8
Ofloxacin	0.12–>32	1	8

^aIncludes B. distasonis (one isolate), B. ovatus (one isolate), B. thetaiotaomicron (two isolates), and B. vulgatus (two isolates). 

^bIncludes B. levii and B. ureolyticus (five isolates each). 

^cIncludes C. cadaveris (one isolate), C. innocuum (one isolate), C. malenominatum (one isolate), C. oroticum (one isolate), C. perfringens (one isolate), C. sordellii (three isolates), C. sporogenes (one isolate), and a Clostridium species (one isolate). 

^dIncludes P. anaerobius (3 isolates), P. asaccharolyticus (4 isolates), P. magnus (7 isolates), P. micros (18 isolates), P. prevotii (4 isolates), and a Peptostreptococcus species (1 isolate). 

^eIncludes Actinomyces odontolyticum (nine isolates), Actinomyces species (five isolates), Eubacterium limosum (one isolate), Eubacterium nodatum (one isolate), a Eubacterium species (three isolates), Lactobacillus jensenii (two isolates), Lactobacillus species (nine isolates), Propionibacterium acnes (four isolates), and a Propionibacterium species (one isolate). 

^fIncludes P. bivia (three isolates), P. buccae (seven isolates), P. denticola (two isolates), P. disiens (one isolate), P. intermedia (eight isolates), P. loescheii (eight isolates), P. melaninogenica (eight isolates), P. oralis (one isolate), P. oris (one isolate), and Prevotella species (four isolates). 

TABLE 2.

Geometric mean MICs of fluoroquinolone agents for anaerobic bacteria

Organism (no. of strains)	MIC (μg/ml) of:
	Levofloxacin	Ofloxacin	Ciprofloxacin
B. fragilis (6)	1.1	2	3.2
Other B. fragilis group species (6)	2.2	4.5	9
Other Bacteroides species (10)	0.7	1.4	0.98
Clostridium species (9)	1	1.5	1.4
F. nucleatum (18)	0.86	1.7	1.8
Peptostreptococcus species (37)	0.69	1.3	0.96
Gram-positive (non-spore-forming) rods (35)	0.69	1.7	1.5
Prevotella species (43)	1	1.9	1.7
Veillonella species (10)	0.4	0.76	0.18

There have been a limited number of studies of the in vitro efficacy of levofloxacin against anaerobes. In one study of clinical isolates, levofloxacin was effective against B. fragilis (15 strains) (MIC₉₀, 1 μg/ml) and Clostridium perfringens (5 strains) (MIC₉₀, 0.25 μg/ml); the MIC₉₀ was 4 μg/ml for Clostridium difficile (5 strains). The activities of ciprofloxacin and sparfloxacin were comparable in this study and the newer agents DU 6859 and trovafloxacin (CP 99,219) were more effective (2). In another study of fresh clinical isolates collected from laboratories around the United States from 1990 to 1991 the MIC₉₀ of levofloxacin was 2.0 μg/ml for B. fragilis (n = 39), compared to 8.0 μg/ml for ofloxacin and ciprofloxacin. The MIC₉₀s of the three agents for Peptostreptococcus were 8.0, 16.0 and 4.0 μg/ml, respectively (6).

Weighted mean MICs were calculated for several quinolones from multiple published studies (4); all isolates were from hospital or community sources, and at least three studies were used to calculate the mean MICs. The mean MICs of levofloxacin, ofloxacin, and ciprofloxacin, respectively (weighted to reflect the number of strains tested in the different studies), were as follows: for B. fragilis, 3.5, 6.5, and 14.9 μg/ml; for C. difficile, 5.0, 12.5, and 14.4 μg/ml; for C. perfringens, 0.75, 1.2, and 2.5 μg/ml; and for Peptostreptococcus, 4.6, 9.4, and 5.6 μg/ml (4). In a study of 194 aerobic and anaerobic strains isolated from bite wound infections (primarily oral or skin isolates) (7), levofloxacin was active against all aerobic isolates (MIC₉₀, <1.0 μg/ml). Among the anaerobes isolated, levofloxacin was active against 90% of Peptostreptococcus strains and against most of the Prevotella, Porphyromonas, and Bacteroides spp. (MIC₉₀, 4.0 μg/ml; no higher MICs were seen). Levofloxacin was less active against the fusobacteria (MIC₉₀, 64 μg/ml).

Quinolones with improved antianaerobic activity are being developed. The promising in vitro activity of these agents against certain groups of anaerobes will need confirmatory clinical data to support their use as therapeutic options against infections involving anaerobes. Levofloxacin has good activity against certain groups of anaerobic isolates (non-B. fragilis Bacteroides species, Veillonella species, Prevotella species, and Porphyromonas species), and its utility in specific anaerobic clinical infections warrants further study.