Dry Powder versus Intravenous and Nebulized Gentamicin in Cystic Fibrosis and Bronchiectasis

A Pilot Study
N. RENE CROWTHER LABIRIS, ANNE M. HOLBROOK, HENRY CHRYSTYN, STUART M. MACLEOD, and MICHAEL T. NEWHOUSE*
Centre for Evaluation of Medicines, Father Sean OSullivan Research Centre and Barnett Medical Aerosol Research Laboratory, McMaster University, Hamilton, Ontario, Canada; and School of Pharmacy, University of Bradford, United Kingdom

Aminoglycosides are a mainstay of therapy for patients with cystic fibrosis (CF) or non-CF bronchiectasis who are infected with Pseudomonas aeruginosa (Psa). Traditionally, aerosolized antibiotics are delivered by liquid nebulization. The objective of this study was to determine whether a gentamicin dry powder inhaler (DPI) is as microbiologically active and potentially safe as gentamicin inhaled via a small-volume nebulizer (SVN) or given intravenously. The study was done according to a randomized, single-dose, and triple crossover protocol. Ten patients with CF or non-CF bronchiectasis and chronically infected with Psa were recruited. Patients received a single dose of either gentamicin 160 mg via DPI or SVN, or gentamicin at 5 mg/kg by intravenous infusion. In seven of the 10 patients, the minimum inhibitory concentration (MIC) was achieved in sputum after DPI and SVN, with mean (95% confidence interval) gentamicin concentrations at 2 h after administration of 13.1 g/g sputum (range: 2.2 to 23.9 g/g) and 97.2 g/g sputum (range: 0.3 to 194.2 g/g), respectively, whereas gentamicin levels in the sputum after intravenous administration failed to reach the MIC. Gentamicin given by DPI and SVN significantly decreased the sputum Psa density (p 0.05), by almost one order of magnitude. No significant decline in bacterial counts was observed after intravenous gentamicin. When gentamicin was inhaled, blood concentrations were minimal, and were below concentrations known to cause systemic toxicity. For treatment of Psa infections susceptible to gentamicin, gentamicin administration by DPI appeared to be as efficient as by SVN, despite the delivery of a 7-fold lower dose to the airways. Crowther Labiris NR, Holbrook AM, Chrystyn H, Macleod SM, Newhouse MT. Dry powder versus intravenous and nebulized gentamicin in AM J RESPIR CRIT CARE MED 1999;160:17111716. cystic fibrosis and bronchiectasis: a pilot study.

Lower respiratory tract infections with Pseudomonas aeruginosa (Psa) are a major cause of morbidity and mortality among patients with cystic fibrosis (CF) and non-CF bronchiectasis (15). Patients infected with Psa have frequent acute exacerbations of respiratory infection characterized by increased sputum production and tenacity; increased sputum purulence; increased frequency and severity of cough; malaise; deterioration of pulmonary function; weight loss; fever; and

(Received in original form October 21, 1998 and in revised form May 12, 1999 ) * Currently Director of Medical Affairs, Inhale Therapeutics Systems, San Carlos, CA and Visiting Professor, Stanford Medical School, Palo Alto, CA. Supported by Innovata Biomed Ltd. (Respiratory Division of ML Laboratories, PLC), St. Albans, UK. Ms. Crowther Labiris is supported by a Centre for Evaluation of Medicines/Canadian Drug Manufacturing Association/Medical Research Council of Canada Studentship. A.M. Holbrook is the recipient of an Ontario Ministry of Health research personnel award. Correspondence should be addressed to Rene Crowther Labiris, Centre for Evaluation of Medicines, 50 Charlton Avenue East, Martha Wing H312, Hamilton, ON, L8N 4A6 Canada. E-mail: renee.crowther@utoronto.ca Am J Respir Crit Care Med Vol 160. pp 17111716, 1999 Internet address: www.atsjournals.org

leukocytosis (6). Many patients are given long-term treatment with inhaled and/or systemic antibiotics in the hope of maintaining their physical well-being, quality of life, weight, and lung function, as well as to decrease the number of disease exacerbations and hospital admissions (79). Antibiotic therapy has become a mainstay of treatment for patients with CF or non-CF bronchiectasis who are infected with Psa, and has been credited with the prolonged survival of CF patients (1012). Aminoglycosides, including gentamicin, are considered among the most useful classes of antibiotics for treating Psa infections. The major drawback of aminoglycosides is the need for their relatively high-dose intravenous administration, which carries the potential for systemic toxicity (1315). Consequently, when gentamicin is given intravenously in maximum safe doses, only relatively low sputum aminoglycoside concentrations are achievable. These limitations can be circumvented by direct delivery of antibiotic aerosol to the airways. Inhalation offers the potential benefit of producing high concentrations of antibiotic at the site of infection in the airways, without the risk of systemic toxicity. Numerous studies have shown that inhaled antibiotics can improve pulmonary function and reduce the number of acute hospital admissions in CF and non-CF bronchiectasis, with negligible drug hypersensitivity and little apparent increase in

1712

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

VOL 160

1999

bacterial drug resistance or fungal superinfection (9, 1620). Thus, inhaled antibiotics have the potential to provide more effective antimicrobial therapy with fewer adverse effects. Traditionally, aerosolized antibiotics have been administered by means of small-volume nebulizers (SVN). This form of administration has been officially sanctioned by the U.S. Food and Drug Administration, which recently approved a preservative-free and more concentrated solution of tobramycin for nebulization (TOBI; Pathogenesis Corporation, Seattle, WA). Drug administration by SVN is cumbersome, time consuming, and costly (21). For treating reversible airflow obstruction with bronchodilators and steroids, SVN have been increasingly replaced by pressurized metered dose inhalers (pMDI) or dry powder inhalers (DPI) because they are much more efficient, provide more rapid and easier drug administration, and are more cost-effective. For antibiotic therapy, which requires delivery of much larger doses of medication to the lungs than in the case of bronchodilators or steroids, DPIs are likely to be more effective than pMDI, since they can potentially deliver up to 20 mg of powder readily, while PMDI cannot efficiently deliver more than 5 mg per puff (22, 23). We therefore developed a DPI containing a 50% gentamicin formulation suitable for antibiotic administration by inhalation, in order to allow direct comparison of the DPI and SVN inhalation systems for gentamicin with the gold standard of intravenous administration. Our study compared the potential systemic and oropharyngeal safety, sputum drug concentration, and delivery efficiency of gentamicin administered via a novel DPI with those of gentamicin given by SVN intravenously in stable CF or non-CF bronchiectactic patients chronically infected with Psa.

mg of powder formulation from the Clickhaler DPI, to produce a nominal gentamicin dose of 160 mg. Each actuation dose contained 5 mg micronized gentamicin and 5 mg of 100 M lactose formulation. Lactose was added as a flow aid to facilitate effective aerosol discharge. An Investigational New Drug approval was obtained for the Clickhaler gentamicin DPI from the Therapeutic Products Programme of Health Canada. The fine particle dose ( 6.4 m) of gentamicin delivered by the DPI, as measured with the glass twin impinger (24) at a flow rate of 60 L/min, was 14.9 4.3% (mean SD) (1.9 mg) of the nominal dose. Patients were trained to use the Clickhaler, and were supervised while they inhaled rapidly and forcefully at a flow rate of 30 to 60 L/min as reflected by a DPI clinical flow monitor (25, 26). Gentamicin sulfate solution was nebulized at resting tidal volume through use of a Pari LC jet nebulizer (Pari-Werk GmbH, Starnberg, Germany) driven by compressed air at a flow rate of 8 L/min. The nebulizer produced particles with a mass median aerodynamic diameter (MMD) of 6.6 m, with 49.5% and 35.9% of the aerosol mass having an aerodynamic diameter of 6.4 m and 5 m, respectively. The geometric standard deviation (GSD) of the particle size was 2.0 (MasterSizer X laser droplet and particle analyzer; Malvern Instruments Ltd, Malvern, UK). Eight percent of the aerosol mass was 2 m, with a high probability of reaching the lower respiratory tract (27).

Patient Selection
Clinically stable patients (18 yr of age or older) with CF (confirmed by a sweat chloride concentration 60 mEq/L) (n 3) or with non-CF bronchiectasis (n 7) (confirmed with a computed tomographic scan or chest X-ray) were recruited for the study. All produced more than 15 mL/d of purulent sputum that was chronically infected with Psa, which was either sensitive or showed intermediate susceptibility to gentamicin. Patients were excluded if their FEV1 or FVC decreased by 15% upon aerosol inhalation, their serum creatinine was above 120 mol/L, or their blood urea was above 9 mmol/L, or if they were hypersensitive to gentamicin or lactose. The study was approved by the St. Josephs Hospital Research Committee. Written informed consent was obtained from each patient before initiation of the study.

METHODS
SVN and DPI Delivery Devices and Formulations: Aerosol Characterization
The gentamicin powder delivery system used in our study was developed from the Clickhaler DPI (Innovata Biomed, Ltd. [Respiratory Division of ML Laboratories PLC], St. Albans, UK) (Figure 1). Gentamicin administration consisted of 32 actuations, each providing 10

Drug Administration
On three study days, separated from one another by a minimum of 2 wk, patients received single doses of 160 mg gentamicin via the DPI or the SVN, or of 5 mg/kg given intravenously (Garamycin injectable; Schering, Plough, Montreal, PQ, Canada) according to a computerdetermined randomization order. All doses were administered over periods of 30 min.

Outcome Measures
Gentamicin assay: sputum, blood, urine. The primary outcome measures in the study were sputum gentamicin concentrations and changes in quantitative sputum Psa cultures at 2 h after treatment as compared with baseline. To prevent contamination of the sputum sample with oral deposits of gentamicin, patients consumed food and/or beverages before sputum collection. The concentrations of gentamicin in serum, urine, and expectorated sputum were measured with an automated fluorescence polarization immunoassay (TDX; Abbott Laboratories, Diagnostic Division, Irving, TX). The assays lower limit of sensitivity was 0.3 g/ml, and both the intrarun and interrun coefficients of variation were 5% (28). Serum and sputum gentamicin levels were assayed 2 h after treatment. Urine gentamicin levels were measured after a 6-h collection period. Serum samples were assayed as collected; urine samples were diluted as necessary. Sputum samples were stored at 20 C before being analyzed. Sputum gentamicin concentrations were measured according to a procedure developed by the Clinical Chemistry Department of St. Josephs Hospital. The sputum was liquefied by adding an equal volume of 10% dithiothreitol (Sputolysin; Calbiochem, La Jolla, CA) in 0.1 M phosphate buffer (Abbott Laboratories Diagnostics Division), and was vortexed, sonicated, and centrifuged. The supernatant was analyzed for gentamicin. The sputum preparation procedure and assay were developed and validated by adding known amounts of gentamicin to antibiotic-free sputum from bronchiectactic patients. Samples containing more than

Figure 1. Clickhaler DPI used for aerosol delivery of gentamicin powder formulation.

Crowther Labiris, Holbrook, Chrystyn, et al.: Gentamicin Powder Aerosol in CF and Bronchiectasis
TABLE 1 GENTAMICIN CONCENTRATIONS OBTAINED FROM SPUTUM, BLOOD, AND URINE WITH A SINGLE DOSE OF GENTAMICIN DELIVERED VIA DRY POWDER INHALER, SMALL-VOLUME NEBULIZER (160 mg GENTAMICIN), OR BY INTRAVENOUS INFUSION (5 mg/kg GENTAMICIN) OVER A 30-min PERIOD (n 10)
Mean Gentamicin Concentration (95% CI) Dosage Form DPI SVN IV Sputum (mg/g sputum) 13.1 (2.2, 23.9) 97.2 (0.3, 194.2) 0.7 (0.4, 0.9) Serum (mg/L) Not detected 0.3 (0.1, 0.4) 9.1 (8.2, 10.1) Urine (mg/L) 0.9 (0.6, 1.2) 13.1 (1.6, 24.5) 574.5 (337.5, 811.5)

1713

Definition of abbreviations: CI confidence interval; DRP dry powder inhaler; IV intravenous; SVM small-volume nebulizer. The specimens were obtained 2 h (sputum and blood) and 6 h (urine) after drug administration.

10 g/ml gentamicin were diluted with 0.1 M phosphate buffer to bring the sample concentration into the assay range.

Quantitative Psa Cultures

All sputum samples were screened for salivary contamination with a Gram stain technique (29). An acceptable specimen had a ratio of polymorphonuclear cells to squamous epithelial cells 2:1. Bacterial concentrations were quantified via a method adapted from the technique of Wong and coworkers (30).

Patient Preference and Adverse Effects

Patients preferences for and ability to use the different antibiotic delivery methods were determined with a 10-cm visual analogue scale adapted from the instrument devised by Boe and colleagues (31). Overall satisfaction, treatment efficacy, treatment duration, ease of learning to use the delivery device, device handling and convenience, taste, and oropharyngeal irritation were evaluated at the end of each treatment visit. At the last visit, patients were asked to name the delivery method that they would prefer for taking their antibiotic therapy. At the end of each treatment visit, patients were asked if they had experienced any adverse effects. The development of resistance to gentamicin was also considered an adverse event. The sensitivity of Psa to gentamicin was determined at the screening visit and at the end of the study. Possible nephrotoxicity was monitored by measurements of serum creatinine and blood urea at the end of the study. The occurrence of nephrotoxicity was defined as a 50% increase in serum creatinine levels. To detect any short-term changes in pulmonary function that might be attributed to the mode of drug delivery, FEV1 was measured before, 15 min after, and 2 h after treatment. A decrease in FEV1 of 10% after treatment was defined as bronchospasm.

Statistical Evaluation
One-way analysis of variance (ANOVA) was used to compare sputum gentamicin concentrations, quantitative sputum Psa cultures, serum and urine gentamicin concentrations, and outcomes on the patient preference questionnaire. The pulmonary function data were analyzed via ANOVA, using the two grouping factors of treatment and measurement time. The analysis was based on inter-group and on intra-group, intersubject variance (32, 33). When a statistically significant result was found, a post hoc comparison was performed.

RESULTS
Of 50 patients invited to participate in the study, only 14 met the study entry criteria. Four of these patients did not complete the study. The reasons for withdrawal were cessation of sputum production (one patient), infection with Burkholderia cepacia (one patient), infection with commensal flora only (one patient), and voluntary withdrawal. The age of the 10 study subjects (eight female and two male), three with CF and seven with bronchiectasis, was 51.9 20.6 yr (mean SD). Percent predicted FEV1 and FVC values were 40.8 23.5% and 50.7 25.7%, respectively. On the basis of these vari-

ables, one subject was considered to have mild disease (FEV1: 80 to 100% predicted), four were considered to have moderate disease (FEV1: 60 to 79% predicted), and five were classified as having severe disease (FEV1: 59% predicted). Two patients used prophylactic oral ciprofloxacin, one used oral cloxacillin, one used oral amoxicillin, and two used aerosolized tobramycin. In all cases, sputum was chronically infected with Psa (four sensitive strains with MIC 4 g/ml, five intermediate strains with MIC 8 g/ml, and one indeterminate strain). The mean (95% confidence interval [CI]) sputum and serum gentamicin concentrations are shown in Table 1, together with the amount of gentamicin excreted in the urine over the 6 h after treatment. After inhalation from the DPI, gentamicin concentrations of 13.1 g/g sputum were achieved, as compared with 97.2 g/g sputum with SVN delivery and 0.7 g/g sputum after intravenous administration. Thus, on average, the inhaled route provided a minimum of about 20-fold (DPI) and a maximum of over 100-fold (SVN) higher sputum gentamicin concentrations than were observed with intravenous administration. Sputum concentrations of gentamicin after DPI or SVN dosing were greater than the MIC of 4 g/ml of sensitive Psa in seven of the patients, and ranged from 1 to 2 g/ml (adequate MIC for 50% of isolates) in two other patients (7). In no patients was the MIC in sputum achieved with intravenous dosing. In serum, gentamicin was below the limit of detection (0.3 g/ml) after all DPI treatments, and after SVN administration in four patients. Urinary excretion of gentamicin was significantly different with the three methods of antibiotic administration. Urine gentamicin levels were approximately 40- to 600-fold greater after intravenous than after SVN or DPI dosing administration, respectively (p 0.01). All sputum specimens, with the exception of one patients samples during nebulized gentamicin treatment, were acceptable (i.e., derived from the lower respiratory tract). After DPI treatment, the number of colony-forming units (cfu) of Psa decreased in eight of the 10 patients, with a mean 6-fold reduction in colony counts (Table 2). Psa density also decreased after SVN administration of gentamicin, with a 10fold reduction in bacterial counts. Psa density decreased after intravenous dosing in seven of the 10 patients, but only a 2-fold decline in Psa density was noted. No statistically significant difference in bacteria counts could be demonstrated with any of the three different treatment methods. We observed no significant effect of treatment order on the change in bacterial density (p 0.27). Bacterial sensitivity to gentamicin had a significant effect on the change in Psa density regardless of the treatment method (p 0.01). The mean decrease in log10 cfu/ml sputum from patients with Psa sensitive to gentamicin (MIC 4 g/ml) was 1.2 (95% CI: 0.7 to 1.8) (n 18 specimens from six patients), compared with a

1714
TABLE 2

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

VOL 160

1999

Pseudomonas aeruginosa COLONY COUNTS AT BASELINE AND 2 h AFTER DELIVERY OF 160 mg OF GENTAMICIN VIA DRY POWDER INHALER, SMALL-VOLUME NEBULIZER (160 mg, BOTH), INTRAVENOUS INFUSION (5 mg/kg) (n 10)
Pseudomonas aeruginosa Colony Forming Units cfu/ml Sputum (SD) Drug Treatment DPI gentamicin SVN gentamicin IV gentamicin Pretreatment 1.2 10 (1.3 10 ) 8.1 107 (1.1 108) 1.6 108 (2.2 108)
8 8

2 h Posttreatment 6.3 107 (1.1 108)* 4.3 107 (1.0 108)* 7.0 107 (1.1 108)

Definition of abbreviations: DRP dry powder inhaler; IV intravenous; SVM small-volume nebulizer. * Colony forming units were significantly fewer, as compared with baseline, after a single dose of gentamicin; p 0.05.

mean decrease of 0.03 (95% CI: 0.2 to 0.3) (n 9 specimens from three patients) in patients with an intermediate sensitivity to gentamicin (MIC 8 g/ml). The data were adjusted for the effect of sensitivity on the change in cfu in sputum, and the effect of each dosage form (DPI, SVN, intravenous) was then reanalyzed, with no change in results. No bronchospasm was detected after gentamicin administration. No significant difference was observed in pulmonary function as compared with baseline function at 15 min or 2 h after each of the three types of treatment. At the start of SVN aerosol inhalation, however, the gentamicin solution induced persistent cough in most patients. The results of the patient preference questionnaire indicated that nebulization of gentamicin also caused significantly more mouth (p 0.02) and throat irritation (p 0.03), induced coughing (p 0.002) and unpleasant taste (p 0.006) than did administration by DPI. No resistant strains of Psa emerged during the study, and there was no evidence of nephrotoxicity or other serious druginduced systemic adverse events. Patient preference for the different antibiotic delivery methods was evaluated (data not shown). Overall, and despite having to take 32 inhalations over a 30-min period, patients liked the DPI system significantly more than either the nebulizer (p 0.003) or intravenous (p 0.001) treatment. Against the background of the patients lifestyles, use of a DPI to deliver antibiotic was considered more convenient than either SVN (p 0.006) or intravenous administration (p 0.001). Patients perceptions were generally that all three drug delivery methods examined in the study would be effective for treating respiratory infections. The SVN and DPI devices were considered to be equally successful at delivering gentamicin to the lungs. However, the SVN was judged significantly more reliable than the DPI (p 0.04).

DISCUSSION
This study showed that despite the unoptimized formulation with which it was charged, the Clickhaler DPI can deliver aerosolized gentamicin to the lungs of CF and non-CF bronchiectactic patients, and can produce sputum concentrations at 2 h after aerosol administration of gentamicin that are approximately 20-fold greater than those produced by intravenous administration. However, this particular DPI produced sputum gentamicin levels that were 7-fold lower than with a current standard wet nebulization device: the Pari LC jet nebulizer. DPI delivery was accomplished with minimal systemic absorption of gentamincin and no apparent bronchial irritation. The clinical relevance of the sputum gentamicin concentration was emphasized by the overall 6-fold reduction in

Psa colony counts with DPI treatment. Achieving sputum concentrations above the reported MIC for both sensitive and intermediate susceptible strains of Psa by delivering gentamicin directly to the site of infection in the airways may allow effective treatment of lung infections normally classified as antibiotic resistant. Furthermore, effective bacterial killing may not require the relatively large doses of gentamicin currently advocated (up to 600 mg three times daily) (3), if relatively efficient drug delivery can be achieved. This, in turn, raises important pharmacoeconomic issues. Although the delivery of medications by inhalation provides a much greater dose of antibiotic than that achieved by intravenous administration, a potential drawback is that the inhaled antibiotic may be deposited poorly or may be virtually absent in regions of the lung where airflow is minimal and where infection is presumably greatest (34, 35). Furthermore, inhaled antibiotic is unlikely to have a beneficial effect on the associated chronic Psa sinusitis that may act as a reservoir of infection of the lower respiratory tract. It is in these regions that penetration of antibiotic is most critical and most challenging to accomplish by inhalation, in contrast to intravenous administration. However, since bronchial blood flow in bronchiectasis may be greatly augmented (from 2% to as much as 30% of cardiac output), and since the endobronchial circulation is recirculated to the peripheral airways and lung parenchyma via the bronchial veins and right atrium, inhaled antibiotics may be redistributed from the central airways to otherwise poorly accessible peripheral areas of the lung (36, 37). Furthermore, gentamicin could diffuse from central to more peripheral airways along a diffusion gradient within obstructed airways, although no data are available to support this contention. Although total gentamicin deposition in the lungs was not measured directly in our study, the decline in Psa colony counts suggests that a single dose of gentamicin delivered via DPI is effectively deposited in the airways. The sputum concentration of gentamicin was relatively high with both nebulized and DPI treatments, exceeding the MIC required to kill Psa in the majority of patients in both cases. The mean sputum gentamicin level achieved after DPI delivery of a nominal 160mg dose of gentamicin was 13.1 g/g sputum, with no detectable gentamicin found in the blood. These results are comparable with the findings of Goldman and colleagues, who reported a mean of 9.3 mg/L of gentamicin in bronchoalveolar lavage fluid and no detectable gentamicin in the blood after delivery of a single 180 mg dose of gentamicin through a Rotahaler (38). Despite the large difference in sputum drug concentration with the two methods, the mean gentamicin concentration achieved after DPI inhalation was not significantly different from that after use of the SVN. This finding is clearly the result of the marked variation in drug concentrations that occurs with both devices, and is supported by two previous single-dose studies that compared a dry powder formulation with nebulization and reported similar sputum gentamicin levels with these two approaches (38, 39). Sputum glycoproteins, DNA, and divalent cations have been shown to inactivate aminoglycosides, thereby reducing the amount of free active drug available (4042). We did not specifically measure the amount of unbound gentamicin in sputum. However, the reduction in Psa colony counts of almost a full order of magnitude observed with a mean sputum gentamicin concentration of 13.1 g/g sputum after DPI delivery suggests that a sufficient amount of free active drug was available. A significant decrease in the bacterial count was found for both inhalation treatments. However, results were extremely variable among patients within the same treatment group.

Crowther Labiris, Holbrook, Chrystyn, et al.: Gentamicin Powder Aerosol in CF and Bronchiectasis

1715

During DPI treatment, the change in log10cfu/ml ranged from a slight increase in counts, of 0.03 log10, to a decrease of 1.6 log10. This variation may be explained by different Psa sensitivities to gentamicin or by major variations in aerosol delivery to the airways. The effect of Psa sensitivity on the change in bacterial density was significant (p 0.01). A mean decline of 1.2 log10cfu/g sputum was observed in patients who were sensitive to gentamicin (MIC 4 g/ml), as compared with a mean reduction of only 0.03 log10cfu/g sputum in patients with an intermediate sensitivity (MIC 8 g/ml). Ramsey and colleagues (9) reported that the density of Psa in sputum samples decreased during each of three 28-d periods when 600 mg of inhaled tobramycin was administered (300 mg BID). The greatest treatment effects were observed during the first two treatment cycles, with a mean reduction of 2.2 log10cfu/g of sputum at Week 2. At Week 20, the mean reduction in Psa density had decreased only by 0.8 log10cfu/g sputum. Over this same period there was a trend toward an increase in the MIC of tobramycin for Psa isolates. Although Ramsey and colleagues discounted the increase in resistance to tobramycin as a possible cause of the reduction in antimicrobial effect of the drug, we have shown that there is a correlation between decreased sensitivity to an aminoglycoside and reduction in Psa density in sputum, making increased resistance of Psa to the aminoglycoside the most likely explanation for a decrease in sensitivity. The DPI appears to be a safe device for inhalation of gentamicin, producing no apparent bronchial irritation or systemic absorption of the drug, thus almost completely eliminating the risk of nephro- and ototoxicity. The DPI was rated significantly higher than either of the two other gentamicin delivery methods by all of the patients in our study, who considered it superior in convenience and freedom from side effects despite having to inhale 32 puffs in 30 min. All patients in the study indicated that if given a choice, they would have preferred to receive antimicrobial therapy with a DPI. Although we recognized that the large number of inhalations required as a result of the limitations of the particular gentamicin formulation used in this study would be somewhat impractical, future formulations and delivery systems having greater drug delivery efficiency are not unrealistic, giving DPI devices promise as practical alternatives to SVNs for antimicrobial drug delivery. The majority of our patients had used a nebulizer previously, but accepted the possibility that the DPI might be as effective for controlling their pulmonary infection as was the SVN. Our study had several limitations. Only 10 patients participated in the study. Consequently, we may not have been able to detect a significant difference among treatments, if any existed. Nevertheless, the study has provided preliminary information indicating that the further study of antibiotic delivery with DPI devices is warranted when optimized drug formulations and more efficient delivery systems become available. Potential adverse effects may have been undetectable because of the small sample size in our study. Oropharyngeal contamination of expectorated sputum with gentamicin as an explanation for the drug concentrations measured in sputum is highly unlikely, since 2 h had elapsed since inhalation of the drug, and patients had eaten and consumed fluids before providing the sputum sample. Furthermore, sputum concentrations would have been much higher with DPI than with SVN administration had the sample been contaminated with oropharyngeal gentamicin, since it is a reasonable assumption that oropharyngeal contamination is proportional to the orally delivered dose of a drug. A larger proportion of the dose of gentamicin aerosolized by the DPI than by the SVN must have been deposited by impaction in the mouth and oropharynx, owing to the larger particles (greater MMAD) and greater inspiratory

flow rate with the DPI. Sputum gentamicin concentrations may be misleading in terms of drug delivered to the lower respiratory tract, but a more accurate measure was not available. Measurements obtained in sputum samples may be either underestimates caused by dilution of drug with saliva during expectoration, or overestimates of the concentration of gentamicin at the actual infection site, due to deposition in large proximal airways. However, on average, sputum concentrations of gentamicin were sufficient to cause approximately a full order of magnitude reduction in Psa cfus after either DPI or SVN administration. In retrospect, the antibiotic formulation and DPI delivery system used in the study were suboptimal. The Clickhaler, an efficient delivery system for bronchodilators and corticosteroids, was not designed for delivering large doses of antibiotic. Its overall efficiency, extrapolated from particle size and sputum concentration data, was only about 15% that of the SVN. However, even with a sputum concentration much lower than that achieved with the SVN, bacterial killing was substantial, at least with susceptible organisms. This suggests that with a dose-ranging study, the plateau of the doseresponse curve for bacterial killing might have been reached at a somewhat lower concentration of gentamicin than those achieved with the more efficient SVN.
Conclusions

This pilot study suggests that the development of an efficient and clinically effective gentamicin DPI should be feasible. A gentamicin DPI would eliminate the obvious problems of discomfort and poor pulmonary delivery with intravenous administration, and should be as effective as nebulization for inhaled drug delivery, while being potentially much more acceptable to patients. Further refinement of the drug formulation to be used, in order to achieve a smaller particle size and improvements in the efficiency of the gentamicin DPI delivery system, should be undertaken before reevaluating its potential role in antibiotic prophylaxis. A drug deposition study done with radiolabeled gentamicin would be a valuable follow-up investigation after the drug formulation has been refined and the aerosol delivery platform has been optimized to provide a much greater fine particle fraction ( 3 m).
Acknowledgment : The authors thank the Clinical Chemistry Department of St. Josephs Hospital for developing the technique used to measure sputum gentamicin concentrations, Mark Parry-Billings for his useful comments and suggestions, and Schering Canada, Inc. for donating the Garamycin Injectable used in the SVN and intravenous treatments.

References
1. Truchsis, M., and M. N. Swartz. 1991. Bronchiectasis: a current view. Curr. Clin. Top. Infect. Dis. 11:170205. 2. Nicotra, M. B. 1994. Bronchiectasis. Semin. Respir. Infect. 9:3140. 3. Ramsey, B. W., H. L. Dorkin, J. D. Eisenberg, R. L. Gibson, I. R. Harwood, R. M. Kravitz, D. V. Schidlow, R. W. Wilmott, S. J. Astley, M. A. McBurnie, K. Wentz, and A. L. Smith. 1993. Efficacy of aerosolized tobramycin in patients with cystic fibrosis. N. Engl. J. Med. 328: 17401746. 4. Horrevorts, A. M., O. M. J. Driessen, M. F. Michel, and K. F. Kerrebijn. 1988. Pharmacokinetics of antimicrobial drugs in cystic fibrosis. Aminoglycoside antibiotics. Chest 94(Suppl. 2):120s125s. 5. Kulczycki, L. L., T. M. Murphy, and J. A. Bellanti. 1978. Pseudomonas colonization in cystic fibrosis: a study of 160 patients. J.A.M.A. 240: 3034. 6. Rubio, T. T. 1986. Infection in patients with cystic fibrosis. Am. J. Med. 81(Suppl. 1A):7377. 7. van Klingeren, B. 1988. Antibiotic resistance in Pseudomonas aeruginosa, Hemophilus influenzae and Staphylococcus. Chest 94(Suppl. 2): 103s109s. 8. Mukhopadhyay, S., M. Singh, J. I. Cater, S. Ogston, M. Franklin, and R. E. Olver. 1996. Nebulized antipseudomonal antibiotic therapy in cystic fibrosis: a meta-analysis of benefits and risks. Thorax 51:364368.

1716

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

VOL 160

1999

9. Ramsey, B. W., M. S. Pepe, J. M. Quan, K. L. Otto, A. Montgomery, J. Williams-Warren, K. M. Vasiljev, D. Borowitz, C. M. Bowman, B. C. Marshall, S. Marshall, and A. L. Smith. 1999. Intermittent administration of inhaled tobramycin in patients with cystic fibrosis. N. Engl. J. Med. 340:2330. 10. Boat, T. F. 1988. Cystic fibrosis. In J. F. Murray and J. A. Nadel, editors. Textbook of Respiratory Medicine. W.B. Saunders, Philadelphia. 11261152. 11. Kepron, W. 1996. Cystic fibrosis in the young adult. Medicine North America. March:2832. 12. Hoiby, N., B. Friis, K. Jensen, C. Koch, N. E. Moller, S. Stovring, and M. Szaff. 1982. Antimicrobial chemotherapy in cystic fibrosis patients. Acta. Paediatr. Suppl. 301:75100. 13. Chan, G. L. C. 1989. Alternative dosing strategy for aminoglycosides: impact on efficacy, nephrotoxicity, and ototoxicity. Drug Intell. Clin. Pharm. 23:788793. 14. Moore, R. D., P. S. Lietman, and C. R. Smith. 1987. Clinical response to aminoglycoside therapy: importance of the ratio of peak concentration to minimal inhibitory concentration. J. Infect. Dis. 155:9399. 15. Barclay, M. L., and E. J. Begg. 1994. Aminoglycoside toxicity and relation to dose regimen. Adverse Drug React. Toxicol. Rev. 13:207234. 16. Hodson, M. E., A. R. L. Penketh, and J. C. Batten. 1981. Aerosol carbenicillin and gentamicin treatment of Pseudomonas aeruginosa infection in patients with cystic fibrosis. Lancet November 21:11371139. 17. Penketh, A. R. L. 1982. Long-term use of aerosol antibiotics in adults with cystic fibrosis. In Proceedings of the 11th Annual Meeting of the European Working Group in Cystic Fibrosis. Belgian Cystic Fibrosis Association, Brussels, Belgium. 18. Wall, M. A., A. B. Terry, J. Eisenberg, M. McNamara, and R. Cohen. 1983. Inhaled antibiotics in cystic fibrosis. Lancet 1(8337):1325. 19. Littlewood, J. M., M. G. Miller, A. T. Ghoneim, and C. H. Ramsden. 1985. Nebulized colomycin for early Pseudomonas colonization in cystic fibrosis. Lancet 1(8433):865. 20. Stead, R. J., M. E. Hodson, and J. C. Batten. 1987. Inhaled ceftazidime compared with gentamicin and carbenicillin in older patients with cystic fibrosis infected with Pseudomonas aeruginosa. Br. J. Dis. Chest 81: 272278. 21. Eisenberg, J., M. Pepe, J. Williams-Warren, M. Vasiliev, A. B. Montgomery, A. L. Smith, B. W. Ramsey, and the Aerosolized Tobramycin Study Group. 1997. A comparison of peak sputum tobramycin concentration in patients with cystic fibrosis using jet and ultrasonic nebulizer systems. Chest 111:955962. 22. Newman, S. P., A. R. Clark, N. Talaee, and S. W. Clarke. 1991. Lung deposition of 5mg Intal from a pressurized metered dose inhaler assessed by radiotracer technique. Int. J. Pharmaceutics 74:203208. 23. Irani, F. A., N. L. Jones, M. Gent, and M. T. Newhouse. 1972. Evaluation of disodium cromoglycate in intrinsic and extrinsic asthma. Am. Rev. Respir. Dis. 106:179185. 24. Apparatus A. British Pharmacopoeia 1993, Appendix XVII. 1998. Her Majestys Stationery Office, London. 25. Nantel, N. P., M. T. Newhouse, and M. R. Sears. 1996. Pediatric inspiratory flows through a novel multi-dose DPI (ClickhalerTM) in pediatric

26. 27. 28.

29.

30.

31.

32.

33. 34.

35.

36. 37. 38.

39.

40.

41.

42.

patients with asthma. In R. N. Dalby, P. R. Byran, and S. T. Farr, editors. Respiratory Drug Delivery, 5th ed. InterPharm Inc., Buffalo Grove, IL. 386388. Nantel, N. P., J. C. Smerek, and M. T. Newhouse. 1995. Flow target system for DPI clinical trials (abstract). J. Aerosol Med. 8:A103. Newhouse, M. T. 1998. The current laboratory determination of respirable mass is not clinically relevant. J. Aerosol Med. 11(Suppl.):S122S132. Bosso, J. A., M. V. Relling, P. L. Townsend, and L. M. Strand. 1987. Intrapatient variation in aminoglycoside disposition in cystic fibrosis. Clin. Pharm. 6:5458. Morin, S., J. Tetrault, L. James, J. E. Hoppe-Bauer, and M. Pezzlo. 1994. Specimen acceptability: evaluation of specimen quality. In H. D. Isenberg, editor. Clinical Microbiology Procedures Handbook. American Society of Microbiology, Washington, DC. Wong, K., M. C. Roberts, L. Owens, M. Fife, and A. L. Smith. 1984. Selective media for the quantitation of bacteria in cystic fibrosis sputum. J. Med. Microbiol. 17:113119. Boe, J., G. Stiksa, K. Svensson, and E. Asbrink. 1992. New method of evaluating patient preference for different inhalation delivery systems. Ann. Allergy 68:255260. Kleinbaum, D. G., L. L. Kupper, and K. E. Miller. 1988. Applied Regression Analysis and Other Multivariable Methods, 2nd ed. Wadsworth Publishing Company, Belmont, CA. Norman, G. R. and D. L. Streiner. 1994. Biostatistics: The Bare Essentials. Mosby-Year Book Inc., Toronto. Lourenco, R. V., R. Loddenkemper, and R. W. Carton. 1972. Patterns of distribution and clearance of aerosols in patients with bronchiectasis. Am. Rev. Respir. Dis. 106:857866. Touw, D. J., R. W. Brimicombe, M. E. Hodson, H. G. M. Heijerman, and W. Bakker. 1995. Inhalation of antibiotics in cystic fibrosis. Eur. Respir. J. 8:15941604. Ryrfeldt, A. 1990. The bronchial circulationa significant local distribution system in the lung, in inhalation therapy? J. Aerosol. Med. 3:165168. Deffebach, M. E., N. B. Charan, S. Lakshminarayan, and J. Butler. 1987. The bronchial circulation. Am. Rev. Respir. Dis. 135:463481. Goldman, J. M., S. M. Bayston, S. OConnor, and R. E. Meigh. 1990. Inhaled micronised gentamicin powder: a new delivery system. Thorax 45:939940. Conway, S. P., A. Watson, M. N. Pond, S. Bayston, R. Booth, and A. Ghonheim. 1993. Sputum gentamicin levels after delivery by rotahaler and nebulizer. Pediatr. Pulmonol. 9(Suppl.):250. Mendelman, P. M., A. L. Smith, J. Levy, A. Weber, B. Ramsey, and R. L. Davis. 1985. Aminoglycoside penetration, inactivation, and efficacy in cystic fibrosis sputum. Am. Rev. Respir. Dis. 132:761765. Levy, J., A. L. Smith, M. A. Kenny, B. Ramsey, and F. D. Schoenknecht. 1983. Bioactivity of gentamicin in purulent sputum from patients with cystic fibrosis or bronchiectasis: comparison with activity in serum. J. Infect. Dis. 148:10691076. Bryant, R. E. and D. Hammond. 1974. Interaction of purulent material with antibiotics used to treat Pseudomonas infections. Antimicrob. Agents Chemother. 6:702707.