INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pseudomonas aeruginosa is the most common respiratory pathogen infecting patients with cystic fibrosis (CF) and is cultured from 60% of CF patients in the United States (1). Chronic infection with P. aeruginosa in CF patients is associated with poorer median survival (2) and a lower mean FEV₁ (3) than in those who have never been infected with this pathogen.

Recent randomized, placebo-controlled trials have shown that combination antibiotic therapy with two antipseudomonal antibiotics is superior to antibiotic monotherapy, and results in a significant reduction in bacterial colony counts and a more prolonged clinical remission in CF patients who are infected with P. aeruginosa (4, 5). Unfortunately, many CF-associated P. aeruginosa isolates are now multiresistant to single antibiotics. Traditional microbiology laboratory susceptibility results for single antibiotics are often inadequate for these multiresistant bacteria, with the result that the choice of a combination antibiotic regimen for CF patients infected with multiresistant P. aeruginosa can often only be made empirically. The danger with this approach is that empirically chosen antibiotic combinations may not be bactericidal against multiresistant P. aeruginosa and may result in a less than optimal treatment outcome.

In response to the challenge of multiresistant bacteria in patients with CF, our laboratory has designed a method to systematically test bacterial isolates against multiple combinations of antibiotics in order to determine sensitivity patterns. Multiple-combination bactericidal testing (MCBT) involves a modified time-versus-kill-curve method that provides combination sensitivity data to clinicians within 48 to 72 h of isolation of the bacterial species from sputum. In the past 11 yr, a number of CF centers across North America have referred multiresistant isolates to our laboratory for evaluation. The purpose of the present study was to analyze MCBT data collected from 75 consecutively received multiresistant isolates of P. aeruginosa from 44 CF patients in order to determine in vitro susceptibility patterns to single-, double-, and triple-antibiotic combinations.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Definitions

Multiresistance was defined as resistance to all of the antibiotics in two of the following three antibiotic classes: (1) -lactams, including piperacillin, aztreonam, and imipenem. (2) Aminoglycosides, including amikacin, gentamicin and tobramycin. (3) Fluoroquinolones, and particularly ciprofloxacin.

This definition was developed at the Microbiology and Infectious Diseases Consensus Conference sponsored by the Cystic Fibrosis Foundation in 1994 (6).

Bactericidal activity was defined as the absence of growth on subculture (from a nonturbid microtiter well) of an organism in the presence of antibiotics.

Nonbactericidal activity was defined as the growth of an organism as exemplified by well turbidity or growth on subculture despite the presence of antibiotics.

Antagonism was defined as the growth of an organism when an additional antibiotic was added to a previously bactericidal combination.

Source of Clinical Isolates

From June 1988 to June 1999, physicians referred P. aeruginosa isolates, recovered from sputum samples of CF patients, that were considered multiresistant according to single agent antibiotic susceptibility testing. Isolates were received from 14 centers situated in four Canadian provinces (Quebec, Ontario, Alberta, and British Columbia) and one state in the United States (Pennsylvania). A total of 104 isolates of P. aeruginosa from 52 patients were consecutively received; however, 29 isolates from eight patients were excluded from our analysis because they did not meet the criteria of multiresistance as defined earlier (Consensus Conference 1994 [6]).

Seventy-five P. aeruginosa isolates from 44 CF patients were included in the final analysis. Seventy percent of the isolates were of a mucoid phenotype. The mean age of the 44 patients at the time of first isolate referral was 24 ± 8 (mean ± SD) yr, and 25% of the patients were male.

Identification of Species of Isolates

Isolates were sent by courier to our laboratory after phenotypic identification and single agent susceptibility testing in the 14 referral centers' own laboratories. In 16 cases, more than one isolate of P. aeruginosa was cultured from a single sputum sample. Multiple P. aeruginosa isolates from the same patient were determined to be distinct from one another if they had different phenotypes and antibiograms. In eight cases, multiple isolates from the same patient were referred to our reference laboratory on two different dates. The average time between successive instances of MCBT was 22 ± 24 mo. The isolates tested were considered distinct if they had different antibiograms.

MCBT

Antibiotics. With the exception of high-dose tobramycin, antibiotic concentrations for MCBT testing were chosen on the basis of published estimates of the average peak levels seen in serum after standard single-dose intravenous administration (7). The concentration of high-dose tobramycin for MCBT (200 µg/ml) was chosen to reflect concentrations of tobramycin found in the sputum and airway-lining fluid following administration of nebulized tobramycin (8). Antibiotics used for MCBT testing, and their concentrations, have been previously described (9).

Test procedure. MCBTs were performed according to previously published methods (9).

Data Analysis

Significance of differences between two antibiotic groups was determined with McNemar's test for correlated proportions, and p values were reported for clinically significant comparisons. Results were reported as mean ± 1 SD unless otherwise stated.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Susceptibility of P. aeruginosa to Single Antibiotics

Each isolate of P. aeruginosa was tested against 11.5 ± 2.5 (mean ± SD) single antibiotics. The most effective single antibiotics were meropenem and tobramycin (4 µg/ml), which were bactericidal against 44% (24 of 54), and 16% (12 of 75) of isolates, respectively (Figure 1). Meropenem showed significantly more in vitro bactericidal activity than did the other intravenous antibiotics (p < 0.05).

View larger version (31K): [in this window] [in a new window]   Figure 1.   Bactericidal activity of single antibiotics against multiply resistant P. aeruginosa. The most effective drugs used alone were high-dose tobramycin (200 µg/ml), meropenem, low-dose tobramycin (4 µg/ ml), and cefepime, which were bactericidal against 72%, 44%, 16%, and 11%, respectively, of in vitro isolates of the organism.

Although low dose tobramycin (4 µg/ml) was bactericidal against only 16% of isolates, tobramycin tested in high concentrations (200 µg/ml) was found to be significantly more effective, and was bactericidal against 72% (54 of 75) of isolates (p < 0.0001).

Overall, 53% (40 of 75) of isolates were resistant to all single antibiotics including tobramycin tested at 4 µg/ml. However, when high-dose tobramycin was included in the test panel, the proportion of isolates resistant to all single antibiotics dropped to 13% (10 of 75).

Susceptibility of P. aeruginosa to Double-Antibiotic Combinations (Excluding High Dose Tobramycin)

Isolates were tested against 63 ± 23 double-antibiotic combinations (excluding high-dose tobramycin). Meropenem/ciprofloxacin was the most effective intravenous double-antibiotic combination, and was bactericidal against 85% (45 of 53) of isolates. The next most effective combinations were meropenem/ tobramycin (4 µg/ml), meropenem/cefepime, meropenem/amikacin, and meropenem/ceftazidime (Table 1). In total, double combinations containing meropenem were bactericidal in 59% (348 of 589) of tests performed.

Susceptibility of P. aeruginosa to Double-Antibiotic Combinations Containing High-Dose Tobramycin

The bactericidal activity of high-dose tobramycin was significantly enhanced by the addition of a second antibiotic. The most effective double-antibiotic combinations contained high-dose tobramycin (200 µg/ml) plus meropenem, piperacillin/tazobactam, or ciprofloxacin, and were bactericidal against 88% to 94% of isolates (Table 2).

Each isolate was susceptible to at least one double-antibiotic combination containing high-dose tobramycin. Among all the isolates, double-antibiotic combinations that contained high-dose tobramycin were bactericidal in 83% (645 of 774) of tests performed. In contrast, double-antibiotic combinations containing low-dose tobramycin (4 µg/ml) were significantly less effective, and were bactericidal in only 23% (183 of 780) of tests performed (p < 0.001).

Antagonism with Addition of a Second Antibiotic to a Bactericidal Single Antibiotic

Fifty percent (12 of 24) of isolates sensitive to meropenem as a single agent showed antagonism (growth of the isolate when a second antibiotic was added to a previously bactericidal antibiotic). The most common antagonistic combinations were meropenem/azithromycin, meropenem/minocycline, and meropenem/ticarcillin/clavulanic acid (Table 3).

Of the 54 isolates sensitive to high-dose tobramycin as a single agent, 11% (six of 54) showed antagonism. The antibiotic combinations exhibiting this antagonism are shown in Table 3.

Overall, antagonism was seen in 29% of the 61 isolates that were sensitive to tobramycin (200 µg/ml) and/or meropenem monotherapy.

Susceptibility of P. aeruginosa to Triple-Antibiotic Combinations

Overall, triple-antibiotic combinations were bactericidal in 49% (7,937 of 16,274) of the tests performed, as compared with 32% (1,760 of 5,501) of double-antibiotic combinations in tests of the latter (p < 0.001). However, we were unable to detect a significant increase in bactericidal activity upon addition of a third antibiotic to any particular double-antibiotic combination (Table 4).

Triple-antibiotic combinations containing high-dose tobramycin were bactericidal in 85% (3,340 of 3,924) of tests performed, as compared with 83% (645 of 774) of double-antibiotic combinations in tests of the latter (p = 0.15).

Triple-antibiotic combinations containing meropenem (excluding high-dose tobramycin) were bactericidal in 66% (1,942 of 2,940) of tests performed, as compared with 59% (348 of 589) of double antibiotic combinations containing meropenem (p < 0.001). The most successful bactericidal triple-antibiotic combination was meropenem/ciprofloxacin/cefipime; however, this combination was not significantly more effective than the double-antibiotic combination of meropenem/ciprofloxacin (91% versus 85% bactericidal activity, respectively; p = 0.45) (Table 4).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Multiresistant P. aeruginosa present a great challenge to physicians in decisions about antibiotic therapy. Traditional antibiotic susceptibility testing methods are of limited value with multiresistant organisms, since they test susceptibility only to single agents. When done on CF isolates of P. aeruginosa, these traditional susceptibility tests often reveal resistance to many, if not all, of the single antibiotics tested. Empiric antibiotic selection is the usual option in the face of a multiresistant P. aeruginosa isolate.

Our study demonstrated that choosing antibiotic combinations empirically can be potentially counterproductive. In the case of 29% of the P. aeruginosa isolates that were sensitive to a single antibiotic, adding a second, empirically chosen antibiotic resulted in antagonism with loss of bactericidal effect. These results underscore the importance of using MCBT to reveal a nonantagonistic combination of antibiotics. With MCBT, isolates can be tested against numerous combinations of antibiotics, with results available within 48 to 72 h after isolation of the bacterial species from sputum. Hence, if sputum is cultured at the time of a pulmonary exacerbation of CF, MCBT results can be used to select an appropriate bactericidal combination of antibiotics and to discontinue nonbactericidal antibiotic therapy within days of the exacerbation. Susceptibilities in MCBT can also be determined when the CF patient is clinically stable, to allow clinicians to decide on appropriate antibiotic combination therapy in advance of the patient's next pulmonary exacerbation. Each multiple combination bactericidal test consumes approximately 8 to 10 h of technician time, and costs approximately U.S. $300 per isolate.

Results of the present study suggest that double-antibiotic combinations are more likely than single antibiotics to be bactericidal in vitro against multiresistant isolates of P. aeruginosa. However, the addition of a third antibiotic to a double-antibiotic combination did not significantly increase in vitro bactericidal effects against multiresistant P. aeruginosa. These results are in contrast to those of our previous study of MCBT of Burkholderia cepacia organisms isolated from CF sputum, in which we found that addition of a third antibiotic significantly enhanced in vitro bactericidal activity against B. cepacia (9).

In a previous study, by Saimen and colleagues, resistance to low concentrations of tobramycin was found to be related to specific permeability mutations that impeded the ability of aminoglycodes to enter P. aeruginosa organisms (10). This resistance mechanism can sometimes be overcome by higher aminoglycoside concentrations. Unfortunately, these higher concentrations of aminoglycosides are above those obtainable in the serum without producing serious systemic side effects.

In the present study, P. aeruginosa isolates were tested against 200 µg/ml of tobramycin in MCBT. Although this concentration is not achievable with intravenous therapy, sputum concentrations of tobramycin above this level are readily achievable by aerosol administration of this drug. Jet nebulization of 300 mg of tobramycin in CF patients has been shown to result in mean tobramycin concentrations of 489 to 1,498 µg/ml sputum (8). Neubulization therapy with high doses of inhaled aminoglycosides is an attractive option, since it produces higher local levels of antibiotic while avoiding the systemic toxicities associated with intravenous administration.

In our study, low dose tobramycin (4 µg/ml) as a single agent was bactericidal against only 16% of isolates; however, 72% of isolates were killed with high-dose tobramycin (200 µg/ ml). Overall, high-dose tobramycin (200 µg/ml) combined with another antipseudomonal antibiotic resulted in the highest bacterial kill rates in vitro. Double-antibiotic combinations containing high dose tobramycin were significantly more bactericidal than were double-antibiotic combinations containing low-dose tobramycin. Our results suggest that in CF patients colonized with multiresistant P. aeruginosa, the most frequently effective bactericidal activity was achieved by using inhaled tobramycin plus a second intravenous antibiotic (meropenem, piperacillin/ tazobactam, ciprofloxacin, aztreonam, or ceftazidime).

At present, clinical trials have not yet been performed to determine whether inhaled antibiotics are beneficial in acute exacerbations of CF with multiresistant P. aeruginosa; however, clinical trials have shown benefit with chronic intermittent suppressive therapy. Trials of intermittent administration of inhaled tobramycin in CF patients colonized by P. aeruginosa have revealed a significant decrease in the density of P. aeruginosa in sputum, an increase in lung function, and decreased hospital admission in the treatment group (11, 12). No significant systemic toxicity was associated with inhaled tobramycin therapy (13).

Aside from high-dose tobramycin, meropenem was the most effective bactericidal antibiotic against the P. aeruginosa isolates in our study. As a single agent, meropenem had greater in vitro bactericidal activity than imipenem or the cephalosporins that we tested. Furthermore, meropenem in combination with ciprofloxacin was the most effective double combination of intravenous antibiotics in our in vitro tests. The superiority of meropenem may reflect its duration of use in the CF population. The drug is relatively new, and development of significant mechanisms of resistance to it may not yet have emerged in North America (14).

The results of the present study suggest that in vitro testing of the bactericidal activity of combinations of antibiotics may afford clinicians the ability to select more effective antibiotic combinations and avoid antagonistic ones. Further work, including randomized controlled clinical trials, is required to fully evaluate the usefulness of MCBT and to determine whether the routine use of MCBT results in decreased bacterial sputum density and improved clinical outcomes with antibiotic therapy in patients with CF who are colonized with multiresistant P. aeruginosa.