INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Ethambutol is an important orally administered drug that is used for the treatment of tuberculosis. It is recommended as a fourth drug with isoniazid, rifampin, and pyrazinamide (PZA) until the susceptibility of the organism is determined (7). The usual dose is 15 to 25 mg/kg of body weight administered as a single daily dose. The elimination half-life in humans is approximately 12 h (19). After administration of a single 25-mg/kg dose under fasting conditions, the mean maximum concentration of ethambutol in serum has been reported to be 4.5 µg/ml, with a range of 1.8 to 6.9 µg/ml, and the mean time to maximum concentration was 2.5 h, with a range of 1.5 to 4 h (19). Previous studies have suggested that the absorption of antimycobacterial agents, including ethambutol, is impaired in patients with AIDS (15, 20, 23; S. E. Berning, G. A. Huitt, M. D. Iseman, and C. A. Peloquin, Letter, N. Engl. J. Med. 327:1817-1818, 1992; C. A. Peloquin, A. A. MacPhee, and S. E. Benning, Letter, N. Engl. J. Med. 329:1122-1123, 1993). However, other reports have not confirmed this observation (8, 11, 12, 18, 26). The effects of gender and AIDS on the steady-state plasma and pulmonary kinetics of ethambutol have not been reported.

Ethambutol is moderately active against tubercle bacilli growing within cultured human macrophages (21). At a concentration of 6 µg/ml, a 1-log decrease (90% killing) in viable organisms was observed after 7 days of incubation (24). Mycobacterium tuberculosis enters and replicates within macrophages, has the ability to evade macrophage-mediated intracellular bactericidal mechanisms, and ultimately kills the cell (14). It is likely, although not proven, that the intrapulmonary concentrations of ethambutol are relevant to the drug's effectiveness in the treatment of tuberculosis. The in vivo penetration of ethambutol into pulmonary macrophages and pulmonary epithelial lining fluid (ELF) in humans has not been reported.

We (9-11) and others (2-4) have developed techniques for the measurement in vivo of the concentration of drugs in pulmonary ELF and alveolar cells (AC). We have also developed a sensitive and specific liquid chromatographic-mass spectrometric (LC-MS) method for determining ethambutol concentrations in plasma, AC, and ELF (J. E. Conte, Jr., E. Lin, and E. Zurlinden, submitted for publication). The purpose of this study was to compare the steady-state plasma and intrapulmonary ethambutol concentrations in normal volunteers and men and women with AIDS.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Study design and subjects. This was a prospective, nonblinded, controlled study of the effects of gender and AIDS on the concentrations of ethambutol in plasma, AC, and ELF. All subjects gave written informed consent and were required to be 18 years of age or older and to be within 10% of acceptable weight for height according to the Metropolitan Life height/weight tables (1). The research complied with all relevant federal guidelines and institutional policies. The evaluation included a medical history, physical examination, and a purified protein derivative (PPD) skin test. Baseline laboratory testing included a complete blood count (CBC), including CD4 counts for subjects with AIDS, platelet count, blood urea nitrogen, serum creatinine, aspartate aminotransferase, alanine aminotransferase, alkaline phosphatase, and total bilirubin. Women were required to be not pregnant and not lactating. Subjects were excluded who had a history of asthma requiring daily therapy, tuberculosis, or a positive PPD skin test (greater than 10 mM induration for normal subjects and greater than 5 mM induration for subjects with AIDS); intolerance to ethambutol or lidocaine; presence of clinically significant organ dysfunction; were required to take chronic medications other than self-prescribed vitamins, birth control pills, or thyroid replacement therapy; abnormal serum creatinine level; or other screening laboratory values outside the normal range (greater than twice normal for subjects with AIDS). The diagnosis of AIDS was based upon the revised Centers for Disease Control and Prevention criteria (6). Subjects with AIDS were permitted to continue all prescribed medications for their care. Subjects with AIDS were required to have (i) less than four soft stools per day without hematochezia, (ii) no abdominal pain or cramping, (iii) no nausea or vomiting, (iv) no symptoms of acute respiratory infection, and (v) a negative chest X ray within 4 months of enrollment. If an X ray had not been done, it was performed as part of the study. Ten men with AIDS, 10 healthy men, 10 women with AIDS, and 10 healthy women were enrolled. The age (mean ± standard deviation [SD]) of the 40 volunteers was 36.4 ± 7.8 years. The subjects with AIDS (men and women) were older than the subjects without AIDS (40.8 ± 5.9 versus 32.0 ± 7.0 years) when the two groups were compared (P  0.0001).

Bronchoscopy and BAL. Standardized bronchoscopy and bronchoalveolar lavage (BAL) (9-13) were performed 4 h after the administration of the last dose. Blood pressure, heart rate, and respiratory rate were recorded prior to and at the completion of bronchoscopy and as clinically indicated following the procedure. Nasal oxygen was administered throughout the procedure, and fingertip oximetry was monitored in all subjects.

Specimen handling. Blood samples were kept on ice until centrifuged. The plasma was separated and then frozen at 70°C until assay. The cells from the BAL were volumetrically resuspended in water to a 10-fold concentration of the lavage fluid, which was centrifuged to produce the cell pellet. The cell suspension was sonicated for 2 min in a Model 550 Sonic Dismembrator (Fisher Scientific, Santa Clara, Calif.). Samples were prepared by a deproteinization step with acetonitrile, which contained an internal standard.

Ethambutol assay. Ethambutol was measured in plasma, BAL fluid, and AC by a new high-pressure column liquid chromatography-tandem mass spectrometric technique (Conte, Jr., et al., submitted). Briefly, the mobile phase containing 80% acetonitrile, 4 mM ammonium acetate, and 0.10% trifluoroacetic acid was run through a Hypersil silica column (4.6 mm [inside diameter] by 50 mm; particle size, 5 µm) at a flow rate of 0.8 ml/min with using a Shimadzu LC-10 AD pump (Shimadzu, Columbia, Md.). Extracts from samples were injected into the system with a Waters intelligent Sample Processor 717 Plus (Waters, Milford, Mass.). The retention time for ethambutol was 2 min. The retention times for neostigmine and propranolol (used as internal standards) were 1.4 and 1.1 min, respectively, with a total run time of 2.8 min. Peak detection and area determinations for some plasma and BAL were carried out with a PE Sciex API III (Perkin-Elmer, Foster City, Calif.). Mass spectrometry settings and conditions were as follows. (i) The multiple reaction monitor scanning mode was set at 205 to 116 m/z for ethambutol and 209 to 71 m/z for neostigmine. (ii) Atmospheric pressure chemical ionization (APCI)-positive ionization was used. (iii) The sample inlet used a heated nebulizer at 450°C. (iv) The discharge current was +3 µA. (v) The gas curtain flow was 1.2 liter/min (N₂ = 99.999%). (vi) The nebulizer pressure was 551.4 kPa. (vii) The collision gas consisted of a 9.99% nitrogen-90.01% argon mixture (set at 250 × 10¹² molecules/cm²). Peak detection for alveolar cells and some plasma and BAL specimens was carried out on a Micromass Quattro LC (Micromass Co., Manchester, England). The following mass spectrometry conditions were used. (i) The reaction channel was 205.35 to 116.10 m/z for ethambutol and 260.18 to 115.95 m/z for propranolol. (ii) Electrospray-positive ionization was used. (iii) The sample inlet utilized a heated nebulizer: the sample cone was set to 25 V for ethambutol and 35 V for propranolol. (iv) The energy collision was set to 15.0 eV for both ethambutol and propranolol. A Macintosh Quadra 800 computer was used for peak integration and analysis. The detection limits for ethambutol were 0.05 µg/ml for plasma and 0.005 µg/ml for BAL supernatants and alveolar cell suspensions. The mean (± SD) coefficients of variation and ranges of the assay for intraday and inter-day determinations together for plasma, BAL supernatants, and alveolar cells were 7.81% ± 2.02% (range, 3.9 to 10.14%), 6.46% ± 3.69% (range, 1.42 to 11.42%), and 12.67% ± 4.59% (range, 6.0 to 20.0%), respectively. The mean (± SD) recoveries and ranges of the assay for intraday and interday determinations together in plasma, BAL supernatants, and alveolar cells were 105.91% ± 7.73% (range, 93.3 to 119.0%), 95.94% ± 10.43% (range, 80.0 to 106.88%), and 105.48% ± 3.60% (range, 100 to 110%), respectively. The accuracy ranges for all determinations in plasma, BAL supernatants, and alveolar cells were 6.67 to 19.0%, 20.0 to 6.88%, and 0.0 to 10.0%, respectively.

Quantitation of volume of ELF and concentration of antibiotics in ELF and AC. The amount of ELF recovered was calculated by the urea dilution method described by Rennard et al. (22) and as reported in our previous pharmacokinetic studies (9-12). The concentration of urea in serum was analyzed by the clinical laboratory at the University of CaliforniaSan Francisco by using a coupled urease-glutamate dehydrogenase enzymatic method (25) modified by the Boehringer Mannheim Corporation (Indianapolis, Ind.). Measurements were made at a fixed time interval permitting automated analysis with a BM 747 Analyzer (Boehringer Mannheim). Urea was measured in BAL supernatant with a modified enzymatic assay (blood urea nitrogen kit UV-66; Sigma, St Louis, Mo.) as previously reported (9-13). Controls were included with every run, and if not within 10% of the known value, the standard curve, controls, and specimen assays were repeated.

Statistical analysis. PROPHET software, version 6.0 (Division of Research Resources, National Institutes of Health, Bethesda, Md., and Abtech Corporation, Charlottesville, Va.), was used to compute descriptive statistics and sample sizes and to perform the linear regression analyses. Analysis of variance (ANOVA) with a two-factor factorial model was used to assess the effects of gender and AIDS status on subject physical characteristics, clinical laboratory values, drug dosage, drug concentrations, AC recovery, ELF recovery, and AC/plasma and ELF/plasma ratios. The ethambutol concentrations in ELF and AC were compared for subjects with and without AIDS by one-way ANOVA. The two-sample equal-variances t test (two sided) was used to compare AC and ELF recovery and drug concentrations in plasma, AC, and ELF between the groups of women with AIDS who were smokers and nonsmokers. The equalities of variances of the smoking and nonsmoking groups were calculated by using the F test (Levene's test). The two-sample Mann-Whitney rank-sum test (two sided) was used to compare the daily doses in men and women, CD4 counts in men and women, and the serum creatinine determinations between healthy subjects and men or women with AIDS. The Shapiro-Wilk test was used to evaluate the normality of the distributions of the data sets prior to comparison. P < 0.05 was regarded as significant.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

The CD4 counts (shown as mean ± SD, median, and range, respectively) for the 10 men with AIDS were 370 ± 208, 300, and 113 to 697 and for the 10 women with AIDS were 330 ± 127, 304, and 86 to 550 and were not significantly different (P > 0.05). All of the serum creatinine determinations were within normal limits; however, the values were greater for all 20 men (0.95 ± 0.25 mg/dl) than for all 20 women (0.71 ± 0.16 mg/dl) (P < 0.05). AIDS status was without effect (P > 0.05) on the serum creatinine determinations. Seven of the 10 female subjects with AIDS were cigarette smokers; the remainder of the subjects were nonsmokers.

Forty-eight subjects were recruited and signed informed consent for the study. Of the 48, 5 did not keep their study appointments and were dropped, 1 withdrew consent, 1 was discontinued due to intercurrent illness, and 1 refused bronchoscopy on day 5. The 40 remaining subjects successfully completed the bronchoscopy and BAL. One subject developed pneumonia postbronchoscopy and was successfully treated with antibiotics. There were no other major adverse events, and all of the subjects returned to their normal duties. One subject experienced mild, self-limited chest discomfort, and the temperature was transiently elevated in 1 subject. Mild cough or lightheadedness occurred in 4 subjects and 16 subjects, respectively; both of these symptoms were self-limited and did not require treatment.

The number (mean ± SD, median, and range) of AC recovered from BAL in the 40 subjects was 3.4 × 10⁸ ± 6.4 × 10⁸ cells/liter; AC recovery (Table 1) was not affected by gender or smoking history (P > 0.05), but was greater in subjects with AIDS (5.5 × 10⁸ cells/liter) than in the healthy subjects (1.2 × 10⁸ cells/liter) (P = 0.01). The percentage of cells in the monocyte/macrophage class was not affected by AIDS status or gender (P > 0.05) and was not significantly different when the smoking women (n = 7) were compared to the nonsmoking subjects (n = 33) (93% ± 9.9% versus 86% ± 14.9%). The volume (mean ± SD) of ELF recovered from the 20 subjects with AIDS (1.3 ± 0.6 ml) was greater than that recovered from healthy subjects (1.0 ± 0.4) (P = 0.03). Gender or smoking had no effect on the recovery of ELF (P > 0.05).

Plasma ethambutol concentrations. For all 40 subjects combined, there was no significant difference between the concentrations (mean ± SD) of ethambutol in plasma at 2 h (2.1 ± 1.2 µg/ml) or 4 h (2.1 ± 0.8 µg/ml) after the last dose (Table 2), and there was no correlation between the 2- and 4-h concentrations (r = 0.07, P = 0.65). There was no significant effect of gender or AIDS status on the concentrations of ethambutol in plasma at 2 h following the last dose. Plasma ethambutol concentrations at 4 h following the last dose were greater in the 20 men than in the 20 women (2.4 ± 0.9 versus 1.8 ± 0.6 µg/ml) (P < 0.05). There was no correlation between the weights of the subjects and the concentrations of ethambutol at 2 h (r = 0.27, P = 0.09) or 4 h (r = 0.27, P = 0.18).

AC concentrations. AC concentrations were greater in women with AIDS than in the other three groups (Table 2). The greater AC ethambutol concentrations in this group were associated with smoking. The AC ethambutol concentration (mean ± SD) in the smoking women (97.2 ± 32.1 µg/ml) with AIDS was more than twice the concentration in nonsmoking women (46.4 ± 34.5 µg/ml) with AIDS (P = 0.05) and more than twice the concentration in all other nonsmoking subjects (45.2 ± 16.8 µg/ml) combined (P < 0.05).

ELF concentrations. There was no significant effect of gender or AIDS status on the concentrations of ethambutol in ELF (Table 2), although the volume of ELF recovered was approximately 30% greater in patients with AIDS (Table 1). When the male and female data were grouped to create a larger sample size (n = 20 in each group), the ELF concentrations (mean ± SD) of ethambutol in subjects with (2.1 + 0.8 µg/ml) and without (2.3 ± 1.3 µg/ml) AIDS were not significantly different (P > 0.05). For all 40 subjects, the ELF/plasma ratios at 2 and 4 h were 1.4 and 1.1 (P > 0.05). There was no correlation between the weights of the subjects and the concentrations of ethambutol in ELF (r = 0.27, P > 0.05). ELF ethambutol concentrations (mean ± SD) in smoking (1.8 ± 0.5 µg/ml) and nonsmoking (2.0 ± 0.6 µg/ml) women with AIDS were not significantly different (P > 0.05). Again, the sample sizes for the smoking and nonsmoking groups were small, and a larger sample size might have detected a difference. The 2-h plasma ethambutol concentrations were not correlated with the ELF ethambutol concentrations (r = 0.06, P = 0.7); however, there was a weak positive correlation between the 4-h plasma and ELF ethambutol concentrations (r = 0.52, P = 0.0006).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Concentrations of ethambutol in plasma at 2 and 4 h following the last dose and AC and ELF ethambutol concentrations were not affected by gender or the presence of AIDS, as defined in our subjects. Within the range of CD4 counts in our patients and within the limitations of the sample size of patients with AIDS (n = 20), CD4 cell counts were not correlated with ethambutol concentrations in plasma, AC, or ELF. Further investigation will be necessary to determine whether absorption of ethambutol would be more impaired in patients with greater degrees of immunosuppression (lower CD4 cell counts) than those included in this study. With all subjects combined, plasma drug concentrations at 2 h were not significantly different from the concentrations at 4 h. The time to maximum concentration of drug in serum (T_max) for orally administered ethambutol has been reported to be a mean of 2.5 h with a range of 1.5 to 4.0 h (19), indicating considerable interpatient variability in this parameter. Since we drew blood specimens at 2 and 4 h after administration of the oral dose, it is likely that the actual C_max was not detected in our subjects. This interpretation is supported by the lack of correlation, in our subjects, between the 2- and the 4-h measurements and by the wide range of T_max that has been reported (19).

MICs of ethambutol for sensitive strains of Mycobacterium tuberculosis, tested by the BACTEC method, have been reported to be 2 µg/ml (21) and approximately 2 to 4 µg/ml (17). For clinical purposes, 2.5, 5.0, and 10.0 µg/ml, respectively, have been recommended as breakpoints for designating strains of M. tuberculosis as susceptible, intermediate, or resistant (21). It is of note that 27 (68%), 19 (48%), and 22 (55%) of the 40 subjects had 2-h plasma, 4-h plasma, and ELF drug concentrations below the MIC (2.0 µg/ml) that has been recommended as the laboratory breakpoint for susceptible strains of M. tuberculosis. A larger dose of ethambutol (e.g., 25 mg/kg), rather than the 15 mg/kg used in this study, likely would have resulted in greater plasma, AC, and ELF drug concentrations. More of the subjects would have had 2- and 4-h plasma and ELF drug concentrations that exceeded the published MICs of ethambutol for M. tuberculosis. The dose used in this study, 15 mg/kg once daily, was at the low end of the range recommended for clinical purposes (15 mg/kg to 25 mg/kg/day) and was chosen in order to minimize potential toxicity in volunteer subjects.

AC ethambutol concentrations in all 40 subjects were greater than 10 µg/ml. The high drug concentrations in AC relative to plasma and ELF were similar to those that we have described with macrolides (9, 10). Neither the 2-h nor the 4-h plasma concentrations were correlated with the AC ethambutol concentrations. The AC/plasma concentration ratio was 21 ± 13 at the time the bronchoscopy was performed (4 h following the last dose) in the 33 nonsmoking men and women. The AC/plasma concentration ratio at 4 h was 61 ± 30 in the seven smoking women (P < 0.05). This finding indicates that alveolar cells in smoking subjects concentrate ethambutol greater than in nonsmoking subjects and is consistent with the prior demonstration that the active transport ex vivo of ethambutol by alveolar macrophages is increased in smokers (16). Our ability to detect differences concentrations of ethambutol in plasma, AC, and ELF between the subgroups of smoking and nonsmoking women with AIDS or to measure effect of gender apart from smoking was limited by the small sample size of nonsmoking women. As with macrolides, considerable antibacterial activity resides in the AC. For example, for organisms with MICs of 0.25 to 2.0 µg/ml, inhibitory ratios between 30 to 1 and 240 to 1 would be present in AC.

High inhibitory and killing ratios are viewed as desirable in the treatment of infectious diseases. For concentration-dependent antibiotics such as fluoroquinolones, C_MAX/MIC ratios of 10 in plasma have been associated with maximum efficacy and delay in emergence of resistance (24, 27). Whether these pharmacokinetic characteristics apply to ethambutol and the treatment of tuberculosis is unknown. However, it is likely, but not proven, that high intrapulmonary concentrations of antituberculous drugs would be related to their therapeutic efficacy.

The ELF/plasma ratio at the time of bronchoscopy (4 h after the last dose) was 1. ELF ethambutol concentrations were not affected by AIDS status or gender and were not significantly different from the plasma drug concentrations. These observations suggest that, at steady state, this compartment was in equilibrium with the intravascular compartment. The volume of ELF recovered in the AIDS patients was approximately 30% greater than that in the healthy subjects. This increased ELF recovery should not affect the calculated concentration of ethambutol in ELF, since the ELF concentration is also dependent upon the volume of BAL recovered and the concentration of ethambutol measured in BAL.

This study was not designed to detect interactions between ethambutol and other drugs taken by our AIDS patients. The clinical significance of the intrapulmonary ethambutol concentrations described in our study is unknown and requires further investigation.