1 4 7 Let te rs t o t h e Ed i to rs

[9]. All of the above reasons led us to focus on the toxin B gene in order to detect all possible genetic situations in toxigenic strains. In contrast to some other published genetic methods based on hybridizations or single amplifications, we have designed a nested PCR which has allowed us to achieve 100% sensitivity. Specificity has been optimal, since no amplifications were detected, either on other species of Clostridium or in the entero- pathogens included in the study. In other published techniques, C. sordellii, whose toxin is highly homo- logous with the one produced by C. d @ d e [l], has offered false-positive results by cross-reactions. In our case a positive amplification band has never been obtained within this microorganism.

All strains included in this study were classified identically by both techniques, with the exception of strain no. 110, which was PCR negative and a toxin producer according to the cytotoxicity assay. Never- theless, its cytopathic effect on human fibroblasts is atypical and it was not inhibited by toxin B antitoxin, so it could have been the result of a cause unrelated to C. df ic i le toxin B production. This finding still remains controversial; nevertheless, strain 110 was considered as non-cytotoxic and PCR negative when calculating sensitivities, predictive values, and false result rates. The 1125-kb band corresponding to the first amplification reaction was occasionally seen. Probably, the amounts of this PCR product are at the limit of detection by agarose gel electrophoresis and so it is observed occasionally in cases where the reaction yield has been optimal. Usually in these cases the amount of the main amplification product (322-bp band) is slightly reduced, probably due to inhibition of the nested amplification reaction by an excess of DNA template in it [lo]. All our studied isolates of C. d@ale, with the exception of C. d@cile ATCC 9689, were isolated fiom significant clinical specimens. This may explain the high rate of toxigenicity encountered (over 93% of all C. df ic i le clinical isolates included).

Our approach for the detection of toxigenic strains of C. d@ale is easy both to perform and to interpret and offers excellent sensitivity and specificity. Some attempts have also been made on a reduced number of fecal specimens which showed adequate yields. Further studies are now ongoing.

R. Alonso, C. Muiioz ?: Pelciez, E. Cercenado

M. Rodriguez-Creixems, E. B o u z a Servicio de Microbiologia Clinica y

Enfermedades Infecciosas, Hospital General Universitario ‘Gregorio Maraiibn’,

C/Doctor Esquerdo, 46, 28007 Madrid, Spain

References 1. h o o p FC, Owens M, Crocker IC. Clostridium dficile:

clinical hsease and hagnosis. Clin Microbiol Rev 1993; 6:

2. Gumerlock PH, Tang YJ, Weiss JB, Silva J Jr. Specific detection of toxigenic strains of Clostridium dficile in stool specimens. J Clin Microbiol 1993; 31 : 507-1 1.

3. Wren B, Clayton C, Tabaqchali S . Rapid identification of toxigenic, Clostridium dfiile by polymerase chain reaction. Lancet 1990; 335: 423.

4. Murray F’, Baron JOE, Pfaller MA, Tenover FC, Yolken RH. Manual of clinical microbiology, 6th edn. Washngton DC: American Society of Microbiology (ASM), 1995.

5. Alonso R, Nicholson PS, Pitt TL. Rapid extraction of high purity chromosomal DNA from Serratia marcescens. Let Appl Microbiol 1993; 16: 77-9.

6. Barroso LA, Wang S Z , Phelps CJ, Johnson JL, Wilkins TD. Nucleotide sequence of Clostridium dficile toxin B gene. Nucleic Acids Res 1990; 18: 4004.

7. MacMihn DE, Muldrow LL, Leggette SJ, Abdulahi Y, Ekanemesang UM. Molecular screening of Clostridium dtjiale toxins A and B. Genetic determinants and identifications of mutant strains. FEMS Microbiol Lett 1991; 62: 75-80.

8. Borriello SP, Wren BW, Hyde S, et al. Molecular, immuno- logical and biological characterization of a Toxin A-negative, toxin B-positive strain of Clostridium dfi le . Infect Immun

9. Depitre C, Delmee M, Avesani V et al. Serogroup F strains of Clostridium dfi le produce toxin B but not toxin A. J Med Microbiol 1993; 38: 434-41.

10. Innis MA, Gelfand DH, Sninsky JJ, White TJ, eds. PCR protocols. A guide to methods and applications. San Diego, C A Academic Press, 1990.

251-65.

1991; 60: 4192-9.

Selection of rifampicin-resistant Mycobacferium fuberculosis does not occur in the presence of low concentrations of rifaximin

C l i n Microbiol Znfect 1997; 3: 147-151

Rifaximin, an insoluble derivative of rifamycin-SV [l] that is very poorly absorbed fiom the gastrointestinal tract [2], has been proposed for the oral therapy of several bacterial enteric diseases [3-61. In fact, rifaximin is particularly efficacious in eradicating strains of Salmonella and Shigella spp. and Escherichia coli possess- ing different virulence factors [7]. Staphylococci, streptococci and other Gram-positive as well as Gram- negative, aerobic and anaerobic species are also inhibited by the compound [8-lo]. Mycobacteria are also killed in vitro, as would be expected of an antibiotic sharing the properties of the rifamycin

Since rifampicin is widely used for treating tuber- culosis and other types of mycobacterial infections which are increasing worldwide [12], the aim of our

family [ll].

1 4 8

1000

800

X

C

z

600 - -z 6 400

200

0

X Q) U S z - P 6

1 000

800

600

400

200

0

Cl in ica l M i c r o b i o l o g y a n d In fec t ion , V o l u m e 3 N u m b e r 1, F e b r u a r y 1997

A. RlFAXlMlN 6 ng/mL

* I I

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 192021 2223

Days

B. RlFAXlMlN 20 n'g/mL

. . .

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1 9 2 0 2 1 2223

Days

Figure 1 Growth rates of strains 1 to 5 of M. tuberculosis, expressed as growth indices as a function of time, when tested with four rifaximin concentrations (A, 6 ng/mL; B, 20 ng/mL; C, 90 ng/mL, D, 270 ng/mL), compared with growth indexes of controls (C1 to C5) in the absence of antibiotic. Treated cultures, solid lines; control, dotted lines.

L e t t e r s t o t h e E d i t o r s

C. RlFAXlMlN 90 ng/mL

1 4 9

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 192021 2223

Days

D. RlFAXlMlN 270 ng/mL

1000 I r r m - - A L

Days

Figure 1 continued

1 5 0 Clinical Microbio logy and Infection, Volume 3 Number 1, February 1997

work was to investigate whether rifaximin could select mutants of Mycobacteriurn tuberculosis (showing cross- resistance to rifampicin) when used for the treatment of enteric infections in patients harboring M. tuber- culosis. Five M. tuberculosis strains, recently isolated in the &agnostic laboratory of our institute, were tested three came &om patients with pulmonary and two &om patients with renal tuberculosis.

Sterile stock solutions of rifaximin and rifampicin (provided by Alfa Wassermann, Bologna, Italy) were pre-diluted in methanol and then in saline. Cultures grown on Lowenstein-Jensen solid medium (a modi- fied Middlebrook 7H12 medium) were inoculated into 12B Bactec liquid medium, following standardized procedures [13]. M e r appropriate incubation in Bactec 12B, each culture was resuspended in the special dduting fluid recommended in the standardized method [13] and adjusted to a concentration of 1 x lo8 CFU/mL. (McFarland number 1). These suspensions were used to inoculate all vials containing drugs (rifampicin and rifaximin) and drug-fiee control vials.

Assessment of susceptibility or resistance to rifam- picin and rifaximin was performed in 12B medium (in 4-mL amounts, in 20-mL glass vials with rubber septa and caps) [14]. Minimal inhibitory concentrations (MICs) of rifaximin and rifampicin were determined by exposing the microorganisms to graded concen- trations of drugs (&om 0.06 to 4 pg/mL). Vials were then incubated for 6 days in Bactec 460. M e r ths period, the control cultures demonstrated the maxi- mum growth index value (999) and the MIC was recorded as the antibiotic concentration which had prevented any bacterial growth [15].

In order to simulate the conditions deemed to occur in vivo during an average course of rifaximin therapy (400 mglday, in three doses per day for 10 days) a suitable inoculum of each of the microorganisms was seeded into 12B vials containing 6, 20, 90 and 270 ng/ml of rifiximin. The times taken to reach the maximum growth index value in the absence (control) and in the four different concentrations of the selecting drug are depicted in Figure 1. At 6 ng/mL and 20 n g / d of rifaximin the growth rates of control and treated cultures were approximately the same (Figure lA, and 1B). With an increase in the concentration of rifaximin to 90 ng/mL, a substantial lag of 4 to 5 days was apparent in the time needed for drug-treated cultures to reach the same growth index as that of the control broths (Figure 1C). At the highest concen- tration of rifaximin employed (270 ng/mL) growth in the first 10 to 13 days was extremely slow or altogether absent. Only after about 18 to 23 days did the treated cultures reach the growth index attained in about 10 days by the untreated controls (Figure 1D).

In our experimental approach the artificially high amounts of rifaximin employed were in large excess of the concentrations to be expected &om intestinal absorption, during an average course of therapy for enteric infections. Their remarkable effects on the physiology of the pathogens tested, as apparent h m the slower growth rates observed in treated as opposed to control cultures, clearly indicate that these levels of drug act as sub-minimal inhibitory concentrations and might therefore express a powerful stimulus for the selection of resistant mycobacterial strains. In fact, with several pathogens, and possibly with mycobacteria, mutational resistance is not a consequence of a one-step event, and high-level resistance to drugs can be acquired only by a series of successive exposures to increasing concentrations of drugs [16]. Therefore we analyzed whether the different rifaximin concen- trations used in our work could have selected M. tuberculosis strains resistant to rifaximin and rifampicin.

For this purpose the MICs of rifampicin and rifaximin were determined for the five M. tuberculosis strains before and after treatment with rifaximin. The results obtained (data not reported) showed that the MICs of rifampicin and rifaximin were identical for each strain, examined before and afier treatment with the Uerent concentrations of rifaximin. The data clearly show that no pathogens resistant to both rifampicin and rifaximin were selected under our test conditions. Since the rifiximin concentrations used in our in vitro experiments were higher than those expected &om intestinal absorption, it is probable that the minimal amounts of rifaximin possibly resorbed into the systemic fluids after oral adrmnistration do not represent a level which wdl induce the selection of drug-resistant microorganisms.

Acknowledgments This investigation was supported in part by a FATMA grant (17/93.00241) &om the Italian National Research Council and by Alfa Wassermann SPA, Bologna, Italy.

Ornella Soro, Adelaide Pesce, Monica Raggi, Eugenio A. Debbia, Gian Carlo Schito

Institute of Microbiology, School of Medicine,

University of Genoa, Largo Rosanna Benzi, 10,

16132 Genoa, Italy

References 1. Venturini AE? Pharrnacokinetics of LAOS, a new rifamycin,

in rats and dogs, after oral administration. Chemother 1983; 29: 1-3.

L e t t e r s t o t h e E d i t o r s 1 5 1

2. Descombe JJ, Dubourg D, Picard M, Palazzini E. Pharma- cokinetic study of rifaximin after oral administration in healthy volunteers. Int J Clin Pharmacol Res 1994; 14:

3. Corazza GR, Ventrucci M, Strocchi A, et al. Treatment of small intestine bacterial overgrowth with rifaximin, a nonabsorbable rifamycin. J Int Med Res 1988; 16: 312- 16.

4. Di Febo G, Calabrese C, Matassoni E New trends in non- absorbable antibiotics in gastrointestinal disease. I d J Gastroenterol 1992; 24: 10-13 .

5. Fera G, Agostinaccho F, Nigro M, Schiraldi 0, Ferrieri A. fifaximin in the treatment of hepatic encephalopathy. Eur J Clin Res 1993; 4: 57-66.

6. Vinci M, Gatto A, Giglio A, et al. Double-blind clinical trial on infectious diarrhoea therapy: rifaximin versus placebo. Curr Ther Res 1984; 36: 92-9.

7. Venturini AF’, Marchi E. In vitro and in vivo evaluation of L/105, a new topical intestinal rifamycin. Chemioterapia

8. Lamanna A, Orsi A. In vitro activity of rifaximin and rifampicin against some anaerobic bacteria. Chemioterapia 1984; 3: 365-7.

9. Hoover W, Gerlach EH, Hoban DJ, Eliopoulos GM, Pfaller MA, Jones RN. Antimicrobial activity and spectrum of rifaximin, a new topical rifamycin derivative. Diagn Microbiol Infect Dis 1993; 16: 111-18.

51-6.

1986; 5: 257-62.

10. Eftimiadi C, De Leo C, Schito GC. Treatment of hepatic encephalopathy with L/105, a new non-absorbable rifamycin. Drugs Exp Clin Res 1984; 10: 691-6.

11. Garrod LP, Lambert HF’, O’Grady F, Waterworth PM. Antibiotics and chemotherapy. New York Churchdl Living- stone, 1989: 217-37.

12. Enarson DA, Grosset J, Mwinga A, et al. The challenge of tuberculosis: statements on global control and prevention. Lancet 1995; 346: 809-19.

13. Inderlied CB, Salfinger M. Antimicrobial agents and susceptibdity tests: mycobacteria. In Murray PR, Baron EJ, Pfaller MA, Tenover FC, Yoken R H eds. Manual of clinical microbiology; Washington, DC: American Society of Mcrobiology, 1995: 1385-404.

14. Siddiqi SH, Hawlans JE, Laszlo A. Interlaboratory drug susceptibihty testing of Mycobacteriurn tuberculosis by radio- metric and two conventional methods. J Clin Microbiol

15. National Committee for Clinical Laboratory Standards. htimycobacterial susceptibility testing. Proposed standard M24-P, Vol. 10 No. 10. Villanova, PA: NCCLS, 1990.

16. Murray BE, Hodel-Christian SL. Bacterial resistance: theoretical and practical considerations, mutations to anti- biotic resistance, characterization of R plasmid, and detection of plasmid-specific genes. In Lorian V, ed. Antibiotics in laboratory medicine. Baltimore: Wfiams and Wilkins,

1985; 22: 919-23.

1991: 556-98.