Infect Dis Clin N Am 17 (2003) 599–613

Antimicrobial tissue concentrations Ping Liu, PhD, Hartmut Derendorf, PhD* Department of Pharmaceutics, College of Pharmacy, University of Florida, 1600 SW Archer Road, PO Box 100494, Gainesville, FL 32610, USA

In the past, pharmacokinetic (PK) and pharmacodynamic (PD) assessment of antimicrobial agents was based mostly on measuring plasma or serum concentrations. More attention has been given to the respective tissue concentrations, however, because most infections occur at tissue sites. Moreover, most bacterial pathogens are extracellular, and it has been realized that only unbound antibiotic concentrations at the infected sites are responsible for antibacterial activity. Insufficiently high free tissue concentrations of antibiotics may provide an explanation for some of the clinical failures in which the antibiotics showed desired plasma PK profiles and in vitro sensitive susceptibility to targeted pathogens. So far, the optimal dosing regimens for antimicrobial agents are still poorly defined because the treatment of bacterial infections is frequently based on clinical experience rather than a rational scientific approach. The clinical outcome of anti-infection treatment, however, is determined by both PK and PD properties of an antibiotic. A PK-PD link allows one to relate the PK properties and their antibacterial activity of the antimicrobial agents. Currently, the most common PK-PD approaches for anti-infective agents rely on plasma concentration as the PK input value and in vitro minimum inhibitory concentration (MIC) as the PD input value, and several PK-PD indices have been used extensively for making dosing decisions [1]. For instance, the time above MIC is proposed for b-lactams, AUC24-MIC is proposed for quinolones, and Cmax-MIC is proposed for aminoglycosides. The use of both total plasma concentrations and in vitro MIC values, however, is not ideal. Relating the PKs of antimicrobial plasma concentrations to its MIC values seriously compromises the PK-PD link. This article focuses mainly on the application of antimicrobial free tissue concentrations as PK target for anti-infective therapy.

* Corresponding author. E-mail address: [email protected]fl.edu (H. Derendorf). 0891-5520/03/$ - see front matter Ó 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0891-5520(03)00060-6

600

P. Liu, H. Derendorf / Infect Dis Clin N Am 17 (2003) 599–613

Plasma concentrations versus tissue concentrations Previously, because of the limited sampling techniques, plasma concentrations were monitored routinely and used as a surrogate marker for pharmacologic effect in most PK studies and clinical practice. With several recent breakthroughs in tissue sampling techniques, more tissue data are available and it is possible to obtain a clear picture of drug tissue distribution. Still, several issues, such as cost and convenience of the techniques, limited their applications. This is the major reason that plasma concentration measurement is still dominant in PK studies. Because most infections do not occur in plasma but rather in tissue sites (extracellular fluid), the ability of antibiotics to reach the target sites is a key determinant of clinical outcome. It is very important to realize that only free (unbound) antibiotic concentrations in the interstitial fluid at the target site are responsible for the antibacterial activity. Free antimicrobial tissue concentrations are more relevant than plasma concentrations in predicting therapeutic efficacy. Tissues are not homogenous compartments. The distribution of drug molecules in plasma and tissue depends on their physicochemical properties (Fig. 1). In plasma, some drug molecules bind to plasma proteins or blood cells, or diffuse into the blood cells, and some also remain unbound in plasma and can move freely in the body. A similar scenario also occurs in the tissue. Some drug molecules bind to tissue proteins or tissue cells, and some stay unbound in tissue fluid. The distribution of free drug in plasma and tissues reaches equilibrium. Theoretically, at steady state, free drug levels in plasma and tissues should be equal assuming that tissue distribution of drug molecules only depends on passive diffusion. Based on these assumptions, the total plasma

Fig. 1. Schematic diagram of drug distribution in plasma and tissues.

P. Liu, H. Derendorf / Infect Dis Clin N Am 17 (2003) 599–613

601

concentrations and plasma protein binding values can be used to predict the free tissue concentrations. This statement does not always hold true because many studies have shown the discrepancy between free drug levels in plasma versus tissue [2–5]. The difference between total plasma concentrations and free tissue concentrations can be significant in many situations, such as the drug with high or nonlinear plasma protein binding. Tissue penetration and tissue metabolism also affect the PK profiles of the antibiotics in the tissues. Total plasma concentration is not an ideal PK input value for rational dosing of antibiotics, and free antibiotic concentrations at the infection site should be considered. Protein binding Because protein binding plays an important role in drug distribution, many analytic techniques have been developed to determine the plasma protein binding of the compounds. There are several conventional methods including equilibrium dialysis, ultrafiltration, and ultracentrifugation, and chromatographic methods [6]. Equilibrium dialysis is the classic method, which is based on establishing equilibrium between a protein compartment and a buffer compartment that are separated by a dialysis membrane. Ultrafiltration produces a separation of the free drug from bound drug by using a pressure gradient that forces the small molecules through a semipermeable membrane. Ultrafiltration has been used widely to measure plasma protein binding of compounds in clinical laboratories because of its several advantages (eg, fast, simple, commercially available kits; lack of dilution effects and volume shifts; and so forth). The major concern of this method is the stability of the binding equilibrium during the separation process. Ultracentrifugation can avoid problems associated with membrane effects and can separate free drug from unbound drug in a natural environment. Possible floating lipoproteins in the supernatant are the major concern. Chromatographic methods (eg, affinity chromatography, size-exclusion chromatography, capillary electrophoresis, and fluorescence spectroscopy) could offer more precise and reproducible binding data compared with conventional methods. Cost, inconvenience, and timeconsuming properties keep chromatographic methods from becoming the routine method. There is a misconception that drugs in the same class with similar total plasma PK profiles have similar therapeutic effect. Only the unbound drug molecules are able to distribute throughout the body and only free drug concentrations at the target site are responsible for pharmacologic effect. Conclusions or judgments based on total plasma PK profiles could be misleading in many cases. It is understood that only the unbound drug molecules are able to distribute freely throughout the body. In many cases, free plasma levels are equal to free tissue levels after fast equilibrium, where

602

P. Liu, H. Derendorf / Infect Dis Clin N Am 17 (2003) 599–613

the protein binding values can be used to link total plasma levels with free tissue levels. A skin blister study of amoxicillin and flucloxacillin showed that total plasma concentrations of these two drugs were similar, whereas the simultaneously measured unbound concentrations in blister fluid showed significant difference because of different protein binding values [7]. When data from this study were fitted with a two-compartment PK model, free concentrations of both antibiotics in skin blisters were in agreement with the respective calculated free concentrations in the peripheral compartment, based on total plasma concentrations and protein binding values (Fig. 2) [8].

Fig. 2. Antibiotic concentrations in plasma (d) and blister fluid ( ) after intravenous administration of 1-g dose (mean  SD). The dashed lines are the free tissue concentrations predicted from a two-compartment pharmacokinetic model. (A) Amoxicillin. (B) Flucloxacillin. (From Derendorf H. Pharmacokinetic evaluation of beta-lactam antibiotics. J Antimicrob Chemother 1989;24:407–13; with permission.)

P. Liu, H. Derendorf / Infect Dis Clin N Am 17 (2003) 599–613

603

In this case, it is possible to perform indirect modeling of free tissue concentrations from total plasma concentrations. This indirect modeling of tissue concentrations from plasma concentrations relies on the assumptions that free plasma levels are equal to free tissue levels at equilibrium and that tissue distribution is fast. It cannot be applied unconditionally, however, because of the discrepancy between free concentrations in tissues versus plasma.

Issues in tissue distribution It is commonly believed that the unbound antibiotic concentrations in plasma and tissue fluids are equal at equilibrium assuming that the driving force of distribution is passive diffusion only [9]. Time to equilibrium may range from minutes to days, however, depending on the ratio of surface area of the capillary to volume of the tissue fluid compartment and the physicochemical properties of the compound [9]. It has been realized that tissue distribution is also affected by anatomic barriers, such as the bloodbrain barrier in the central nervous system, the eye, and the prostate gland [10]. The active transport system also plays an important role in tissue distribution; a typical example is P-glycoproteins in the central nervous system. Besides, anesthesia may affect tissue distribution because it is possible that the interaction between the anesthetic and the drug affect the disposition of the drug in the body. Also, it is possible that the physiologic function of the elimination organs (eg, liver, and kidney) decreases under the anesthesia condition. Many studies have shown impaired tissue penetration of antibiotics at different infection sites, such as epidermal infections [11], ear infections [12], tonsillitis [13], liver infections [14], urinary tract infections [15,16], and respiratory infections [2–5]. These phenomena cannot be explained by barrier mechanisms. One speculation is that structural resistance of the capillary wall is based on alterations in local blood flow, capillary density, capillary permeability, interstitial diffusion coefficients, and transcapillary osmotic pressure gradients [17–19]. It is also possible that active transporter plays a role. Some studies have shown impaired tissue penetration of antibiotics at normal tissue sites, such as soft tissue and lung [20,21]. Direct measurement of free tissue levels is necessary.

Techniques for free tissue concentration measurement In the last few decades, several techniques (eg, skin blisters, saliva, imaging techniques, and microdialysis) were used to monitor free drug concentrations in extracellular fluids in animal and human studies [22,23]. Traditionally, total tissue homogenate from biopsy was used to determine the tissue concentrations and the results from this approach offered

604

P. Liu, H. Derendorf / Infect Dis Clin N Am 17 (2003) 599–613

misleading information because a tissue homogenate represents a mixture of total and free drug in the tissue. Currently, in vivo imaging techniques and microdialysis make it possible to monitor directly the real-time course of free drug levels in the extracellular fluids in animals and humans. Skin blister and saliva techniques Skin blister and saliva techniques are traditional techniques that provide approximate information about free tissue concentrations. These two techniques are relatively simple in terms of requirement of technical skills and cost. Skin blister and saliva samples have been used as surrogate markers for free antibiotic tissue levels [24–28]. The skin blister technique has been used in many PK studies, although it has several major disadvantages. For instance, the blisters are manipulated by suction with pressure, which only mimics the extracellular fluid to a certain extent. Results showed that the sampling fluid is not completely free of proteins. Many factors (eg, the site of the blisters, the time and the pressure of the suction used) affect the composition of the blister fluid. The risk of inducing inflammation is high. Saliva sampling is a noninvasive method, but is suitable only for a few compounds. In most cases, results obtained from this technique are not reliable because there is a significant difference in the composition of saliva versus extracellular fluids in the tissue. Because saliva usually has a more acidic pH than blood, it usually shows high concentration of basic compounds and low concentration of acidic compounds. Microdialysis Previously, microdialysis has been used extensively in the neurosciences to monitor neurotransmitter release, and now is applied for various tissue PK studies. Microdialysis is a reliable sampling technique, which is able directly and continuously to monitor the free drug levels in different tissues and organs in both animals and humans [20,29–34]. Many endogenous compounds and a large variety of exogenous compounds have been studied. Valuable information about tissue distribution of many antibiotics is available [20,34]. The basic principle of microdialysis is to mimic the function of blood capillary by perfusing a thin dialysis probe implanted into the tissue with physiologic solution at a very low flow rate. The semipermeable probe tip excludes the large molecules from sampling and the dialysate (outlet of the solution) reflects the composition of the interstitial fluid over time. Based on the analytical technique, either average free tissue concentrations during fixed interval are measured or on-line analysis is performed. Mu¨ller’s group has conducted many microdialysis studies to monitor the pharmacokinetics of several antibiotics, such as ciprofloxacin [35],

P. Liu, H. Derendorf / Infect Dis Clin N Am 17 (2003) 599–613

605

moxifloxacin [36], fleroxacin [37], phenoxymethylpenicillin [37], dirithromycin [37], cefodizime [38], cefpirome [38], and piperacillin [39] in soft tissues of healthy volunteers. Also, the effects of surgery, intensive care procedures, and septicemia on peripheral distribution of piperacillin in patients [39,40] and the effect of inflammation on antibiotic penetration into foot lesions in diabetics [41] and dermatologic patients [37] were investigated. For instance, microdialysis study of phenoxymethylpenicillin in patients with cellulitis showed that there was no significant difference in the time course of free tissue levels in infected and noninfected dermis. Other groups have also studied the tissue PK profiles of different antibiotics in humans using microdialysis, such as gentamicin [42] and penciclovir (antiviral agent) [43]. The antibiotic drug penetration into the interstitial space fluid of human brain also was investigated [44]. The results of these studies further demonstrated that antimicrobial concentrations at the effected site may be subinhibitory, although effective concentrations are attained in plasma. In the authors’ group, microdialysis studies of several antibiotics (eg, piperacillin [45], ceftriaxone [46], piperacillin-tazobactam combinations [47], cefaclor [21,48], cefpodoxime, and cefixime [49]) in animals and humans have been performed. When piperacillin and ceftriaxone were compared, results showed that the difference between free muscle concentrations of ceftriaxone and its total plasma concentrations was much more significant than that of piperacillin (Fig. 3) [45,46]. These findings are consistent with their protein binding values (ceftriaxone: up to 98%; piperacillin: 40% to 50%). In both studies, a two-compartment PK model could fit the data well, and comparisons between calculated free concentrations in the peripheral compartment and measured free tissue concentrations revealed excellent agreement. This study also showed that rat is a suitable animal model for tissue penetration study of antibiotics. Orally available cephalosporins have been studied. A modified release formulation of cefaclor was evaluated for its therapeutic efficacy [21]. The PK profiles of 500 mg and 750 mg single doses of cefaclor in this formulation in plasma and muscle were evaluated in 12 healthy male subjects using microdialysis (Fig. 4). It showed that the oral absorption of cefaclor could be sustained, but only up to 3 hours because of the presence of an absorption window. Free muscle concentrations of cefaclor were sufficient for therapeutic levels, but lower than respective total and free plasma concentrations. The discrepancy between the free drug levels in plasma and soft tissue is probably caused by an active transport mechanism or metabolism of cefaclor at the tissue sites. The kinetic profiles of cefaclor in the muscle confirmed that this sustained-release formulation could have sufficient clinical efficacy for anti-infective treatment. In another study, cefpodoxime and cefixime with different protein binding values (about 25% and 65% for cefpodoxime and cefixime, respectively) have similar plasma kinetic profiles after oral administration. Using microdialysis, a direct comparison of the kinetic profiles of these two

606 P. Liu, H. Derendorf / Infect Dis Clin N Am 17 (2003) 599–613

Fig. 3. Plasma concentrations (n) and free, unbound concentrations in muscle (d) after intravenous administration of 120 mg/kg of piperacillin (left) and 50 mg/kg of ceftriaxone (right) in male Wistar rats (N = 6). The lines are the fitted curves based on a two-compartment pharmacokinetic model. Values are mean  SD (Data from references 45 and 46.)

P. Liu, H. Derendorf / Infect Dis Clin N Am 17 (2003) 599–613

Fig. 4. Concentrations of cefaclor in plasma (n) and free, unbound concentrations in muscle ( ) after given 500 mg of modified-release tablets (left) and 750 mg of modified-release tablets (right) in 12 healthy male volunteers. The lines are the fitted curves based on a two-compartment pharmacokinetic model. Values are mean  SD (From de la Pen˜a A, Brunner M, Eichler HG, Rehak E, Gross J, Thyroff-Friesinger U, et al. Comparative target site pharmacokinetics of immediate- and modified-release formulations of cefaclor in humans. J Clin Pharmacol 2002;42:403–11; with permission.)

607

608

P. Liu, H. Derendorf / Infect Dis Clin N Am 17 (2003) 599–613

cephalosporins in soft tissue given the same oral dose was conducted in six healthy male volunteers [49]. The ability to penetrate the tissue is a key determinant for clinical efficacy. Cefpodoxime with low protein binding of 25% had more than twice higher peak concentration in the muscle than cefixime (2.1 mg/L versus 0.9 mg/L) (Fig. 5). The average tissue penetration (AUCtissue,free /AUCplasma,total) of cefpodoxime (70%) was much higher than that of cefixime (29%). These results were consistent with their protein binding values, which indicate that the higher the protein binding, the lower the tissue penetration. The authors’ results suggest that cefpodoxime produces higher tissue levels than cefixime with the same dosing regimen for the treatment of uncomplicated soft tissue infections. Also, as with other antibiotic studies in humans, the free tissue levels of these two cephalosporins were lower than the respective total plasma levels. These findings confirmed that using total plasma concentrations overestimates the target site concentrations and likely their clinical efficacy. This might explain the therapeutic failure of some antibiotics, which have high in vitro antibacterial activity (expressed as low MIC values). The subinhibitory levels at the target site are one of the major reasons for many therapeutic failures of antiinfective treatment, and resistance development. Imaging techniques Several imaging techniques based on radiopharmaceuticals or nuclear magnetic drug effects have been developed for the study of drug distribution in humans, such as planar c-scintigraphy, single photon emission CT, positron emission tomography, and MR spectroscopy [50–55]. Imaging techniques are only applicable for a small group of compounds with special functional groups. The most significant advantage of microdialysis is that it can be applied to monitor a large variety of compounds in all kinds of tissues with relatively low cost. Compared with microdialysis, imaging techniques are very expensive and labor-intensive, which is not suitable for clinical routine settings.

Issues in minimum inhibitory concentration measurements The MIC is defined as the lowest antibiotic concentration allowing no visible bacterial growth after 20-hour incubation, and determined by macrodilution method using twofold dilutions. Obviously, antibacterial activity of an antibiotic is a dynamic process, whereas MIC is only a threshold value, a one-point measurement with poor precision. Although MIC is a good predictor of potency for antibiotics, it offers little information about the time course of the antibacterial activity of an antibiotic. The in vitro MIC values are determined in the presence of free antibiotic concentrations and the protein binding of the antibiotic is frequently not taken into account. Nath et al [56] studied the effect of plasma protein binding on antibacterial

P. Liu, H. Derendorf / Infect Dis Clin N Am 17 (2003) 599–613

Fig. 5. Plasma concentrations (m) and free, unbound concentrations in muscle (n) after 400-mg oral dose of cefpodoxime (left) and cefixime (right) in six healthy male volunteers. The dashed lines are the calculated free plasma concentrations based on average protein binding values (cefpodoxime, 25%; cefixime, 65%). Values are mean  SD (From Liu P, Muller M, Grant M, Webb AI, Obermann B, Derendorf H. Interstitial tissue concentrations of cefpodoxime. J Antimicrob Chemother 2002;50:19–22; with permission.)

609

610

P. Liu, H. Derendorf / Infect Dis Clin N Am 17 (2003) 599–613

activities of two antibiotics: ceftriaxone (protein binding: 90% to 95%) and cefotaxime (protein binding: 45% to 50%). Results confirmed that the free plasma concentration of an antibiotic has better correlation with its antibacterial activity. Using total plasma concentration to compare with MIC values overcomes the true subinhibitory levels at the infection site so that it causes treatment failure. Also, subinhibitory concentration at the infection site is one of major reasons for the emergence of resistance. Food and Drug Administration guidance in 1997 stated that PD studies should include relating the concentrations at the site of action to the in vitro susceptibility of the target microorganism [57]. To evaluate the antibacterial activity in this way, the free drug concentrations at the target tissue are appropriate PK input values for rational dosing of antibiotics. Summary The clinical outcome of anti-infective treatment is determined by both PK and PD properties of the antibiotic. Only the free tissue concentrations of antibiotics at the target site, which are usually lower than the total plasma concentrations, are responsible for therapeutic effect. The free antibiotic concentrations at the site of action are a more appropriate PK input value for PK-PD analysis. The unbound tissue concentrations can be measured directly by microdialysis. Using plasma concentrations overestimates the target site concentrations and its clinical efficacy. The optimal dosing regimens of antibiotics have an impact on patients’ outcome and cost of therapy, and reduce the emergence of resistance. References [1] Craig WA. Pharmacodynamics of antimicrobials: general concepts and applications. In: Nightingale CH, Murakawa T, Ambrose PG, editors. Antimicrobial pharmacodynamics in theory and clinical practice. New York: Marcel Decker; 2002. p. 1–23. [2] Nordbring F. Tissue penetration of antibiotics. Introduction. Focus on some problems involved in the treatment of infectious diseases. Scand J Infect Dis Suppl 1978;14:21–2. [3] Fischman AJ, Babich JW, Bonab AA, Alpert NM, Vincent J, Callahan RJ, et al. Pharmacokinetics of [18F]trovafloxacin in healthy human subjects studied with positron emission tomography. Antimicrob Agents Chemother 1998;42(8):2048–54. [4] Jain RK. The next frontier of molecular medicine: delivery of therapeutics. Nat Med 1998;4:655–7. [5] Heikkinen T, Laine K, Neuvonen PJ, Ekblad U. The transplacental transfer of the macrolide antibiotics erythromycin, roxithromycin and azithromycin. Br J Obstet Gynaecol 2000;107:770–5. [6] Oravcova J, Bohs B, Lindner W. Drug-protein binding sites: new trends in analytical and experimental methodology. J Chromatogr B Biomed Appl 1996;677:1–28. [7] Wise R, Gillett AP, Cadge B, Durham SR, Baker S. The influence of protein binding upon tissue fluid levels of six beta-lactam antibiotics. J Infect Dis 1980;142:77–82. [8] Derendorf H. Pharmacokinetic evaluation of beta-lactam antibiotics. J Antimicrob Chemother 1989;24:407–13.

P. Liu, H. Derendorf / Infect Dis Clin N Am 17 (2003) 599–613

611

[9] Barza M, Cuchural G. General principles of antibiotic tissue penetration. J Antimicrob Chemother 1985;15(suppl A):59–75. [10] Barza M. Anatomical barriers for antimicrobial agents. Eur J Clin Microbiol Infect Dis 1993;12(suppl 1):S31–5. [11] Mertin D, Lippold BC. In-vitro permeability of the human nail and of a keratin membrane from bovine hooves: penetration of chloramphenicol from lipophilic vehicles and a nail lacquer. J Pharm Pharmacol 1997;49:241–5. [12] Harrison CJ, Belhorn TH. Antibiotic treatment failures in acute otitis media. Pediatr Ann 1991;20:600–1603–8. [13] Savolainen S, Mannisto PT, Gordin A, Antikainen R, Haataja H, Tuominen RK, et al. Tonsillar penetration of erythromycin and its 2’-acetyl ester in patients with chronic tonsillitis. J Antimicrob Chemother 1988;21(suppl D):73–84. [14] Schentag JJ, Heller AS, Hardy BG, Wels PB. Antibiotic penetration in liver infection: a case of tobramycin failure responsive to moxalactam. Am J Gastroenterol 1983;78:641–4. [15] Naber KG. Renal lymph concentrations of antibiotics. Scand J Infect Dis Suppl 1978;14:164–5. [16] Schierholz JM, Beuth J, Pulverer G. ‘‘Difficult to treat infections’’ pharmacokinetic and pharmacodynamic factors–a review. Acta Microbiol Immunol Hung 2000;47:1–8. [17] Renkin EM. Multiple pathways of capillary permeability. Circ Res 1977;41:735–43. [18] Michel CC. Filtration coefficients and osmotic reflexion coefficients of the walls of single frog mesenteric capillaries. J Physiol 1980;309:341–55. [19] Curry FE. Determinants of capillary permeability: a review of mechanisms based on single capillary studies in the frog. Circ Res 1986;59:367–80. [20] de la Pen˜a A, Liu P, Derendorf H. Microdialysis in peripheral tissues. Adv Drug Deliv Rev 2000;45:189–216. [21] de la Pen˜a A, Brunner M, Eichler HG, Rehak E, Gross J, Thyroff-Friesinger U, et al. Comparative target site pharmacokinetics of immediate- and modified-release formulations of cefaclor in humans. J Clin Pharmacol 2002;42:403–11. [22] Wise R. Protein binding of beta-lactams II: the effects on activity and pharmacology particularly tissue penetration. Studies in man. J Antimicrob Chemother 1983;12: 105–18. [23] Benfeldt E, Serup J, Menne T. Microdialysis vs. suction blister technique for in vivo sampling of pharmacokinetics in the human dermis. Acta Derm Venereol 1999;79:338–42. [24] Simon C, Malerczyk V, Klaus M. Absorption of bacampicillin and ampicillin and penetration into body fluids (skin blister fluid, saliva, tears) in healthy volunteers. Scand J Infect Dis Suppl 1978;14:228–32. [25] Hoffstedt B, Walder M, Forsgren A. Comparison of skin blisters and implanted cotton threads for the evaluation of antibiotic tissue concentrations. Eur J Clin Microbiol 1982;1:33–7. [26] Blaser J, Rieder HL, Luthy R. Interface-area-to-volume ratio of interstitial fluid in humans determined by pharmacokinetic analysis of netilmicin in small and large skin blisters. Antimicrob Agents Chemother 1991;35:837–9. [27] Freeman CD, Nightingale CH, Nicolau DP, Belliveau PP, Banevicius MA, Quintiliani R. Intracellular and extracellular penetration of azithromycin into inflammatory and noninflammatory blister fluid. Antimicrob Agents Chemother 1994;38:2449–51. [28] Brunner M, Schmiedberger A, Schmid R, Jager D, Piegler E, Eichler HG, et al. Direct assessment of peripheral pharmacokinetics in humans: comparison between cantharides blister fluid sampling, in vivo microdialysis and saliva sampling. Br J Clin Pharmacol 1998;46:425–31. [29] Rittenhouse KD, Pollack GM. Microdialysis and drug delivery to the eye. Adv Drug Deliv Rev 2000;45:229–41. [30] Verbeeck RK. Blood microdialysis in pharmacokinetic and drug metabolism studies. Adv Drug Deliv Rev 2000;45:217–28.

612

P. Liu, H. Derendorf / Infect Dis Clin N Am 17 (2003) 599–613

[31] Chu J, Gallo JM. Application of microdialysis to characterize drug disposition in tumors. Adv Drug Deliv Rev 2000;45:243–53. [32] Hammarlund-Udenaes M. The use of microdialysis in CNS drug delivery studies: pharmacokinetic perspectives and results with analgesics and antiepileptics. Adv Drug Deliv Rev 2000;45:283–94. [33] Sawchuk RJ, Elmquist WF. Microdialysis in the study of drug transporters in the CNS. Adv Drug Deliv Rev 2000;45:295–307. [34] Mu¨ller M. Microdialysis in clinical drug delivery studies. Adv Drug Deliv Rev 2000;45:255–69. [35] Brunner M, Hollenstein U, Delacher S, Jager D, Schmid R, Lackner E, et al. Distribution and antimicrobial activity of ciprofloxacin in human soft tissues. Antimicrob Agents Chemother 1999;43:1307–9. [36] Mu¨ller M, Stass H, Brunner M, Moller JG, Lackner E, Eichler HG. Penetration of moxifloxacin into peripheral compartments in humans. Antimicrob Agents Chemother 1999;43:2345–9. [37] Mu¨ller M, Haag O, Burgdorff T, Georgopoulos A, Weninger W, Jansen B, et al. Characterization of peripheral-compartment kinetics of antibiotics by in vivo microdialysis in humans. Antimicrob Agents Chemother 1996;40:2703–9. [38] Mu¨ller M, Rohde B, Kovar A, Georgopoulos A, Eichler HG, Derendorf H. Relationship between serum and free interstitial concentrations of cefodizime and cefpirome in muscle and subcutaneous adipose tissue of healthy volunteers measured by microdialysis. J Clin Pharmacol 1997;37:1108–13. [39] Brunner M, Pernerstorfer T, Mayer BX, Eichler HG, Muller M. Surgery and intensive care procedures affect the target site distribution of piperacillin. Crit Care Med 2000;28:1754–9. [40] Joukhadar C, Frossard M, Mayer BX, Brunner M, Klein N, Siostrzonek P, et al. Impaired target site penetration of beta-lactams may account for therapeutic failure in patients with septic shock. Crit Care Med 2001;29:385–91. [41] Mu¨ller M, Brunner M, Hollenstein U, Joukhadar C, Schmid R, Minar E, et al. Penetration of ciprofloxacin into the interstitial space of inflamed foot lesions in non-insulin-dependent diabetes mellitus patients. Antimicrob Agents Chemother 1999;43:2056–8. [42] Lorentzen H, Kallehave F, Kolmos HJ, Knigge U, Bulow J, Gottrup F. Gentamicin concentrations in human subcutaneous tissue. Antimicrob Agents Chemother 1996; 40:1785–9. [43] Borg N, Gotharson E, Benfeldt E, Groth L, Stahle L. Distribution to the skin of penciclovir after oral famciclovir administration in healthy volunteers: comparison of the suction blister technique and cutaneous microdialysis. Acta Derm Venereol 1999;79:274–7. [44] Mindermann T, Zimmerli W, Gratzl O. Rifampin concentrations in various compartments of the human brain: a novel method for determining drug levels in the cerebral extracellular space. Antimicrob Agents Chemother 1998;42:2626–9. [45] Nolting A, Dalla Costa T, Vistelle R, Rand KH, Derendorf H. Determination of free extracellular concentrations of piperacillin by microdialysis. J Pharm Sci 1996;85:369–72. [46] Kovar A, Dalla Costa T, Derendorf H. Comparison of plasma and free tissue levels of ceftriaxone in rats by microdialysis. J Pharm Sci 1996;86:52–6. [47] Dalla Costa T, Nolting A, Kovar A, Derendorf H. Determination of free interstitial concentrations of piperacillin-tazobactam combinations by microdialysis. J Antimicrob Chemother 1998;42:769–78. [48] de la Pen˜a A, Dalla Costa T, Talton JD, Rehak E, Gross J, Thyroff-Friesinger U, et al. Penetration of cefaclor into the interstitial space fluid of skeletal muscle and lung tissue in rats. Pharm Res 2001;18:1310–4. [49] Liu P, Muller M, Grant M, Webb AI, Obermann B, Derendorf H. Interstitial tissue concentrations of cefpodoxime. J Antimicrob Chemother 2002;50:19–22. [50] Wiebe LI, Stypinski D. Pharmacokinetics of SPECT radiopharmaceuticals for imaging hypoxic tissues. Q J Nucl Med 1996;40:270–84.

P. Liu, H. Derendorf / Infect Dis Clin N Am 17 (2003) 599–613

613

[51] Fischman AJ, Alpert NM, Babich JW, Rubin RH. The role of positron emission tomography in pharmacokinetic analysis. Drug Metab Rev 1997;29:923–56. [52] Gibson RE, Burns HD, Hamill TG, Eng WS, Francis BE, Ryan C. Non-invasive radiotracer imaging as a tool for drug development. Curr Pharm Des 2000;6:973–89. [53] Spencer SS, Theodore WH, Berkovic SF. Clinical applications: MRI, SPECT, and PET. Magn Reson Imaging 1995;13:1119–24. [54] Bujenovic LS, Perry JR, McCartney WH. Comparison of planar, SPECT, and 3-D synthetic reprojection images: a case study. Clin Nucl Med 1995;20:302–5. [55] Phelps ME. PET: the merging of biology and imaging into molecular imaging. J Nucl Med 2000;41:661–81. [56] Nath SK, Foster GA, Mandell LA, Rotstein C. Antimicrobial activity of ceftriaxone versus cefotaxime: negative effect of serum albumin binding of ceftriaxone. J Antimicrob Chemother 1994;33:1239–43. [57] FDA-CDER. Guidance for industry: evaluating clinical studies of antimicrobials in the division of anti-infective drug products, Food and Drug Administration. Available at: www.fda.gov/cder. 1997.