Journal of Antimicrobial Chemotherapy (1992) 30, 321-326
Comparative activity of glycopeptide antibiotics against coagulasenegative stapfaylococci embedded in fibrin clots R. Cunningham and J. Cbeesbrough*
The susceptibilities of 16 strains of coagulase-negative staphylococti to vancomycin and teicoplanin were determined by three microdilution methods run in parallel; (i) a reference method in which the medium was Iso-Sensitest broth, (ii) a method in which the organisms were suspended in plasma and (iii) one in which the bacteria were incorporated into a fibrin clot. In comparison with the reference method there was a less than two-fold change in the geometric mean MICs of both antibiotics in plasma, while in clot the increase was 12-fold for vancomycin and 28-fold for teicoplanin. MBCs showed a similar trend although the increase was greater for teicoplanin. These results suggest that protein binding may have a much greater impact on antimicrobial activity when the mobility of the protein is reduced within a dot.
Introduction Coagulase-negative staphylococci (CNS) are the leading cause of nosocomial bacteraemia in North America and Europe (Ponce de Leon & Wenzel, 1984). Intravascular catheters, particularly long-term, central venous catheters, are the major predisposing factors. Morphological studies have shown that the colonizing bacteria are often embedded in a fibrin/platelet matrix deposited on the catheter surface (Cheesbrough, Elliott & Finch, 1985). This matrix might impair antimicrobial penetration which may lead to discrepancies between in-vitro susceptibility and clinical efficacy. The glycopeptide antibiotics, vancomycin and teicoplanin, are used widely to treat CNS infections. These agents are large, poorly diffusing compounds with very similar spectra of antimicrobial activity. Teicoplanin, however, is more heavily protein-bound than vancomycin (Krogstad, Moellering & Greenblatt, 1980; Assandri & Bernareggi, 1987) and consequently its activity in vitro may be affected to a greater degree by dissolved and clotted protein. To explore this possibility we have modified the conventional microdilution sensitivity test to determine whether incorporation of the organism into a fibrin clot interferes with the activities of vancomycin and teicoplanin against CNS. 'Corresponding author. 0305-7453/92/090321 +06 $08.00/0
321 © 1992 The British Society for Antimicrobial Chemotherapy
Downloaded from http://jac.oxfordjournals.org/ at University of Tasmania Library on May 21, 2015
Department of Medical Microbiology, Duncan Building, Royal Liverpool University Hospital, Liverpool L7 8XP, UK
322
R. Cmmlngtuun and J. Cbeesbrongh Materials and methods
Antibiotics Vancomycin and teicoplanin were obtained as pure substances of known potency from the following sources: teicoplanin, Marion Merrell Dow, Winnersch, Berkshire, UK; vancomycin, Eli Lilly Ltd, Basingstoke, UK. Media
Bacterial strains Sixteen strains of coagulase-negative staphylococci were randomly selected from a collection of 72 isolates from clinically significant Hickman line infections as previously described (Smith et al., 1989). Thirteen strains were identified as Staphylococcus epidermidis, two as Staphylococcus haemolyticus and one as Staphylococcus simulans. An overnight culture of each strain in ISB was diluted 1:100 in the ISB and FFP. Aliquots of 150 /iL were dispensed into microtitre trays giving a mean final inoculum of
MIC and MBC determinations Three microdilution assays were run in parallel. One hundred and fifty microlitre volumes of doubling dilutions of each antibiotic were dispensed on to ISB, FFP and clot preparations of each organism. The range of the final concentrations was 0-25-2048 mg/L. The microtitre trays were incubated overnight in air at 37°C. MICs were determined by observing turbidity in broth and plasma and the presence of microcolonies in the clots. This was facilitated by an inverted microscope and x 10 objective. The MIC was defined as the highest dilution of each antibiotic which inhibited visible growth. Supernatant liquid wasremovedfrom the clots with a vacuum aspirator with a sterile tip; an equivalent volume (150 /JL) was removed from the wells containing broth and plasma. One hundred and fifty /zL of sterile 1-25% trypsin (Gibco Ltd, Paisley, UK) was then added to each well to lyse the fibrin and the trays were incubated at 37°C for another hour. The trypsin had previously been shown not to affect the viability of CNS. Following incubation any residual clot was disrupted with a sterile rod and 50 ph aliquots of the lysate, broth and plasma cultures were spread on to plates containing 5% horse blood agar. These were incubated overnight at 37°C; the MBC was defined as the lowest concentration of antibiotic killing 99-9% of the original inoculum. The first four assays were run in quadruplicate. As there was minimal intra-assay variation subsequent assays were run singly.
Downloaded from http://jac.oxfordjournals.org/ at University of Tasmania Library on May 21, 2015
Bacterial strains were suspended in the following media: Iso-Sensitest broth (ISB) (Oxoid Ltd, Basingstoke, UK), thawed human fresh frozen plasma (FFP) and fibrin clots. The clots were prepared in microtitre trays by adding 10 /iL of 100 IU/mL sterile bovine thrombin (Sigma Chemical Co., Poole, UK) to 150 /iL of FFP in which the organisms had been suspended. This led to the formation of uniform stable clots within 30 s. FFP without thrombin remained unclotted.
Activity of glycopeptitles against staphylococd in fibrin dots
323
Table I. Vancomycin MICs and MBCs (mg/L) determined in broth, plasma and clot Broth Bacterial strain
Clot
MBC
MIC
MBC
MIC
MBC
1 2 2 2 2 2 2 2 2 2 2 4 2 2 1 2
4 2 8 4 8 2 2 2 4 4 2 8 2 4 2 4
2 4 4 4 4 4 2 4 2 2 2 4 2 2 2 4
4 4 8 4 4 4 2 4
16 32 32
16 32 32 32 16 16 32 32 32 16 32 32 32 16 32 64
2 8 8 8 2 2 4 4
32 16 16 32 32 32 16 32 16 32 8 16 32
Results The MICs and MBCs of vancomycin and teicoplanin for 16 strains of CNS, determined under the three assay conditions, are given in Tables I and II respectively. The geometric mean MICs and MBCs are illustrated in the figure. All strains tested were more sensitive to vancomycin than to teicoplanin, irrespective of the assay condition. While the mean teicoplanin MIC was the same in broth and plasma the mean MIC of vancomycin increased from 1-9 mg/L in broth to 2-8 mg/L in plasma. A more marked difference was observed between the MICs determined in broth and those determined
Tabk n. Teicoplanin MICs and MBCs (mg/L) determined in broth, plasma and clot Broth Bacteria] strain A B C
D E F G H I J K L M N 0 P
Clot
Plasma
MIC
MBC
MIC
MBC
MIC
MBC
4 4 8 8 4 4 8 4 8 4 4 8 4 4 4 8
8 8 16 8 16 16 8 8 8 4 8 32 8 8 32 8
8 4 8 4 8 16 4 8 4 2 4 8 4 2 8 4
16 8 16 8 16 64 8 16 32 8 16 32 8 16 64 32
256 128 128 128 128 512 128 256 256 64 64 128 128 128 128 128
256 128 256 128 128 512 256 256 512 64 256 256 128 512 256 512
Downloaded from http://jac.oxfordjournals.org/ at University of Tasmania Library on May 21, 2015
A B C D E F G H I J K L M N O P
Plasma
MIC
R. Cunningham and J. Cbeesbrongh
324
1000
Ptosmo
Clot
Broth
Rotmo
Clot
MBC
10-4
17-5
234-6
3-4
4
26-9
MIC
5-2
5-2
145-7
1-9
2-8
22-6
Figure. Companion of geometric mean MICs ( • ) and MBCs ( • ) of vancomycin and teicoplanin for 16 strains of coagulase-negative staphylococci in broth, plasma and dot
in the clot. In the latter, the mean MIC of teicoplanin was 145-7 mg/L (cf. 5-2 mg/L in broth) and that of vancomycin was 22-6 mg/L (cf. 1-9 mg/L in broth). Thisrepresentsa 12-fold increase in the vancomycin MIC and a 28-fold increase in the teicoplanin MIC; there was a similar increase in the MBCs of both antibiotics. MBCs in broth and plasma were generally higher than the corresponding MICs. However, in the clot the difference was two dilutions or less for all isolates tested, with the exception of isolate K for which the MIC of teicoplanin was 64 mg/L and the MBC 256 mg/L. Discussion
The effect of protein binding on antimicrobial efficacy in vivo remains controversial. While several very highly protein-bound antibiotics are apparently as clinically active as less protein-bound analogues (Gerding & Peterson, 1985); Kunin (1966) showed that the bactericidal activity of penicillin was directly proportional to the concentration of free drug. Other investigators performing similar experiments in nutrient broth containing 25% or 50% human blood found no difference in the activities of various highly protein-bound antibiotics (Lacey & Barr, 1985). Many of these differences may be attributed to methodology; whether whole blood, plasma or serum is employed and the type of anticoagulant used may all account for the discrepancies. For example, blood which had been anticoagulated with citrate phosphate dextrose adenine did not inhibit the activity of flucloxacillin but when lithium heparin was used the MIC increased four-fold (George, 1986). Joos, ICading & Hall (1967) found that the activities of a wide range of antibiotics against staphylococci were not significantly affected when the organisms were trapped
Downloaded from http://jac.oxfordjournals.org/ at University of Tasmania Library on May 21, 2015
Broth
Activity of glycopeptides against staphyiococd infibrindots
325
Acknowledgements We thank the Mersey Blood Transfusion Service for supplying FFP, Dr Murko Zuzel for technical advice, Dr Godfrey Smith for his helpful comments and Mr Richard Little of Eli Lilly for arranging financial support.
Downloaded from http://jac.oxfordjournals.org/ at University of Tasmania Library on May 21, 2015
within a fibrin clot. Although glycopeptides were not tested, their observation did apply to novobiocin which is 99-2% protein-bound (Kuccrs & Bennett, 1987). In-vivo models have also failed to resolve this question. When fibrin clots were implanted subcutaneously into rabbits, an inverse relationship was demonstrated between penetration into fibrin and serum protein binding (Barza, Samuelson & Weinstein, 1974). The net effect of protein binding remains unclear, however, as the half-life of highly protein-bound antibiotics within clots was also prolonged. Teicoplanin is approximately 90% bound to albumin in plasma, whereas vancomycin is only 55% bound. This difference was not matched by a corresponding increase in MICs determined in plasma in our study; however, when the bacterial inoculum was incorporated into an identical amount of protein in clot there was a significant difference between vancomycin and teicoplanin with respect to both the MICs and MBCs. This finding suggests that the binding of teicoplanin to protein in a gel phase restricts its diffusion to the site of action. Autoradiographic studies showing poor penetration of teicoplanin into cardiac vegetations support this theory (Cremieux et al., 1989); however, vancomycin penetration was not evaluated in this model. There are obviously other factors which might explain the relative increase in the MIC and MBC of organisms in clots. The micro-environment of staphylococci growing as a micro-colony within clots may be very different to that in a broth culture. Crowding, lack of nutrients and a build-up of waste products may lead to prolongation of generation time or other unexpected changes which alter bacterial susceptibility to antibiotics. Slime, the external glycocalyx produced by many CNS, may be an additional barrier to penetration, especially when it engulfs a colony retained within a clot However, as the mode of action of both antibiotics is similar, it is unlikely that these factors alone could account for the differences observed between the two agents. The similar clinical response rate of patients with infected Hickman catheters to either glycopeptide (Smith et al., 1989; Van der Auwera, Aoun & Meunier, 1991) is not necessarily incompatible with our results. In intraluminal infections, even the high MICs we observed would be greatly exceeded by the concentrations of the antibiotic solutions infused through the catheter. Indeed, in this situation the higher protein binding of teicoplanin may confer an advantage as the result of the greater and prolonged concentrations of the drug present at the catheter surface. Our model may have more clinical relevance to subcutaneous tunnel infections which respond less well to antibiotics than do intraluminal infections. In this instance interstitial fluid concentrations may be more relevant. Plasma concentrations obtained following standard doses of teicoplanin would rarely, if ever, exceed the mean MIC of the antibiotic in clot, while peak vancomycin concentrations would often exceed its mean clot MIC. In conclusion, the method we have described provides a simple, reproducible model of antibiotic penetration and activity within fibrin clots. Our observations suggest that CNS multiplying within a clot of human plasma are partly protected from inhibition and killing by glycopeptide antibiotics and that this effect is more pronounced for teicoplanin than for vancomycin.
326
R. Cmrntngham and J. Cbeesbrough
(Received 10 March 1992; accepted 13 May 1992)
Downloaded from http://jac.oxfordjournals.org/ at University of Tasmania Library on May 21, 2015
References Assandri, A. & Bcrnareggi, A. (1987). Binding of tcicoplanin to human scrum albumin. European Journal of Clinical Pharmacology 33, 191-5. Barza, M., Samuelson, T. & Weinstein, L. (1974). Penetration of antibiotics into fibrin loci in vivo. II. Comparison of nine antibiotics: effect of dose and degree of protein binding. Journal of Infectious Diseases 129, 66-72. Cheesbrough, J. S., Elliott, T. S. J. & Finch, R. G. (1985). A morphological study of bacterial colonization of intravenous cannulae. Journal of Medical Microbiology 19, 149-57. Cremieux, A. C , Maziere, B., Vallois, J. M., Ottaviani, M., Azancot, A., Raffoul, H. et al. (1989). Evaluation of antibiotic diffusion into cardiac vegetations by quantitative autoradiography. Journal of Infectious Diseases 159, 938-44. George, R. H. (1986). The influence of protein binding on the antistaphylococcal activity of antibiotics. Journal of Antimicrobial Chemotherapy 17, 539—40. Gerding, D. N. & Petersen, L. R. (1985). Serum protein binding and extravascular distribution of antimicrobials. Journal of Antimicrobial Chemotherapy 15, 136-8. Joos, R. W., Kading, W. H. & Hall, W. H. (1967). Effect of antibiotics on growth of staphylococci in plasma clots. American Journal of the Medical Sciences 253, 305-11. Krogstad, D. J., Moellering, R. C. & Grcenblatt, D. J. (1980). Single-dose kinetics of intravenous vancotnycin. Journal of Clinical Pharmacology 20, 197-201. Kucers, A. & Bennett, N. McK, (1987). The Use of Antibiotics, 4th cdn, p. 117. Heinemann Medical Books, London. Kunin, C. M. (1966). Clinical pharmacology of the new penicillins I. The importance of serum protein in determining antimicrobial activity and concentration in berum. Clinical Pharmacology and Therapeutics 7, 166-79. Lacey, R. W. & Barr, K. (1985). Does the protein binding of antibiotics affect their antistaphylococcal activity? Journal of Antimicrobial Chemotherapy 15, 786-7. Ponce de Leon, S. & Wenzel, R. P. (1984). Hospital-acquired bloodstream infections with Staphylococcus epidermidis. Review of 100 cases. American Journal of Medicine 71, 639-44. Smith, S. R., Cheesbrough, J., Spearing, R. & Davies, J. M. (1989). Randomized prospective study comparing vancomycin with teicoplanin in the treatment of infections associated with Hickman catheters. Antimicrobial Agents and Chemotherapy 33, 1193-7. Van der Auwera, P., Aoun, M. & Meunier, F. (1991). Randomized study of vancomycin versus teicoplanin for the treatment of Gram-positive bacterial infections in immunocompromised hosts. Antimicrobial Agents and Chemotherapy 35, 451-7.
Comparative activity of glycopeptide antibiotics against coagulasenegative stapfaylococci embedded in fibrin clots R. Cunningham and J. Cbeesbrough*
The susceptibilities of 16 strains of coagulase-negative staphylococti to vancomycin and teicoplanin were determined by three microdilution methods run in parallel; (i) a reference method in which the medium was Iso-Sensitest broth, (ii) a method in which the organisms were suspended in plasma and (iii) one in which the bacteria were incorporated into a fibrin clot. In comparison with the reference method there was a less than two-fold change in the geometric mean MICs of both antibiotics in plasma, while in clot the increase was 12-fold for vancomycin and 28-fold for teicoplanin. MBCs showed a similar trend although the increase was greater for teicoplanin. These results suggest that protein binding may have a much greater impact on antimicrobial activity when the mobility of the protein is reduced within a dot.
Introduction Coagulase-negative staphylococci (CNS) are the leading cause of nosocomial bacteraemia in North America and Europe (Ponce de Leon & Wenzel, 1984). Intravascular catheters, particularly long-term, central venous catheters, are the major predisposing factors. Morphological studies have shown that the colonizing bacteria are often embedded in a fibrin/platelet matrix deposited on the catheter surface (Cheesbrough, Elliott & Finch, 1985). This matrix might impair antimicrobial penetration which may lead to discrepancies between in-vitro susceptibility and clinical efficacy. The glycopeptide antibiotics, vancomycin and teicoplanin, are used widely to treat CNS infections. These agents are large, poorly diffusing compounds with very similar spectra of antimicrobial activity. Teicoplanin, however, is more heavily protein-bound than vancomycin (Krogstad, Moellering & Greenblatt, 1980; Assandri & Bernareggi, 1987) and consequently its activity in vitro may be affected to a greater degree by dissolved and clotted protein. To explore this possibility we have modified the conventional microdilution sensitivity test to determine whether incorporation of the organism into a fibrin clot interferes with the activities of vancomycin and teicoplanin against CNS. 'Corresponding author. 0305-7453/92/090321 +06 $08.00/0
321 © 1992 The British Society for Antimicrobial Chemotherapy
Downloaded from http://jac.oxfordjournals.org/ at University of Tasmania Library on May 21, 2015
Department of Medical Microbiology, Duncan Building, Royal Liverpool University Hospital, Liverpool L7 8XP, UK
322
R. Cmmlngtuun and J. Cbeesbrongh Materials and methods
Antibiotics Vancomycin and teicoplanin were obtained as pure substances of known potency from the following sources: teicoplanin, Marion Merrell Dow, Winnersch, Berkshire, UK; vancomycin, Eli Lilly Ltd, Basingstoke, UK. Media
Bacterial strains Sixteen strains of coagulase-negative staphylococci were randomly selected from a collection of 72 isolates from clinically significant Hickman line infections as previously described (Smith et al., 1989). Thirteen strains were identified as Staphylococcus epidermidis, two as Staphylococcus haemolyticus and one as Staphylococcus simulans. An overnight culture of each strain in ISB was diluted 1:100 in the ISB and FFP. Aliquots of 150 /iL were dispensed into microtitre trays giving a mean final inoculum of
MIC and MBC determinations Three microdilution assays were run in parallel. One hundred and fifty microlitre volumes of doubling dilutions of each antibiotic were dispensed on to ISB, FFP and clot preparations of each organism. The range of the final concentrations was 0-25-2048 mg/L. The microtitre trays were incubated overnight in air at 37°C. MICs were determined by observing turbidity in broth and plasma and the presence of microcolonies in the clots. This was facilitated by an inverted microscope and x 10 objective. The MIC was defined as the highest dilution of each antibiotic which inhibited visible growth. Supernatant liquid wasremovedfrom the clots with a vacuum aspirator with a sterile tip; an equivalent volume (150 /JL) was removed from the wells containing broth and plasma. One hundred and fifty /zL of sterile 1-25% trypsin (Gibco Ltd, Paisley, UK) was then added to each well to lyse the fibrin and the trays were incubated at 37°C for another hour. The trypsin had previously been shown not to affect the viability of CNS. Following incubation any residual clot was disrupted with a sterile rod and 50 ph aliquots of the lysate, broth and plasma cultures were spread on to plates containing 5% horse blood agar. These were incubated overnight at 37°C; the MBC was defined as the lowest concentration of antibiotic killing 99-9% of the original inoculum. The first four assays were run in quadruplicate. As there was minimal intra-assay variation subsequent assays were run singly.
Downloaded from http://jac.oxfordjournals.org/ at University of Tasmania Library on May 21, 2015
Bacterial strains were suspended in the following media: Iso-Sensitest broth (ISB) (Oxoid Ltd, Basingstoke, UK), thawed human fresh frozen plasma (FFP) and fibrin clots. The clots were prepared in microtitre trays by adding 10 /iL of 100 IU/mL sterile bovine thrombin (Sigma Chemical Co., Poole, UK) to 150 /iL of FFP in which the organisms had been suspended. This led to the formation of uniform stable clots within 30 s. FFP without thrombin remained unclotted.
Activity of glycopeptitles against staphylococd in fibrin dots
323
Table I. Vancomycin MICs and MBCs (mg/L) determined in broth, plasma and clot Broth Bacterial strain
Clot
MBC
MIC
MBC
MIC
MBC
1 2 2 2 2 2 2 2 2 2 2 4 2 2 1 2
4 2 8 4 8 2 2 2 4 4 2 8 2 4 2 4
2 4 4 4 4 4 2 4 2 2 2 4 2 2 2 4
4 4 8 4 4 4 2 4
16 32 32
16 32 32 32 16 16 32 32 32 16 32 32 32 16 32 64
2 8 8 8 2 2 4 4
32 16 16 32 32 32 16 32 16 32 8 16 32
Results The MICs and MBCs of vancomycin and teicoplanin for 16 strains of CNS, determined under the three assay conditions, are given in Tables I and II respectively. The geometric mean MICs and MBCs are illustrated in the figure. All strains tested were more sensitive to vancomycin than to teicoplanin, irrespective of the assay condition. While the mean teicoplanin MIC was the same in broth and plasma the mean MIC of vancomycin increased from 1-9 mg/L in broth to 2-8 mg/L in plasma. A more marked difference was observed between the MICs determined in broth and those determined
Tabk n. Teicoplanin MICs and MBCs (mg/L) determined in broth, plasma and clot Broth Bacteria] strain A B C
D E F G H I J K L M N 0 P
Clot
Plasma
MIC
MBC
MIC
MBC
MIC
MBC
4 4 8 8 4 4 8 4 8 4 4 8 4 4 4 8
8 8 16 8 16 16 8 8 8 4 8 32 8 8 32 8
8 4 8 4 8 16 4 8 4 2 4 8 4 2 8 4
16 8 16 8 16 64 8 16 32 8 16 32 8 16 64 32
256 128 128 128 128 512 128 256 256 64 64 128 128 128 128 128
256 128 256 128 128 512 256 256 512 64 256 256 128 512 256 512
Downloaded from http://jac.oxfordjournals.org/ at University of Tasmania Library on May 21, 2015
A B C D E F G H I J K L M N O P
Plasma
MIC
R. Cunningham and J. Cbeesbrongh
324
1000
Ptosmo
Clot
Broth
Rotmo
Clot
MBC
10-4
17-5
234-6
3-4
4
26-9
MIC
5-2
5-2
145-7
1-9
2-8
22-6
Figure. Companion of geometric mean MICs ( • ) and MBCs ( • ) of vancomycin and teicoplanin for 16 strains of coagulase-negative staphylococci in broth, plasma and dot
in the clot. In the latter, the mean MIC of teicoplanin was 145-7 mg/L (cf. 5-2 mg/L in broth) and that of vancomycin was 22-6 mg/L (cf. 1-9 mg/L in broth). Thisrepresentsa 12-fold increase in the vancomycin MIC and a 28-fold increase in the teicoplanin MIC; there was a similar increase in the MBCs of both antibiotics. MBCs in broth and plasma were generally higher than the corresponding MICs. However, in the clot the difference was two dilutions or less for all isolates tested, with the exception of isolate K for which the MIC of teicoplanin was 64 mg/L and the MBC 256 mg/L. Discussion
The effect of protein binding on antimicrobial efficacy in vivo remains controversial. While several very highly protein-bound antibiotics are apparently as clinically active as less protein-bound analogues (Gerding & Peterson, 1985); Kunin (1966) showed that the bactericidal activity of penicillin was directly proportional to the concentration of free drug. Other investigators performing similar experiments in nutrient broth containing 25% or 50% human blood found no difference in the activities of various highly protein-bound antibiotics (Lacey & Barr, 1985). Many of these differences may be attributed to methodology; whether whole blood, plasma or serum is employed and the type of anticoagulant used may all account for the discrepancies. For example, blood which had been anticoagulated with citrate phosphate dextrose adenine did not inhibit the activity of flucloxacillin but when lithium heparin was used the MIC increased four-fold (George, 1986). Joos, ICading & Hall (1967) found that the activities of a wide range of antibiotics against staphylococci were not significantly affected when the organisms were trapped
Downloaded from http://jac.oxfordjournals.org/ at University of Tasmania Library on May 21, 2015
Broth
Activity of glycopeptides against staphyiococd infibrindots
325
Acknowledgements We thank the Mersey Blood Transfusion Service for supplying FFP, Dr Murko Zuzel for technical advice, Dr Godfrey Smith for his helpful comments and Mr Richard Little of Eli Lilly for arranging financial support.
Downloaded from http://jac.oxfordjournals.org/ at University of Tasmania Library on May 21, 2015
within a fibrin clot. Although glycopeptides were not tested, their observation did apply to novobiocin which is 99-2% protein-bound (Kuccrs & Bennett, 1987). In-vivo models have also failed to resolve this question. When fibrin clots were implanted subcutaneously into rabbits, an inverse relationship was demonstrated between penetration into fibrin and serum protein binding (Barza, Samuelson & Weinstein, 1974). The net effect of protein binding remains unclear, however, as the half-life of highly protein-bound antibiotics within clots was also prolonged. Teicoplanin is approximately 90% bound to albumin in plasma, whereas vancomycin is only 55% bound. This difference was not matched by a corresponding increase in MICs determined in plasma in our study; however, when the bacterial inoculum was incorporated into an identical amount of protein in clot there was a significant difference between vancomycin and teicoplanin with respect to both the MICs and MBCs. This finding suggests that the binding of teicoplanin to protein in a gel phase restricts its diffusion to the site of action. Autoradiographic studies showing poor penetration of teicoplanin into cardiac vegetations support this theory (Cremieux et al., 1989); however, vancomycin penetration was not evaluated in this model. There are obviously other factors which might explain the relative increase in the MIC and MBC of organisms in clots. The micro-environment of staphylococci growing as a micro-colony within clots may be very different to that in a broth culture. Crowding, lack of nutrients and a build-up of waste products may lead to prolongation of generation time or other unexpected changes which alter bacterial susceptibility to antibiotics. Slime, the external glycocalyx produced by many CNS, may be an additional barrier to penetration, especially when it engulfs a colony retained within a clot However, as the mode of action of both antibiotics is similar, it is unlikely that these factors alone could account for the differences observed between the two agents. The similar clinical response rate of patients with infected Hickman catheters to either glycopeptide (Smith et al., 1989; Van der Auwera, Aoun & Meunier, 1991) is not necessarily incompatible with our results. In intraluminal infections, even the high MICs we observed would be greatly exceeded by the concentrations of the antibiotic solutions infused through the catheter. Indeed, in this situation the higher protein binding of teicoplanin may confer an advantage as the result of the greater and prolonged concentrations of the drug present at the catheter surface. Our model may have more clinical relevance to subcutaneous tunnel infections which respond less well to antibiotics than do intraluminal infections. In this instance interstitial fluid concentrations may be more relevant. Plasma concentrations obtained following standard doses of teicoplanin would rarely, if ever, exceed the mean MIC of the antibiotic in clot, while peak vancomycin concentrations would often exceed its mean clot MIC. In conclusion, the method we have described provides a simple, reproducible model of antibiotic penetration and activity within fibrin clots. Our observations suggest that CNS multiplying within a clot of human plasma are partly protected from inhibition and killing by glycopeptide antibiotics and that this effect is more pronounced for teicoplanin than for vancomycin.
326
R. Cmrntngham and J. Cbeesbrough
(Received 10 March 1992; accepted 13 May 1992)
Downloaded from http://jac.oxfordjournals.org/ at University of Tasmania Library on May 21, 2015
References Assandri, A. & Bcrnareggi, A. (1987). Binding of tcicoplanin to human scrum albumin. European Journal of Clinical Pharmacology 33, 191-5. Barza, M., Samuelson, T. & Weinstein, L. (1974). Penetration of antibiotics into fibrin loci in vivo. II. Comparison of nine antibiotics: effect of dose and degree of protein binding. Journal of Infectious Diseases 129, 66-72. Cheesbrough, J. S., Elliott, T. S. J. & Finch, R. G. (1985). A morphological study of bacterial colonization of intravenous cannulae. Journal of Medical Microbiology 19, 149-57. Cremieux, A. C , Maziere, B., Vallois, J. M., Ottaviani, M., Azancot, A., Raffoul, H. et al. (1989). Evaluation of antibiotic diffusion into cardiac vegetations by quantitative autoradiography. Journal of Infectious Diseases 159, 938-44. George, R. H. (1986). The influence of protein binding on the antistaphylococcal activity of antibiotics. Journal of Antimicrobial Chemotherapy 17, 539—40. Gerding, D. N. & Petersen, L. R. (1985). Serum protein binding and extravascular distribution of antimicrobials. Journal of Antimicrobial Chemotherapy 15, 136-8. Joos, R. W., Kading, W. H. & Hall, W. H. (1967). Effect of antibiotics on growth of staphylococci in plasma clots. American Journal of the Medical Sciences 253, 305-11. Krogstad, D. J., Moellering, R. C. & Grcenblatt, D. J. (1980). Single-dose kinetics of intravenous vancotnycin. Journal of Clinical Pharmacology 20, 197-201. Kucers, A. & Bennett, N. McK, (1987). The Use of Antibiotics, 4th cdn, p. 117. Heinemann Medical Books, London. Kunin, C. M. (1966). Clinical pharmacology of the new penicillins I. The importance of serum protein in determining antimicrobial activity and concentration in berum. Clinical Pharmacology and Therapeutics 7, 166-79. Lacey, R. W. & Barr, K. (1985). Does the protein binding of antibiotics affect their antistaphylococcal activity? Journal of Antimicrobial Chemotherapy 15, 786-7. Ponce de Leon, S. & Wenzel, R. P. (1984). Hospital-acquired bloodstream infections with Staphylococcus epidermidis. Review of 100 cases. American Journal of Medicine 71, 639-44. Smith, S. R., Cheesbrough, J., Spearing, R. & Davies, J. M. (1989). Randomized prospective study comparing vancomycin with teicoplanin in the treatment of infections associated with Hickman catheters. Antimicrobial Agents and Chemotherapy 33, 1193-7. Van der Auwera, P., Aoun, M. & Meunier, F. (1991). Randomized study of vancomycin versus teicoplanin for the treatment of Gram-positive bacterial infections in immunocompromised hosts. Antimicrobial Agents and Chemotherapy 35, 451-7.
