Blood, Tissue, and Intracellular Concentrations of Erythromycin and Its Metabolite Anhydroerythromycin during and after Therapy

Abstract

For macrolides, clinical activity but also the development of bacterial resistance has been attributed to prolonged therapeutic and subtherapeutic concentrations. Although erythromycin is a long-established antimicrobial, concomitant determination of the pharmacokinetics of erythromycin and its metabolites in different compartments is limited. To better characterize the pharmacokinetics of erythromycin and its anhydrometabolite (anhydroerythromycin [AHE]) in different compartments during and after the end of treatment with 500 mg of erythromycin four times daily, concentration-time profiles were determined in plasma, interstitial space of muscle and subcutaneous adipose tissue, and white blood cells (WBCs) at days 1 and 3 of treatment and 2 and 7 days after end of therapy. In WBCs, concentrations of erythromycin exceeded those in plasma approximately 40-fold, while free concentrations in plasma and tissue were comparable. The observed delay of peak concentrations in tissue might be caused by fast initial cellular uptake. Two days after the end of treatment, subinhibitory concentrations were observed in plasma and interstitial space of both soft tissues, while 7 days after the end of treatment, erythromycin was not detectable in any compartment. This relatively short period of subinhibitory concentrations may be advantageous compared to other macrolides. The ratio of erythromycin over AHE on day 1 was highest in plasma (2.81 ± 3.45) and lowest in WBCs (0.27 ± 0.22). While the ratio remained constant between single dose and steady state, after the end of treatment the concentration of AHE declined more slowly than that of the parent compound, indicating the importance of the metabolite for the prolonged drug interaction of erythromycin.

INTRODUCTION

Macrolides account for 10 to 15% of the worldwide consumption of antibiotics (25, 31). Although in many countries erythromycin has been replaced by newer relatives like azithromycin and clarithromycin, prescription of erythromycin still has a substantial place in less developed countries but also certain regions within Europe (15, 40). However, bacterial resistance against macrolides dramatically escalated in many countries during the last decades (14, 23, 35). Total consumption of antimicrobials on the one hand and the presence of long periods of exposure of bacterial populations to subinhibitory concentrations of antibiotics on the other hand are considered critical factors for selecting bacterial resistance (2, 3, 16).

Erythromycin is known for its highly variable bioavailability after oral ingestion and its susceptibility toward degradation under acidic conditions. Although its half-life in plasma is short and the pharmacokinetic (PK) profile intraindividually heterogeneous, it is very well distributed throughout body tissues and accumulates in leukocytes (11). While clinical studies have shown the effectiveness of erythromycin in skin and soft tissue infections (33, 39), the in vivo penetration and the resulting free concentrations in interstitial space of soft tissues, i.e., the most important site of bacterial infections, have not been reported until now. High cellular uptake, relatively short retention, and fast back release from white blood cells (WBCs) are considered important properties of the PK of erythromycin (1). For macrolides, intracellular accumulation was considered to prolong their antibacterial effects (6). However, a clearer picture of the PK of erythromycin in the totality of different compartments seems necessary to understand the complex pharmacokinetic-pharmacodynamic (PK/PD) interactions of macrolides with regard to antimicrobial effects as well as development of bacterial resistance.

Erythromycin is hydrolyzed to anhydro forms (anhydroerythromycin [AHE] and other metabolites), and this process is potentiated by acidic conditions (21). AHE is microbiologically inactive but inhibits drug oxidation in the liver and thus is considered to be an important factor for the unwanted drug-drug interactions of erythromycin (20, 36). In vitro AHE is a more potent inhibitor of the cytochrome P450 3A subfamily than erythromycin, thereby prolonging elimination of benzodiazepines, immunosuppressants, statins, and many more drugs (36). Acidic conditions in stomach increase the level of AHE, but high levels of AHE were also found in inflamed tonsils (4).

Thus, due to the lack of currently available information on the target site penetration of erythromycin and its main metabolite, we for the first time determined the concentration-time profiles of erythromycin and AHE simultaneously in the interstitial space fluid of muscle and subcutaneous adipose tissue and in leukocytes. We compared PK parameters of those compartments with plasma concentrations over 3 days of active treatment and after the end of the active treatment period up to 7 days after the last drug administration.

MATERIALS AND METHODS

Regulatory issues.

This was a prospective, open-labeled PK study (EUDRACT 2009-015678-37), conducted at the Department of Clinical Pharmacology at the General Hospital of Vienna and performed in accordance with the actual ICH-GCP guidelines and the Declaration of Helsinki. The clinical study was approved by the Ethical Committee of Medical University of Vienna (EK no. 767/209) and was authorized by the Austrian Agency for Health and Food Safety (AGES).

Subjects.

Within 2 weeks of the start of the study, volunteers underwent a screening visit including a physical examination, blood samples (hematology, clinical chemistry, virology, and coagulation test), electrocardiogram (ECG), and noninvasive arterial pressure in order to identify 6 eligible subjects. Key inclusion criteria were as follows: male, age between 18 and 50 years, body mass index (BMI) between 18 and 30, no regular concomitant medication within the last 2 weeks prior to the study day, and written informed consent given by volunteers. Key exclusion criteria were known allergy or hypersensitivity against erythromycin and clinically relevant abnormal physical or laboratory findings.

Study days.

Starting with study day 1, subjects received 500 mg of erythromycin (erythromycinethylsuccinat) four times daily (EryHexal; Hexal AG, Germany) for three consecutive days. Erythromycin had to be taken fasted, i.e., at least 2 h after last intake of food and 1 h before intake of food. PK sampling was performed on study day 1 and day 3 (i.e., after a single dose and at steady state) and on study days 5 and 10 (i.e., two or 7 days after end of therapy) for plasma, muscle tissue, subcutaneous adipose tissue, and WBCs as described below. Subjects were confined at the department of clinical pharmacology for the duration of PK sampling on the respective study days. On study days 5 and 10, microdialysis was performed for only three subjects each day (subjects were randomly assigned to undergo microdialysis on study day 5 or 10). All samples were snap-frozen at approximately −20°C. Thereafter, samples were stored at approximately −80°C until analysis.

Plasma collection.

Blood (4.5 ml) was drawn from an antecubital vein for determination of erythromycin concentrations in plasma at baseline, 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 6, and 8 h after drug administration on days 1 and 3, and three times on days 5 and 10. Blood samples were kept on ice for a maximum of 60 min and were centrifuged at +4°C and 4,020 × g for 10 min to obtain plasma.

Determination of erythromycin in muscle tissue and subcutis.

An in vivo microdialysis perfusion technique was performed as previously described (28, 29, 37, 43, 45). Microdialysis is based on sampling of analytes from the extracellular space by a semipermeable membrane at the tip of a microdialysis probe. Once the probe is implanted into the tissue, it is constantly perfused at a low flow rate. Substances present in the extracellular fluid diffuse from the extracellular fluid into the dialysate.

On study days, two microdialysis probes (CMA63; CMA Microdialysis AB, Solna, Sweden) were aseptically inserted into skeletal muscle and subcutaneous adipose tissue of the thigh by use of a steel guidance split cannula. The probes then were perfused with saline solution (NaCl 0.9%; Meditrade, Austria) at a flow rate of 1.5 μl/min by a microinfusion pump (CMA 107). After 30 min for equilibration of tissue trauma on days 1 and 3, baseline sampling was performed for 30 min and then sampling was performed in 30-min intervals up to 3 h followed by hourly intervals up to 10 h after the morning dose of erythromycin. On days 5 and 10, microdialysis sampling was performed for two 2-h intervals. At the end of each microdialysis sampling period, the calibration was performed according to the retrodialysis method.

The principle of retrodialysis relies on the fact that the diffusion process through the semipermeable membrane is quantitatively equal in both directions. The in vivo recovery values were thus calculated from the following: in vivo recovery (%) = 100 − $\left(\frac{100\mathrm{·\; analyte\; concentration\; out}}{\mathrm{analyte\; concentration\; in}}\right)$. Interstitial concentrations were calculated according to the following equation: interstitial concentration = 100 · $\left(\frac{\mathrm{sample\; concentration}}{\mathrm{in\; vivo\; recovery}}\right)$. Microdialysis probes were removed at the end of each study day.

Isolation of WBCs.

Isolation of WBCs was performed using the Polymorphprep (Axis-Shield PoC AS, Norway) according to the manufacturer's instructions at baseline and 2, 6, and 10 h after the morning dose of erythromycin on days 1 and 3 and once on days 5 and 10. For each sample, determination of WBC count was performed by hematology analyzer (Sysmex analyzer CE series 2100). To obtain the average concentration of erythromycin and its metabolite in WBCs, the concentration measured in the sample was subsequently corrected for the total cell count and the cell volume as previously described (30).

Analytical methods. (i) Sample preparation. (a) Human plasma.

All sample handling was done at +4°C and daylight protection. Human plasma samples (100 μl) were stabilized by 0.05 ml ammonium acetate buffer (50 mM) and were deproteinized by the addition of 250 μl of acetonitrile containing the internal standard. After thorough mixing, the samples were centrifuged at 11,000 rpm for 5 min at approximately +4°C. Following centrifugation and dilution of the supernatant with 5 mM ammonium acetate buffer, 20 μl of each sample was injected.

(b) Microdialysate.

All sample handling was done at +4°C and daylight protection. Human microdialysate samples (10 μl) were diluted by 90 μl of ammonium acetate buffer (5 mM) containing the corresponding internal standard. After thorough mixing, 10 μl of each sample was injected. Samples with concentration above the upper quantification limit were prediluted with isotonic saline solution.

(ii) LC/MS.

The liquid chromatography system consisted of a binary high-performance liquid chromatography (LC) pump (1200 Series; Agilent Technologies, Waldbronn, Germany). Detection was performed using an AB SCIEX API 5000 triple quadrupole mass spectrometer (MS) (AB SCIEX, Concord, Ontario, Canada; in Germany supplied by AB SCIEX Germany GmbH, Darmstadt) with TurboIonSpray interface. The Analyst version 1.4.2 (AB SCIEX Germany GmbH, Darmstadt, Germany) was used for evaluation of chromatograms. The internal standard was oleandomycin.

Each sample was chromatographed on a reversed-phase column eluted with an isocratic solvent system consisting of buffer and acetonitrile (50:50, vol/vol) and monitored by LC-MS/MS with a standard reference method (SRM) as follows: precursor → product ion for erythromycin m/z 734.3 → m/z 158.4, anhydroerythromycin m/z 716.5 → m/z 158.4, and for the internal standard m/z 688.2 → m/z 158.4; all analyses were in positive mode. Under these conditions, erythromycin, anhydroerythromycin, and the internal standard were eluted after approximately 1.4 min, 1.5 min, and 1.3 min, respectively.

Plasma samples were measured against a calibration curve prepared in human drug-free plasma. Calibration standards were prepared by adding the defined amounts of standard solution of analyte to drug-free plasma. Samples were measured against a calibration curve prepared in 0.9% NaCl. Calibration standards were prepared by adding the defined amounts of standard solution of analyte to 0.9% NaCl. Calibration was performed by weighted (1/concentration²) linear regression. Spiked quality controls (SQCs) were prepared for determination of interassay variation by the addition of defined amounts of the stock solution of the analyte or the spiked control of higher concentration to defined amounts of tested drug-free human plasma, 0.9% NaCl, or human dialysate.

(a) Human plasma.

The limits of quantification (LOQ) of erythromycin A and anhydroerythromycin A in human plasma were set to 4.00 and 4.02 ng/ml, respectively. The linearity (coefficient of correlation) was 4.00 to 2,000 ng/ml (1.000) for erythromycin A and 4.02 to 2,006 ng/ml (≥0.995) for anhydroerythromycin A. The intraday precision (accuracy) for erythromycin A and anhydroerythromycin A in human plasma was determined as 0.8 to 4.6% (98.9 to 100.5%) and 2.2 to 5.9% (99.0 to 104.8%), respectively. The interday precision (accuracy) for erythromycin A and anhydroerythromycin A in human plasma was determined as 2.5 to 3.5% (99.3 to 100.4%) and 6.0 to 8.7% (98.3 to 102.0%), respectively.

(b) Microdialysate.

The linearity of erythromycin A in 0.9% NaCl was shown to be between 4.00 and 2,000 ng/ml (LOQ, 4.00 ng/ml) for the analyses of the muscle retrodialysis samples and between 1.00 and 500 ng/ml (LOQ, 1.00 ng/ml) for the analyses of the muscle dialysis and subcutis dialysis and retrodialysis samples and for the analyses of the leukocytes between 1.50 and 500 ng/ml (LOQ, 1.50 ng/ml).

The linearity of anhydroerythromycin A in 0.9% NaCl was shown to be between 4.01 and 2,006 ng/ml (LOQ, 4.01 ng/ml) for the analyses of the muscle retrodialysis samples and between 1.00 and 502 ng/ml (LOQ, 1.00 ng/ml) for the analyses of the muscle dialysis and subcutis dialysis and retrodialysis samples and for the analyses of the leukocytes between 2.01 and 502 ng/ml (LOQ, 2.01 ng/ml).

The intraday precision (accuracy) for erythromycin A in 0.9% NaCl, muscle retrodialysis, muscle dialysis, and subcutis dialysis as well as in leukocytes was determined as 4.7 to 9.5% (99.8 to 104.1%), 4.7% (100.3%), 5.5% (99.6%), and 1.7% (102.0%), respectively. The intraday precision (accuracy) for anhydroerythromycin A in 0.9% NaCl, muscle retrodialysis, muscle dialysis, and subcutis dialysis as well as leukocytes was determined as 1.4 to 8.2% (92.3 to 107.2%), 2.4% (99.9%), 7.7% (102.2%), and 3.0% (100.4%), respectively.

The interday precision (accuracy) for erythromycin A in 0.9% NaCl, muscle retrodialysis, and muscle dialysis and subcutis dialysis was determined as 6.3 to 8.4% (97.4 to 106.4%), 7.9% (100.1%), and 6.7% (103.9%), respectively. The interday precision (accuracy) for anhydroerythromycin A in 0.9% NaCl, muscle retrodialysis, and muscle dialysis and subcutis dialysis was determined as 7.2 to 8.9% (93.7 to 102.5%), 7.0% (98.6%), and 6.9% (101.2%), respectively.

PK and statistical analyses.

PK parameters were calculated using a commercially available computer program (Kinetika 3.0; Innaphase): maximum concentration (C_max), time to maximum concentration (T_max), terminal elimination half-life (t_1/2), area under the concentration-time curve (AUC) for the dosing period of 6 h (AUC_0–6), AUC_0–24, and the total AUC (AUC_0–∞) were calculated. The AUC values were calculated from nonfitted data by employing the trapezoidal rule, while AUC_0–24 was calculated from AUC_0–6 multiplied by 4 for days 1 and 3 and by multiplying C_avg by 24 for day 5. C_avg was defined as the average concentration observed at the three sampling points on day 5. For plasma, in addition, apparent total body clearance (CL) and apparent volume of distribution (V) were calculated. V was based on the dose corrected for bioavailablity divided by the product of AUC_0–∞ and the elimination constant obtained from plasma. AUC_0–24/MIC was calculated for relevant pathogens. The free fraction of erythromycin in plasma was estimated to be 10% based on data derived from the literature (41).

Statistical analysis was performed using SPSS 16.0 (IBM). For comparison of parameters between compartments, Wilcoxon matched-pairs test was employed. All data are presented as means ± standard deviation (SD).

RESULTS

Six healthy male volunteers (age 29.17 ± 7.68 years [mean ± SD], weight 72.48 ± 11.49 kg, height 179.17 ± 4.02 cm, and BMI 22.57 ± 3.23 kg/m²) were included in the study. Erythromycin was well tolerated, and all six subjects completed the study according to the protocol. Gastric pain, pain at microdialysis probe insertion, loss of appetite, meteorism, leucocytopenia, and thrombocytopenia were observed in one subject each; all adverse events were considered mild, and no serious adverse events were reported.

The comparative mean concentration-time profiles of erythromycin during one dosing interval of 6 h are shown in Fig. 1a and b for day 1 and day 3, respectively, for different compartments, including the calculated free fraction in plasma. For all compartments, higher mean concentrations of erythromycin were observed on day 3 of active treatment in comparison to day 1. Two days after the end of treatment, a significant decrease of mean erythromycin concentrations was present, while after 7 days the erythromycin concentration was under the limit of detection in all compartments. The descriptive PK parameters of erythromycin on days 1, 3, and 5 of the study are presented in Tables 1, 2, and 3 respectively. In WBCs, concentrations of erythromycin were significantly higher than in the other compartments and exceeded concentrations in plasma by approximately 40-fold. While total concentrations in plasma exceeded concentrations in interstitial space of soft tissues by approximately 10-fold, the free fraction of erythromycin in plasma was in good agreement with concentrations in muscle and subcutaneous adipose tissue.

Fig 1.

Comparative mean (SD) concentration-time profiles of erythromycin in different compartments on day 1 (a) and day 3 (b) for plasma (squares), calculated free fraction in plasma (open squares), muscle (triangles), subcutaneous adipose tissue (inverted triangles), and WBCs (diamonds) during one dosing interval of 6 h.

Table 1.

PK parameters of erythromycin at day 1

Sample	Cmax (ng/ml)		Tmax (h)		AUC0–6(ng · h/ml)		AUCtotal (ng · h/ml)		t1/2 (h)		CL (liter/h)		V (liter)
	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD
Plasma	1,371	653	1.75	0.61	3,092	1,150	3,549	1,377	1.67	0.25	58.9	45.5	135	90.5
Muscle	41.9	27.3	3.13	0.88	129	84.2	266	155	3.68	1.43	ND	ND	ND	ND
Subcutis	55.5	23.9	2.50	0.27	147	56.6	218	123	1.98	0.87	ND	ND	ND	ND
WBCs	38,542	32,402	2.00	0.00	116,858	74,953	189,846	138,843	3.49	3.59	ND	ND	ND	ND

C_max, maximum concentration; T_max, time to maximum concentration; AUC, area under the concentration time curve; t_1/2, terminal elimination half-life; CL, total body clearance; V, apparent volume of distribution.

Table 2.

Pharmacokinetic parameters of erythromycin at day 3

Sample	Cmax (ng/ml)		Tmax (h)		AUC0–6(ng · h/ml)		AUCtotal (ng · h/ml)		t1/2(h)		CL (liter/h)		V (liter)
	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD
Plasma	2,544	1,266	2.17	0.52	8,537	5,419	11,782	9,811	2.26	0.84	18.9	8.47	53.9	17.8
Muscle	154	86.6	2.88	1.58	560	369	1,158	835	4.13	1.11	ND	ND	ND	ND
Subcutis	209	139	2.67	0.72	682	465	1,026	543	2.85	1.28	ND	ND	ND	ND
WBCs	51,120	26,925	2.00	0.00	190,440	101,452	ND	ND	ND	ND	ND	ND	ND	ND

C_max, maximum concentration; T_max, time to maximum concentration; AUC, area under the concentration time curve; t_1/2, terminal elimination half-life; CL, total body clearance; V, apparent volume of distribution.

Table 3.

Pharmacokinetic parameters of erythromycin at day 5

Sample	Cavg(ng/ml)		AUC0–24(ng · h/ml)
	Mean	SD	Mean	SD
Plasma	34.2	54.1	822	1,298
Muscle	3.58	6.20	85.9	149
Subcutis	5.11	7.99	123	192
WBCs	7,321	7,928	175,700	190,268

C_avg, average concentration on day 5; AUC, area under the concentration time curve.

Figure 2 shows the concentration-time profile of the metabolite AHE. Although the overall PK profiles of parent compound and metabolite (Fig. 1 and 2, respectively) looked very similar, the ratio of both compounds differed significantly during and after the end of treatment. On day 1, the ratio of erythromycin over AHE on was highest in plasma (2.81 ± 3.45), around 1 in interstitial space of muscle and subcutaneous adipose tissue, and lowest in WBCs (0.27 ± 0.22) (Fig. 3). While the ratio remained relatively constant between single dose and steady state, the proportion of AHE gets higher after end of treatment.

Fig 2.

Comparative mean (SD) concentration-time profiles of the erythromycin metabolite AHE on day 1 (a) and day 3 (b) for plasma (squares), muscle (triangles), subcutaneous adipose tissue (inverted triangles), and WBCs (diamonds) during one dosing interval of 6 h.

Fig 3.

Mean (SD) ratio of erythromycin parent compound over the metabolite AHE during day 1, day 3, and day 5 for plasma (squares), muscle (triangles), subcutaneous adipose tissue (inverted triangles), and WBCs (diamonds).

The PK/PD parameter AUC_{0–24 h}/MIC was used to estimate the clinical efficacy of erythromycin on day 1 and day 3 of treatment as well as 2 days after the end of treatment against example pathogens with an MIC₉₀ of 0.125 μg/ml (erythromycin-susceptible Streptococcus pneumoniae or Streptococcus pyogenes) (Table 4) and an MIC₉₀ of 0.5 μg/ml (erythromycin-susceptible Staphylococcus aureus or coagulase-negative Staphylococcus spp.) (Table 5) (18).

Table 4.

AUC_0–24/MIC of erythromycin for the 0.125-μg/ml value (MIC₉₀ of erythromycin-susceptible S. pneumoniae and erythromycin-susceptible S. pyogenes)

Day	AUC0–24/MIC [mean (SD)] for:
	Plasmaa	Subcutis	Muscle	WBC
1	9.90 (3.68)	4.13 (2.69)	4.71 (1.81)	3,739 (2,399)
3	27.3 (17.3)	17.9 (11.8)	21.8 (14.9)	6,094 (3,246)
5	0.82 (0.13)	0.71 (1.23)	1.24 (1.82)	703 (1,233)

^aThe calculated free fraction in plasma was used.

Table 5.

AUC_0–24h/MIC of erythromycin for the 0.5-μg/ml value (MIC₉₀ of erythromycin-susceptible S. aureus and erythromycin-susceptible coagulase-negative Staphylococcus spp.)

Day	AUC0–24/MIC [mean (SD)] for:
	Plasmaa	Subcutis	Muscle	WBC
1	2.47 (0.92)	1.03 (0.67)	1.18 (0.45)	935 (600)
3	6.83 (4.34)	4.48 (2.95)	5.46 (3.72)	1,524 (812)
5	0.20 (0.33)	0.18 (0.31)	0.31 (0.46)	176 (308)

^aThe calculated free fraction in plasma was used.

DISCUSSION

During the last decades, PK/PD considerations have gained a central role in optimizing clinical efficacy of antibiotic therapies and in preventing bacterial resistance (6, 8, 27, 34). While for erythromycin various studies have determined its concentrations in blood including concentrations in WBCs, until now no study has investigated the PK in the interstitial space fluid of soft tissues, perhaps the most important compartment of bacterial infections (7, 32).

In the present study, for all compartments, mean C_max and AUC values were significantly higher at steady state than after the first dose. However, while AUC_0–6 values for interstitial space fluid of muscle and subcutaneous adipose tissue increased more than 4-fold, the concentration in WBCs increased only by the factor 1.6. Based on these data, one might speculate that erythromycin first rapidly penetrates into leukocytes and later, when the accumulation capacity of WBCs is saturated, the extracellular concentration increases. Indeed, saturable uptake of erythromycin by WBCs was previously described (10). In line with previous in vitro experiments, after the end of therapy a fast decrease of concentrations was observed for all compartments including WBCs in the present study (24). Seven days after the end of treatment, erythromycin concentrations were below the limit of detection in all compartments including WBCs, suggesting that no strong intracellular binding of erythromycin takes place (10).

In vitro cellular uptake of erythromycin was previously extensively investigated (5, 24, 42). Active and passive transport through cell membranes resulting in accumulation of erythromycin in the intracellular compartment has been described previously (42). The cellular/extracellular ratios were constant over a wide range of concentrations (24). Independent of the cell type, antimicrobial active erythromycin rapidly effused when cells were incubated in antibiotic-free medium.

Thus, the observed accumulation of erythromycin of approximately 40-fold compared to plasma is in good agreement with previous data (22, 26). However, a limitation of the present study is that all investigations were performed in healthy, noninflammed tissues. Inflammation and the associated low pH might lead to reduced uptake of erythromycin into WBCs, resulting in higher extracellular concentrations (19). On the other hand, acidification leads to a striking reduction of antimicrobial activity of erythromycin, so the overall impact of major local inflammation remains unclear (17).

Acidification is also associated with acceleration of the metabolism of erythromycin into inactive metabolites like AHE (4). AHE is microbiologically inactive but seems competent for modulating inflammation and may have other biological effects, including a potent role as inhibitor of oxidative metabolism (36). For many inflammatory airway diseases like diffuse panbronchiolitis or cystic fibrosis, the immune-modulatory effect of erythromycin is considered responsible for beneficial effects rather than its antimicrobial action (12). The ratio of erythromycin over AHE on day 1 was >1 in plasma, around 1 in muscle and subcutaneous adipose tissue, and <1 in WBCs (Fig. 3). During treatment, almost 4-fold higher concentrations of AHE compared to parent erythromycin was found in WBCs, possibly due to enhanced intracellular formation of AHE in acidic endosomal vesicles like lysosomes. While the ratio remained relatively constant between single dose and steady state, after the end of therapy absolute concentrations of AHE declined more slowly than concentrations of the parent compound, indicating a role of the metabolite for prolonged drug interaction.

For macrolides, time above MIC or AUC_0–24/MIC predicts antimicrobial activity (13, 34, 38, 44). Based on animal data, the bacteriostatic effects of erythromycin can be expected for free AUC_0–24/MIC ratios greater than 20 in plasma; however, this value is rarely achieved in clinical practice (9). In the present study, we calculated the AUC_0–24/MIC ratio for relevant pathogens like the highly susceptible S. pneumoniae or S. pyogenes and the moderately susceptible Staphylococcus spp. (Tables 4 and 5) (18). For WBCs, the AUC_0–24/MIC ratios easily exceeded the value of 20 from the beginning up to 2 days after end of treatment for all chosen pathogens. However, for other compartments, the threshold of approximately 20 could not be achieved for pathogens with an MIC of 0.5 mg/liter on any study day and was reached only at day 3 of treatment for MIC values of 0.125 mg/liter. It has to be noted that the observed thresholds were obtained by calculations from plasma and therefore may not be readily transferred to other compartments.

Overall, for day 3, considerably better AUC_0–24/MIC ratios were achieved than on the first day of treatment. Overcoming the initially low concentrations in tissues by a loading dose, leading to a more rapid saturation of erythromycin uptake into WBCs, might be an interesting option for a drug like erythromycin with good clinical tolerability. Two days after the end of treatment, we detected subinhibitory concentrations in all compartments except for WBCs; however, the subinhibitory concentrations disappeared before day 10. Compared to newer macrolides, the relatively short period of subinhibitory concentrations may be advantageous (23).

In conclusion, the present study confirmed the previously described fast uptake of erythromycin into WBCs but additionally for the first time demonstrated that sufficient levels of free erythromycin might be delayed until the accumulation in WBCs is widely saturated. Higher initial doses of erythromycin might be considered to accelerate sufficient free concentrations in tissue. Despite this accumulation in WBCs, subinhibitory concentrations of erythromycin in the human body seem to persist for relatively short times. High concentrations of the metabolite AHE compared to those of parent erythromycin might be responsible for prolonged drug interactions after the end of treatment.