Romain Béraud, Louis Huneault, Dave Bernier, Francis Beaudry, Ann Letellier, Jérôme R.E. del Castillo. Abstract. Résumé

Transcription

1 Article Comprison of the selection of ntimicroil resistnce in fecl Escherichi coli during enrofloxcin dministrtion with locl drug delivery system or with intrmusculr injections in swine model Romin Bérud, Louis Huneult, Dve Bernier, Frncis Beudry, Ann Letellier, Jérôme R.E. del Cstillo Astrct This study evluted, for the first time, the selection of ntiiotic resistnce in fecl Escherichi coli, potentil reservoir of genes of resistnce, during the prolonged exposure to fluoroquinolones fter the implnttion of locl drug delivery system (LDDS) in swine model. Fourteen pigs were rndomly ssigned to group IM (5 mg/kg/dy of intrmusculr enrofloxcin EFX) or LD (surgicl implnttion of EFX-polymethyl-methcrylte peri-femorl implnts). Blood smples were collected dily for determintion of plsm EFX nd ciprofloxcin (CFX) concentrtions. Fecl smples were collected dily to determine the E. coli counts nd the susceptiility ptterns of its isoltes s evluted y ntiiotic disk diffusion tests. In oth groups, EFX dministrtion significntly reduced the cteril counts fter 2 dys. During recoloniztion, the cteril counts remined lower thn seline in group IM ut not significntly, nd lmost reched pre-tretment levels in group LD. Susceptiility to EFX, CFX, nd nlidixic cid of recolonizing E. coli in LD pigs slightly decresed ut remined within the limit of susceptile isoltes. In contrst, quinolone susceptiility of recolonizing E. coli in IM pigs dropped drmticlly (P, ). In ddition, intrmusculr exposure to fluoroquinolones significntly decresed the susceptiility of E. coli to mpicillin nd trimethoprimsulfmethoxzole (P, 0.05). In conclusion, the use of dosing regimen tht minimized the intestinl output of fluoroquinolones lso minimized the selection of resistnce to severl clsses of ntiiotics. This could represent nother dvntge of LDDS usge compred to long-lsting systemic dministrtion of fluoroquinolones. Résumé Pour l première fois une étude évlué l sélection de résistnce ux ntiiotiques chez des isolts fécux d Escherichi coli, un réservoir potentiel de gènes de résistnce, durnt une exposition prolongée ux fluoroquinolones près l implnttion d un système d dministrtion locle de médicments (LDDS) dns un modèle porcin. Qutorze porcs ont été ssignés u hsrd u groupe IM (5 mg/kg/jour d enrofloxcin [EFX] pr voie intrmusculire) ou LD (mise en plce chirurgicle d un implnt péri-fémorl d EFX-polyméthyl-méthcrylte). Des échntillons snguins ont été prélevés quotidiennement fin de déterminer les concentrtions plsmtiques d EFX et de ciprofloxcin (CFX). Des échntillons de fèces ont été collectés quotidiennement fin de dénomrer les E. coli et déterminer les ptrons de sensiilité des isolts tels qu évlués pr l méthode de diffusion en gélose. Dns les deux groupes, l dministrtion d EFX réduit significtivement les comptes ctériens près 2 jours. Au cours de l recolonistion, les comptes sont demeurés inférieurs de mnière non-significtive à l vleur de se dns le groupe IM, et ont presque tteint les niveux prétritement dns le groupe LD. Chez les porcs LD l sensiilité à l EFX, CFX et l cide nlidixique des E. coli recolonisnt diminué légèrement mis est demeurée à l intérieur de l limite pour les isolts «sensiles». À l opposé, chez les porcs du groupe IM l sensiilité ux quinolones des E. coli recolonisnt chuté drmtiquement (P, 0,0001). De plus, l exposition intrmusculire ux fluoroquinolones diminué de mnière significtive l sensiilité de E. coli à l mpicilline et u trimethroprime-sulfméthoxzole (P, 0,05). En conclusion, l utilistion d un régime doseur qui minimise l excrétion intestinle de fluoroquinolones minimise églement l sélection de résistnce à plusieurs clsses d ntiiotiques. Ceci pourrit représenter un utre vntge de l utilistion de LDDS comprtivement à l dministrtion systémique de fluoroquinolones à longue ction. (Trduit pr Docteur Serge Messier) Deprtment of Clinicl Sciences (Bérud), Deprtment of Clinicl Sciences, Groupe de Recherche sur les Mldies des Animux de Compgnie (GRMAC) (Huneult), Deprtment of Biomedicine (Bernier, Beudry), Deprtment of Pthology nd Microiology, Reserch Chir in Met Sfety, Groupe de Recherche sur les Mldies Infectieuses du Porc (GREMIP), Centre de Recherche en Infectiologie Porcine (CRIP) (Letellier), Deprtment of Biomedicine, Groupe de Recherche sur les Mldies Infectieuses du Porc (GREMIP), Centre de Recherche en Infectiologie Porcine (CRIP), Groupe de Recherche en Phrmcologie Animle du Quéec (GREPAQ) (del Cstillo); Fculté de Médecine Vétérinire, Université de Montrél. Address ll correspondence to Dr. Romin Bérud; telephone: (450) ; fx: (450) ; e-mil: romin.erud@umontrel.c Dr. Bérud s current ddress is Centre Hospitlier Universitire Vétérinire, Université de Montrél, CP 5000, St-Hycinthe, Quéec J2S 7C6. Received June 19, Accepted Septemer 21, ;72: The Cndin Journl of Veterinry Reserch 311

2 Introduction Musculoskeletl infections cn e difficult to tret with systemic ntimicroils ecuse of their limited penetrtion into the devsculrized tissues nd iofilms tht my e present t the infection site (1,2). The current therpeutic pproch for such infections includes n ggressive surgicl deridement, removl or chnge of fixtion devices without impiring one stility, soft tissue coverge, nd long-term systemic ntiiotic therpy, with success rte vrying etween 80% nd 90% (3,4). To improve this tretment response, locl drug delivery systems (LDDS) hve een developed, to llow sustined high-level locl ntiiotic concentrtion t the infection site (1,2,5). Although elution chrcteristics cn vry gretly mong different ntiiotic nd LDDS comintions, the systemic sorption of the chosen ntiiotic is generlly poor, which leds to sustined low plsm drug concentrtions nd miniml risk of systemic toxicity (2,5,6). The systemiclly sored ntiiotic, however, my provide concentrtions within the mutnt selection window (MSW) of vrious cteri, theory tht hs een proposed to explin how exposure to ntiiotics leds to the selective enrichment of resistnt cell supopultions (7 11). In the cse of veterinry fluoroquinolones, the risk of resistnce selection my e significnt, ecuse P-glycoprotein hs een shown in vivo to contriute to the intestinl excretion of this clss of ntiiotics (12). Hence, prenterlly dministered fluoroquinolones my exert selective pressure on the intestinl microiot, including Escherichi coli nd other cteril species with zoonosis potentil. While severl studies hve investigted the potentil role of LDDS in the selection of ntiiotic resistnce mong pthogens (13 20), none to our knowledge hs evluted the impct of the prolonged systemic ntiiotic concentrtions ner MSW on commensl flor. Resistnce selection mong pthogenic cteri is of concern s it increses the risk of infections tht re refrctory to clssicl tretments, which limits the therpeutic possiilities of prctitioners. Moreover, the selection of resistnce mong commensl flor is eqully of concern ecuse of its insidious role s reservoir of genes of resistnce (21,22). Mny reports in humn nd veterinry medicine hve highlighted the potentil for these resistnt commensl cteri to trnsfer resistnce genes to ny incoming pthogen or to themselves, leding to the possiility of infection (23 26). The ojective of this study ws to investigte the selection of cteril resistnce in the fecl E. coli flor of swine model during locl dministrtion of enrofloxcin (EFX) y LDDS, nd to compre the results to intrmusculr (IM) EFX injections. It ws hypothesized tht the EFX-loded LDDS used in this model would led to plsm drug concentrtions of EFX tht would e low enough to void ny cteril resistnce selection in the fecl flor, s opposed to tht from the IM dministrtion of EFX. Mterils nd methods Animls nd housing conditions Fourteen, 4 8-week-old helthy pigs ( kg) were used tht were known not to hve hd previous exposure to ntiiotics. These pigs were rndomly ssigned to 2 equl groups: IM (intrmusculr EFX), nd LD [locl delivery with EFX-loded polymethylmethcrylte (PMMA) implnts]. The pigs were housed individully in seprte pens, nd did not hve ny direct contct with pigs in the other group. The pigs were fed twice dily with n ntiiotic-free commercil pig feed, nd hd unlimited ccess to wter. Pigs were llowed to cclimtize for 4 d efore experiment initition on dy 1. Opertors lwys went through rigorous disinfection procedure efore entering nd exiting the unit nd etween mnipultions to void cross-contmintion. The protocol ws in greement with the Cndin Council on Animl Cre guidelines nd ws pproved y the Institutionl Animl Cre nd Use Committee of the Université de Montrél. Drugs nd implnts An injectle solution of enrofloxcin (Bytril; Byer, Toronto, Ontrio), 100 mg/ml, ws used for EFX dministrtion. The EFXloded PMMA implnts were prepred in sterile fshion y dding 6 g of EFX powder (Sigm-Aldrich, Okville, Ontrio) to 40 g of powdered polymer (Surgicl Simplex P Rdiopque Bone Cement; Howmedic, Limerick, Irelnd). The comintion ws homogenized, the liquid monomer ws dded, nd the soft EFX-PMMA mixture ws poured into 3-mL syringe rrels. Plungers were then inserted nd depressed to the 0.7 ml mrk to expel ir nd excess cement (27). The cement in the syringe ws llowed to hrden to produce 500 mg cylindricl EFX-PMMA implnts tht were 9 mm in dimeter, 7 mm long, nd contined ~65 mg EFX. Experimentl procedures On dy 7, pigs received IM injections of zperone [2 mg/kg ody weight (BW)], ketmine (8 mg/kg BW), nd hydromorphone (0.1 mg/kg BW). Anesthesi ws induced using fce msk delivering 1 3% isoflurne in 100% O 2, nd then continued vi n endotrchel tue. No ntiiotic prophylxis ws pplied ecuse this would interfere with the study s ojective. An indwelling ctheter ws plced in the left externl jugulr vein for seril lood collection. In the LD group, the lterl spect of the left femur ws pproched to plce 7 peri-diphysel EFX-PMMA implnts/10 kg BW. From dys 7 to 21, pigs in the IM group received IM injections of EFX t dose of 5 mg/kg BW, q24h, djusted twice week sed on increse in ody weight. Pigs in the LD group received no other ntiiotic fter their peri-femorl EFX-PMMA implnts. In ll pigs, lood ws collected in EDTA tues on 8 occsions during the first 32 h of tretment (from dy 7 to dy 8), nd then ws collected once dily etween dys 9 nd 21, efore the morning EFX IM injections. Plsm ws hrvested fter centrifugtion nd stored t -70 C pending nlysis. Fecl smples (20 g) were collected once dily from fresh feces or rectlly. Smples tken etween dys 1 nd 6 were used to determine the seline conditions, nd smples tken etween dys 7 nd 21 were used to document the time-course of the E. coli popultion size nd ntiiotic susceptiility of isoltes during exposure to EFX. After collection, ll smples were kept t 4 C until homogenized nd plted (within the next 6 h). At dy 21, ll pigs were euthnized y riturte overdose injected intrvenously. For prcticl resons, experiments were 312 The Cndin Journl of Veterinry Reserch 2000;64:0 00

3 divided into 2 seprte time locks, such tht 2 sugroups of 3 or 4 pigs from ech tretment group were studied simultneously. Plsm enrofloxcin nd ciprofloxcin concentrtions A liquid chromtogrphy, electrospry-tndem mss spectrometry (LC-ESI/MS/MS) method ws developed to mesure the plsm concentrtions of EFX nd its ctive metolite ciprofloxcin (CFX) simultneously. Briefly, the EFX nd CFX were extrcted y protein precipittion from plsm. One hundred microliters of plsm smple were mixed with 500 ml of cetonitrile fortified with 250 ng/ml of reserpine (internl stndrd). The smples were centrifuged t g for 10 min nd 10 ml of superntnt ws injected into the HPLC system, fitted with Wters Symmetry C mm column (3.5 mm prticule size). The moile phse ws 80:20 mixture of cetonitrile nd 0.5% formic cid in wter nd the flow rte ws 0.8 ml/min. The nlytes were mesured y mss spectrometry using PESciex API 31 instrument equipped with n electrospry ion source (ESI). The nlyte responses were mesured in selected rection monitoring mode (SRM) using mss trnsitions of m/z (EFX), m/z (CFX), nd m/z (reserpine). The quntifiction ws sed on the pek-re rtios of ech nlyte with the internl stndrd, nd the clirtion curve ws constructed using liner regression model (weighted 1/concentrtion). The lower limit of detection of the method ws 2 mg/l for oth nlytes, with overll ccurcies nd precisions of 100.3%, 10.5% for EFX nd 98.7%, 8.4% for CFX. Microiologicl nlyses The dily fecl Escherichi coli counts were performed ccording to stndrd procedures on McConkey gr. For ech smple, 20 g of feces were homogenized in sterile g y hnd mixing for 1 min, nd 5 g were diluted in sterile sline (1:10). The direct plting method (DPM) for detection of ntimicroil-resistnt E. coli ws crried out (28). Ech dilution (100 ml) ws plted onto McConkey Agr plte (Quel Lortories, Montrel, Queec) to yield isolted colonies. Bcteril growth ws exmined fter overnight incution t 37 C for lctose-fermenting colonies with morphologicl ppernce typicl of E. coli. For ech fecl smple, 3 such colonies were selected t rndom nd purified on 5% sheep lood gr (Quel Lortories) nd confirmed with triple iron sugr gr, citrte gr, indole nd urese tests (29). One out of 10 isoltes with typicl morphology or with typicl iochemicl ptterns of E. coli ws further chrcterized on API20E glleries (Biomérieux, Hzelwood, Missouri, USA) (30). The ntimicroil susceptiility of E. coli isoltes ws determined using the disk diffusion method, stndrdized ccording to the Clinicl nd Lortory Stndrds Institute guidelines (31). The following ntiiotics were used: EFX, CFX, nlidixic cid, ceftiofur, moxicillin, mpicillin, streptomycin, gentmicin, prmycin, trimethoprimsulfmethoxzole (TMS), florfenicol, nd tetrcycline. After overnight incution t 37 C for 16 to 18 h, the inhiition zone dimeters (mm) of ech ntiiotic were mesured using stndrd ruler. Phrmcokinetic nd sttisticl nlyses The time-course of plsm EFX nd CFX concentrtions in pigs dosed IM or with the LDDS were suject to sttisticl moment PK nlyses using WinNonlin 5.1 stndrd edition (Phrsight, Plo Alto, Cliforni, USA) (32). The liner trpezoid lgorithm ws pplied to estimte the res under the concentrtion (AUC) nd first moment curves (AUMC) ssocited with the first IM dosing, nd with the LDDS implnttion. Extrpoltion to infinity of AUC nd AUMC ws performed with stndrd formuls fter estimtion of the terminl decy slope (l z ). The following PK prmeters were clculted: pprent systemic clernce (CL/F) nd volume of distriution (V/F), men residence time (MRT), nd hlf-life of the terminl decy slope (t1/2 2 l z ). In ddition, the overll systemic exposure to EFX nd CFX ws estimted with the mesured AUC for the whole study period. Sttisticl nlyses were performed with the SAS system, version 9.1 (SAS Institute, Cry, North Crolin, USA), using the 0.05 level of significnce for ll nlyses. All dily fecl E. coli counts were log10 trnsformed to normlize their distriutions nd include in the nlysis. In ech pig, 1 to 4 smples were rndomly chosen per period for susceptiility testing, depending on the durtion of ech period nd the presence of recovered colonies. Three colonies were tested per smple nd verged. Inhiition zone dimeters were lso log10 trnsformed, nd the ssumptions of normlity necessry to perform prmetric nlysis were fulfilled. The dt re still presented without trnsformtion [men nd stndrd error (S x )] in the text or figures. Accounting for the dimeter of the ntimicroil disks, the miniml vlue for reported inhiition zone dimeter ws 6 mm. The time-course of the size of E. coli popultion nd the ntiiotic susceptiility of their isoltes ws divided in 4 periods, sed on the oserved fecl E. coli counts: (1) seline, dys 1 to 6, (2) ntiiotic first exposure effect, dy 7 to time of undetectle cteril count, (3) ctericidl equilirium, durtion of undetectle cteril counts, nd (4) regrowth equilirium, time of cteril resurgence to dy 21. The effect of EFX dosing route on the durtion of periods 2 nd 3 ws ssessed with survivl nlysis, using the log-rnk test t the level to focus on the lte differences in E. coli recovery (33). Liner mixed-effect models for repeted mesures, with time periods 1, 2, nd 4 s within-suject fctor, tretment s etweensuject fctor, nd pig within tretment s rndom vrile, were used to ssess the chnges in E. coli count nd inhiition dimeters over time. The time 3 tretment interction ws lso ssessed s fixed effect. Afterwrds, lest-squre mens were used to exmine differences in the time-course of the size of the cteril popultion nd of the inhiition zone dimeters, s well s the differences recorded for ech tretment group. Results re presented s lestsqure mens 6 stndrd errors unless otherwise specified. Person correltion coefficients were determined etween ntimicroils. Results Pigs gined kg during the experiment nd none demonstrted ny mjor systemic or locl dverse rections. Plsm enrofloxcin nd ciprofloxcin concentrtions Men plsm EFX nd CFX concentrtions in groups LD nd IM re presented in Figure 1, nd selected PK prmeters re presented in Tle I. Plsm EFX concentrtions in LD pigs were on verge 2000;64:0 00 The Cndin Journl of Veterinry Reserch 313

4 Figure 1. Men plsm enrofloxcin nd ciprofloxcin concentrtions in IM nd LD pigs. 4.5% s high s those in IM pigs, nd plsm CFX concentrtions were on verge 8.5% s high s those of EFX in ech tretment group. Plsm EFX concentrtion peked t similr times in oth tretment groups, ut its decy in LD pigs ws clerly iphsic, with shrp decrese during the 2 d following t mx, nd slope with 8.76-dy hlf-life tht lsted for the rest of the experiment (Tle I). In oth tretment groups, the time-course of plsm CFX concentrtions ws similr to tht of EFX ut the plsm CFX concentrtions were elow the quntifiction limit in some LD pigs. Fecl Escherichi coli counts The course of fecl E. coli counts is depicted in Figure 2. During the seline period (from dys 1 to 6), this cteril popultion verged nd CFU/g in LD nd IM pigs, respectively, which ws significntly different (P = 0.019). The dministrtion of EFX, either intrmusculrly (IM) or loclly (LD), mrkedly reduced the size of the fecl E. coli popultion during the next 2 dys (first exposure effect period), resulting t its end in counts elow the limit of detection of 100 CFU/g. The durtion of the first effect period ws not significntly different etween groups IM nd LD ( d nd d, respectively, survivl nlysis P = 0.17). However, the durtion of the ctericidl equilirium in LD pigs ws significntly less thn in IM pigs: d versus d, respectively (survivl nlysis, P = 0.037). Finlly, the fecl E. coli counts during the dys of the regrowth equilirium period were lower (ut not significntly lower) thn those of seline vlues in group IM (P = ), nd lmost reched pre-tretment levels in group LD (P = ). Though not sttisticlly significnt, the numer of pigs per group for which cteril counts remined elow the detection level during the entire experiment differed etween group LD (0/7 pigs) nd IM (3/7 pigs). Escherichi coli susceptiility to EFX simultneously The results of fecl E. coli isoltes susceptiility to EFX, s estimted from growth inhiition zone dimeters, re presented in Figure 3. Significnt effects of the groups nd time periods were identified (P, 0.001). In group LD, susceptiilities during period 2 were not significntly different from seline vlues ( mm, P = ), ut reched sttisticl significnce in period 4 (P = 0.05). However, the isoltes remined susceptile to EFX ccording to CLSI guidelines. In group IM, EFX susceptiility in period 2 ws not sttisticlly different from seline ( mm; P = 0.494), ut dropped drmticlly during period 4 (P, ). The E. coli isoltes recovered from IM pigs during the regrowth phse were resistnt to EFX. The significnt time 3 tretment interction on the susceptiility of E. coli isoltes to EFX ws restricted to period 4 (P, ). Escherichi coli susceptiility to other quinolones The results of fecl E. coli isoltes susceptiility to CFX nd nlidixic cid were similr to those descried for EFX for oth groups. Significnt effects of tretment, time, nd tretment 3 time were recorded (P, for oth drugs). Compred with seline vlues, the susceptiility of E. coli isoltes to CFX nd nlidixic cid in the LD pigs experienced no significnt vrition t period 2 (P. 0.78), ut significnt decreses were recorded t period 4 (P # 0.025). The verge inhiition zone dimeters of CFX nd nlidixic cid for the ltter isoltes, however, were still. 21 mm nd. 18 mm, respectively, indicting tht they were still susceptile. Similr findings were recorded for IM pigs: seline inhiition zone dimeters were mm for CFX nd mm for nlidixic cid, within-group differences for period 2 were not sttisticlly significnt (P $ 0.352), ut significnt decreses were recorded t period 4 ( mm nd mm respectively, with P, for oth). The sttisticlly significnt time 3 tretment interction ws limited to inhiition zone dimeter differences for the regrowth period (P, ). Person correltion coefficients were excellent mong ll 3 fluoroquinolones tested: 0.98 etween EFX nd CFX, 0.96 etween EFX nd nlidixic cid, nd 0.95 etween CFX nd nlidixic cid. Escherichi coli susceptiility to other ntiiotics The results of fecl E. coli isoltes susceptiility to moxicillin nd mpicillin re presented in Figures 4 nd 5, respectively. Susceptiility for moxicillin nd mpicillin were suject to significnt time effect (P = nd P # , respectively). In the LD nd IM pigs, no significnt difference in moxicillin inhiition zone dimeters ws recorded for exposure periods 1, 2, nd 4 (P nd P , respectively). In contrst, mpicillin susceptiility of E. coli isoltes recovered from IM pigs decresed t period 4 (P, ) while mpicillin inhiition zone dimeters from LD pigs decresed slightly, ut significntly, during period 4 compred with period 2 (P = ). Person correltion coefficients were 0.29 etween EFX nd moxicillin, nd 0.56 etween EFX nd mpicillin. The results of fecl E. coli isoltes susceptiility to TMS re presented in Figure 6. Tretment, time nd tretment 3 time effects were lso found for TMS (P, 0.02). Inhiition zone dimeters for the E. coli isoltes recovered from period 4 were significntly smller thn those of isoltes from period 1 (P = for LD nd P = for IM), nd the difference etween periods 2 nd 4 ws significnt for IM pigs (P = ). The tretment 3 time interction ws limited to period 4, where isoltes from LD pigs hd inhiition zone dimeters close to seline vlues, s opposed to those from IM pigs, in which inhiition zone dimeters were mostly less thn seline vlues (P, ). Person correltion coefficient etween EFX nd TMS ws The Cndin Journl of Veterinry Reserch 2000;64:0 00

5 Tle 1. Selected phrmcokinetic prmeters in groups LD nd IM Enrofloxcin Ciprofloxcin Prmeter Units LD IM LD IM AUC 0 24 h h (mg/l) AUC h h (mg/l) l z /h C mx mg/l t mx h MRT h t1/2 2 l z h CL/F L/(h kg) V/F L/kg Norml distriution. Hrmonic men nd jck-knife estimte of stndrd devition. AUC 0 24h Are under the curve for the first 24 h; AUC 0 336h Are under the curve for the entire project; l z terminl decy slope; C mx mximl concentrtion; t mx time to C mx ; MRT men residence time; CL/F pprent systemic clernce; V/F volume of distriution. E. coli count [log10(cfu/g)], Inhiition zone dimeter (mm),, c Enrofloxcin exposure phse Figure 2. Evolution of the fecl Escherichi coli popultion size (lest-squre mens, nd stndrd errors). LOD lower limit of cteril detection., lest-squre mens with different superscripts significntly differ Susceptiility to ceftiofur, gentmicin, streptomycin, tetrcycline, prmycin, nd florfenicol were lso tested, nd no significnt vritions were oserved for these ntiiotics. Excellent Person correltion coefficients were found etween moxicillin nd ceftiofur (0.87) or mpicillin (0.86). Discussion During the lst decdes, LDDS hve gined populrity in the prophylxis nd tretment of humn nd niml osteomyelitis nd infected rticulr prosthesis. However, there hs een more nd more concern out their ility to select ntiiotic resistnce in pthogenic cteri (13 17), prolem tht is fvored y the production of iofilms on the surfce of surgicl implnts (18 20,34). The focus in this study ws on the effect of prolonged low systemic ntiiotic concentrtions following LDDS implnttion on the E. coli popultion from the commensl fecl flor, which to our knowledge hs never een investigted. Enrofloxcin exposure phse Figure 3. Evolution of Escherichi coli susceptiility to enrofloxcin (lestsqure mens, nd stndrd errors).,,c lest-squre mens with different superscripts or suscripts significntly differ (sttisticl nlysis ws performed on log trnsformed vlues) Our LDDS ws designed to chieve 2 ojectives: 1) to chieve low systemic ntiiotic concentrtions over prolonged period of time, nd 2) to mimic clinicl situtions where, most of the time, clinicins design nd shpe their own LDDS implnts ccording to cteril susceptiility nd room ville for implnttion. Enrofloxcin nd PMMA re cliniclly relevnt choices for this purpose: EFX is fluoroquinolone efficient ginst stphylococci nd grm-negtive cteri, which re the min infectious gents implicted in posttrumtic osteomyelitis; PMMA is still considered the LDDS mtrix gold stndrd until n ffordle nd convenient resorle mtrix ecomes ville (1,3,35 37). Moreover, low serum EFX concentrtions were mesured throughout the study with our EFX-PMMA LDDS, similr to most LDDS descried in the literture (35,37). It is known tht the mteril used s mtrix, the shpe, porosity nd volume of the implnts, s well s the nture nd concentrtion of the chosen ntiiotic ll influence the elution properties of LDDS (5,6,27,38 40). The plsm EFX concentrtions mesured fter implnttion typiclly followed iphsic pttern with quick 2000;64:0 00 The Cndin Journl of Veterinry Reserch 315

6 Inhiition zone dimeter (mm) Inhiition zone dimeter (mm),,, c Amoxicillin exposure phse Figure 4. Evolution of Escherichi coli susceptiility to moxicillin lestsqure mens nd stndrd errors (sttisticl nlysis ws performed on log trnsformed vlues). Ampicillin exposure phse Figure 5. Evolution of Escherichi coli susceptiility to mpicillin (lestsqure mens, nd stndrd errors).,,c lest-squre mens with different superscripts or suscripts significntly differ (sttisticl nlysis ws performed on log trnsformed vlues) Inhiition zone dimeter (mm),,,c,c,,c Trimethoprim-sulfmethoxzole exposure phse Figure 6. Evolution of Escherichi coli susceptiility to trimethoprimsulfmethoxzole lest-squre mens, nd stndrd errors.,,c,d lest-squre mens with different superscripts or suscripts significntly differ (sttisticl nlysis ws performed on log trnsformed vlues) relese during the first 24 h, likely resulting from solvtion of the frction of ntiiotic tht is loosely dhered to the implnt surfce. Afterwrds, elution from deeper lyers my hve sustined the plsm EFX concentrtion. In vitro nd locl in vivo elution studies nd microscopic porosity evlution could hve een performed to 1) more precisely descrie the chrcteristics of our EFX-PMMA LDDS, nd 2) verify tht our LDDS ws providing higher concentrtions t the site of interest thn the IM route for the tretment durtion. Vrious studies hve reported systemic or locl in vivo elution chrcteristics of different LDDS nd hve compred them to locl concentrtions otined y systemic ntiiotic dministrtion (6,38,40). Adms et l (6) reported good grnultion tissue nd one CFX concentrtions (ove rek point sensitivity level) 28 d fter peri-tiil implnttion of CFX impregnted PMMA eds (6 g CFX/40 g PMMA powder). Resistnce to fluoroquinolones cn occur minly y 3 mechnisms: 1) decresed permeility of the cteril cell wll cused y c d ltertions of the hydrophilic pores (outer memrne proteins), 2) n efflux pump, which ctively trnsports the fluoroquinolone molecule out of the cell s it pproches or psses through the cteril memrne, nd 3) muttions in the genes encoding the GyrA nd GyrB enzyme suunits of DNA gyrse nd the PrC nd PrE suunits of topoisomerse IV, thus ltering the quinolone molecule s inding site (14,35,41,42). The genetic determinnts ssocited with our recorded chnges in ntiiotic susceptiility ptterns will e reported in seprte communiction. For resistnt mutnts to develop in the environment, fvorle events like n exposition of the cteril popultion to su-ctericidl concentrtions of the incriminted ntiiotic, sufficient energy metolism to support plsmid cquisition nd positive iologic lnce of muttion, or oth, re required, mong other possiilities. The mutnt selection window (MSW) is n ntimicroil concentrtion rnge extending from the miniml concentrtion required to lock the growth of wild-type cteri up to tht required to inhiit the growth of the lest susceptile, single-step mutnt (7,8). It hs een demonstrted tht plcing ntimicroil concentrtions inside the MSW enriches resistnt mutnt supopultions selectively, wheres plcing concentrtions ove the upper oundry of the window, lso clled the mutnt prevention concentrtion, restricts selective enrichment (7,8). We oserved temporry ut significnt first ntiiotic exposure effect of our LDDS on the totl E. coli count from the 2nd to the 4th dy post-implnttion. This effect could e explined y the EFX pek relese ( mg/l) reched in the first 3 to 4 h, which is ner the E. coli minimum inhiitory concentrtion (MIC) ( mg/l) (35). Through intestinl excretion of EFX vi P-glycoproteins, this initil urst effect could hve hd temporry impct on the fecl flor. However, only slight chnge in EFX isoltes susceptiility of the E. coli popultion ws oserved fter implnttion of the LDDS. This could e explined y the very short durtion of the systemic pek EFX concentrtion following LDDS implnttion, presumly not enough to llow sustined selection. Although slight decrese in susceptiility ws oserved during the period of ctericidl equilirium, it styed lrgely ove the intermedite sensitivity threshold. Even if this my not seem cliniclly relevnt, 316 The Cndin Journl of Veterinry Reserch 2000;64:0 00

7 these results demonstrte tht chnge in the E. coli isoltes EFX susceptiility did occur, nd tht the use of LDDS should e well justified, especilly for prophylxis where excessive usge hppens more often. In contrst, the susceptile E. coli popultion ws completely replced y presumly highly resistnt vrints during IM dministrtion of EFX. This result ws expected, nd ws in greement with those of Wiuff et l (43). The EFX resistnt E. coli isoltes were most likely selected from pre-existing mutnts initilly present in low frequencies (43). This would men tht IM dministrtion of 5 mg/kg BW EFX led, for some of the pigs, to fecl concentrtions ove the MSW (the 3 pigs for which the E. coli popultion ws completely nd durly depleted), wheres for others it presumly led to concentrtions inside the MSW (the 4 pigs for which the E. coli isoltes popultion recovered with resistnt strins) (9 11). Although mild decrese in susceptiility to vrious ntiiotics ws found fter the use of our LDDS, this chnge ws drmticlly more importnt during the IM dministrtion of EFX. For similr or etter efficcy for treting severe one infections, where ntiiotics re mndtory, LDDS seem to e more dvntgeous thn the IM route of dministrtion for the ntimicroil resistnce selection of resident microflor (2,4,37). Furthermore, sed on comprisons of resistnce selection etween the IM nd orl dministrtions of EFX previously pulished, this my lso e more dvntgeous thn the orl route of dministrtion (43). All pigs were euthnized t the end of the study, preventing ny follow-up of the susceptiility pttern of the E. coli popultion fter stopping EFX IM dministrtion. However, this hs een studied y Wiuff et l (43) who reported tht the resistnt phenotype of fecl E. coli would persist nd predominte for t lest 2 wk or even longer fter only 3 consecutive dys of orl or IM EFX dministrtion. It is known tht resistnce to 1 quinolone frequently results in resistnce to ll (35), s seen in our study with CFX nd nlidixic cid. However, some significnt vritions of E. coli susceptiility to other clsses of ntiiotics tested during EFX dministrtion were oserved. Some muttions tht confer resistnce to the fluoroquinolones, vi ltertions in permeility or ctivtion of the efflux pump cn confer resistnce to other ntimicroil gents; however, previously reported cross-resistnces during fluoroquinolone dministrtion concerned the cephlosporins nd tetrcyclines (35). To our knowledge, this is the first report of mpicillin nd TMS resistnce development induced y the dministrtion of EFX. As previously reported, the mechnisms implicted in these cross-resistnces could potentilly concern the mrrab regulon, responsile for the numer of porines in the externl memrne or the AcrAB-TolC efflux pump system in the cytoplsmic memrne (41). We were surprised y the discrepncies of cteril susceptiilities etween moxicillin nd mpicillin during period 4, s these 2 ntiiotics elong in the sme clss nd re often used interchngely. Although this could potentilly result from type sttisticl error, it could lso e explined y difference in their respective ffinity for the previously mentioned porines or efflux pump. Another hypothesis could e tht, in fluoroquinolone-resistnt mutnts, modifiction of penicillin-inding protein (PBP) with reduced drug ffinity or reduced cteril permeility hs occurred (44). Given the different ffinities for PBPs etween mpicillin (PBPs 2 nd 3) nd moxicillin (PBPs 1A nd 2), modifiction of one prticulr PBP could theoreticlly ffect cteril susceptiility for one penicillin ut not the other (45). However, these hypotheses need to e confirmed. Our next communiction will report the precise nture of the muttions involved in the mechnism of resistnce to ssess their reltive loction/interctions with those involved in resistnce ginst fluoroquinolones. The IM nd locl dministrtion of EFX logiclly resulted in mrked difference in the systemic exposure to this ntiiotic nd its metolite CFX, s estimted y the re under the curve (AUC; Figure 1, Tle I). This resulted in difference in the exposure of the fecl flor to oth EFX nd CFX, ut ecuse none of these drugs ws dministered intrvenously, it is not possile t this stge to ssess their sorption, metolism, nd disposition phrmcokinetics, or their respective contriutions to the selection of fecl resistnt cteri. However, it ws felt tht hving clinicl-sed model with different overll EFX doses (which is inherent to the specific mode of ction of the LDDS nd its dministrtion route) would e more representtive nd cliniclly relevnt thn phrmcologiclsed model where oth IM nd locl overll doses would hve een equl. Moreover, the dosing regimens herein reveled tht the mutnt selection window hs lower oundry in vivo, oth in terms of mount nd durtion of drug exposure. A negtive control group ws not used in our experimentl design ecuse the rte of fluoroquinolone resistnce muttion in E. coli ws expected to e low in group LDDS nd not detectle in the sence of selection pressure. Moreover, given the highly controlled environment of the pigs during the study (room temperture nd humidity, light cycle, disinfection procedure, etc.) we felt tht hving ech individul eing its own negtive control would e vlule nd more ccurte. Since the difference in seline E. coli counts etween oth groups ws sttisticlly significnt, higher proility of muttion could hve theoreticlly ised group IM nd ffected our results. Genetic muttions responsile for ntimicroil resistnce, however, re rndom events occurring t very low frequencies (10-6 to for the fluoroquinolones), which suggests tht the risk of is in our study ws negligile (46). Finlly, we re wre tht conclusions regrding the rte of exposure of fecl E. coli to EFX nd CFX produced with our implnts re difficult to trnspose to other LDDS. As previously mentioned, ech comintion of ntiiotic nd mtrix presents specific elution properties tht my vry gretly, nd dditionlly e suject to host effects. Moreover, fluoroquinolone resistnce usully occurs through stepwise ccumultion of point muttions, which is quite different from most other resistnces minly cused y onestep cquisition of resistnce gene (14,35,41,42). Future reserch is wrrnted on other ntiiotic/mtrix ssocitions to confirm our oservtions nd increse our understnding of this topic. This study evluted, for the first time, the development of ntiiotic resistnce in E. coli isoltes from the commensl flor, indictor micro-orgnism nd potentil reservoir of genes of resistnce, following the use of LDDS. We showed tht the implnttion of EFX-loded PMMA implnts hs miniml potentil to select for ntimicroil resistnce within the studied fecl E. coli isoltes in swine model. Although implnttion of our LDDS led to temporry decrese in the totl count of the fecl E. coli popultion, the intestines were rpidly recolonized y popultion of E. coli with 2000;64:0 00 The Cndin Journl of Veterinry Reserch 317

8 very slightly decresed ntimicroil susceptiility pttern. Even if the chnges in susceptiility were minor, this finding dicttes proper justifiction of LDDS usge for prophylxis. Also, the lrgely inferior vritions in susceptiility for group LD versus IM could represent nother dvntge of LDDS usge compred to long-lsting systemic dministrtion of fluoroquinolones. Finlly, further studies implying other ntiiotic-mtrix comintions nd flor would e interesting to complement our oservtions. Acknowledgments This study ws supported y Snté Cnd nd the Fondtion du XXIII e Congrès Mondil Vétérinire. The uthors thnk Guy Beuchmp for ssistnce in the sttisticl tests, Mrs. Mrinne Turgeon-Plouffe nd Vlérie Normnd for technicl ssistnce, nd the Lortoire d Hygiène Vétérinire et Alimentire (Agriculture et Agro-limentire Cnd, St-Hycinthe) for use of their niml reserch fcilities. References 1. Knellkopoulou K, Gimrellos-Bouroulis EJ. Crrier systems for the locl delivery of ntiiotics in one infections. Drugs 2000;59: Strepp HK, Singer MJ, Budserg SC. Applictions of locl ntimicroil delivery systems in veterinry medicine. J Am Vet Med Assoc 2001;219: Brden TD. Posttrumtic osteomyelitis. Vet Clin North Am Smll Anim Prct 1991;21: Mder JT, Ortiz M, Clhoun JH. Updte on the dignosis nd mngement of osteomyelitis. Clin Poditr Med Surg 1996;13: Henry SL, Gllowy KP. Locl nticteril therpy for the mngement of orthopedic infections. Phrmcokinetic considertions. Clin Phrmcokinet 1995;29: Adms K, Couch L, Cierny G, Clhoun J, Mder JT. In vitro nd in vivo evlution of ntiiotic diffusion from ntiioticimpregnted polymethylmethcrylte eds. Clin Orthop Relt Res 1992;278: Drlic K. The mutnt selection window nd ntimicroil resistnce. J Antimicro Chemother 2003;52: Zho X, Drlic K. Restricting the selection of ntiiotic-resistnt mutnts: A generl strtegy derived from fluoroquinolone studies. Clin Infect Dis 2001;33 Suppl 3:S147 S Linde HJ, Lehn N. Mutnt prevention concentrtion of nlidixic cid, ciprofloxcin, clinfloxcin, levofloxcin, norfloxcin, ofloxcin, sprfloxcin or trovfloxcin for Escherichi coli under different growth conditions. J Antimicro Chemother 2004;53: Olofsson SK, Mrcusson LL, Komp Lindgren P, Hughes D, Crs O. Selection of ciprofloxcin resistnce in Escherichi coli in n in vitro kinetic model: Reltion etween drug exposure nd mutnt prevention concentrtion. J Antimicro Chemother 2006;57: Psquli F, Mnfred G. Mutnt prevention concentrtion of ciprofloxcin nd enrofloxcin ginst Escherichi coli, Slmonell Typhimurium nd Pseudomons eruginos. Vet Microiol 2007;119: Cvet ME, West M, Simmons NL. Fluoroquinolone (ciprofloxcin) secretion y humn intestinl epithelil (Cco-2) cells. Br J Phrmcol 1997;121: Snzen L, Wlder M. Antiiotic resistnce of cogulse-negtive stphylococci in n orthopedic deprtment. J Hosp Infect 1988; 12: Weer FA, Lutench EE. Revision of infected totl hip rthroplsty. Clin Orthop Relt Res 1986;211: Kendll RW, Duncn CP, Smith JA, Ngui-Yen JH. Persistence of cteri on ntiiotic loded crylic depots. A reson for cution. Clin Orthop Relt Res 1996;329: Thomes B, Murry P, Bouchier-Hyes D. Development of resistnt strins of Stphylococcus epidermidis on gentmicin-loded one cement in vivo. J Bone Joint Surg Br 2002;84: vn de Belt H, Neut D, Schenk W, vn Horn JR, vn der Mei HC, Busscher HJ. Infection of orthopedic implnts nd the use of ntiiotic-loded one cements. A review. Act Orthop Scnd 2001;72: Arciol CR, Cmpocci D, Montnro L. Effects on ntiiotic resistnce of Stphylococcus epidermidis following dhesion to polymethylmethcrylte nd to silicone surfces. Biomterils 2002;23: Arizono T, Og M, Sugiok Y. Incresed resistnce of cteri fter dherence to polymethyl methcrylte. An in vitro study. Act Orthop Scnd 1992;63: Nylor PT, Myrvik QN, Gristin A. Antiiotic resistnce of iomteril-dherent cogulse-negtive nd cogulse-positive stphylococci. Clin Orthop Relt Res 1990;261: Sørum H, Sunde M. Resistnce to ntiiotics in the norml flor of nimls. Vet Res 2001;32: Gustfsson I, Sjolund M, Torell E, et l. Bcteri with incresed muttion frequency nd ntiiotic resistnce re enriched in the commensl flor of ptients with high ntiiotic usge. J Antimicro Chemother 2003;52: Amyes SG, Tit S, Thomson CJ, et l. The incidence of ntiiotic resistnce in eroic fecl flor in south Indi. J Antimicro Chemother 1992;29: Shnhn PM, Wylie BA, Adrin PV, Koornhof HJ, Thomson CJ, Amyes SG. The prevlence of ntimicroil resistnce in humn fecl flor in South Afric. Epidemiol Infect 1993;111: Sunde M, Fossum K, Solerg A, Sørum H. Antiiotic resistnce in Escherichi coli of the norml intestinl flor of swine. Micro Drug Resist 1998;4: Thomson CJ. The glol epidemiology of resistnce to ciprofloxcin nd the chnging nture of ntiiotic resistnce: A 10-yer perspective. J Antimicro Chemother 1999;43 Suppl A: Ethell MT, Bennett RA, Brown MP, Merritt K, Dvidson JS, Trn T. In vitro elution of gentmicin, mikcin, nd ceftiofur from polymethylmethcrylte nd hydroxyptite cement. Vet Surg 2000;29: Brtoloni A, Brtlesi F, Mntell A, Dell Amico E, Roselli M, Strohmeyer M, et l. High prevlence of cquired ntimicroil resistnce unrelted to hevy ntimicroil consumption. J Infect Dis 2004;189: The Cndin Journl of Veterinry Reserch 2000;64:0 00

9 29. Lidin-Jnson G, Kijser B, Lincoln K, Olling S, Wedel H. The homogeneity of the fecl coliform flor of norml school-girls, chrcterized y serologicl nd iochemicl properties. Med Microiol Immunol 1978;164: Kronvll G, Lrsson M, Borén C, et l. Extended ntimicroil resistnce screening of the dominnt fecl Escherichi coli nd of rre resistnt clones. Int J Antimicro Agents 2005;26: Ntionl Committee for Clinicl Lortory Stndrds. CLSI, Performnce Stndrds for Antimicroil Disk nd Dilution Susceptiility Tests of Bcteri Isolted from Animls; pproved stndrd M31-S1. Villnov, Pennsylvni, Ymok K, Nkgw T, Uno T. Sttisticl moments in phrmcokinetics. J Phrmcokinet Biophrm 1978;6: SAS Online doc SAS Institute Inc. Cry, North Crolin, USA [pge on the Internet] Aville from com/documenttion/onlinedoc/91pdf/index_913.html Lst ccessed My 10, Olson ME, Ceri H, Morck DW, Buret AG, Red RR. Biofilm cteri: Formtion nd comprtive susceptiility to ntiiotics. Cn J Vet Res 2002;66: Wlker RD. Fluoroquinolones. In: Prescott JF, Bggot JD, Wlker RD, eds. Antimicroil Therpy in Veterinry Medicine. 3rd ed. Ames, Iow: Iow Stte Univ Pr, 2000: Criou M, Boulouis HJ, Moissonnier P. Inclusion of mrofloxcin in PMMA orthopedic cement: An in vitro experimentl study. Vet Comp Orthop Trumtol 2006;19: Huneult LM, Lussier B, Dureuil P, Chouinrd L, Désévux C. Prevention nd tretment of experimentl osteomyelitis in dogs with ciprofloxcin-loded crosslinked high mylose strch implnts. J Orthop Res 2004;22: DiMio FR, O Hllorn JJ, Qule JM. In vitro elution of ciprofloxcin from polymethylmethcrylte cement eds. J Orthop Res 1994;12: Rmos JR, Howrd RD, Plesnt RS, et l. Elution of metronidzole nd gentmicin from polymethylmethcrylte eds. Vet Surg 2003;32: Anguit-Alonso P, Rouse MS, Piper KE, Jcofsky DJ, Osmon DR, Ptel R. Comprtive study of ntimicroil relese kinetics from polymethylmethcrylte. Clin Orthop Relt Res 2006;445: Hooper DC. Mechnisms of fluoroquinolone resistnce. Drug Resist Updt 1999;2: Piddock LJ. Mechnisms of fluoroquinolone resistnce: An updte Drugs 1999;58 Suppl 2: Wiuff C, Lykkesfeldt J, Svendsen O, Arestrup FM. The effects of orl nd intrmusculr dministrtion nd dose escltion of enrofloxcin on the selection of quinolone resistnce mong Slmonell nd coliforms in pigs. Res Vet Sci 2003;75: Prescott JF. Bet-lctm ntiiotics: Penm penicillins. In: Prescott JF, Bggot JD, Wlker RD, eds. Antimicroil Therpy in Veterinry Medicine. 3rd ed. Ames, Iow: Iow Stte Univ Pr, 2000: vn Heijenoort J. Mécnismes moléculires de l ctéricidie: Bet-lctmines. In: Courvlin P, Drugeon H, Flndrois J-P, Goldstein F, eds. Bctéricidie. Aspects théoriques et thérpeutiques. Pris, Frnce: Editions Mloine, 1990: Acr JF, Goldstein FW. Trends in cteril resistnce to fluoroquinolones. Clin Infect Dis 1997;24 Suppl 1:S67 S ;64:0 00 The Cndin Journl of Veterinry Reserch 319