Determination of Susceptibilities of 26 Leptospira sp. Serovars to 24 Antimicrobial Agents by a Broth Microdilution Technique

Leptospirosis is a worldwide zoonotic infection that is predominantly found in tropical climates. The genus Leptospira consists of 16 species and over 250 serovars, with the virulent strains producing infections that range from clinically inapparent to life threatening. The typical symptomatic presentation of leptospirosis is a mild febrile illness without pathognomonic findings. Definitive diagnosis relies upon either culture which has a low yield and requires incubation of 4 to 6 weeks or serological testing of acute- and convalescent-phase samples. Therefore, initial therapy of leptospirosis is often empirical, based upon a broad differential diagnosis that includes leptospirosis and other etiologies of acute febrile illness in the community.

Clinical trials of antimicrobial agents against leptospirosis are limited to penicillin, doxycycline, and ceftriaxone (4, 7, 11, 15). Although no survival benefit has been shown with these agents, symptom improvement, including resolution of fever and leptospiruria, has been documented. The ever-growing armamentarium of antimicrobial agents used as empirical therapies for acute febrile illnesses necessitates the assessment of newer antimicrobial agents against leptospirosis. In vitro antibiotic susceptibility testing, typically done on a limited scale using macrodilution techniques and small numbers of serovars, has been reported (3, 6, 8, 10, 12-14). We have developed an in vitro broth microdilution technique by which a more efficient evaluation of a greater number of antimicrobial agents and Leptospira serovars can be accomplished (9). In this study we assess the in vitro activity of 24 antimicrobial agents against 26 Leptospira serovars.

(This work was presented in part at the 52nd Annual Meeting of the American Society of Tropical Medicine and Hygiene, Philadelphia, Penn., December 2003.)

Twenty-six Leptospira serovars, representing seven species and 18 serogroups, were obtained from the Veterinary Command Food and Drug Analysis Laboratory, Fort Sam Houston, Texas (Table 1). These strains originated at the USDA National Veterinary Services Laboratories, Ames, Iowa. Although the majority of these were initially recovered from human infections, all have been maintained by subculture as laboratory strains for many years. The organisms were maintained and the inocula were prepared as previously described (9).

Broth microdilution susceptibility testing was performed as previously reported with antibiotic concentrations ranging from 25.0 to 0.01 μg/ml (U/ml for penicillin G) (9). Amoxicillin, ampicillin, cefotaxime, chloramphenicol, doxycycline, erythromycin, penicillin G, and tetracycline were purchased from Sigma-Aldrich (St. Louis, Mo.). Other antibiotics were obtained from their manufacturers (cefdinir and clarithromycin from Abbott Laboratories, Abbott Park, Ill.; telithromycin from Aventis, Bridgewater, N.J.; ciprofloxacin and moxifloxacin from Bayer Corporation, West Haven, Conn.; aztreonam, cefepime, garenoxacin, and gatifloxacin from Bristol-Myers Squibb, Wallingford, Conn.; ceftriaxone from Hoffmann-La Roche Inc., Nutley, N.J.; cefoxitin, ertapenem, imipenem, and cilastatin from Merck & Co., Inc., Rahway, N.J.; levofloxacin from Ortho-McNeil Pharmaceutical, Inc., Raritan, N.J.; and ampicillin, sulbactam, and azithromycin from Pfizer, Groton, Conn.). Ampicillin-sulbactam and imipenem-cilastatin were combined in 2:1 and 1:1 ratios, respectively, with the reported MICs being based on the concentration of the first agent in each combination. Each serovar-drug combination was tested in triplicate, and median MICs were used for comparisons. Cumulative susceptibility results across all serovars are expressed as MIC₉₀, the concentration at which 90% of the Leptospira isolates are inhibited.

The median MICs from three experiments using the strains indicated in Table 1 are reported in Table 2. The observed reproducibility of drug-serovar combinations revealed that 84% (527 of 624) of test results fell within two dilutions of each other. The reproducibility increases to 94% if amoxicillin, ampicillin, and penicillin G are excluded from this comparison.

Ampicillin-sulbactam, the carbapenems, cefepime, cefotaxime, cefoxitin, ceftriaxone, the fluoroquinolones, the macrolides, and telithromycin all produced MIC₉₀s of ≤0.39 μg/ml. Cefepime, clarithromycin, erythromycin, imipenem-cilastatin, erythromycin, and telithromycin were all found to have MIC₉₀s below the limit of testing (≤0.01 μg/ml). The MIC₉₀s of amoxicillin, aztreonam, cefdinir, chloramphenicol, and penicillin G across serovars were ≥3.13 μg/ml.

Delays inherent in the diagnosis of leptospirosis, a potentially fatal infection, often dictate empirical therapy for acute febrile illness. As the differential diagnosis of an acute febrile illness in any particular setting may be expansive, numerous antimicrobial agents are often used as initial therapies. Only three drugs have been evaluated for leptospirosis in randomized human trials (4, 7, 11, 15). Determination of the antileptospiral activity of newer agents is needed. Currently there is no standard method to assess in vitro antimicrobial agents for antileptospiral activity; however, we have recently described a broth microdilution method which is more rapid than the traditional macrodilution method (9). Our assessment by broth microdilution of 24 antimicrobial agents against 26 Leptospira serovars found that many antimicrobial agents have excellent in vitro activity.

Included in our testing were the traditional antileptospiral drugs penicillin G, ceftriaxone, and doxycycline and representative agents from many other classes of agents. As classes of antimicrobial agents, carbapenems, cephalosporins (with the exception of cefdinir), fluoroquinolones, and macrolides (and telithromycin) had lower MICs than the traditional agents. All agents with the exception of aztreonam had MICs equal to or less than that of penicillin G, the accepted treatment alternative for severe disease. The overall reproducibility seen in our study was excellent, although amoxicillin, ampicillin, and penicillin G had greater variability than the other antimicrobial agents tested. Serovar-specific diagnosis is virtually never available for therapy of acute infection, so an assortment of Leptospira spp. were included in our study to allow observation of potential strain-to-strain and species-to-species variability. We noted no clinically significant strain-to-strain or species-to-species variability.

A variety of antimicrobials have shown variable therapeutic activity in small in vitro and in vivo (animal) studies; however, these studies are limited by the numbers of antimicrobials and/or isolates of Leptospira spp. examined at one time (1-3, 5, 6, 8, 10, 12-14, 16). Through our broth microdilution study, we delivered reproducible results and allowed more convenient testing with greater throughput than previous macrodilution methods. Presently correlation of in vitro susceptibility to outcome is lacking in treatment trials for leptospirosis. Future study, including animal and human treatment or prophylaxis trials, can now be pursued to determine if any of these agents will be useful in the therapy of leptospirosis.