Antibiotic Tissue Penetration in Diabetic Foot Infections. A Review of the Microdialysis Literature and Needs for Future Research

Abstract

Although many antimicrobial agents display good in vitro activity against the pathogens frequently implicated in diabetic foot infections, effective treatment can be complicated by reduced tissue penetration in this population secondary to peripheral arterial disease and emerging antimicrobial resistance, which can result in clinical failure. Improved characterization of antibiotic tissue pharmacokinetics and penetration ratios in diabetic foot infections is needed. Microdialysis offers advantages over the skin blister and tissue homogenate studies historically used to define antibiotic penetration in skin and softtissue infections by defining antibiotic penetration into the interstitial fluid over the entire concentration versus time profile. However, only a select number of agents currently recommended for treating diabetic foot infections have been evaluated using these methods, which are described herein. Better characterization of the tissue penetration of antibiotic agents is needed for the development of methods for maximizing the pharmacodynamic profile of these agents to ultimately improve treatment outcomes for patients with diabetic foot infections.

Diabetic foot infections are some of the most debilitating problems that the diabetic population may encounter and are associated with a substantial rate of limb loss, financial hardship, and mortality. [1] Most diabetic foot infections are polymicrobial, with gram-positive cocci being the most predominant pathogen (Table 1). [2] Successful treatment of these infections can be hindered by the emergence of antimicrobial drug resistance, such as methicillin-resistant Staphylococcus aureus (MRSA) and resistant gram-negative pathogens, as well as reduced perfusion to the infected limb secondary to high rates of peripheral arterial disease in the diabetic population. [4,5] Without adequate perfusion to the site of infection, the effectiveness of antibiotic drug therapy is diminished, which may be of even greater concern as antibiotic resistance continues to evolve. Many antibiotic failures have already been linked to impaired target-site penetration, particularly in soft-tissue infections, osteomyelitis, and orthopedic surgery. [3]

Table 1. Empirical Therapy Options for Diabetic Foot Infections³

Table 1. Empirical Therapy Options for Diabetic Foot Infections³

Given the significant implications of a treatment failure, it is essential to gain a better understanding of the lower-extremity tissue penetration to the site of infection of commonly used antibiotics for diabetic foot infections (Table 1) to better predict and improve treatment outcomes for diabetic foot infections. Historically, there is a lack of well-characterized tissue penetration data in this population, largely due to limitations surrounding older methods of studying tissue penetration. However, microdialysis studies can overcome many of these limitations and provide a better estimate of the true level of tissue penetration for antibiotic drugs used in treating diabetic foot infections. The purpose of this review was to explain the microdialysis process and its role in improving the characterization of tissue penetration, to evaluate and describe the existing body of evidence that uses microdialysis to determine antibiotic tissue penetration in diabetic patients, and to identify gaps in the literature where further research is needed.

Role of Pharmacokinetics and Pharmacodynamics in Predicting Antibiotic Efficacy

Pharmacokinetics is used to describe what happens to a drug in the body through absorption, distribution, metabolism, and elimination of drugs, which are described using various pharmacokinetic parameters. [6] When given a specific dosing regimen for an antimicrobial drug, it is these processes that ultimately determine the concentration versus time profile of the drug in a specific compartment of the body, such as serum, plasma, various tissues, and body fluids. In addition, corrections for protein binding need to be taken into consideration when assessing pharmacokinetic profiles because it is typically the free or unbound concentration of drug that is available to elicit its mechanism of action at the target site of infection. In recent years, methods of characterizing the pharmacokinetic profile of antimicrobial drugs in tissue have improved substantially from tissue homogenate and skin blister studies to more recent microdialysis methods; each will be discussed in greater detail in the subsequent section.

Pharmacodynamics describes the relationship between antimicrobial drug concentration versus time profiles derived from pharmacokinetic studies and antimicrobial drug efficacy. [6] In antimicrobial drug pharmacodynamics, there are three parameters than can be used to predict efficacy: the ratio of the area under the free drug concentration versus time profile (ƒAUC) to the mean inhibitory concentration (MIC) of the offending pathogen, the ratio of the maximum free drug concentration to the MIC of the pathogen, and the percentage of the dosing interval in which free drug concentrations remain above the MIC of the given pathogen (% ƒT > MIC). [6] For each class of antibiotics, one of these pharmacodynamic parameters typically correlates best with efficacy. Target values for the parameters are determined from studies of each individual antimicrobial agent, where efficacy is assessed across a variety of different drug exposures, or concentration versus time profiles, to derive a value at which clinical efficacy is most likely. Determination of these target values has historically been based on pharmacokinetic data obtained in plasma or serum. Because concentration versus time profiles in tissue often do not mimic those in serum or plasma, determining pharmacodynamic targets from these data can be misleading. [7] Using accurately collected tissue pharmacokinetic data to derive these targets would likely be more appropriate for determining these targets. However, this has not yet been studied, and targets based on plasma pharmacokinetic data are commonly used in the interim for predicting efficacy, albeit with potential limitations.

Traditional Methods of Characterizing Tissue Penetration

The first step to accurately characterizing the pharmacokinetic-pharmacodynamic relationship of antibiotic drugs is to gain a precise understanding of the concentration versus time profile at the site of infection and to determine the degree of penetration from the central compartment (serum or plasma) into the peripheral compartment of interest (the interstitial fluid of infected tissue). Historically, antimicrobial tissue penetration and pharmacokinetics have been determined from drug concentrations obtained from homogenized tissue samples and cantharis- or suction-induced skin blisters, both of which carry significant limitations. [3]

Tissue homogenate studies consist of simultaneous collection of plasma and tissue samples at a given point in time after administration of a dose of an antibiotic; the samples are then analyzed for drug concentration. Because the entire tissue sample is homogenized before drug concentration determination, the resultant concentration will represent a combination of intracellular and extracellular, as well as protein-bound and unbound, drug concentrations, which does not accurately reflect the free drug concentration available in the interstitium of skin and soft tissue. Furthermore, the best method for calculating the tissue penetration is to compare drug exposures across the entire concentration versus time profile by comparing the AUC in tissue with that in plasma or serum (AUC_0→24tissue:AUC_{0→24serum or plasma}). [8] However, due to the limited number of sampling points in tissue homogenate studies, penetration ratios are typically calculated for single points in time (concentration_tissue:concentration_{serum or plasma}).

Cantherous-induced skin blister studies consist of measuring drug concentrations in an experimentally induced skin blister containing secretory fluid, which is thought to mimic the characteristics of the interstitial space. These techniques allow for more frequent, albeit not continuous, sampling of blister fluid for determination of drug concentrations. [1,7] Although the skin blister allows for obtaining antibiotic concentrations at more points across a concentration versus time profile, the concentrations obtained are still only an estimate of drug concentration in interstitial fluid and are often influenced by blister size and surface to volume ratios. [1,7] Furthermore, many of the antibiotic skin blister studies published to date were conducted in healthy volunteers, when it is well-known that peripheral arterial disease, as well as local and systemic inflammation, can affect plasma and tissue pharmacokinetics of antimicrobial drugs in diabetic patients. [1] It is these differences that likely account for discrepancies observed in tissue pharmacokinetic data from skin blister studies and newer microdialysis methods (Table 2).

Table 2. Antibiotic Tissue Penetration Data Available From Skin Blister and Microdialysis Studies

Table 2. Antibiotic Tissue Penetration Data Available From Skin Blister and Microdialysis Studies

Microdialysis

Microdialysis is an extracellular fluid sampling technique that uses hollow fibers to serve as artificial blood capillaries to monitor chemical events in tissue specimens. [43] It was first used in the 1960s to measure neurotransmitter release in the rat brain. [3,7,43,44] By the 1980s, microdialysis was being used in human clinical trials to measure glucose levels in adipose tissue. [7] Currently, there are a variety of different commercially available microdialysis probes that allow for less invasive, more efficient, and accurate sampling of interstitial fluid in a variety of human tissue. [43,44]

In vivo microdialysis begins with the implantation of a probe containing a semipermeable membrane into the tissue being evaluated. [7,43,44] The semipermeable membrane contains inflow and outflow channels, which allow for a physiologically compatible perfusion fluid (perfusate) to be constantly delivered (usually lactated Ringer's solution at flow rates of 0.1–5.0 μL/min). The solutes of interest in interstitial fluid, which are typically drugs or biomarkers, are able to cross the semipermeable membrane through passive diffusion following a concentration gradient. The outflow solution, referred to as the dialysate, is collected continuously and sampled at various intervals and is then analyzed for free drug concentration. Because new perfusate is constantly being infused, a true equilibrium of solute across the probe membrane is never achieved. However, a steady-state exchange rate can be achieved and needs to be accounted for in analyzing recovery concentrations. This is done by calibrating each microdialysis probe using a technique called retrodialysis, with drug concentrations corrected accordingly. During retrodialysis, drug is added to the perfusate, and the rate of disappearance from the perfusate into the interstitial fluid across the semipermeable membrane is determined. Because diffusion is quantitatively equal in both directions, the rate of drug disappearance from the perfusate is considered equivalent to the rate of in vivo recovery. Drug concentrations determined from dialysate represent free or unbound drug (the portion of drug that is actually available to exert an effect) because proteins do not diffuse into the dialysate because of their large molecular weight. [3,7,43,44]

Microdialysis has emerged as the method of choice for determining tissue penetration of various drugs due to its ability to accurately characterize the entire concentration versus time profile rather than at a limited number of points along this curve, as was historically done with tissue homogenate and skin blister studies, because it allows for continuous sampling of free (unbound) drug from the site of infection. [3,7,43,44] Microdialysis is the only method available that samples directly from the extracellular space (interstitial fluid), whereas skin blister studies provide only an approximation of this space, typically created from otherwise healthy tissue, and homogenized tissue samples represent intracellular and extracellular drug concentrations. [3,7,44] In addition, subcutaneous implantation of the microdialysis probe can be less painful, less invasive, and easier to use in clinical settings than previously used methods. The biggest limitation of microdialysis is that in vivo recovery rates are highly sensitive to changes in flow rate, which often require a certain degree of fine tuning to achieve adequate recovery concentrations. [7] Furthermore, the analytical assays typically used for determination of drug concentrations from samples (high-performance liquid chromatography and liquid chromatography tandem mass spectrometry are most common) carry their own limitations, emphasizing the need to pay close attention to lower limits of detection and requisite sample volumes, relative to the volume and concentrations that can be realistically recovered during microdialysis. [7]

The improvements that microdialysis has brought in characterizing tissue pharmacokinetics compared with older techniques are important for furthering our understanding of antimicrobial tissue penetration in the diabetic population. Recent research evaluating antimicrobial pharmacokinetics in diabetic foot infections has begun to use microdialysis methods to better define the penetration of various antimicrobial agents at the site of infection and to determine whether adequate antibiotic drug exposures can be achieved to elicit a therapeutic response based on previously defined serum pharmacodynamic targets.

Literature Search

We searched the PubMed database for primary literature characterizing antibiotic tissue penetration data through microdialysis methods. The initial search was for articles published in English through December 2013. Search terms included microdialysis and all of the antibiotic medications recommended by the 2012 Infectious Diseases Society of America Clinical Practice Guideline for the Diagnosis and Treatment of Diabetic Foot Infections. [2] Abstracts from the literature search were reviewed, and studies were included if they used microdialysis methods to characterize the tissue pharmacokinetic profile of relevant antimicrobials in skin and soft tissue. Because tissue penetration for ertapenem has not been evaluated to date in a diabetic population, a study conducted in healthy volunteers has been included for completeness, albeit with limitations as noted in the following section. Comparative data from skin blister studies were extracted from a previously published review of these data. [1]

Literature Review

Historically, tissue penetration data to support the use of these antibiotics in this setting has consisted of limited penetration ratios derived from tissue homogenate and skin blister studies in healthy volunteers and, to some extent, infected patients. [1] Microdialysis can be used to determine whether antibiotic drug exposures are, in fact, appropriate at the site of infection through extrapolation of pharmacodynamic targets derived from serum pharmacokinetic-pharmacodynamic data. [45] Although more than 20 antibiotic tissue penetration studies using microdialysis were identified, only a small percentage of these used an infected diabetic patient study population. The strengths and limitations of the existing data are discussed herein.

Cefazolin

Cefazolin is often used as antimicrobial prophylaxis for elective surgery and as treatment for diabetic foot infections caused by susceptible gram-positive and gram-negative pathogens. A study of 12 patients undergoing an abdominal aortic aneurysm repair evaluated the perioperative prophylactic dosing of cefazolin. [46] Two grams of intravenous cefazolin was administered 30 min before the procedure. Plasma, urine, and subcutaneous interstitial fluid samples were collected at 30-min intervals during the preoperative and postoperative periods. A single dose of cefazolin administered 30 min before incision had interstitial fluid and plasma concentrations in excess of the MIC for susceptible S aureus isolates. Although this trial provides useful insight into the penetration of cefazolin into healthy tissue in the perioperative setting, it is crucial to also characterize penetration differences in infected tissue of diabetic patients with lower-limb infections.

A recent microdialysis study sought to characterize the steady-state pharmacokinetics and penetration of cefazolin into infected tissue in seven diabetic patients with foot infections. [9] Six patients received a dosing regimen of 1 g every 8 hours, and one received an adjusted regimen of 2 g every 24 hours secondary to renal impairment. Serum and interstitial fluid samples were collected at various time points over the course of a dosing interval after steady state was achieved. The mean ± SD tissue to serum ƒAUC_0→24 ratio was 1.06 ± 0.78 for infected tissue, suggesting that the ƒT > MIC for cefazolin in infected tissue would be similar to that achieved in serum. Therefore, current dosing strategies for cefazolin are adequate for treating diabetic foot infections caused by susceptible pathogens.

Piperacillin-Tazobactam

Piperacillin-tazobactam is often used in the treatment of diabetic foot infections because of its broad-spectrum activity, including coverage of Pseudomonas aeruginosa. The penetration of piperacillin-tazobactam was investigated in the interstitial fluid of inflamed and noninflamed soft tissue of six diabetic patients using microdialysis. [10] Study participants received 4.5 g every 8 hours, with samples taken on the third day of therapy. Piperacillin was found to penetrate the interstitial fluid of inflamed and noninflamed tissues similarly (mean ± SD AUC_0→8tissue:AUC_0→8plasma: 0.45 ± 0.22 and 0.42 ± 0.36, respectively). However, penetration of tazobactam was much lower in noninflamed tissue (mean ± SD AUC_{0→∞tissue}:AUC_{0→∞plasma}: 1.36 ± 0.75 in inflamed versus 0.53 ± 0.17 in noninflamed tissue). Furthermore, although piperacillin concentrations persisted throughout the dosing interval, tazobactam levels were undetectable within 4 hours after intravenous administration, a finding that correlates with previous studies. [47,48] Based on this level of drug exposure in inflamed tissue and a plasma pharmacodynamic target of 50% ƒT > MIC, piperacillin-tazobactam will likely demonstrate efficacy against pathogens with a MIC of 16 μg/mL or less. [10,11]

Ertapenem

Ertapenem is approved for the treatment of diabetic foot infections without osteomyelitis and complicated skin and soft-tissue infections. Although ertapenem has broad-spectrum activity against gram-positive and gram-negative organisms, it lacks MRSA and pseudomonal coverage. Although there is no published data to date describing the tissue pharmacokinetics of ertapenem in diabetic patients, a microdialysis study has been conducted in six healthy volunteers to determine the tissue penetration of ertapenem into the interstitial space of two peripheral target sites—skeletal muscle and subcutaneous adipose tissue—compared with plasma concentrations. [13] Study participants were given a single 1-g intravenous dose of ertapenem. The mean ± SD ertapenem ƒAUC_0→∞ was 39.7 ± 24.8 μg*h/mL for skeletal muscle and 18.6 ± 4.6 μg*h/mL for subcutaneous adipose tissue, which corresponded to mean ± SD tissue penetration ratios (AUC_tissue:AUC_plasma) of 0.13 ± 0.09 and 0.05 ± 0.01, respectively. However, the investigators used total plasma concentration and free drug concentration in tissue to calculate the tissue penetration ratio. Given the high level of protein binding for ertapenem (4%–16% free drug in plasma), it is not surprising that the resulting penetration level was much lower compared with other beta-lactam antibiotics where penetration ratios are calculated using free drug concentration in plasma and tissue. Owing to the time-dependent nature in which ertapenem exerts its bactericidal effect, the free drug concentration should exceed the MIC for 30% to 40% of the dosing interval for efficacy. [14] Muscle tissue concentrations in the preceding study were found to exceed current MIC₉₀ values for at least 50% of the dosing interval for S aureus, Streptococcus spp, extended-spectrum β-lactamase–producing Enterobacteriaceae, and Bacteroides fragilis, and adipose tissue concentrations exceeded the MIC of these pathogens for at least 30% of the dosing interval. [15,16] These data were derived from healthy volunteers and may not accurately reflect those from infected tissue in patients with diabetes or those with reduced perfusion to the lower limb. Although clinical data currently support the use of ertapenem in the management of diabetic foot infections, accurate characterization of tissue pharmacokinetics in infected diabetic patients will be needed to inform future use of this agent in the event of increasing antimicrobial resistance.

Vancomycin

Vancomycin is frequently used as part of an empirical therapy regimen for treating severe, polymicrobial diabetic foot infections when MRSA is suspected. One microdialysis study in particular looked at measuring vancomycin tissue concentrations in six diabetic patients and six nondiabetic patients after cardiac surgery, all of whom were not infected. [19] Vancomycin was administered as a continuous infusion of 80 to 120 mg/h. Steady-state vancomycin concentrations were then measured in soft tissue and plasma. Vancomycin tissue concentrations in diabetic patients were significantly lower than concentrations in nondiabetic patients, as demonstrated by median (range) penetration ratios of 0.44 (0.08–2.02) and 1.45 (0.35–3.72), respectively (P = .002). This finding is extremely relevant in treating complex diabetic foot infections because insufficient tissue concentrations of drugs can contribute to the failure of an antibiotic regimen. Vast differences in the degree of peripheral arterial disease may have likely accounted for the wide variability in penetration observed between the two groups; however, similar to most microdialysis studies conducted to date, ankle brachial index/pulse volume wave form studies were not performed.

A more recent microdialysis study also sought to evaluate the tissue pharmacokinetics of vancomycin in diabetic patients with lower-limb infections. [20] Patients were given clinically relevant doses of vancomycin. Once steady state was achieved, serum and interstitial fluid samples from the site of infection were simultaneously collected throughout the dosing interval and were analyzed for drug concentration. The mean ± SD ƒAUC_0→t for vancomycin was 232.8 ± 75.7 μg*h/mL in infected tissue, which corresponded to a mean ± SD tissue penetration ratio of 0.8 ± 0.2.

The observed penetration ratio in this lower-limb study was substantially higher than the average ratio of 0.44 observed in the previous study of diabetic patients undergoing cardiac surgery. [19] However, note that unlike patients with lower-limb infection, patients undergoing cardiac surgery were not infected; moreover, more than half of the patients required concurrent treatment with catecholamine during the study. Both of these factors may have contributed to the lower vancomycin penetration observed in the initial diabetic population studied. [19]

Based on our current understanding of the required exposures for efficacy (ie, ƒAUC:MIC target defined in plasma as 200), the degree of tissue penetration observed in patients with lower-limb infections supports the clinical utility of vancomycin for most methicillin-susceptible S aureus (MSSA) and MRSA isolates as the MIC₉₀ of these organisms is 1 μg/mL. [21,22] Although vancomycin may be effective for most staphylococci, it is important to recognize that MSSA and MRSA with MICs of 2 μg/mL or greater have been noted with increasing frequency and that this phenotypic profile may mitigate the clinical utility of vancomycin when present.

Linezolid

Linezolid is an antimicrobial therapy that is approved for use in complicated skin and soft-tissue infections. It has been shown to have a broad range of activity against gram-positive organisms and is available in both oral and parenteral formulations. Early tissue penetration studies evaluating linezolid pharmacokinetics in the interstitial space of adipose tissue and skeletal muscle using microdialysis were conducted in healthy volunteers and in patients with sepsis. [49,50] Linezolid demonstrated good distribution into the interstitial fluid of skeletal muscle and adipose tissue in these two studies but displayed high interpatient variability in the critically ill population. The major downfalls to these studies were the relatively small sample sizes and that the study groups did not consist of infected diabetic patients.

Three studies to date have analyzed linezolid tissue penetration in diabetic populations. [23,51,52] The first study evaluated linezolid penetration in two diabetic patients with lower-extremity ulcerations. [51] Tissue pharmacokinetics were assessed at the ulceration site using in vivo microdialysis before and after hyperbaric oxygen therapy. Linezolid was evaluated before and after an 8-week course of hyperbaric oxygen therapy in diabetic patients with ulcerations defined as Wagner grade 3. Dialysate and serum concentrations were collected hourly over a 12-hour dosing interval after a single 600-mg dose. Tissue penetration ratios for these two patients increased from 47.4% and 47.9%, respectively, before hyperbaric oxygen therapy to 95.0% and 75.7%, respectively, after therapy. Limitations of this study include the small sample size (n = 2) and that penetration was evaluated only after a single oral dose and not at steady state.

A second microdialysis study conducted in three diabetic patients with foot infections demonstrated a greater level of tissue penetration for linezolid after steady-state concentrations were achieved, with mean ± SD tissue penetration ratios of 1.32 ± 0.09 and 1.12 ± 0.22 in healthy and inflamed tissue, respectively. [52] This improved level of tissue penetration at steady-state concentrations in healthy and infected wound tissue was confirmed by a third study conducted in nine diabetic patients with chronic lower-extremity infections. [23] Hourly plasma and dialysate samples were obtained over a 12-hour dosing interval, following three to four doses of linezolid (600 mg intravenously every 12 hours). Linezolid achieved a mean ± SD ƒAUC_0→12 of 92.5 ± 60.4 μg*h/mL in healthy tissue and 82.8 ± 59.0 μg*h/mL in infected tissue from the same extremity. Linezolid penetrated equally well into healthy thigh tissue and infected tissue of diabetic patients, as evidenced by mean (range) penetration ratios of 1.42 (1.08–2.23) and 1.27 (0.86–2.26), respectively (P = .6). Furthermore, the drug exposures achieved in this study are sufficient for achieving the total drug AUC:MIC target of 80 to 120 against isolates with a MIC up to 2 μg/mL, which is the current MIC₉₀ for contemporary MRSA and MSSA isolates. [24,25,26]

Daptomycin

Daptomycin is used in the treatment of complicated skin and soft-tissue infections of the lower extremities. It has activity against gram-positive organisms, including the following resistant species: MRSA, vancomycin intermediate-resistant S aureus, vancomycin-resistant S aureus, and vancomycin-resistant enterococci. Two microdialysis studies have been completed to date evaluating the tissue pharmacokinetics of daptomycin in diabetic patients. The first study sought to compare the pharmacokinetic profiles of daptomycin in the interstitial fluid of soft tissues in healthy versus diabetic adult volunteers using in vivo microdialysis. [28] Twelve individuals were included in the study, with half being diabetic and the other half nondiabetic, healthy volunteers. Participants received a single 4-mg/kg dose of daptomycin intravenously, and plasma and interstitial fluid from healthy tissue samples were collected over a 24-hour period. The degree of tissue penetration in the lower extremities was similar between healthy and diabetic volunteers, given a mean ± SD ƒAUC_0→∞ of 33.5 ± 8.1 and 45.1 ± 40.6 μg*h/mL in nondiabetic and diabetic patients, respectively (P = .8). Furthermore, this translated to mean ± SD penetration ratios of 0.74 ± 0.09 and 0.93 ± 0.61 in healthy and diabetic patients, respectively. The major limitation to this study was that participants were not actively infected because this can have a significant effect on tissue penetration for certain drugs. [53,54]

Similar levels of tissue penetration were also noted in a second study evaluating daptomycin tissue pharmacokinetics in infected patients. [29] Daptomycin penetration into healthy and infected subcutaneous adipose tissue, as well as bone, was evaluated after the administration of daptomycin intravenously, 6 mg/kg, for 4 consecutive days in nine diabetic patients who presented with bacterial foot infections and osteomyelitis. Within 3 hours of a 30-min infusion, free daptomycin in plasma equilibrated completely with soft tissues and bone. Furthermore, inflammation did not adversely affect daptomycin's penetration to the target site of infection, as represented by similar mean penetration ratios (ƒAUC_0→24tissue:ƒAUC_0→24plasma) of 1.54, 1.06, and 1.17 in healthy subcutis, inflamed subcutis, and metatarsal bone, respectively. The 6-mg/kg dose used in this study is higher than the 4-mg/kg dose currently recommended and resulted in higher overall ƒAUC_0→24 for plasma and tissue but may be warranted in the setting of osteomyelitis or pathogens with elevated MICs. At steady state, the free concentration of daptomycin was found to exceed the MIC for relevant gram-positive bacteria. Although pharmacokinetic-pharmacodynamic targets have not yet been identified in tissue, an ƒAUC_0→24:MIC ratio of approximately 100 in plasma is frequently used. [30] This target ratio would likely be achieved against most MRSA and coagulase-negative staphylococci given the MIC₉₀ of 0.25 to 0.5 μg/mL, as well as for β-hemolytic streptococci based on a MIC₉₀ of 0.25 μg/mL. [31]

Ciprofloxacin

Ciprofloxacin can be used in the treatment of diabetic foot infections, primarily for its gram-negative coverage against P aeruginosa. Its lack of reliable activity against gram-positive organisms often requires combination with additional antimicrobial agents, mainly clindamycin. An early microdialysis study sought to characterize the free drug pharmacokinetic profile of ciprofloxacin in the interstitial space of skeletal muscle and adipose tissue in eight healthy volunteers. [33] Participants received a single dose of ciprofloxacin, 200 mg, intravenously. The mean ± SD AUC after this single dose was 155.7 ± 9.9 in serum, 110.5 ± 16.7 in skeletal muscle, and 85.1 ± 15.2 in subcutaneous adipose tissue. These results demonstrated that interstitial ciprofloxacin concentrations in subcutaneous tissue were significantly lower than total drug serum concentrations. If protein binding is accounted for, which is approximately 30% for ciprofloxacin, these results make sense as penetration ratios approach 1 when calculated using ƒAUC_tissue:ƒAUC_serum (mean ± SD tissue penetration ratios were 1.23 ± 0.24 and 0.89 ± 0.16 for muscle and adipose tissue, respectively). However, when the observed serum and interstitial fluid concentration versus time profiles from this study were simulated in an in vitro model against four microorganisms—Proteus mirabilis, Klebsiella pneumoniae, P aeruginosa, and S aureus (MICs not reported)—the level of bacterial reduction was reduced in models simulating interstitial versus serum pharmacokinetics. [33] These experiments demonstrate the discordance in using serum versus tissue pharmacokinetics to predict efficacy at the target site. Although these data support the theory that ciprofloxacin concentrations may be subinhibitory at the target site despite what seem to be effective serum concentrations, they do not represent infected diabetic patients. Tissue biopsy studies have historically reported much higher tissue penetration ratios for ciprofloxacin. [55,56,57] However, these discrepancies are explained by the fact that ciprofloxacin has been shown to accumulate intracellularly, leading to overestimation of the amount of unbound drug available in the interstitium as homogenate measurements contain intracellular and extracellular total drug concentrations. [58]

Additional research has compared ciprofloxacin tissue pharmacokinetics obtained through microdialysis versus cantherous-induced skin blisters. [34] Eight healthy volunteers were given a single dose of ciprofloxacin (400 mg intravenously or 500 mg orally). Concentrations were measured in skeletal muscle, subcutaneous adipose tissue, skin blisters, saliva, and capillary and venous plasma. The ciprofloxacin ƒAUC_0→12 in the interstitial space of skeletal muscle was significantly higher after intravenous administration compared with an oral dose. However, no significant difference was observed between dosing modalities in adipose tissue, suggesting a possible difference in pharmacokinetics between these two peripheral compartments. Furthermore, the ƒAUC_0→12 of ciprofloxacin was three to four times higher in skin blister fluid than in subcutaneous tissue for both dosing modalities, supporting the concept that fluoroquinolones accumulate in inflammatory fluid.

Improving on the methods of the studies mentioned previously herein, a third study investigated the ciprofloxacin concentration in six patients with infected diabetic foot ulcers. [35] Study volunteers were given a single dose of ciprofloxacin, 200 mg, intravenously. Microdialysis samples compared inflamed tissue versus unaffected subcutaneous tissue. Ciprofloxacin exposure in the inflamed tissue was significantly lower than that in the serum, with a mean ± SD ƒAUC_{0→∞lesion}:ƒAUC_0→∞serum of 0.78 ± 0.08 in inflamed tissue. In addition, there was no difference in drug distribution between the inflamed and unaffected tissues (mean ± SD ƒAUC_{0→∞lesion}:ƒAUC_{0→∞uninflamed tissue}: 0.99 ± 0.15), a finding that deviates from previous reports that ciprofloxacin may distribute to sites of inflammation through phagocytes.

Tigecycline

Tigecycline is an antibiotic approved for use in the treatment of complicated skin and skin structure infections. Although tigecycline is used for MSSA and MRSA, pharmacokinetic data at the target site of infection was historically lacking. The steady-state tissue penetration of a standard dose of tigecycline (100 mg, followed by 50 mg every 12 hours) into the interstitial fluid of infected wound tissue versus uninfected tissue in eight diabetic patients was evaluated using microdialysis. [39] Reliable tissue penetration was achieved in healthy and infected tissue, with mean ± SD tissue penetration ratios (ƒAUC_0→24tissue:ƒAUC_0→24plasma) of 0.99 ± 0.53 and 1.00 ± 0.44, respectively (P = .96). In addition, tigecycline penetration into infected and uninfected tissue occurred within 4 hours of administration, and concentrations remained at constant levels through the dosing interval. Furthermore, the mean ± SD ƒAUC_0→24 of tigecycline in infected tissue was 2.6 ± 1.02 μg*h/mL. This level of tigecycline exposure would likely achieve the requisite ƒAUC:MIC target needed for efficacy of 5.7 for most contemporary S aureus isolates, given a recent MIC₉₀ estimate of 0.12 μg/mL. [40,41] Based on these data, tigecycline penetrated equally well into infected and uninfected tissues of the same extremity, supporting the use of tigecycline in diabetic lower-limb infections.

Conclusions

Microdialysis is a minimally invasive technique that allows for continuous sampling of drug concentration from the site of infection; it has replaced tissue homogenate and skin blister studies as the method of choice for determining antibiotic tissue penetration in skin and soft-tissue infections. Although it can provide valuable insight into antibiotic tissue penetration, most of the current research has been conducted in the nondiabetic population, with very limited studies conducted in infected diabetic patients. Tissue pharmacokinetics from infected diabetic patients characterized by in vivo microdialysis is currently limited to cefazolin, piperacillin/tazobactam, vancomycin, linezolid, daptomycin, ciprofloxacin, and tigecycline. Data from microdialysis studies to date demonstrate that interstitial drug exposures achieved with the antibiotics discussed herein are adequate for contemporary pathogens based on pharmacodynamic predictors of efficacy, with the exception of ciprofloxacin. However, in light of increasing MICs and resistance concerns among pathogens isolated from diabetic foot infections, accurate characterization of the pharmacokinetic profile of antibiotic agents used in this setting is paramount for determining the pharmacodynamic effects and level of drug exposure needed for sustaining positive clinical outcomes.

The current body of microdialysis literature in diabetic foot infections desperately needs to be expanded to help ensure that our current antimicrobial agents are being used in a manner that will elicit good clinical outcomes. First, tissue pharmacokinetics and penetration data for other contemporary agents used in managing skin and soft-tissue infections in the diabetic population are needed. Second, determining the tissue pharmacokinetic profile of novel antibiotic therapies in diabetic populations using microdialysis is crucial in early drug development so that informed decisions can be made regarding the pharmacodynamic profile and anticipated efficacy for these agents. Third, current microdialysis studies demonstrate wide variability in tissue penetration estimates. This may be due, in part, to the small sample size of existing studies, as well as differences in the degree of vascular disease in included patients; however, ankle brachial index/pulse volume wave form studies have not been consistently conducted as part of tissue penetration studies. Assessment of the degree of vascular disease in patients included in microdialysis studies is needed so that we can attempt to correlate interstitial fluid data with the level of peripheral vascular disease. Last, although significant advancements have been made in the methods for obtaining antibiotic pharmacokinetic data at the site of infection, it has not been determined whether pharmacodynamic targets historically defined in plasma are the most accurate for predicting efficacy in infected skin and soft tissue or whether peripherally derived pharmacodynamic targets are more appropriate. Further research in these four areas is ultimately needed to improve the antimicrobial management of diabetic foot infections.