Microdialysis Study of Imipenem Distribution in Skeletal Muscle and Lung Extracellular Fluids of Noninfected Rats

The aim of this study was to investigate the imipenem distributionin muscle and lung interstitial fluids by microdialysis in ratsand to compare the free concentrations in tissue with the freeconcentrations in blood. Microdialysis probes were insertedinto the jugular vein, hind leg muscle, and lung. Imipenem recoveriesin these three media were determined in each rat by retrodialysisby drug period before drug administration. Imipenem was infusedintravenously at a dose of 120 mg · kg^–1 over 30min, and microdialysis samples were collected for 150 min. Thewhole study was conducted with nonhydrated rats (n = 4) andhydrated rats (n = 6) while the animals were under isofluraneanesthesia. The decay of free concentrations in blood, muscle,and lung with time were monoexponential; and the concentrationprofiles in these three media were virtually superimposed inboth groups. Accordingly, the ratios of the area under the curve(AUC) for tissue (muscle or lung) to the AUC for blood werealways virtually equal to 1. Compared to values previously determinedwith awake rats, clearance was reduced by 2 and 1.5 in nonhydratedand hydrated rats, respectively, but the volume of distributionwas unchanged. By combining microdialysis in blood and tissues,it was possible to demonstrate that free imipenem concentrationswere virtually identical in blood, muscle, and lung.

INTRODUCTION

Top

Introduction

As for many other drugs, pharmacokinetic studies of antibioticsmost often rely on plasma data. However, since most infectionsoccur in tissue extracellular fluid, it has been realized thatfree antibiotic concentrations in the interstitial fluid atthe target site are responsible for the antibacterial effect.Although from a theoretical point of view unbound drug concentrationsin blood and interstitial fluid of peripheral tissues withoutphysiological barriers such as muscle or lung should be identicalat equilibrium (26), the relevance of using plasma antibioticconcentrations rather than tissue antibiotic concentrationshas frequently been challenged (21). For many years skin blisterstudies have been very popular for the monitoring of free antibioticconcentrations in extracellular fluids, but more recently, microdialysishas appeared to be the technique of choice (10). Amino ß-lactamantibiotics have gained considerable attention in recent years,and because unbound concentrations in muscle are relativelyaccessible and should constitute reasonable predictors of theunbound concentrations in more therapeutically relevant tissuessuch as lung tissue, many microdialysis studies have been conductedin muscles in both rats (7, 11, 19, 22, 25) and humans (5, 17,28, 29). However, only two animal studies (11, 22) have beenconducted to compare free antibiotic concentrations in muscleand lung.

Imipenem (IPM) is an antibiotic used to treat severe infectionsof all body systems (3, 18, 32), particularly lower pulmonaryinfections (1). It is frequently used in intensive care unitsfor the treatment of nosocomial infections (6). A recent microdialysisstudy in muscle and subcutaneous tissue has shown that the IPMtissue distribution in seriously ill patients is reduced comparedwith that in healthy volunteers (28), confirming previous observationswith piperacillin (5). Several hypotheses including reducedtissue perfusion could be proposed to explain the impaired distributionof IPM in tissue. However, because of the numerous differencesbetween healthy subjects and patients, including age, kidneyand cardiovascular function, as well as drug treatment, no clearexplanation could be proposed for this reduced tissue distribution(28).

Because experiments with animals allow a better control of suchconfounding parameters, our objective was to investigate thetissue IPM distribution, initially in healthy rats, with a specialfocus on the comparison between muscle and lung free interstitialfluid concentrations as well as free blood levels.

MATERIALS AND METHODS

Top

Materials and Methods

Chemicals. IPM monohydrate-sodium cilastatin salt (Tienam; Merck Sharp& Dohme-Chibret Laboratories, Paris, France) was used toprepare imipenem solutions in 0.9% NaCl or Ringer solution forintravenous administration and probe perfusion, respectively.All chemicals used were of analytical grade, and all solventswere of high-performance liquid chromatography (HPLC) grade.

Animals. Ten male Sprague-Dawley rats from Janvier Laboratories (Le Genest-St-Isle,France) weighing 288 ± 20 g (mean ± standard deviation[SD]) were used and divided into two groups, a nonhydrated group(n = 4) and a hydrated group (n = 6). All animals were acclimatizedin wire cages in a 12-h light-dark cycle for a minimum of 5days before the beginning of the experiment to allow them toadjust to the new environment. During this period, they hadfree access to food (A03, Safe; Villemoisson-sur-Orge, France)and water. Ethical approval was obtained from the local animalethics committee.

Surgery. (i) Vein probe implantation. On the day of the experiment, the rats were anesthetized byisoflurane (Forene; Abbot, Rungis, France) inhalation. Theywere placed in a hermetic enclosure which was supplied withan air-isoflurane (3.9%) mixture at a flow rate of 500 ml ·min^–1 (Anaesthesia Unity, Univentor 400; Phymep, Paris,France). When the animals were sleeping, a mask was placed onthe muzzle, and the concentration of isoflurane was decreasedto 1.9% during the insertion of the probe into the vein andmuscle.

The rats were then placed in the dorsal position with the tailtoward the experimenter. An incision was made in the skin toexpose the right pectoral muscle. An introducer (correspondingto a needle inserted into a tube) was inserted through the pectoralmuscle into the right jugular vein. By removing the needle,the blood rose in the tube and a microdialysis CMA/20 probe(polycarbonate; cutoff, 20,000 Da; membrane length, 10 mm; CMAMicrodialysis; Phymep) was inserted through the tubing. Duringsurgery, the vein probe was connected to a CMA 102 microdialysispump (CMA Microdialysis; Phymep) and perfused with 0.1% low-molecular-weightheparin solution at a flow rate of 3.5 µl · min^–1.After insertion, the vein microdialysis probe was flushed at10 µl · min^–1 for approximately 15 min toremove bubbles; the flow rate was then decreased to 3.5 µl· min^–1 until the end of the surgery. The probewas then secured by suturing it to the pectoral muscle, andthe tubing was removed.

(ii) Muscle probe implantation. The right hind leg muscle (biceps femoris) was used for theinsertion of the CMA/20 microdialysis probe (polycarbonate;cutoff, 20,000 Da; membrane length, 10 mm; CMA Microdialysis;Phymep). The muscle was exposed, and the probe was insertedwith the help of an introducer, which was removed after insertion.The probe was finally secured by suturing it to the muscle.During insertion, the muscle probe was connected to a CMA 102microdialysis pump (CMA Microdialysis; Phymep) and perfusedwith Ringer solution (perfusion fluid T1 for peripheral tissues;CMA Microdialysis; Phymep) at a flow rate of 3.5 µl ·min^–1. As for the vein microdialysis probe, the muscleprobe was flushed at 10 µl · min^–1 for approximately15 min to remove the bubbles. The flow rate was then decreasedto 3.5 µl · min^–1 until the end of surgery.

(iii) Implantation of vein femoral catheter. A polyethylene catheter constituted with the connection of asmall-diameter catheter (inner diameter, 0.26 mm; outer diameter,0.61 mm; Phymep) with a larger one (inner diameter, 0.58 mm;outer diameter, 0.96 mm; Harvard, Les Ulis, France) was implantedin the left femoral vein of the anesthetized rat. The femoralcatheter, together with the inlets and the outlets of the probes,was passed subcutaneously so that they exited at the nape. Forthe hydrated group, saline was administered at a flow rate of1.2 ml · h^–1 directly after the implantation ofthe femoral catheter and until the beginning of IPM infusion.

(iv) Tracheotomy. The animals were immobilized in a supine position with cervicalhyperextension, and the rate of isoflurane flow was increasedto 3.5% in order to make the anesthesia deeper. A median incisionwas made from the suprasternal hollow over 2 cm in the directionof the muzzle. To expose the trachea, the adipose tissue andthe muscles were spread. Two silk threads were placed underthe trachea, which was incised between the second and the thirdtracheal rings. The tracheal tube was then inserted and fixedwith the previously placed threads. The rats were artificiallyventilated with an air-isoflurane mixture (3.5%) at a flow rateof 460 ml · min^–1 throughout the experiment byusing a rodent respirator (model 683 apparatus; Harvard), witha frequency of 65 to 68 min^–1 and a volume of 2.25 to2.5 ml. After the tracheotomy and before the thoracotomy, theamount of isoflurane was decreased to 2.2%.

(v) Lung probe implantation. The rats were placed on the left lateral decubitus. For thethoracotomy, an incision was made 3 cm below the right foreleg,and the right mean pulmonary lobe was individualized. An introducer,corresponding to a spinal needle (Yale Spinal; 20 gauge; 0.9by 90 mm; Becton Dickinson S.A., Madrid, Spain), was insertedhorizontally through the right median pulmonary lobe. A custom-madelinear microdialysis probe (LMP 5.35.35; polyether sulfone;cutoff, 6,000 Da; membrane length, 5 mm; outer diameter, 0.6mm; Microbiotech, Stockholm, Sweden) was inserted through thisspinal needle, which was then removed. After the insertion,the pulmonary probe was connected to a CMA 102 microdialysispump (CMA Microdialysis; Phymep) and perfused with Ringer solutionat a flow rate of 2 µl · min^–1. The pulmonarylobe was replaced in the rib cage, which was then superficiallyclosed.

Design of infusion experiments. After insertion, the blood, muscle, and lung probes were perfusedwith blank Ringer solution at a flow rate of 1 µl ·min^–1 for 30 min for a stabilization period.

Recovery calculations. After stabilization, the experiment started with a retrodialysisby drug period, during which the probes were perfused at 2 µl· min^–1 with Ringer containing IPM (10 µg· ml^–1) for 45 min and then at 1 µl ·min^–1 for 15 min to equilibrate the system. After thisequilibration period, microdialysate samples were collectedautomatically with a CMA/140 microfraction collector (CMA Microdialysis;Phymep) for 40 min by fractions corresponding to 10-min intervals.To determine the in vivo recovery by loss, IPM concentrationsin the perfusate (C_in) and in dialysates (C_out) were determinedby HPLC. This in vivo relative recovery by loss was expressedas a percentage (RL_{in vivo}) and was calculated for each intervalof time as (1)

(1)
The in vivo recoveryused to correct the dialysate concentrations was the mean valueobtained from the four individual determinations. A washoutperiod of 1 h (45 min at 2 µl · min^–1 and15 min at 1 µl · min^–1) with blank Ringersolution perfusion was allowed before the start of intravenousIPM administration to remove the IPM from the probes. The flowrate was maintained at 1 µl · min^–1 for therest of the study (pharmacokinetic investigation).

Imipenem administration. An intravenous infusion of IPM (Tienam) at a dose of 120 mg· kg^–1 and at an infusion rate of 10 ml ·h^–1 was performed over 30 min.

Microdialysis experiment. In each group, at 14 min after the beginning of the intravenousIPM administration (which was time needed to flush the deadvolume), the dialysates in muscle, lung, and blood were collectedover 150 min, at 10-min intervals during the first 60 min and15-min intervals during the last 90 min.

Directly after collection, the microdialysis dialysates werediluted (1:3 [vol/vol] for dialysates collected at 10-min intervalsand 1:2 [vol/vol] for the rest of the intervals) with a stabilizer(0.5 M HEPES buffer, pH 6.8; ethylene glycol; HPLC-grade water[1:0.5:0.5; vol/vol/vol]).

Stability of imipenem probe recoveries with time. Three dedicated rats were used to test the stability with timeof the in vivo relative recovery by loss of IPM. Probes wereperfused at 2 µl · min^–1 with Ringer solutioncontaining IPM (10 µg · ml^–1) for 45 minand at 1 µl · min^–1 for 15 min to equilibratethe system. After this equilibration period, microdialysatesamples were collected automatically over 150 min by fractionscorresponding to 15-min intervals (flow rate, 1 µl ·min^–1). The in vivo relative recovery by loss was calculatedas described above (equation 1) for each interval of time.

Microdialysis sample analysis. Analysis of IPM in the dialysates was performed by an adaptationof a previously described HPLC assay (12, 13). Standard curveswere prepared with IPM in water at concentrations ranging from0.78 to 100 µg · ml^–1. The diluted microdialysateswere directly injected onto a Novapack C₁₈ column (150 by 3.9mm [inner diameter]; Waters, France). The chromatographic systemconsisted of a Shimadzu (Croissy Beaubourg, France) LC-10ATpump and a Waters 717 plus autosampler connected to a UV detector(SPD 10A Shimadzu UV detector) at 313 nm. Data were recordedand analyzed on a Kromasystem integrator (Bio-Tek, St. Quentinen Yvelines, France). The mobile phase consisted of 0.2 M aqueousborate buffer, pH 7.2; and the flow rate was 1 ml · min^–1.The between-day variability of the IPM concentration was characterizedat 10 µg · ml^–1, with a coefficient of variationequal to 4.1% (n = 14) and an accuracy equal to 102.6% (n =14).

Noncompartmental pharmacokinetic analysis. Pharmacokinetic parameters were determined in each individualrat by a noncompartmental approach according to standard proceduresand with the software WinNonLin 4.0.1 (Pharsight Corporation,Mountain View, Calif.). Total body clearance (CL) was calculatedas CL = dose/AUC_blood, where AUC_blood is the total area underthe free blood concentration-versus-time curve, calculated byusing the trapezoidal rule. The area remaining under the curveafter the last measured concentration, C(last)_blood, was determinedfrom C(last)_blood/k_el,blood. The elimination rate constant (k_el,_blood)and its corresponding half-life (t_1/2,blood) were estimatedby least-squares fit of the datum points (log concentration-time)in the terminal phase of the decline. The volume of distribution(V) was obtained from CL/k_el,_blood. The area under the curve(AUC) and t_1/2 in tissues were also estimated by the same procedureand are referred as AUC_muscle, AUC_lung, t_1/2,muscle, and t_1/2,lung.Note that although microdialysis allows determination of theunbound drug concentrations in plasma water or tissue extracellularfluid, for simplicity, the subscripts (blood, muscle, and lung)refer to the medium in which the probe was implanted.

Statistical analysis. Concentrations were expressed as means ± SDs. Comparisonsof AUC, t_1/2, and the maximum concentration (C_max) obtainedfor blood, muscle, and lung tissue between rats in the samegroup were performed by the nonparametric Kruskall-Wallis test,with significance being a P level of <0.05. The ratio ofAUC_muscle/AUC_blood, the ratio of AUC_lung/AUC_blood, CL, and Vbetween the two groups were compared by a Mann-Whitney test,with significance being a P level of <0.05.

RESULTS

Top

Results

Stability of imipenem probe recoveries with time. The relative recovery by loss (RL) of IPM over a period of 150min determined in the three dedicated rats varied between 24.1%± 1.1% and 49.6 ± 2.7% in blood, between 15.4%± 4.0% and 16.3% ± 1.7% in muscle, and between5.9% ± 1.5% and 16.0% ± 4.6% in lung tissue. Therecovery by loss of IPM differed in blood, muscle, and lungin a particular rat and for a particular probe between experimentsbut was stable during the duration of any particular experiment(150 min) (Fig. 1).

View larger version (15K):
[in this window]
[in a new window]
FIG. 1. Relative recovery by loss of imipenem (%) in blood, muscle, and lung in a representative rat obtained after probe perfusion with imipenem solution (10 µg · ml^–1) at a flow rate of 1 µl · min^–1 for 150 min.

Imipenem distribution study. (i) Nonhydrated group. The decay of free concentrations in blood, muscle, and lungtissue with time were monoexponential; and the concentrationprofiles in these three media were almost superimposed (Fig.2), with the concentrations at the end of infusion equal to253 ± 27, 292 ± 25, and 285 ± 80 µg· ml^–1, respectively. Greater between-rat variabilitywas observed in lung tissue than in muscle or blood. The valuesof the pharmacokinetic parameters obtained for the nonhydratedgroup are presented in Table 1. Half-lives and AUCs were notstatistically different between blood, muscle, and lung, withthe AUC ratios being close to 1 (Table 1).

View larger version (20K):
[in this window]
[in a new window]
FIG. 2. Unbound imipenem concentrations in nonhydrated rats (n = 4) after a 30-min intravenous infusion of 120 mg · kg^–1 of imipenem: superimposed mean concentrations in the three media (blood, muscle, and lung) (a), mean ± SD concentrations in blood (b), mean ± SD concentrations in muscle (c), and mean ± SD concentrations in lung (d).

View this table:
[in this window]
[in a new window]
TABLE 1. Values of pharmacokinetic parameters obtained for blood, muscle, and lung of nonhydrated rats (n = 4) after a 30-min intravenous infusion of imipenem at a dose of 120 mg · kg^–1

(ii) Hydrated group. The decay of free concentrations in blood, muscle, and lungtissue with time were monoexponential; and the concentrationprofiles in these three media were also virtually superimposed(Fig. 3). The values of the pharmacokinetic parameters obtainedin the hydrated group are presented in Table 2. Half-lives andAUCs were not statistically different between blood, muscle,and lung tissue, with AUC ratios still being close to 1 (Table2). However, the AUCs and half-lives of IPM in each tissue weresignificantly lower in the hydrated group than in the nonhydratedgroup (Table 1 and Table 2). CL was increased by a factor of1.3 between the hydrated group and the nonhydrated group (10.4± 0.6 ml · min^–1 kg^–1 versus 8.0 ±0.9 ml · min^–1 kg^–1), and this differencewas statistically significant. V was not significantly differentbetween the two groups (326 ± 35 ml · kg^–1in the nonhydrated group versus 330 ± 25 ml ·kg^–1 in the hydrated group).

View larger version (18K):
[in this window]
[in a new window]
FIG. 3. Unbound imipenem concentrations in hydrated rats (n = 6) after a 30-min intravenous infusion of 120 mg · kg^–1 of imipenem: superimposed mean concentrations in the three media (blood, muscle, and lung) (a), mean ± SD concentrations in blood (b), mean ± SD concentrations in muscle (c), and mean ± SD concentrations in lung (d).

View this table:
[in this window]
[in a new window]
TABLE 2. Values of pharmacokinetic parameters obtained for blood, muscle, and lung of hydrated rats (n = 6) after a 30-min intravenous infusion of imipenem at a dose of 120 mg · kg^–1

DISCUSSION

Top

Discussion

Numerous studies have recently been conducted with microdialysisto investigate tissue antibiotic distributions in rats (7, 11,19, 22, 25). However, although the unbound concentrations weredetermined in the tissue interstitial space by using microdialysis,the total concentrations were measured in plasma, followed bycorrection with an averaged unbound fraction value. Intravascularmicrodialysis is a technique of choice for the direct determinationof free drug concentrations in blood (15, 31). It does not requirerepeated blood sampling, which eventually alters hematocritand plasma protein concentrations and, therefore, possibly,drug disposition characteristics (10), which was preferred forthis study.

Only two articles on the kinetics of antibiotics in rat lungsdetermined by microdialysis have been published: one with cefaclorand CMA/10 probes (11) and the other with cefpodoxime and CMA/20probes (22). Our attempts with these probes were not successfulbecause of tissue damage, which was suppressed because of theuse of a 5-mm linear flexible probe. Yet, under these conditionsIPM recovery in lung tissue ranged from 5.03% ± 0.05%to 16.8% ± 0.2%, which is much less than that in muscle(19.9% ± 3.2% to 53.4% ± 2.8%) or blood (21.9%± 2.1% to 55.9% ± 1.1%) because of the reducedlength and/or the different polymer type of the probe. As aconsequence, a small experimental error in recovery estimateresults in a relatively larger error in drug concentration estimates(4), which is probably responsible for the greater interanimalvariability observed in lung tissue than in the other media(Fig. 2 and Fig. 3).

Lung microdialysis required maintenance of the rats under anesthesiawith a thoracotomy for the duration of the study (about 7 h),which could be responsible for the altered values of the pharmacokineticparameters compared to the values reported previously by ourgroup for awake rats (12, 20). A clearance reduction by halfwas observed in the initial group of nonhydrated rats (8.0 ±0.9 ml · min^–1 · kg^–1 versus 16.4± 1.1 ml · min^–1 · kg^–1), resultingin the impaired renal excretion of IPM. A likely contributionto the clearance reduction was dehydration. It was thereforedecided to continue the experiment with rats receiving an intravascularinfusion of saline to prevent dehydration. A flow rate of 1.2ml · h^–1, adapted from the literature (27), wasselected to account for the fluid volume injected (5 ml) duringimipenem perfusion. The IPM clearance in anesthetized and hydratedrats (10.4 ± 0.6 ml · min^–1 · kg^–1)was closer to but still lower than that previously reportedin awake rats (12, 20). The anesthetic agent may therefore havealso contributed to the reduced IPM clearance, as isofluraneinhalation has been shown to induce a significant and dose-dependentdecrease in blood pressure and cardiac output in dogs (2), with,presumably, decreases in renal blood flow and glomerular filtration,which could explain the reduced IPM clearance in isoflurane-anesthetizedrats. However, by reducing the left ventricular stroke volumeand, consequently, cardiac output, mechanical ventilation mayalso have contributed to this observation (16, 23). As opposedto elimination, the IPM distribution was not altered in theanesthetized rats. Carbapenem antibiotics are usually consideredto be distributed within the extracellular body water (24).Accordingly, the IPM volume of distribution in rats was previouslyfound (12, 20) to be almost equal to the mean value (297 ml· kg^–1) of the extracellular fluid in rats (8).The lack of volume change in anesthetized rats suggests thatthe extracellular fluid volume was not altered, even in nonrehydratedanimals.

The major finding of this study was the observation of virtuallysuperimposed free IPM concentration-versus-time profiles inthe three media investigated, with the values of the pharmacokineticparameters, in particular, those of C_max and AUC, being veryclose to each other and with none being statistically significantlydifferent for both the nonhydrated (Table 1) and the hydrated(Table 2) groups. This result not only is in agreement withtheory but also is consistent with most of the data in the literature.The first study that used microdialysis to investigate the tissuedistributions of ß-lactam antibiotics in rats wasconducted in 1992 by Deguchi and coworkers (9). Although probesrecoveries were not estimated according to current standards,that study demonstrated that total tissue levels could varyby more than 100-fold with the type of tissue but that the freelevels of drug determined in the interstitial fluid were similarfor the various tissues studied (9). In other studies that haveused microdialysis to measure free antibiotic concentrationsin peripheral tissues, total plasma concentrations were presented,but after correction for plasma protein binding, the concentrationsmeasured in these various media should be more or less identical(19, 25).

Yet, similar muscle and lung interstitial fluid concentrationsprofiles were observed with cefaclor (11) and cefpodoxime (22),but the free tissue concentrations were much less than the correspondingunbound plasma concentrations, in particular, with cefaclor,as assessed by the use of a penetration factor that correspondedto the unbound drug concentration in tissue-to-plasma AUC ratioof 0.26 (11). Because cefaclor is rather unstable, chemicaldegradation in muscle was a potential explanation for this lowAUC ratio and apparently limited the tissue distribution. However,some of these authors found that a similar type of AUC ratioworked with the much more stable compound norfloxacin (14).Therefore, the existence of an unidentified active efflux transportsystem from tissue is another potential explanation, which wasalso proposed to explain the limited distribution of amoxicillinin the middle ears of chinchillas (15).

Another important parameter to be considered is the infectionitself, since the unbound interstitial fluid concentrationsof IPM were reduced in critical care patients compared to thosein healthy volunteers (28), suggesting that infection was responsiblefor the impairment of the IPM tissue distribution. Similarly,the free concentrations of meropenem, another carbapenem antibioticstructurally related to IPM, were lower in infected lung thanin presumably noninfected muscle (30).

In conclusion, intravascular microdialysis together with microdialysisof muscle and lung proved useful for investigation of the distributionof IPM and for demonstration that, at least in noninfected rats,the free concentrations in the interstitial fluid of tissueswere almost identical and could be predicted from the correspondingfree blood levels, in agreement with theory.

ACKNOWLEDGMENTS

We thank Pharsight Corporation for supplying WinNonLin throughthe PAL program.

FOOTNOTES

* Corresponding author. Mailing address: EA 3809, Pôle Biologie Santé (PBS), Médecine-Sud, Niveau 1, 40 Avenue du Recteur Pineau, 86022 Poitiers Cedex, France. Phone: 33-5-49-45-43-79. Fax: 33-5-49-45-43-78. E-mail: william.couet{at}univ-poitiers.fr.

C.D. and S.M. contributed equally to this paper.

REFERENCES

Top