TEXT

Top Abstract Text References

Endotoxic lipopolysaccharides (LPS) of gram-negative bacteria have been implicated as a major factor contributing to mortality in patients with gram-negative sepsis (6). It has been demonstrated that a reduction of mortality in experimental animals occurs when they are treated with agents directed against gram-negative LPS (2). Polymyxin B is a cationic cyclic polypeptide antibiotic that binds strongly to the lipid A portion of gram-negative bacterial LPS (4, 5). It has also been shown to reduce mortality in experimental animals with endotoxic shock (3, 9). The use of polymyxin B in humans, however, has been limited because of its nephrotoxicity. In a recent report, polymyxin B was conjugated to dextran 70 (PMX-622), which reduced the toxicity of polymyxin B without interfering with its LPS-neutralizing capacity (1). It was also claimed that this binding to dextran 70 eliminated the antimicrobial activity of polymyxin B (1). Since this conjugate is designed to be used in conjunction with antimicrobial chemotherapy of gram-negative sepsis, the possibility of interaction with other antibiotics must also be considered.

The present study was designed to determine (i) the in vitro antimicrobial activity of PMX-622 against five species of gram-negative bacilli and (ii) the in vitro effect of PMX-622 on the antimicrobial activity of 11 antibiotics commonly used in the treatment of gram-negative bacterial infections.

PMX-622 was provided by Sandoz Research Institute, East Hanover, N.J. The other 11 antibiotics tested in this study (see Table 2) were procured from their U.S. manufacturers or commercial sources.

Clinical isolates of the following species were tested: 30 strains of Escherichia coli, 30 strains of Enterobacter cloacae, 32 strains of Klebsiella spp., 29 strains of Pseudomonas aeruginosa, 30 strains of Salmonella spp., and 5 strains of Staphylococcus aureus.

MICs of each antibiotic were determined by the broth microdilution method outlined by the National Committee for Clinical Laboratory Standards (NCCLS) (7). Eight serial twofold concentrations of each antibiotic were tested, and the ranges selected for each drug included its susceptible and resistant breakpoints. Each drug was tested alone as well as with PMX-622 (0.1 and 2.0 µg of polymyxin B equivalents/ml). PMX-622 was also tested alone at twofold concentrations ranging from 2.0 to 256 µg of polymyxin B equivalent/ml.

MBCs of each drug (with and without added PMX-622) for 27 of the gram-negative isolates representing each of the 5 species tested were determined by the method outlined by the NCCLS (8). The MBC was defined as the lowest drug concentration producing a reduction of viable bacterial counts to <0.1% of the original inoculum (>99.9% kill).

The following quality control organisms were tested in parallel with the test organisms: E. coli ATCC 25922, E. coli ATCC 35218, and P. aeruginosa ATCC 27853. Results with the quality control organisms all fell within the acceptable ranges for each antibiotic.

When tested alone, PMX-622 did exert some antimicrobial activity against isolates of each of the gram-negative species tested (Table 1). However, the PMX-622 MICs were much too high to be of clinical significance. The lowest MIC (8.0 µg of polymyxin B equivalents/ml) occurred for three strains each of E. coli and E. cloacae. For most isolates, the MICs were much higher. Thirty of the gram-negative isolates were tested with polymyxin B alone in parallel with PMX-622; the ratios of polymyxin B/PMX-622 MICs ranged from 1/128 to 1/512. This minimal antimicrobial activity of PMX-622 suggests either that polymyxin B bound to dextran 70 retains a tiny portion of its antimicrobial activity or that a small amount of unbound polymyxin B is present in the preparation.

The effect of PMX-622 on the inhibitory activity of 11 antibiotics on 156 bacterial isolates is summarized in Table 2. The addition of 0.1 µg of polymyxin B equivalents/ml of PMX-622 had no measurable effect on the activity of the 11 antibiotics tested. Of 1,094 pairs of on-scale MIC results (with and without added PMX-622), 852 (77.9%) were identical and 225 (20.5%) differed by one twofold concentrationequally divided. Only 17 (1.6%) results were 2 twofold concentrations above or below the MIC of the antibiotic alone.

When the tests were performed in the presence of 2.0 µg of polymyxin B equivalents/ml of PMX-622, there were 1,069 on-scale MIC pairs. Of these, 718 (67.2%) MIC pairs were identical and 1,051 (97.1%) MIC pairs were within 1 twofold concentration of each other. There was, however, a very minimal skewing of the MICs of some antibiotics with added PMX-622 toward the lower concentration (increased activity). The degree of skewing is small, but could reflect an additive effect of the minimal antimicrobial activity of PMX-622.

The overall effect of PMX-622 at 0.1 and 2.0 µg of polymyxin equivalents/ml on the bactericidal activity of the same 11 antibiotics against 27 gram-negative strains is summarized in Table 3. Of the on-scale paired MBC results, 96.0 and 91.8% were within 1 twofold concentration of the MBC without PMX-622 when tested with 0.1 and 2.0 µg of polymyxin B equivalents of PMX-622/ml, respectively. A slight skewing of the MBC distribution toward the lower concentration was apparent with both concentrations of PMX-622.

We conclude from these data that PMX-622 has negligible antimicrobial activity and that it does not interfere with the in vitro antimicrobial activity of the 11 antimicrobial agents included in this study.