American Journal of Respiratory and Critical Care Medicine

Most Burkholderia cepacia strains are resistant to many, or all, of the antibacterial agents commonly used in cystic fibrosis (CF), and selection of appropriate antibiotics for treatment of pulmonary exacerbations is therefore difficult. We developed a technique for rapid in vitro testing of multiple antibiotic combinations for B. cepacia isolates. For each of 119 multi-drug-resistant isolates of B. cepacia, our multiple combination bactericidal test (MCBT) studied the bactericidal activity of 10 to 15 antimicrobial agents using 225 ± 97 single, double, and triple antibiotic combinations. Of the 119 isolates, 50% were resistant to all single antibiotics tested, 8% were resistant to all two-drug antibiotic combinations, but all were inhibited by at least one bactericidal triple-drug combination. When used alone, meropenem, ceftazidime and high-dose tobramycin (200 μ g/ml) were bactericidal against only 47, 15, and 14% of in vitro isolates, respectively. Using a double antibiotic combination improved bactericidal activity; meropenem-minocycline, meropenem-amikacin, and meropenem-ceftazidime combinations were bactericidal against 76, 73, and 73% of isolates, respectively. However, 47% of isolates demonstrated antagonism (growth of an organism when a second antibiotic was added to a bactericidal single antibiotic). Triple antibiotic combinations that contained tobramycin, meropenem, and an additional antibiotic were most effective, and were bactericidal against 81 to 93% of isolates. We conclude that triple-antibiotic combinations are more likely than double and single antibiotic combinations to be bactericidal against B. cepacia in vitro. MCBT testing is a useful technique to help clinicians decide on appropriate nonantagonistic combination antibiotic therapy for patients with CF infected with B. cepacia.

Burkholderia cepacia bacteria was first recovered from sputum cultures of patients with cystic fibrosis (CF) in the late 1970s (1). By 1984 a syndrome characterized by severe progressive respiratory failure, necrotizing pneumonia, and B. cepacia bacteremia was described and was termed “cepacia syndrome” (2). Since then B. cepacia has been recognized as a significant CF-associated airway pathogen that causes considerable morbidity in affected patients with CF (3).

Recent studies using computer-assisted cluster analysis of whole-cell protein profiles and fatty acid content together with DNA hybridization techniques have demonstrated that organisms identified as B. cepacia by conventional tests actually constitute five distinct genomic species (4). These five species are referred to as genomovars of the B. cepacia complex. Although all five genomovars have been identified from isolates from patients with CF, the majority of B. cepacia isolated from respiratory secretions of patients with CF belong to genomovars II and III (4).

B. cepacia infects a relatively small proportion of patients with CF. Four percent of 17,620 patients with CF registered in the U.S. National Cystic Fibrosis Patient Registry grew B. cepacia from their respiratory secretions in 1996 (5). In Canada the prevalence of B. cepacia colonization is higher; in 1997 7% of 3,200 Canadian patients with CF harbored B. cepacia in their respiratory secretions (6).

Despite the relatively small proportion of patients with CF infected with B. cepacia the organism has a major impact on CF morbidity and mortality. In 1995 the median survival of all American patients with CF was 30 yr; however, for those patients colonized with B. cepacia, median survival was only 21 yr (5). Several case-control studies have established that colonization with B. cepacia is an independent negative prognostic factor that carries with it an increased risk of death and an accelerated decline in pulmonary function (7, 8).

A major aspect of CF patient care has centered around antibiotic therapy to ameliorate the effects of acute pulmonary infections/exacerbations in these patients. Studies published during the past decade have provided evidence that although eradication of organisms from the lower respiratory tract is not possible in patients with CF, decreased bacterial density and decreased bacterial virulence factor production after antibiotic therapy leads to a decrease in inflammation and clinical improvement in these patients (9-13). In addition, recent randomized, placebo-controlled trials have shown that combination antibiotic therapy with two antipseudomonal antibiotics (a β-lactam and an aminoglycoside) is superior to antibiotic monotherapy for patients with CF exacerbations who are colonized with Pseudomonas aeruginosa (9, 10).

Unfortunately, optimal combination antibiotic therapy for patients with CF exacerbations who are colonized with B. cepacia organisms is problematic since most B. cepacia organisms are panresistant to most individual antibiotics. The organisms are intrinsically resistant to aminoglycoside antibiotics and have high levels of beta-lactam resistance because of inducible chromosomal beta-lactamases and altered penicillin-binding proteins (14). To further complicate matters, B. cepacia also possesses an antibiotic efflux pump that mediates resistance to chloramphenicol, quinolones, and trimethroprim (15).

All of these factors make it very difficult for physicians to empirically select a bactericidal antibiotic regimen for patients colonized with B. cepacia based on traditional single-antibiotic microbiology laboratory sensitivity results since these studies usually reveal resistance to many, or all, of the single antibiotics tested. We therefore set out to develop an in vitro method for testing and evaluating single, double, and triple antibiotic combinations for bactericidal activity against B. cepacia isolates. We have systematically studied the activity of a total of 15 antimicrobial agents used in a total of 430 potential combinations using a modification of the time-kill curve method (16). Our method of multiple combination bactericidal testing (MCBT) provides information on bactericidal antibiotic combinations within 48 to 72 h, so that information is available in a timely enough fashion to be used clinically to alter antibiotic therapy for patients with CF exacerbations associated with B. cepacia.

The objective of this study was to report on our results of multiple combination bactericidal antibiotic testing performed on 119 consecutively received B. cepacia isolates over a 10-yr period. It is hoped that results of this study will help to guide clinicians in their choice of empiric combination antibiotic therapy for treatment of CF exacerbations associated with B. cepacia.

Definitions

Multiresistance. Resistance to all agents in two or more of the following three antimicrobial categories: β-lactam antibiotics, aminoglycosides, and fluoroquinolones. This definition was developed at the Microbiology and Infectious Diseases Consensus Conference sponsored by the Cystic Fibrosis Foundation in 1994.

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

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

Antagonism. Growth of an organism when an additional antibiotic is added to a previously bactericidal combination.

Source of Clinical Isolates

From March 1990 through April 1999, physicians caring for patients with CF referred multiply resistant strains of B. cepacia to The Children's Hospital of Eastern Ontario for multiple combination bactericidal testing. A total of 119 multiresistant B. cepacia isolates from 51 patients were referred. Isolates were received from 17 centers situated in six Canadian provinces (Nova Scotia, New Brunswick, Quebec, Ontario, Alberta, and British Columbia) and two U.S. states (North Carolina and Pennsylvania).

The mean age of the 51 patients at the time of first isolate referral was 25.8 ± 9.3 yr. Thirty-four of 51 patients (68%) were male.

Phenotypic Identification of Organisms

The identification of the multiply resistant isolates was confirmed as Burkholderia cepacia both at the referring laboratory and at The Children's Hospital of Eastern Ontario with the use of a combination of a medium selective for B. cepacia and biochemical tests. Isolates were cultured on oxidative-fermentative base, polymyxin B, bacitracin, lactose (OFPBL) selective medium (17) for 24 to 72 h at 35° C for isolation. Identification was confirmed with the following biochemical tests: oxidation of glucose, xylose, 10% lactose and lysine decarboxylase, weakly positive oxidase reaction with the Pathotec cytochrome oxidase strip (Remel, Lenexa, KS), and resistance to polymyxin B (18).

Forty-five percent of patients had more than one isolate of B. cepacia cultured from a single sputum sample. Multiple isolates from the same patient were determined to be distinct from one another if the B. cepacia isolates had different phenotypes and antibiograms. In the case of 14 patients, multiple isolates of B. cepacia from the same patient were referred to our reference laboratory on different dates for MCBT testing. The average length of time between successive MCBT testing for these 14 patients was 13 ± 17 mo. Isolates from the same patient taken over time were determined to be distinct from one another for the purposes of our analysis if each B. cepacia isolate had a unique antibiogram. In two cases, B. cepacia isolates received from the same patient at different times had the same antibiogram. These isolates were included only once in the analysis.

Genotypic Identification of Organisms

B. cepacia isolates from 38 of the 51 patients were typed by random amplified polymorphic DNA (RAPD) analysis to determine strain relatedness. Briefly, genomic DNA was released from the bacteria by mechanical disruption with glass beads, then extracted from extraneous protein by ammonium acetate and chloroform extraction, followed by two ethanol washes to purify the extract. The DNA concentration was calibrated, then subject to RAPD analysis using a 10-base primer no. 270 as described previously (19). RAPD fingerprint profiles were compared visually and with the aid of computer analysis (GelManager for Windows, Biosytematica, Prague, Czech Republic) to standard patterns that previously had been subjected to genomovar analysis (4). Isolates sharing identical RAPD patterns were considered to belong to the same genotype.

Multiple Combination Bactericidal Test Procedures

Antibiotics. When MCBT testing was begun in 1990, 10 antibiotics were tested singly and in combinations for bactericidal activity. However, as new antibiotics became commercially available during the 1990s (such as meropenem, azithromycin) these were added to the testing panel.

The following antibiotics were used singly and in combinations for multiple combination bactericidal testing: ceftazidime (32 μg/ml), tobramycin (200 μg/ml), amikacin (32 μg/ml), cloxacillin (20 μg/ml), and chloramphenicol (20 μg/ml) all from Sigma Pharmaceuticals (St. Louis, MO), azithromycin (0.4 μg/ml) from Pfizer Canada Inc. (Montreal, PQ, Canada), imipenem (10 μg/ml) from Merck Sharp Dohme Canada Inc. (Montreal, PQ, Canada), piperacillin/tazobactam (32/4 μg/ml) from Lederle Cyanamid Canada Inc. (Montreal, PQ, Canada), aztreonam (32 μg/ml) and cefepime (32 μg/ml) from Bristol Meyer Squibb Pharmaceutical (St. Laurent, PQ, Canada), trimethoprim/sulfamethoxazole (10/2 μg/ml) from Roche Laboratories Canada, meropenem (32 μg/ml) from Zeneca Pharma Inc. (Mississauga, ON, Canada), ticarcillin/clavulanic acid (32/10 μg/ml) from SmithKline Beecham Pharma Inc. (Oakville, ON, Canada), minocycline (2.0 μg/ml) from Novopharm Inc. (Toronto, ON, Canada), and ciprofloxacin (2.0 μg/ml) from Miles Laboratories (Etobicoke, ON, Canada).

MCBT procedure. MCBTs were carried out in four 96 well, round-bottomed microtiter plates (Nunc Inc., Roskilde, Denmark). Antibiotic solutions were prepared for each test from stock solutions stored at −80° C. The working antibiotic solutions were prepared in Mueller Hinton II Cation Adjusted Broth (MHB II broth; Becton Dickenson Microbiology Systems, Cockeysville, MD) at 10 times the required concentrations. For each MCT, antibiotic working solutions were made fresh on the day of inoculation. One, two, or three antibiotics were added, each in 10-μl volumes to the appropriate wells. The necessary volume of MHB II was then added to the wells containing one or two antibiotics so that all the wells had a volume of 30 μl before the addition of organism. The organism inoculum consisted of 70 μl of a 100-fold dilution of a 0.5 McFarland turbidity standard prepared from a culture in Tryptone Soya Broth (Oxoid Laboratories, Basingstoke, UK) in the growth phase. This gave a final inoculum concentration in each well of 5 × 105 CFU/ml. Growth and sterility plates (no antibiotics and no organism inoculum, respectively) were run with each MCT procedure as bacteriologic controls. Plates were incubated at 35° C for 48 h. At 24 and 48 h the wells were examined for turbidity. The contents of nonturbid wells at 48 h were subcultured by streaking 10 μl of suspension onto 5% Columbia sheep blood agar plates (PML Microbiologicals, Mississauga, ON, Canada), which was incubated for 24 h at 35° C and examined for 99.9% kill the next day.

Reproducibility of the MCBT results was confirmed by repeated testing of eight selected isolates over a 2-yr period. Each isolate showed identical MCBT susceptibility results with repeated testing (16).

Data Analysis

Data analysis was performed using SPSS software version 8.0 (SPSS Inc., Chicago, IL). Results are reported as means ± 1 SD.

B. cepacia Genotype Analysis

Thirty-eight of the 51 patients (75%) had B. cepacia isolates analyzed by RAPD. Thirty-four patients had isolates that were RAPD 02, two patients had isolates that were RAPD 01, and two patients had isolates that were RAPD 04. One patient had two different isolates belonging to RAPD groups 02 and 04. RAPD groups 01, 02, and 04 are strains belonging to genomovar III. One patient had an isolate that was a member of the B. cepacia complex, but genomovar status could not be determined.

Resistance of B. cepacia to Combination Antibiotic Therapy

Fifty percent of the 119 isolates were resistant to all single antibiotics tested. Eight percent of the isolates were not inhibited by any two-drug antibiotic combination. All the isolates were inhibited by at least one bactericidal triple-drug combination (Figure 1).

Susceptibility of B. cepacia to Single Antibiotics

Each isolate was tested against a minimum of 10 single antibiotics and a maximum of 15. Sixty of the 119 isolates (50.4%) were resistant to all single antibiotics tested. In total, only 93 of 1,277 (7.3%) of single antibiotic tests showed in vitro bactericidal activity against the B. cepacia isolates. The most effective single antibiotics were meropenem, ceftazidime, and high-dose tobramycin (200 μg/ml), which were bactericidal against 47% (41/86), 15% (17/110), and 14% (16/117) of the in vitro isolates, respectively (Figure 2).

Susceptibility of B. cepacia to Double Antibiotic Combinations

Isolates were tested against a mean of 53 ± 17 double-antibiotic combinations. 18.5% of double-antibiotic combinations were bactericidal (1,167/6,322 combinations). Each isolate was sensitive to an average of 9.8 ± 7.9 double antibiotic combinations; however, 8% of the 119 isolates were not inhibited by any two-drug antibiotic combination, and in a further 9% of isolates only one effective bactericidal two-drug combination was found.

The most effective double antibiotic combinations were meropenem-minocycline, meropenem-amikacin, meropenem-ceftazidime, meropenem-chloramphenicol, and meropenem-tobramycin, which were bactericidal against 76% (25/33), 73% (63/86), 73% (54/74), 71% (58/82), and 64% (52/81) of isolates, respectively (Table 1). Among all the B. cepacia strains, a double antibiotic combination that contained meropenem was bactericidal in 558/902 cases (62%). Tobramycin-ceftazidime was the most effective double antibiotic combination that did not contain meropenem, and was bactericidal against 52% (59/113) of isolates.

Table 1. BACTERICIDAL ACTIVITY OF THE MOST EFFECTIVE DOUBLE ANTIBIOTIC COMBINATIONS

Double Antibiotic Combination B. cepaciaIsolates Inhibited (%)
Meropenem-minocycline76
Meropenem-amikacin73
Meropenem-ceftazidime73
Meropenem-chloramphenicol71
Meropenem-tobramycin64
Meropenem-cefipime62
Meropenem-trimethoprim/sulfamethoxasole57
Meropenem-aztreonam56
Meropenem-piperacillin/tazobactam56
Tobramycin-ceftazidime52

Enhancement of Bactericidal Activity with Addition of a Second Antibiotic to a Nonbactericidal Single Antibiotic

The additional bactericidal effect of adding a second antibiotic was examined for the three most effective single antibiotics (meropenem, ceftazidime, and tobramycin). The addition of a second antibiotic to meropenem resulted in bactericidal activity against 39 of 45 isolates (87%) that had been resistant to meropenem monotherapy. Similarly, the addition of a second antibiotic to ceftazidime and to tobramycin resulted in bactericidal activity in 71 of 93 isolates (76%) and 70 of 101 isolates (69%) that had been resistant to ceftazidime and to tobramycin monotherapy, respectively.

A separate analysis was performed for the 45 of 86 isolates that had been tested for and found to be resistant to meropenem monotherapy. These meropenem-resistant isolates were tested against a mean of 62 ± 9 double antibiotic combinations. Each of these isolates was sensitive to an average of 6 ± 5 double antibiotic combinations; however, 13% of the meropenem-resistant isolates were found to be resistant to every double antibiotic combination tested, and an additional 13% were sensitive to only one double antibiotic combination. The most effective bactericidal double antibiotic combinations for isolates that had been resistant to meropenem monotherapy were meropenem-tobramycin, meropenem-ceftazidime, meropenem-amikacin, and tobramycin-ceftazidime, which were bactericidal in 54, 51, 49, and 38% of the 45 isolates, respectively.

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

Fifty-nine of the 119 isolates were sensitive to meropenem, tobramycin or ceftazidime used as monotherapy. Of these 59 isolates, antagonism (growth of an organism when a second antibiotic was added to a previously bactericidal single antibiotic) was observed in 28 isolates (47%). The proportion of isolates that demonstrated antagonism ranged from 65% when a second antibiotic was added to isolates that had been sensitive to ceftazidime alone, to 31% when a second antibiotic was added to isolates that had been sensitive to meropenem or to tobramycin alone.

For those isolates that had been sensitive to meropenem monotherapy the most common antagonistic double antibiotic combination was meropenem-tobramycin. Of the 41 isolates that had been sensitive to meropenem alone, the addition of high-dose tobramycin resulted in antagonism of the bactericidal activity in nine (22%) (Table 2). Similarly, for those isolates that had been sensitive to ceftazidime monotherapy, the addition of imipenem to ceftazidime resulted in antagonism of bactericidal activity among nine of 17 (53%) previously ceftazidime-sensitive isolates (Table 2).

Table 2. ANTAGONISTIC DOUBLE ANTIBIOTIC COMBINATIONS

Single Antibiotic (No. of sensitive isolates)Antagonistic Double Antibiotic CombinationsSensitive Isolates in which Double Antibiotic Combination Was Antagonistic
n(%)
Meropenem, n = 41Meropenem-tobramycin9/4122
Meropenem-azithromycin5/4112
Meropenem-cefipime3/417.3
Meropenem-ceftazidime2/41 5
Meropenem-trimethoprim/sulfamethoxasole1/41 2
Meropenem-aztreonam1/41 2
Meropenem-minocycline1/41 2
Ceftazidime, n = 17Ceftazidime-imipenem9/1753
Ceftazidime-aztreonam5/1729
Ceftazidime-azithromycin5/1729
Ceftazidime-trimethoprim/sulfamethoxasole4/1724
Ceftazidime-cefipime4/1724
Ceftazidime-tobramycin3/1718
Ceftazidime-amikacin3/1718
Ceftazidime-ciprofloxacin2/1711
Ceftazidime-meropenem2/1711
Ceftazidime-piperacillin/tazobactam2/1711
Tobramycin, n = 16 Tobramycin-chloramphenicol3/1619
Tobramycin-ceftazidime2/1613
Tobramycin-minocycline1/16 6

In general, this observed antagonism to two-drug combinations could be overcome if a third antibiotic was added in vitro. For example, eight of the nine isolates that demonstrated antagonism to meropenem-tobramycin could be killed when any third antibiotic was added (the ninth isolate remained resistant to all three-drug combinations containing meropenem-tobramycin). Similarly, all nine isolates that demonstrated antagonism to ceftazidime-imipenem could be inhibited with the addition of a third antibiotic.

Susceptibility of B. cepacia to Triple Antibiotic Combinations

Isolates were tested against a mean of 161 ± 78 triple antibiotic combinations. Each isolate was sensitive to an average of 48 ± 37 triple antibiotic combinations; however, 2.5% of the 119 isolates were sensitive to three or less three-drug combinations. At least one bactericidal triple antibiotic combination was found for all 119 isolates.

Overall, 30% (5,763/19,144) of triple antibiotic drug combinations were bactericidal. The most effective triple antibiotic combinations were tobramycin-meropenem-ceftazidime, tobramycin-meropenem-trimethoprim/sulfamethoxasole, tobramycin-meropenem-chloramphenicol, tobramycin-meropenem-aztreonam, and amikacin-meropenem-ceftazidime, which were bactericidal in 93% (69/74), 88% (76/86), 87% (71/82), 87% (75/ 86), and 87% (64/74) of isolates respectively (Table 3). Among all the B. cepacia strains, a triple antibiotic combination that contained meropenem was bactericidal in 3,129/4,392 cases (71%). Tobramycin-azithromycin-ceftazidime and tobramycin-tazocin-ceftazidime were the most effective triple antibiotic combination that did not contain meropenem, and were bactericidal against 67% (55/82) and 63% (54/86) of isolates, respectively.

Table 3. BACTERICIDAL ACTIVITY OF THE MOST EFFECTIVE TRIPLE ANTIBIOTIC COMBINATIONS

Triple Antibiotic Combination B. cepaciaIsolates Inhibited (%)
Tobramycin-meropenem-ceftazidime93
Tobramycin-meropenem-trimethoprim/sulfamethoxasole88
Tobramycin-meropenem-chloramphenicol87
Tobramycin-meropenem-aztreonam87
Amikacin-meropenem-ceftazidime87
Tobramycin-meropenem-amikacin85
Tobramycin-meropenem-piperacillin/tazobactam85
Tobramycin-meropenem-ticarcillin/clavulinate82
Meropenem-chloramphenicol-ceftazidime84
Amikacin-meropenem-azithromycin73
Tobramycin-azithromycin-ceftazidime67
Tobramycin-piperacillin/tazobactam-ceftazidime63

Enhancement of Bactericidal Activity with Addition of a Third Antibiotic to a Nonbactericidal Double Antibiotic Combination

The additional bactericidal effect of adding a third antibiotic was examined for the six most effective two-drug antibiotic combinations (Table 4). For the six two-drug combinations (meropenem-amikacin, tobramycin-ceftazidime, meropenem-tobramycin, meropenem-chloramphenicol, meropenem-ceftazidime, and meropenem-minocycline) adding in a third antibiotic enhanced bactericidal activity in 78% (18/23), 78% (42/54), 90% (26/29), 83% (20/24), 90% (18/20), and 100% (8/8) of isolates that had been resistant to the two drugs used alone (Table 4).

Table 4. EFFECT OF ADDITION OF A THIRD ANTIBIOTIC TO THE SIX MOST EFFECTIVE BACTERICIDAL DOUBLE ANTIBIOTIC COMBINATIONS

Double Antibiotic CombinationIsolates Tested (n)Isolates Resistant to Double Antibiotic CombinationResistant Isolates in which a Third Antibiotic Resulted in New Bactericidal ActivityIsolates Sensitive to Double Antibiotic CombinationSensitive Isolates in which a Third Antibiotic Resulted in Antagonism
n(%)n(%)n(%)n(%)
Tobramycin-ceftazidime11354/1134842/54 7859/1135228/5948
Meropenem-tobramycin 8129/813626/29 9052/816416/5231
Meropenem-chloramphenicol 8224/822920/24 8358/827118/5831
Meropenem-ceftazidime 7420/742718/20 9054/747322/5441
Meropenem-amikacin 8623/862718/23 7863/867318/6329
Meropenem-minocycline 33 8/3324 8/810025/3376 9/2536

Antagonism with Addition of a Third Antibiotic to a Bactericidal Double Antibiotic Combination

Antagonism (growth of an organism when a third antibiotic was added to a bactericidal combination of two antibiotics) was observed in 52% of the 119 isolates. The proportion of sensitive isolates that demonstrated antagonism ranged from 48% when a third antibiotic was added to a previously bactericidal combination of tobramycin-ceftazidime to 29% when a third antibiotic was added to a previously bactericidal combination of meropenem-amikacin (Table 4).

Treatment of chronic pulmonary disease can be a great challenge in all patients with CF; however, for those patients with CF who are colonized with B. cepacia choosing effective treatment with antimicrobial therapy can be especially problematic. The issue of effective antimicrobial therapy against B. cepacia is made all the more urgent since colonization with B. cepacia is associated with increased mortality for patients at all levels of pulmonary function (20). The MCBT technique that we have developed allows us to rapidly screen multiple combinations of two and three antibiotics for bactericidal activity against B. cepacia.

The value of the MCBT technique is that multiresistant bacteria can be tested for susceptibility against numerous combinations of antibiotics, and that results from the tests are available within 48 to 72 h after bacterial species isolation. If sputum is cultured at the time of a CF pulmonary exacerbation, the MCBT test results can be used to select appropriate bactericidal combinations and discontinue nonbactericidal antibiotic therapy within days of a patient's exacerbation. In addition, B. cepacia MCBT susceptibilities can also be collected when the patient is clinically stable, to allow clinicians to decide on appropriate antibiotic combination therapy in advance of the patient's next pulmonary exacerbation.

Conventional antibiotic susceptibility testing methods are designed to predict therapeutic outcomes with monotherapy in acute infections. These methods assume that therapy requires elimination of the organism from the infection site and that this can be achieved by exposing the target organism to a sufficiently high concentration of one antibiotic for a sufficient time. Unfortunately, we have shown that traditional susceptibility testing methods have limited relevance in this clinical situation since many B. cepacia organisms are often panresistant to antibiotic monotherapy.

In the case of CF lung disease, complete elimination of the organism is usually impossible because of the complex interrelationships between host defenses in patients with CF and these microbes (21). The objectives of treatment in CF are therefore not long-term eradication of the organism but rather control of infection in order to minimize inflammation and damage and thus slow the decline of pulmonary function. Controlling infection and minimizing inflammation can be achieved by antimicrobial therapy that reduces the numbers of organisms (9, 10). In addition to their bactericidal activity, antibiotics have also been shown to suppress bacterial synthesis of pathogenic factors that are potent inciters of inflammation within CF airways (22), and antibiotics can exert antioxidant effects by neutralizing myeloperoxidase release from polymorphonuclear cells in CF sputum (13).

Two recent randomized controlled clinical trials have demonstrated that double antibiotic combination therapy with an aminoglycoside and an antipseudomonal β-lactam significantly decreased sputum bacterial density and improved clinical outcomes compared with antibiotic monotherapy in patients with CF with pulmonary exacerbations who were colonized with Pseudomonas aeruginosa (9, 10). These clinical studies underscore the importance of bactericidal combination therapy in order to decrease the density of airway organisms in patients with CF. However, our current study suggests that a choice of effective double antibiotic therapy cannot be made empirically for patients infected with B. cepacia. Even the most effective double-antibiotic combinations were bactericidal in only 75% of isolates. Furthermore, empiric addition of additional antibiotics was found to be potentially counterproductive. In 47% of the isolates that were sensitive to a single antibiotic, adding in a second empirically chosen antibiotic resulted in antagonism with loss of bactericidal effects. These results illustrate the difficulties inherent in choosing empiric double antibiotic therapy, and they underscore the need to determine effective combination antibiotic therapy by employing MCBT testing on sputum isolated from B. cepacia colonized patients.

Results of our study suggest that triple antibiotic combinations are more likely to be bactericidal against B. cepacia. Adding a third antibiotic to a nonbactericidal double-drug combination increased the likelihood of bactericidal effects. In particular, triple combinations which contained high-dose tobramycin and meropenem were highly effective and were bactericidal in as much as 93% of isolates tested. However, antagonism was again relatively common, with many three-drug combinations, suggesting that empirical addition of an antibiotic can be harmful almost as often as it is helpful. Again, these data underscore the need for careful MCBT susceptibility testing of isolates to avoid potentially antagonistic antibiotic combinations.

It should be noted that we used 200 μg/ml of tobramycin for MCBT testing. Although this concentration is not achievable with intravenous therapy, it is achievable by aerosol administration. Previous studies have shown that nebulization of 300 mg of tobramycin to patients with CF results in mean peak sputum tobramycin concentrations of greater than 600 μg/ml of sputum (23). These achievable sputum tobramycin levels are more than 3-fold higher than we employed with MCBT testing. Results from our in vitro data suggest that inhaled tobramycin should be useful to treat B. cepacia when combined with other parenteral antibiotics. Our study results suggest that in this population of B. cepacia patients, the most frequently effective bactericidal activity was achieved by using inhaled tobramycin plus meropenem plus a second intravenous antibiotic (ceftazidime, chloramphenicol, trimethoprim/ sulfamethoxasole, aztreonam or amikacin).

We initially had also intended to explore whether different genomovars of B. cepacia might have different antibiotic susceptibilities; however, we were unable to perform this analysis since almost all of the patients in this study whose isolates were genotyped were determined to be colonized with genomovar III B. cepacia. This is not surprising since our sample reflects Canadian B. cepacia epidemiology, and the majority of Canadian B. cepacia infections are due to genomovar III organisms. Of 448 patients with CF and B. cepacia whose sputum is stored in the Canadian Burkholderia species repository, 82% have been identified as being infected with genomovar III organisms (24). Similar data from the United States B. cepacia Research Laboratory and Repository (25) suggests that most American B. cepacia CF isolates are due to genomovar III organisms (as much as 58%); however, genomovar II organisms are also prevalent, and account for 37% of U.S. Burkholderia repository isolates compared with only 9% of Canadian repository isolates.

Almost all of the organisms that we studied (89%) were genomovar III, RAPD strain 02. This is not surprising since most of our study patients lived in Ontario, and RAPD strain 02 is seen in 95% of B. cepacia strains cultured from patients with CF from this Canadian province (24). RAPD strain 02 is also the dominant strain seen in eastern Canada, the United Kingdom, and patients who have traveled to/from these areas (19).

Genomovar III B. cepacia strains exhibit clustering and epidemic spread (19). Furthermore, recent studies from Vancouver, BC, Canada suggest that patients with CF colonized with genomovar III B. cepacia are sicker and older than those with genomovar II (D. Speert and E. Mahenthiralingam, unpublished data). Therapeutic options for antimicrobial therapy of patients with genomovar III strains must therefore be considered with some urgency, and MCBT testing of these organisms provides an attractive method of determining appropriate antibiotic therapy.

Although randomized, controlled, clinical studies examining antibiotic combinations have not been performed in patients with CF colonized with B. cepacia, it seems reasonable to infer from studies of patients colonized with P. aeruginosa that effective combination bactericidal antibiotic therapy should decrease sputum bacterial density. It is hoped that clinical utilization of MCBT antibiotic susceptibility testing will result in effective bactericidal antibiotic therapy, with consequent decreases in bacterial sputum density, and improved clinical outcomes, in patients colonized with B. cepacia who present with pulmonary exacerbations of their disease.

Supported in part by an unrestricted grant from Zeneca Pharma Canada Inc.

1. Laraya-Cuasay L., Lipstein M., Huang N.Pseudomonas cepacia in the respiratory flora of patients with cystic fibrosis (abstract). Pediatr. Pulmonol.111977502
2. Isles A., Maclusky I., Corey M.Pseudomonas cepacia infection in cystic fibrosis: an emerging problem. J. Pediatr.1041984206210
3. LiPuma J.Burkholderia cepacia. Management issues and new insights. Clin. Chest Med.191998473486
4. Vandamme P., Holmes B., Vancanneyt M., Coenye T., Hoste B.Occurrence of multiple genomovars of Burkholderia cepacia in cystic fibrosis patients and proposal of Burkholderia multivorans sp. nov. Int. J. Syst. Bacteriol.47199711881200
5. FitzSimmons, S. 1997. National Cystic Fibrosis Patient Registry. Cystic Fibrosis Foundation, Bethesda, MD.
6. Corey, M. 1999. Canadian Cystic Fibrosis Patient Registry. Canadian Cystic Fibrosis Foundation, Toronto, ON, Canada.
7. Lewin L., Byard P., Davis P.Effect of Pseudomonas cepacia colonization on survival and pulmonary function in cystic fibrosis patients. J. Clin. Epidemiol.431990125130
8. Brown P., Butler S., Nelson J.Pseudomonas cepacia in adult cystic fibrosis: accelerated decline in lung function and increased mortality. Thorax481993425429
9. Regelmann W., Elliott G., Warwick W., Clawson C.Reduction of sputum Pseudomonas aeruginosa density by antibiotics improves lung function in cystic fibrosis more than do bronchodilators and chest physiotherapy alone. Am. Rev. Respir. Dis.1411990914921
10. Smith A., Doershuk C., Goldmann D., Gore E., Hilman B.Comparison of a beta-lactam alone versus beta-lactam and an aminoglycoside for pulmonary exacerbation in cystic fibrosis. J. Pediatr.1341999413421
11. Grimwood K., To M., Rabin H., Woods D.Inhibition of Pseudomonas aeruginosa exoenzyme expression by subinhibitory antibiotic concentrations. Antimicrob. Agents Chemother.3319894147
12. Geers T., Baker N.The effect of sublethal levels of antibiotics on the pathogenicity of Pseudomonas aeruginosa for tracheal tissue. J. Antimicrob. Chemother.191987569578
13. Cantin A., Woods D.Protection by antibiotics against myeloperoxidase-dependent cytotoxicity to lung epithelial cells in vitro. J. Clin. Invest.9119933845
14. Hancock R.Resistance mechanisms in Pseudomonas aeruginosa and other nonfermentative gram-negative bacteria. Clin. Infect. Dis.271998S93S99
15. Burns J., Wadsworth C., Barry J.Nucleotide sequence analysis of a gene from Burkholderia (Pseudomonas) cepacia encoding an outer membrane lipoprotein involved in multiple antibiotic resistance. Antimicrob Agent Chemother.33198912471251
16. Ferris W., MacDonald N., MacKenzie A.Multiple combination antibiotic testing for antibiotic synergy in multiresistant Pseudomonas aeruginosa and Burkholderia cepacia in cystic fibrosis (abstract). ICAAC Proceedings11997E147
17. Welch D., Muszynski M., Pai C., Marcon M., Hribar M.Selective and differential medium for recovery of Pseudomonas cepacia from the respiratory tract of patients with cystic fibrosis. J. Clin. Microbiol.25198717301734
18. Henry D., Campbell M., LiPuma J., Speert D.Identification of Burkholderia cepacia isolates from patients with cystic fibrosis and use of a simple new selective medium. J. Clin. Microbiol.351997614619
19. Mahenthiralingam E., Campbell M., Henry D., Speert D.Epidemiology of Burkholderia cepacia infection in patients with cystic fibrosis: analysis by randomly amplified polymorphic DNA fingerprinting. J. Clin. Microbiol.34199629142920
20. Corey M., Farewell V.Determinants of mortality from cystic fibrosis in Canada, 1970–1989. Am. J. Epidemiol.143199610071017
21. Marshall B., Samuelson W.Basic therapies in cystic fibrosis: does standard therapy work? Clin. Chest Med.191998487504
22. Grimwood K., To M., Semple R., Rabin H., Sokol P., Woods D.Elevated exoenzyme expression by Pseudomonas aeruginosa is correlated with exacerbations of lung disease in cystic fibrosis. Pediatr. Pulmonol.151993135139
23. Eisenberg J., Pepe M., Williams-Warren J., Vasiliev M., Montgomery B.A comparison of peak sputum tobramycin concentration in patients with cystic fibrosis using jet and ultrasonic nebulizer systems. Chest1111997955962
24. Henry, D., E. Mahenthiralingam, J. Bischof, M. Campbell, P. Vandamme and D. Speert. 1999. Epidemiology of Burkholderia cepacia in Canadian Cystic Fibrosis Patients. Program of the Sixth Meeting of the International Burkholderia cepacia Working Group, Banff, AB, Canada, April 1999.
25. LiPuma, J. J. 1999. Burkholderia cepacia Research Laboratory and Repository. Program of the Sixth Meeting of the International Burkholderia cepacia Working Group, Banff, AB, Canada, April 1999.
Correspondence and requests for reprints should be addressed to Shawn D. Aaron, M.D., Assistant Professor of Medicine, Division of Respiratory Medicine, The Ottawa Hospital, General Campus, Room LM-17, 501 Smyth Road, Ottawa, ON, K1H 8L6 Canada. E-mail:

Related

No related items
American Journal of Respiratory and Critical Care Medicine
161
4

Click to see any corrections or updates and to confirm this is the authentic version of record