Lack of Effect of Experimental Hypovolemia on Imipenem Muscle Distribution in Rats Assessed by Microdialysis

The aim of this study was to investigate the influence of hypovolemiaon the distribution of imipenem in muscle extracellular fluiddetermined by microdialysis in awake rats. Microdialysis probeswere inserted into the jugular vein and hind leg muscle. Imipenemrecoveries in muscle and blood were determined in each rat byretrodialysis by drug before drug administration. Hypovolemiawas induced by removing 40% of the initial blood volume over30 min. Imipenem was infused intravenously at a dose of 70 mg· kg^–1 over 30 min, and microdialysis samples werecollected for 120 min from hypovolemic (n = 8) and control (n= 8) rats. The decay of the free concentrations in blood andmuscle with time were monoexponential, and the concentrationprofiles in muscle and blood were virtually superimposed inboth groups. Accordingly, the ratios of the area under the concentration-timecurve (AUC) for tissue (muscle) to the AUC for blood were alwaysvirtually equal to 1. Hypovolemia induced a 23% decrease inthe clearance (P < 0.05) of imipenem, with no statisticallysignificant alteration of its volume of distribution. This studyshowed that imipenem elimination was altered in hypovolemicrats, probably due to decreased renal blood flow, but its distributioncharacteristics were not. In particular, free imipenem concentrationsin blood and muscle were always virtually identical.

INTRODUCTION

Top

Introduction

Nosocomial infections are of primary concern in critical caredepartments; and because hospitalized patients suffer from majorphysiological alterations with potential consequences on drugpharmacokinetics, selection of the appropriate antibiotic dosingregimens appears to be very challenging (28). Because infectionsoccur in tissues, measurement of free antibiotic concentrationsin the interstitial fluid of tissues should be the most relevantfor prediction of therapeutic efficacy, at least for extracellularpathogens. Microdialysis is an elegant technique that allowssuch determinations (6) and that has been used to demonstratethat the tissue distributions of several antibiotics, includingpiperacillin (2, 15), cefpirome (16), meropenem (32), and imipenem(IPM) (31), were impaired in critical care patients. However,not only the disease state itself but also iatrogenic procedures,such as surgery or intensive care treatment, might influencedrug kinetics (2, 17); and therefore, the reasons for the altereddistribution are not clear. Yet, it was hypothesized that reducedtissue perfusion contributes to the reduced IPM distribution(31).

Experiments with animals allow better control of these multipleinterfering parameters, and many microdialysis studies havealso been conducted with rats to investigate the distributionof amino ß-lactams in muscle (4, 7, 18, 20, 21, 22,24, 25). Muscle tissue has frequently been chosen because itis relatively accessible and the concentrations in muscle tissueseem to be a reasonable predictor of the unbound concentrationsin more therapeutically relevant tissues, such as lung tissue(7, 20, 21). A microdialysis study conducted by our group withhealthy rats has recently shown that free IPM concentrationsin muscle tissue, lung tissue, and blood are virtually identicalat any time postdosing (24).

The objective of the present study was therefore to investigatethe effect of hypovolemia on the pharmacokinetics of IPM inan experimental rat model, with special focus on its distribution,as assessed by determination of the usual pharmacokinetic parametersderived from blood concentration measurements and by comparisonof the blood and muscle concentrations determined by microdialysis.

MATERIALS AND METHODS

Top

Materials and Methods

Chemicals. IPM monohydrate-sodium cilastatin salt (Tienam; Merck Sharp& Dohme-Chibret Laboratories, Paris, France) was used toprepare IPM solutions in 0.9% NaCl and Ringer solution for intravenousadministration and probe perfusion, respectively. All chemicalsused were of analytical grade, and the solvents were high-performanceliquid chromatography (HPLC) grade.

Animals. Sixteen male Sprague-Dawley rats from Janvier Laboratories (LeGenest-St-Isle, France) weighing 294 ± 20 g (mean ±standard deviation [SD]) were used for the pharmacokinetic experimentsand were divided into two groups: a hypovolemic group (n = 8)and a control group (n = 8). Three extra rats weighing 309 ±5 g were used to evaluate the impact of hypovolemia on IPM recovery.All animals were acclimatized in wire cages in a 12-h lightand 12-h dark cycle for a minimum of 5 days before the beginningof the experiment to allow them to adjust to the new environment.During this period, they had free access to food (A03; Safe,Villemoisson-sur-Orge, France) and water. This work was donein accordance with the Principles of Laboratory Animals Care(25a).

Catheter and probe insertion. On the day before the experiment, the rats were anesthetizedby isoflurane (Forene, Abbot, Rungis, France) inhalation (23,24). Polyethylene cannulas were inserted into the left femoralvein and artery for drug administration and blood sampling,and two CMA/20 probes (polycarbonate; membrane length, 10 mm;CMA microdialysis; Phymep, Paris, France) were inserted intothe right jugular vein and the right hind leg muscle as describedpreviously (24). The rats were allowed to recover to a consciousstate. During the entire night, the muscle probe was perfusedwith blank Ringer solution at a flow rate of 0.5 µl ·min^–1. Food was withdrawn approximately 12 h before theexperiment, but the animals had free access to water until thebeginning of the experiment.

Impact of hypovolemia on IPM recovery. The experiment for determination of the impact of hypovolemiaon IPM recovery was started with retrodialysis by drug period,during which the probes were perfused (CMA 100 microdialysispump; Phymep) for 60 min at 2 µl · min^–1with Ringer solution containing IPM (10 µg · ml^–1)to equilibrate the system. After this equilibration period,microdialysate samples were collected automatically with a CMA/140Microfraction collector (CMA microdialysis; Phymep) for 50 minby fractions corresponding to 10-min intervals. This was followedby two periods of 30 min, one for hypovolemia induction andthe other for stabilization, during which probe perfusion wasmaintained. Then, during a second collection period, sampleswere collected over 110 min. To determine in vivo recoveriesby loss, the IPM concentrations in the perfusate (C_in) and dialysates(C_out) were determined by HPLC. The in vivo recovery by loss(RL_{in vivo}) was calculated for each interval of time and isexpressed as a percentage (1):

(1)

Pharmacokinetic study. (i) Recovery calculations. On the day of the experiment, the experiment started with retrodialysisby drug treatment period, during which the probes were perfused(CMA 100 microdialysis pump; Phymep) at 2 µl ·min^–1 with Ringer solution containing IPM (10 µg· ml^–1) for 60 min to equilibrate the system. Afterthis equilibration period, microdialysate samples were collectedautomatically with a CMA/140 microfraction collector (CMA microdialysis;Phymep) for 40 min in fractions corresponding to 10-min intervals.In vivo recovery by loss was calculated for each interval oftime as described above (equation 1). The mean value obtainedfrom the four individual determinations was calculated to correctthe measured dialysate concentrations. A washout period of 60min with a blank Ringer solution perfusion was allowed beforethe start of intravenous IPM administration to remove IPM fromthe probes. The flow rate was maintained at 2 µl ·min^–1 for the remainder of the study. Directly after collection,dialysates were diluted 1/3 (vol/vol) with a stabilizer (0.5M HEPES buffer, pH 6.8; ethylene glycol; HPLC-grade water [1:0.5:0.5;vol/vol/vol]).

(ii) Hypovolemia induction and determination of plasma total protein concentrations. During the washout period, hypovolemia was induced in rats (n= 8) by removing 40% of the initial blood volume (assumed tobe 60 ml · kg^–1) (5) in six increments (increasingvolumes) over 30 min. The first increment was used to measurethe total protein concentration before hypovolemia. Two extraarterial blood samples (about 0.5 ml, with compensation) werewithdrawn to measure total proteins immediately before IPM infusionand at the end of the experiment. These three blood sampleswere collected in heparinized tubes and were then centrifugedat 3,000 rpm for 10 min at 4°C. The supernatants were keptfrozen at –20°C until protein determinations.

(iii) Imipenem administration. Thirty minutes after removal of the last blood sample, an intravenousinfusion of IPM (Tienam) at a dose of 70 mg · kg^–1and an infusion rate of 3 ml · h^–1 was performedover a 30-min period, which corresponded to an injection volumeof 1.5 ml.

(iv) Microdialysis experiment. In each group, 10 min after the beginning of intravenous IPMadministration (which was the time needed to flush the deadvolume), the dialysates in muscle tissue and blood were collectedover 120 min at 5-min intervals during the first 50 min andat 10-min intervals during the last 70 min. Directly after collection,the microdialysis dialysates were diluted (1:3 [vol/vol] fordialysates collected at 5-min intervals and 1:2 [vol/vol] forthe 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]).

IPM analysis. Analysis of IPM in the dialysates was performed by an adaptationof a previously described HPLC assay (11, 12, 24). Standardcurves were prepared with IPM in water at concentrations rangingfrom 0.78 to 50 µg · ml^–1. The diluted microdialysateswere directly injected onto a Nucleosil C₈ column (250 by 0.4mm [inner diameter]; Interchim, Monluçon, France) inassociation with a Nucleosil C₈ precard (5 µm; Interchim).The chromatographic system consisted of a Shimadzu (CroissyBeaubourg, France) LC-10AT pump and a Waters 717 plus autosamplerat 4°C connected to a UV detector (SPD 10A; Shimadzu) at313 nm. Data were recorded and analyzed on a Kromasystem integrator(Bio-Tek, St. Quentin en Yvelines, France). The mobile phaseconsisted of 0.2 M aqueous borate buffer, pH 7.2, containing10% (vol/vol) methanol, and the flow rate was 1 ml ·min^–1. The retention time of IPM was equal to 5.5 min.The between-day variability of IPM was characterized at 5 µg· ml^–1, with a coefficient of variation equal to7.5% (n = 34) and an accuracy equal to 98.5% (n = 34).

Noncompartmental pharmacokinetic analysis. Pharmacokinetic parameters were determined for 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 also estimated from CL/k_el,blood. The AUC and t_1/2 formuscle tissue were also estimated by the same procedure andare referred as AUC_muscle and t_1/2,muscle, respectively. Notethat although microdialysis allows determination of the unbounddrug concentrations in plasma water or tissue extracellularfluid, for simplicity, the subscripts (blood or muscle) referto 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) obtainedin blood and muscle in the same rat were performed by a pairedt test, with significance being a P level of <0.05. Comparisonsof AUC, t_1/2, C_max, CL, V, and the ratio of AUC_muscle/AUC_bloodbetween the two groups were performed by an unpaired t test,with significance being a P level of <0.05. Comparisons ofplasma total protein concentrations before and after hypovolemiawere performed by a paired t test, with significance being aP level of <0.05.

RESULTS

Top

Results

Impact of hypovolemia on IPM recovery. Hypovolemia had no apparent effect on IPM recovery by loss,which was essentially constant throughout the 220-min studyduration, as illustrated in Fig. 1 for a representative rat.

View larger version (17K):
[in this window]
[in a new window]
FIG. 1. Relative recovery by loss of IPM (in percent) in blood and muscle tissue estimated before and after hypovolemia after probe perfusion with IPM solution (10 µg · ml^–1) at a flow rate of 2 µl · min^–1 in a representative rat.

Pharmacokinetic study. Plasma total protein concentrations were constant throughoutthe study duration in control rats (Table 1), and hypovolemiainduced a statistically significant reduction in plasma totalprotein concentrations, which then remained unchanged for therest of the study (Table 1).

View this table:
[in this window]
[in a new window]
TABLE 1. Total plasma protein concentrations in control rats at the beginning and at the end of the experiment and in hypovolemic rats before the induction of hypovolemia, 30 min after the end of hypovolemia, and at the end of the experiment

RL_{in vivo} varied between 16.1% ± 1.7% and 40.8% ±5.7% in blood and between 15.7% ± 0.8% and 35.2% ±4.1% in muscle tissue.

Control group. The decay of the free IPM concentrations in blood and muscletissue with time were monoexponential, and the concentrationprofiles were almost superimposed (Fig. 2a ₁). The concentrationsat the end of the infusion were equal to 115 ± 23 and111 ± 14 µg · ml^–1 in blood and muscletissue, respectively (P > 0.05) (Fig. 2b₁ and c₁). The valuesof the pharmacokinetic parameters obtained for the control groupare presented in Table 2. Half-lives and AUCs were not statisticallydifferent between blood and muscle, and AUC ratios were closeto 1 (Table 2).

View larger version (27K):
[in this window]
[in a new window]
FIG. 2. Unbound IPM concentrations in control rats (n = 8) (left panels) and hypovolemic rats (n = 8) (right panels) after a 30-min intravenous infusion of 70 mg · kg^–1 of IPM: superimposed mean concentrations in blood (full circles) and muscle tissue (open circles) in control rats (a₁) and hypovolemic rats (a₂), mean ± SD concentrations in blood in control rats (b₁) and hypovolemic rats (b₂), and mean ± SD concentrations in muscle tissue in control (c₁) and hypovolemic rats (c₂).

View this table:
[in this window]
[in a new window]
TABLE 2. Values of pharmacokinetic parameters obtained in blood and muscle tissue of control rats and hypovolemic rats after a 30-min intravenous infusion of imipenem at a dose of 70 mg · kg^–1

Hypovolemic group. The decay of free IPM concentrations in blood and muscle tissuewith time were monoexponential, and the concentration profileswere also virtually superimposed (Fig. 2a₂), with the concentrationsat the end of infusion not being statistically different, andwere equal to 132 ± 23 and 125 ± 22 µg ·ml^–1 for blood and muscle tissue, respectively (Fig. 2b₂and c₂). The values of the pharmacokinetic parameters obtainedfor the hypovolemic group are presented in Table 2. Half-livesand AUCs were not statistically different between blood andmuscle tissue, with AUC ratios still being close to 1 (Table2). However, the AUCs and half-lives of IPM in each tissue weresignificantly higher in the hypovolemic group than in the controlgroup (Table 2). Clearance was significantly decreased by 23%on average in the hypovolemic group compared to that in thecontrol group (11.4 ± 2.1 ml · min^–1 kg^–1versus 14.8 ± 2.4 ml · min^–1 kg^–1).The volume of distribution was not significantly different betweengroups (318 ± 48 ml · kg^–1 in the hypovolemicgroup and 325 ± 67 ml · kg^–1 in the controlgroup).

DISCUSSION

Top

Discussion

In pharmacokinetics, passive bidirectional diffusion is usuallyconsidered to govern the drug distribution within most softtissues, which means that at equilibrium the unbound concentrationsin blood and tissue extracellular fluids (ECFs), and therefore,the corresponding AUCs, should be identical, in theory (29).In fact, this has been confirmed experimentally on many occasions,in particular, during experimental microdialysis studies conductedby several groups with rats and various antibiotics, includingpiperacillin (4, 26); ceftriaxone (18); pazufloxacin, ciprofloxacin,ofloxacin, and ceftazidime (1); imipenem (24); and amoxicillin(25). For similar reasons, ECF concentrations should be identicalin various soft tissues, which in fact constitutes the rationalefor conducting microdialysis studies with muscle tissue, whichis more easily accessible, although possibly less relevant,than lung tissue. Yet, similar free concentrations in bloodand soft tissues have not always been observed experimentally(7, 14, 20, 21). Several phenomena can be proposed to explainthese observations. One is the presence of active efflux transportsystems, which have been proposed as an explanation for thelimited muscle tissue distribution of norfloxacin (14). Suchsystems are found in specific tissues or organs, such as thebrain, and are responsible for tissue AUC-to-blood AUC ratiosless than unity (3, 23); they are also equal to the correspondinginflux distribution-to-efflux distribution clearance ratios(3, 33). However, there is no evidence to support the presenceof such systems in soft tissues, such as muscle tissue. Alterationof local blood flow, capillary density, and capillary permeabilitywere also mentioned as explanations for the cefpodoxime tissuedistribution factor, which ranges from 0.5 to 0.7 (20, 21).Finally, drug degradation within tissue was proposed as an explanationfor the limited tissue distribution of cefaclor (7). Interestingly,it should be kept in mind that this is not related to distribution.In relation to this explanation is the fact that drug disappearancefrom a physiological liquid such as cerebrospinal fluid dueto turnover could also contribute to AUC ratios less than unity(3, 27). Yet, one should keep in mind that such situations shouldbe the exception, since, according to basic pharmacokineticconcepts, peripheral ECF tissue AUC-to-blood AUC ratios shouldbe equal to unity, as a general rule (29).

In order to address these issues under optimal conditions, wehave recently combined microdialysis in both blood and tissuesto measure free antibiotic concentrations (24). The advantagesof this approach are that it allows direct determination offree blood concentrations but also allows multiple concentrationdeterminations without blood sampling, which is particularlyimportant when blood volume must be controlled, as was the casewith this experimental model of hypovolemia. Using this approach,we have recently demonstrated that free IPM concentrations andAUCs in blood and muscle tissue, as well as lung ECFs of rats(24), are virtually similar, which was confirmed during thepresent study, at least for blood and muscle tissue. Yet, thisfinding contradicts the results of Tegeder et al. (31), whoreported a mean total plasma AUC-to-free tissue AUC ratio of2 for healthy subjects, which corresponds to a free tissue AUC-to-totalplasma AUC ratio of 0.5. When it is considered that all of theimipenem was unbound to protein (30), a virtually similar valuewould be obtained for the free tissue AUC-to-free blood AUCratio. Yet, there is no explanation for such a major discrepancy.

The experimental model of hypovolemia used throughout this studywith removal of 40% of the blood volume, which correspondedto about 5 ml of blood, was adapted from the model of de Paepeet al. (8, 9, 10). With the initial model, 30% of the bloodvolume was removed. However, this was considered a model ofmild hypovolemia, and based upon our previous experience (19),we decided to remove 40% of the blood volume for the presentstudy. Under these conditions, a statistically significant reductionof IPM clearance was observed, most likely due to renal bloodflow impairment. However, hypovolemia had no effect on the IPMvolume of distribution (Table 2), which differed from our previousfindings (19). Yet, a much higher IPM dose was administeredin the previous study, and this led to sustained seizures, whichcould have had an effect on the total body water content and,therefore, the IPM distribution. Like other amino-ß-lactamantibiotics, IPM distributes within the total extracellularbody water, which is estimated to be 297 ml · kg^–1in rats (5). Accordingly, the IPM volume of distribution estimate(either total or unbound, since protein binding is limited)was always consistent with that value and, in particular, wasconsistent with that value in the present study (Table 2). Becausethe intravascular water content corresponds to about 10% ofthe total extracellular body water content, removal of 40% ofthe blood volume would initially reduce the total extracellularbody water content by only 4%, which would hardly be detectable.Accordingly, no alteration in the IPM volume of distributioncould be observed in hypovolemic rats.

As another potential effect on drug distribution, blood removalshould lead to shifts in volume between the intravascular andthe extravascular spaces. These redistribution processes havebeen proposed as a possible explanation for the altered volumeof distribution of piperacillin in critical care patients (2).However, volume shifts alone should have an effect on the relativeamounts of IPM present within the intravascular and extravascularspaces but not on the unbound drug concentrations, which atequilibrium should be equal within the intravascular and extravascularspaces, independently of the corresponding volumes (29). Itwas therefore important to confirm experimentally this theoreticalexpectation by using microdialysis and by demonstrating thatfree IPM concentrations in muscle tissue ECFs and blood werealways similar, meaning that hypovolemia did not result in lowertissue antibiotic concentrations, as it could have been suggestedin a previous study, in which it was inappropriately attributedto altered blood flow (31).

Yet, altered tissue blood flow in nonvital organs, such as muscletissue, which could typically be reduced by up to 50% duringhypovolemia (13), constitutes another factor with a potentialeffect of drug distribution. The rate of distribution of IPMin muscle tissue ECF is more likely limited by tissue bloodflow rather than membrane permeability, just as the converseis probably true in brain tissue (29). Then, the time to thepeak concentration in soft tissues may potentially be delayedduring hypovolemia. However, the monoexponetial decay of bloodIPM concentrations with time, together with the rapid peak occurrencein muscle and lung tissues (24), suggests that this distributionis virtually instantaneous and is unlikely to be noticeablydelayed by reduced blood flow, as has recently been documentedwith amoxicillin (25).

In conclusion, the present study has demonstrated that the distributionof IPM in soft tissues was not noticeably altered by using anexperimental model of hypovolemia in rats, whereas its eliminationcharacteristics were modified. More importantly, free IPM AUCsdetermined in blood and muscle tissue ECF were identical inboth control and hypovolemic rats, in agreement with basic pharmacokineticconcepts.

ACKNOWLEDGMENTS

We thank Pharsight Corporation for the free supply of WinNonLinthrough the PAL program.

We thank Brunet Bertrand (Laboratoire de Biochimie, CHU Poitiers)for his assistance during protein determination.

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.

REFERENCES

Top