Journal of Applied Bacteriology 1992, 73, 395-400
Evaluation of antimicrobial interactions between chlorhexidine, quaternary ammonium compounds, preservatives and excipients J.-L. Pons, N. Bonnavelro’, J. Chevalier1and A. Cremleux’ Laboratoire de Microbiologie-Pharmacie, U.F.R.de Medecine-Pharmacie de Rouen, SaintItienne Rouvray and Laboratoire de Microbiologie,Hygiene Microbienne,Immunologie,Facult6 de Pharmacie, Marseille, France
’
4114/02/92: accepted 27 April 1992 CREMIEUX1992. T h e antimicrobial interactions of 49 combinations of chlorhexidine, quaternary ammonium compounds, preservatives and excipients were evaluated by the method of Berenbaum and the checkerboard titration method, with Staphylococcus aureus CIP 53 154 and Escherichia coli CIP 54127 as test strains. MIC determinations were carried out as a preliminary step, and relative growth intensity was used to describe the bacteriostatic activity of surface-active agents (Amonyl 380 BA@, Amonyl 671 SB@). I n the study of combinations, results were interpreted with Fractional Inhibitory Concentration indexes and represented by isobolograms. A fair correlation was shown between the method of Berenbaum and the checkerboard titration method. Combinations between chlorhexidine, cetrimonium bromide and benzalkonium chloride were synergistic or additive ; combinations of antiseptics and preservatives were generally not antagonistic. T h e methods were also well adapted to the study of interactions involving surface-active agents, a critical problem in the formulation of topical antimicrobial agents. J.-L. PONS, N. BONNAVEIRO, J. CHEVALIER AND A.
INTRODUCTION
Antiseptics are widely used in hospitals for hygienic and surgical hand disinfection (Reybrouck 1986) and are usually preferred to antibiotics for topical application in the prevention of superficial infections (Lowbury 1976). Their formulation may be difficult because of physico-chemical and biological interactions between antibacterial agents and excipients or preservatives. Furthermore, they require a high level of activity that cannot be obtained by increasing concentrations of active components because of tolerance side effects. Therefore, the combination of such topical antimicrobial agents with preservatives and excipients must not only avoid antagonism but ideally provide an additive or preferably a synergistic effect. Reports have dealt with combinations of chlorhexidine and other antiseptics with surface-active agents (Crimieux et al. 1983a,b), with antibiotics (Quesnel et al. 1978; Al-Najjar & Quesnel 1979; Barnham & Kerby 1980) or with antifungal agents (Simonetti et al. 1988). Correspondence to :Jean-Louis Pons, Laboratoire de Microbiologie-Pharmacie, U.F.R. de Midecine-Pharmacie de Rouen, B.P.
97, F-76803 Saint-Etienne Rouvray Cedex, France.
Standardized methods to evaluate the bactericidal activities of antiseptics (Anon. 1980, 1981, 1987, 1989) cannot be routinely performed in the development of complex formulations which require numerous assays. More convenient methods are necessary, adaptable to large-scale studies. The present work was an attempt to use the established methods to study combinations of antimicrobial agents in the evaluation of antibacterial interactions involving antiseptics. The method of Berenbaum (1978) and the checkerboard titration method, widely used in the study of antibiotics (Comber et al. 1977), were applied to assess antagonism, additivity, indifference or synergy in 49 combinations of antiseptics, preservatives and excipients. MATERIALS AND METHODS Bacterlal stralns
Staphylococcus aureus CIP 53154 and Escherichia coli CIP 54127 (Collection Institut Pasteur, Paris, France) were used as test organisms. Bacterial inocula were prepared according to AFNOR (Anon. 1989) guidelines. Briefly, strains were grown on Tryptic Soy Agar (Difco) at 37°C for 24 h
396
J . - L . PONS E r A L .
subcultured three times. Bacterial cells were dispersed in a diluent containing Tryptone (Difco) (1 g/l) and sodium chloride (8.5 g/l) in distilled water. The optical density (623 nm) was adjusted so that the suspensions contained 1-3 x 10' colony-forming units (cfu) per ml, as shown by counts on Plate Count Agar (Difco).
Study of associations
Method of Berenbaum (7978) Agents were mixed in test tubes, each at a final concentration corresponding to 2 x MIC. Bacteriostatic activity of these binary combinations was then evaluated by a microdilution technique in microplates as described for single agents. I n addition, the M I C of each component of an association was measured concomitantly as a control.
Compounds
Checkerboard titration method Serial twofold dilutions of each agent were prepared, the median value of seven concentrations being equivalent to the MIC. Pairwise combinations of all concentrations of each agent were then placed in the microplates. Their bacteriostatic activity was evaluated as described below after inoculating with test strains.
The following compounds were investigated alone or in combinations : chlorhexidine digluconate (CHX) and cetrimonium bromide (CET) (ICI-Pharma, Cergy, France) ; benzalkonium chloride (BZA) (Coopkration Pharmaceutique Fragaise, Melun, France) ; phenylethyl alcohol (PEA) (Sigma) ; isopropyl alcohol (IPA), metacresol (mC) and parachlorometacresol (CC) (E. Merck AG, Darmstadt, Germany); Amonyl 380 BA@ (lauryl amidobetaine, Amo BA), Amonyl 671 SB@ (lauryl amido hydroxy sulfobetaine, Amo SB), Kathon CG@ (methyl isothiazolone and chloromethyl isothiazolone, KAT) and Triton C G llO@ (alkyl glucoside, T R I ) (Seppic, Division Cosmktique-Pharmacie, Paris, France).
Expression of results Each result was first expressed as Fractional Inhibitory Concentration index (XFIC), which is the sum of the lowest concentrations of each agent of an association (expressed as a fraction of MIC) inhibiting growth (Anon. 1973; Berenbaum 1978). Then the C F I C values allowed interpretation on the following criteria: XFIC I 0.75 : synergistic; XFIC = 1 : additive; 1 c XFIC c 2 : indifferent; XFIC 2 2 : antagonistic).
MIC determinations
Minimum inhibitory concentration (MIC) determinations were performed by a microdilution technique (serial twofold dilution) in 96 well microplates. The test volume of each well was 0.2 ml. Nutrient Broth (Difco) containing 3 g/l Yeast Extract (Difco) was inoculated with 1-3 x lo5 cfu/ml. Microplates were incubated at 35°C for 48 h. Growth was estimated by reading optical density with a Multiskan spectrophotometer (Titertek, Flow Laboratories, Helsinki) set at 623 nm. M I C was defined as the lowest concentration giving less than 20% growth as compared with the control (Turner et al. 1983).
RESULTS inhibitory activity of antiseptics, preservatives and exciplents
T h e M I C values of antiseptics, preservatives and excipients are listed in Table 1. T h e results agree with those of previous reports on CHX (Quesnel et al. 1978; Al-Najjar & Quesnel 1979; Barnham & Kerby 1980; Brumfitt et al. 1985) and C E T (Brumfitt et al. 1985). T h e MICs of pre-
Table 1 Minimal inhibitory concentrations (MICs) of antiseptics, preservatives and excipients (not combined)
MIC value Antiseptics (pg/ml)
Preservatives (pg/ml)
Excipients (YOv/v)
Test strain
CHX
CET
BZA
mC
CC
KAT
PEA
IPA
AmoBA
AmoSB
TRI
Staphylococcus aureus CIP 53154
2
1
0.5
1000
200
0.5
4000
20
5
0.0125
0.125
Escherichia coli CIP 54127
4
8
5
1000
100
0.125
2000
20
20
0.1
0.5
CHX, Chlorhexidine digluconate ; CET, cetrimonium bromide ; BZA, benzalkoniurn chloride; mC, metacresol ; CC, parachlorometacresol ; KAT, Kathon CG@;PEA, phenylethyl alcohol; IPA, isopropyl alcohol; Amo BA, Amonyl 380 BA@;Amo SB, Amonyl 671 SB@;TRI, Triton CG 1 lo@.
IN VlTRO INTERACTIONS IN ANTISEPTICS 397
Amo BA (70v/v)
Amo SB ( % v/v)
CET ( p g / m l )
Fig. 1 Dose-response curves of antimicrobial inhibitory activity. (a) Amonyl 380 BA@(Amo BA); (b) Amonyl671 SB@(Amo SB); (c) cetrimonium bromide (CET). Test strains: Staphylococcus aureus CIP 53154 Escherichia coli CIP 54127 (m). Growth intensity is expressed as a percentage of growth in drug-free control cultures
(a),
servatives were in the same order of magnitude with both strains, whereas E. coli was less sensitive than Staph. aureus to antiseptics and excipients. The M I C values must be considered as preliminary results, required for the further study of combinations. Non-typical antimicrobial agents (such as surface-active agents) may cause partial growth inhibition over a wide range of concentrations. In such cases MICs give only a partial information. Therefore, dose-response curves were drawn, where growth was expressed as a percentage of growth in drug-free control cultures. Graphs of Amo BA (Fig. la) and Amo SB (Fig. lb) show a partial inhibitory
activity against E. coli, especially with Amo BA; its concentrations from 0625% to 20% (v/v) were associated with a growth varying from 15% to 25%. T h e other compounds gave generally clear-cut M I C end-points, particularly with Staph. aureus. T h e graph of CET (Fig. lc) is given as an example. inhibitory activity of associations
The results of antimicrobial interactions in all pairwise combinations, investigated by the checkerboard titration method and the method of Berenbaum (1978) are shown in
Table 2 Antimicrobial interactions in combinations of antiseptics, preservatives and excipients: comparison of the checkerboard titration method and the method of Berenbaum (result given between parentheses)
Staphylococcus aureus CIP 53 154
Escherichia coli CIP 54127
CHX CET BZA mC
cc
KAT PEA IPA AmoBA Amo SB TRI
CHX
CET
BZA
mC
CC
KAT
PEA
IPA
AmoBA
AmoSB
TRI
***
S(S)'
S(S)2 ad(S) S(S) S(i) S(S) S(ad) S(A) A(A)
***
S(S) S(ad)
i(ad) ad(ad) i(A)
S(S) S(ad) S(A) ad(ad) ad(ad)
ad(A) S(i) i(A) i(S) ad(S) ad(i) i(S)
A(i) A(A) A(S) A(ad) A(ad) S(A) A(S)
S(S) i(ad) ad(S) ad(i) ad(S) i(ad) S(i)
i(i) S(ad) S(i) S(A) S(A) S(A) S(A)
i(ad) S(S)
ad@)
***
S(ad)
i(r)
***
i(ad) i(ad) A(A) ad(ad)
S(A)
ad(i) i(S)
S(ad) S(A) S(S) S(S) A(S)
i(A) S(S) ad(A) A(A)
ad(ad)
ad(ad)
S(ad)
S(ad)
i(A) S(i) S(i)
A(i) S(i)
S(i)
ad(ad) S(A) A(A) i(A) A(A)
***
***
ad(S) S(S) S(i) S(S) S(i) S(i)
ad(i) S(S)
S(A)
ad(A) i(ad) ad(ad)
A(A) S(A)
***
a(a)
S, synergistic (ZFIC 5 0.75); ad, additive (ZFIC = 1); i, indifferent (1 < ZFIC < 2); A, antagonistic (CFIC 2 2). Normal characters : antimicrobial interaction with Staph. aureus; Italic characters : antimicrobial interaction with E . colt. Key as for Table 1. ***, Untested.
398 J.-L. PONS E r A L .
E. coli. A major discordance (synergy with one method,
0.25
0.50
0.75
I
mC or PEA (FIC)
Fig. 2 Isobolograms of the associations of metacresol (mC, 0) or phenylethyl alcohol (PEA,). with Amonyl671 SB@(Amo SB). Test strain: Staphylococcus aureus CIP 53154
Table 2. Interpretation is based upon C F I C values, which were calculated by comparison with the MICs determined concomitantly. Results show a fair correlation between the methods in most of the 49 combinations tested with Staph. aureus and
antagonism with the other one) was noted in 15 cases among 98 tests. An additivity-antagonism discordance was noted in five cases. Antimicrobial interactions may differ according to the test strain. T h e BZA-CC association was antagonistic with Staph. aureus, whereas it was additive or indifferent with E. coli. Associations of CNX or BZA with T R I led to opposite results with the two test strains. Another typical example is given with the PEA-Amo SB association. T h e following trends are worth noting in the checkerboard study : associations including CHX as antimicrobial were often synergistic, especially in the case of CHXquaternary ammonium compound associations. In the study of preservatives, the combination of PEA with phenolic compounds (mC, CC) led to a synergistic effect against both test strains. Considering excipients, it may be noted that TRI-antimicrobial agent associations were often synergistic, whereas Amo BA-antimicrobial agent associations were antagonistic except with KAT; in addition IPA potentiated the activity of antiseptics and preservatives against E. coli. Isobolograms (graphs with points recording equivalent antibacterial effect) provide an additional information to compare two agents whose combinations with the same excipient are synergistic; Fig. 2 shows that synergistic combinations mC-Amo SB and PEA-Amo SB are easily differentiated. As the M I C is a partial result of inhibitory activity for a single compound (see above), the isobologram of FICs of an association is inadequate to quantify relative inhibition at various concentrations. Figure 3 presents a threedimensional isobologram of the Amo BA-CC association, including growth intensity as a third parameter. This type of graph may be particularly helpful for compounds such as surface-active agents which do not give clear-cut M I C endpoints when they are tested singly. DISCUSSION
Fig. 3 Three-dimensional isobologram of the association of
parachlorometacresol (CC) with Amonyl380 BA@(Amo BA). Test strain: Staphylococcus aureus CIP 53154. Growth intensity is expressed as a percentage of growth in drug-free control cultures
T h e aim of the present study was to evaluate antimicrobial interactions between numerous associations with methods that appear well adapted to large scale studies. T h e MICs of individually tested compounds were determined as a first step, but they do not give any predictive information about their activity in complex formulations. I n addition, our results point out that doseresponse curves are necessary to describe the bacteriostatic effect of nontypical antimicrobials such as surface-active agents. M I C determinations, however, are essential as an initial step in the study of combinations and must be verified when testing combinations to avoid misinterpretations of synergistic or antagonistic effects. They are also helpful in the
I N VITRO INTERACTIONS I N ANTISEPTICS
initial choice of the various components of an association because antiseptics require a broad spectrum of activity that may be achieved with more than one antimicrobial agent. Antimicrobial interactions in the associations were evaluated by two different methods whose results showed a fair correlation. In most cases, the discordances were observed for combinations of antimicrobial agents with surface-active agents. It may be that the serial dilutions in the wells of a microplate, in the checkerboard method, are not well adapted to ‘scrubbing’ mixtures. However, the method of Berenbaum (1978) tests combinations of compounds at the same concentrations, whereas the checkerboard method allows combinations of various concentrations of two components. The first method may be considered as a simple and powerful tool to eliminate antagonistic associations in the initial step of a formulation study. Further studies are then necessary with the checkerboard titration method to elaborate a synergistic association. Isobolograms may then serve in discriminating various synergistic combinations. Berenbaum (1978) pointed out that his method was of value to study combinations of any number of agents. It would be perhaps well advised to complete the results by recording growth intensity, as we proposed with the checkerboard titration method. Synergistic combinations of CHX with BAK or C E T have already been reported (Delmotte et al. 1972), and are widely used in antiseptic formulations. Our work confirms that the associations of quaternary ammonium compounds, which are generally known to damage the cytoplasmic membrane of the bacterial cell, with compounds exhibiting other sites of action (CHX, PEA) may be synergistic. Previous studies had also demonstrated that CHX, which acts on the permeability barriers, may enhance the penetration and the intracellular activity of antibiotics such as gentamycin or neomycin on bacterial protein synthesis (Barnham & Kerby 1980). Associations between agents acting on the same biological targets may be merely additive. The associations between antiseptics and preservatives were not antagonistic except in one case (BZA-CC with Staph. aureus). These observations agree with the results of Hugbo (1976) who demonstrated synergistic bactericidal effects between BAK and phenolic compounds. Herein, associations of PEA and phenolic compounds showed a synergistic bacteriostatic activity. They had been previously considered as either additive (Richards 8i McBride 1971) or synergistic (Hugbo 1976) in tests of bactericidal activity. Scrub formulations, especially needed in surgical hand disinfection, include surface-active agents as excipient. Amo BA, Amo SB (amphoteric surface-active agents) and T R I (a non-ionic surface-active agent) were chosen as test excipients because they were physically compatible with the cationic antiseptics included in this work. Our results confirm that physical compatibility may be insufficient to
399
maintain the antibacterial activity of an antiseptic, as evidenced by the antagonistic associations of Amo BA. Furthermore, the exploration of combinations of various concentrations of both components is essential and requires the use of the checkerboard titration method. Isobolograms showing growth intensity may be proposed in the study of interactions involving excipients that often fail to provide a clear-cut end-point of antibacterial activity. T h e two methods chosen in this study are powerful tools to evaluate interactions within topical antimicrobial formulations. They allow large scale studies with other strains representative of the cutaneous microflora.
REFERENCES AL-NAJJAR, A. & QUESNEL, L.B. (1979) Synergism between chlorhexidine and polymyxins against Pseudomonas aeruginosa. Journal of Applied Bacteriology 47, 469-476. A N O N . (1973) Analysis of in vitro antibacterial activities of the combination of trimethoprim and sulfamethoxazole on clinical isolates in Japan (Japan cooperative bacteriological study group for co-trimoxazole). Journal of Infectious Diseases 128 (suppl.), 502-507. A N O N . (1980) Oflcial Methods of Analysis, 13th edn. Washington, DC: Association of Official Analytical Chemists (AOAC). A N O N . (1981) Richtlinienfir die Priifng chemischer Desinfektionsmittel. (Deutsche Gesellschaft fur Hygiene und Mikrobiologie ; DGHM). Stuttgart: G. Fischer Verlag. A N O N. (1987) Prkparations antiseptiques. Pharmacopie Francaise, 10th edn. Additif no. 5. Paris: Adrapharm. ANON. (1989) Antiseptiques et diinfectants. Recueil de Normes francaises, 2nd edn. Versions anglaise et franpise. Paris: Association Franpise de Normalisation (AFNOR). BARNHAM M., & K E R B YJ, . (1980) Antibacterial activity of combinations of chlorhexidine with neomycin and gentamycin. Journal of Hospital Infiction 1, 77-81. BERENBAUM, M.C. (1978) A method for testing for synergy with any number of agents. Journal of Infectious Diseases 137, 122-130. BRUMFITT,W., DIXSON, S. & H A M I L T O N - M I L L E R , J .H . T. (1985) Resistance to antiseptics in methicillin and gentamycin resistant Staphylococcus aureus. Lancet 1, 1442-1443. C O M B E R K.R., , BASKER,M.J., O S B O R N EC.D. , & S U T H E R L A NR.D ,(1977) Synergy between ticarcillin and tobramycin against Pseudomonas aeruginosa and Enterobacteriaceae in vitro and in vivo. Antimicrobial Agents and Chemotherapy 11, 956-964. C R E M I E U XA., , GUIRRAUD-DAURIA HC . , & BENDJELLOUL, D. (1983a) Influence des tensio-actifs non ioniques et ampholytes sur I’activite bactkricide de I’hexachIoroph&-ie, du glutaraldkhyde et de la chlorhexidine. Journal de Pharmacie de Belgique 38,22-26. C R E M I E U XA, , , G U I R R A U D - D A U R I A HC . , & BENDJELLOUL,D . (1983b) Intefierence de deux surfactants anioniques sur I’activitk bactkricide de quelques antiseptiques. Journal de Pharmacie de Belgique 38, 101-104.
400 J.-L. P O N S E T A L .
D E L M O T T EA, , , B E U M E RJ, . , C O T T O N ,E., D E K E Y S E R D E L M O T T EN, . , M I L L E T , M . , VAN D E N ABBEELE, K . G . & YOURASSOWSKY, E. (1972) Etude sur la sensibilite du bacille pyocyanique (Pseudomonas aeruginosa) aux antiseptiques et antibiotiques. IV. Synergie ou antagonisme entre antiseptiques. Therapie 28, 445455. HUGBOP , . G . (1976) Additivity and synergism in vitro as displayed by mixtures of some commonly employed antibacterial preservatives. Canadian Journal of Pharmaceutical Sciences 11, 17-20. L O W B U R YE., J . L . (1976) Prophylaxis and treatment for infection of burns. British Journal of Hospital Medicine 15, 566472. Q U E S N E L , L.B., A L - N A J J A R , A.R. & BUDDHAV U D H I K R A I , P. (1978) Synergism between chlorhexidine and sulphadiazine. Journal of Applied Bacteriology 45, 397405.
REYBROUCK G,. (1986) Unification of the testing of disinfectants in Europe. Zentralblatt f i r Bakteriologie, Parasitenkunde, Infektionskrankenheiten und Hygiene, 1, Abteilung, Originale Reihe B 182, 485498. R I C H A R D SR, . M . E . & M C B R I D ER.J. , (1971) Phenylethanol enhancement of preservatives used in ophthalmic preparations. Journal of Pharmacy and Pharmacology 23 (suppl.), 141-146. S I M O N E T T IN , . , D ’ A U R I A , F . D . , S T R I P P O LV. I, & L U C H E T T I ,G . (1988) Itraconazole: increased activity by chlorhexidine. Drugs Under Experimental and Clinical Research 14, 19-23. T U R N EAR. ,, H A W K E YP, . H . & P E D L E RS.J. , (1983) Computer assisted determination of antibiotic susceptibility in photometer read microtiter plate. Journal of Clinical Microbiology 18,996-998.
Evaluation of antimicrobial interactions between chlorhexidine, quaternary ammonium compounds, preservatives and excipients J.-L. Pons, N. Bonnavelro’, J. Chevalier1and A. Cremleux’ Laboratoire de Microbiologie-Pharmacie, U.F.R.de Medecine-Pharmacie de Rouen, SaintItienne Rouvray and Laboratoire de Microbiologie,Hygiene Microbienne,Immunologie,Facult6 de Pharmacie, Marseille, France
’
4114/02/92: accepted 27 April 1992 CREMIEUX1992. T h e antimicrobial interactions of 49 combinations of chlorhexidine, quaternary ammonium compounds, preservatives and excipients were evaluated by the method of Berenbaum and the checkerboard titration method, with Staphylococcus aureus CIP 53 154 and Escherichia coli CIP 54127 as test strains. MIC determinations were carried out as a preliminary step, and relative growth intensity was used to describe the bacteriostatic activity of surface-active agents (Amonyl 380 BA@, Amonyl 671 SB@). I n the study of combinations, results were interpreted with Fractional Inhibitory Concentration indexes and represented by isobolograms. A fair correlation was shown between the method of Berenbaum and the checkerboard titration method. Combinations between chlorhexidine, cetrimonium bromide and benzalkonium chloride were synergistic or additive ; combinations of antiseptics and preservatives were generally not antagonistic. T h e methods were also well adapted to the study of interactions involving surface-active agents, a critical problem in the formulation of topical antimicrobial agents. J.-L. PONS, N. BONNAVEIRO, J. CHEVALIER AND A.
INTRODUCTION
Antiseptics are widely used in hospitals for hygienic and surgical hand disinfection (Reybrouck 1986) and are usually preferred to antibiotics for topical application in the prevention of superficial infections (Lowbury 1976). Their formulation may be difficult because of physico-chemical and biological interactions between antibacterial agents and excipients or preservatives. Furthermore, they require a high level of activity that cannot be obtained by increasing concentrations of active components because of tolerance side effects. Therefore, the combination of such topical antimicrobial agents with preservatives and excipients must not only avoid antagonism but ideally provide an additive or preferably a synergistic effect. Reports have dealt with combinations of chlorhexidine and other antiseptics with surface-active agents (Crimieux et al. 1983a,b), with antibiotics (Quesnel et al. 1978; Al-Najjar & Quesnel 1979; Barnham & Kerby 1980) or with antifungal agents (Simonetti et al. 1988). Correspondence to :Jean-Louis Pons, Laboratoire de Microbiologie-Pharmacie, U.F.R. de Midecine-Pharmacie de Rouen, B.P.
97, F-76803 Saint-Etienne Rouvray Cedex, France.
Standardized methods to evaluate the bactericidal activities of antiseptics (Anon. 1980, 1981, 1987, 1989) cannot be routinely performed in the development of complex formulations which require numerous assays. More convenient methods are necessary, adaptable to large-scale studies. The present work was an attempt to use the established methods to study combinations of antimicrobial agents in the evaluation of antibacterial interactions involving antiseptics. The method of Berenbaum (1978) and the checkerboard titration method, widely used in the study of antibiotics (Comber et al. 1977), were applied to assess antagonism, additivity, indifference or synergy in 49 combinations of antiseptics, preservatives and excipients. MATERIALS AND METHODS Bacterlal stralns
Staphylococcus aureus CIP 53154 and Escherichia coli CIP 54127 (Collection Institut Pasteur, Paris, France) were used as test organisms. Bacterial inocula were prepared according to AFNOR (Anon. 1989) guidelines. Briefly, strains were grown on Tryptic Soy Agar (Difco) at 37°C for 24 h
396
J . - L . PONS E r A L .
subcultured three times. Bacterial cells were dispersed in a diluent containing Tryptone (Difco) (1 g/l) and sodium chloride (8.5 g/l) in distilled water. The optical density (623 nm) was adjusted so that the suspensions contained 1-3 x 10' colony-forming units (cfu) per ml, as shown by counts on Plate Count Agar (Difco).
Study of associations
Method of Berenbaum (7978) Agents were mixed in test tubes, each at a final concentration corresponding to 2 x MIC. Bacteriostatic activity of these binary combinations was then evaluated by a microdilution technique in microplates as described for single agents. I n addition, the M I C of each component of an association was measured concomitantly as a control.
Compounds
Checkerboard titration method Serial twofold dilutions of each agent were prepared, the median value of seven concentrations being equivalent to the MIC. Pairwise combinations of all concentrations of each agent were then placed in the microplates. Their bacteriostatic activity was evaluated as described below after inoculating with test strains.
The following compounds were investigated alone or in combinations : chlorhexidine digluconate (CHX) and cetrimonium bromide (CET) (ICI-Pharma, Cergy, France) ; benzalkonium chloride (BZA) (Coopkration Pharmaceutique Fragaise, Melun, France) ; phenylethyl alcohol (PEA) (Sigma) ; isopropyl alcohol (IPA), metacresol (mC) and parachlorometacresol (CC) (E. Merck AG, Darmstadt, Germany); Amonyl 380 BA@ (lauryl amidobetaine, Amo BA), Amonyl 671 SB@ (lauryl amido hydroxy sulfobetaine, Amo SB), Kathon CG@ (methyl isothiazolone and chloromethyl isothiazolone, KAT) and Triton C G llO@ (alkyl glucoside, T R I ) (Seppic, Division Cosmktique-Pharmacie, Paris, France).
Expression of results Each result was first expressed as Fractional Inhibitory Concentration index (XFIC), which is the sum of the lowest concentrations of each agent of an association (expressed as a fraction of MIC) inhibiting growth (Anon. 1973; Berenbaum 1978). Then the C F I C values allowed interpretation on the following criteria: XFIC I 0.75 : synergistic; XFIC = 1 : additive; 1 c XFIC c 2 : indifferent; XFIC 2 2 : antagonistic).
MIC determinations
Minimum inhibitory concentration (MIC) determinations were performed by a microdilution technique (serial twofold dilution) in 96 well microplates. The test volume of each well was 0.2 ml. Nutrient Broth (Difco) containing 3 g/l Yeast Extract (Difco) was inoculated with 1-3 x lo5 cfu/ml. Microplates were incubated at 35°C for 48 h. Growth was estimated by reading optical density with a Multiskan spectrophotometer (Titertek, Flow Laboratories, Helsinki) set at 623 nm. M I C was defined as the lowest concentration giving less than 20% growth as compared with the control (Turner et al. 1983).
RESULTS inhibitory activity of antiseptics, preservatives and exciplents
T h e M I C values of antiseptics, preservatives and excipients are listed in Table 1. T h e results agree with those of previous reports on CHX (Quesnel et al. 1978; Al-Najjar & Quesnel 1979; Barnham & Kerby 1980; Brumfitt et al. 1985) and C E T (Brumfitt et al. 1985). T h e MICs of pre-
Table 1 Minimal inhibitory concentrations (MICs) of antiseptics, preservatives and excipients (not combined)
MIC value Antiseptics (pg/ml)
Preservatives (pg/ml)
Excipients (YOv/v)
Test strain
CHX
CET
BZA
mC
CC
KAT
PEA
IPA
AmoBA
AmoSB
TRI
Staphylococcus aureus CIP 53154
2
1
0.5
1000
200
0.5
4000
20
5
0.0125
0.125
Escherichia coli CIP 54127
4
8
5
1000
100
0.125
2000
20
20
0.1
0.5
CHX, Chlorhexidine digluconate ; CET, cetrimonium bromide ; BZA, benzalkoniurn chloride; mC, metacresol ; CC, parachlorometacresol ; KAT, Kathon CG@;PEA, phenylethyl alcohol; IPA, isopropyl alcohol; Amo BA, Amonyl 380 BA@;Amo SB, Amonyl 671 SB@;TRI, Triton CG 1 lo@.
IN VlTRO INTERACTIONS IN ANTISEPTICS 397
Amo BA (70v/v)
Amo SB ( % v/v)
CET ( p g / m l )
Fig. 1 Dose-response curves of antimicrobial inhibitory activity. (a) Amonyl 380 BA@(Amo BA); (b) Amonyl671 SB@(Amo SB); (c) cetrimonium bromide (CET). Test strains: Staphylococcus aureus CIP 53154 Escherichia coli CIP 54127 (m). Growth intensity is expressed as a percentage of growth in drug-free control cultures
(a),
servatives were in the same order of magnitude with both strains, whereas E. coli was less sensitive than Staph. aureus to antiseptics and excipients. The M I C values must be considered as preliminary results, required for the further study of combinations. Non-typical antimicrobial agents (such as surface-active agents) may cause partial growth inhibition over a wide range of concentrations. In such cases MICs give only a partial information. Therefore, dose-response curves were drawn, where growth was expressed as a percentage of growth in drug-free control cultures. Graphs of Amo BA (Fig. la) and Amo SB (Fig. lb) show a partial inhibitory
activity against E. coli, especially with Amo BA; its concentrations from 0625% to 20% (v/v) were associated with a growth varying from 15% to 25%. T h e other compounds gave generally clear-cut M I C end-points, particularly with Staph. aureus. T h e graph of CET (Fig. lc) is given as an example. inhibitory activity of associations
The results of antimicrobial interactions in all pairwise combinations, investigated by the checkerboard titration method and the method of Berenbaum (1978) are shown in
Table 2 Antimicrobial interactions in combinations of antiseptics, preservatives and excipients: comparison of the checkerboard titration method and the method of Berenbaum (result given between parentheses)
Staphylococcus aureus CIP 53 154
Escherichia coli CIP 54127
CHX CET BZA mC
cc
KAT PEA IPA AmoBA Amo SB TRI
CHX
CET
BZA
mC
CC
KAT
PEA
IPA
AmoBA
AmoSB
TRI
***
S(S)'
S(S)2 ad(S) S(S) S(i) S(S) S(ad) S(A) A(A)
***
S(S) S(ad)
i(ad) ad(ad) i(A)
S(S) S(ad) S(A) ad(ad) ad(ad)
ad(A) S(i) i(A) i(S) ad(S) ad(i) i(S)
A(i) A(A) A(S) A(ad) A(ad) S(A) A(S)
S(S) i(ad) ad(S) ad(i) ad(S) i(ad) S(i)
i(i) S(ad) S(i) S(A) S(A) S(A) S(A)
i(ad) S(S)
ad@)
***
S(ad)
i(r)
***
i(ad) i(ad) A(A) ad(ad)
S(A)
ad(i) i(S)
S(ad) S(A) S(S) S(S) A(S)
i(A) S(S) ad(A) A(A)
ad(ad)
ad(ad)
S(ad)
S(ad)
i(A) S(i) S(i)
A(i) S(i)
S(i)
ad(ad) S(A) A(A) i(A) A(A)
***
***
ad(S) S(S) S(i) S(S) S(i) S(i)
ad(i) S(S)
S(A)
ad(A) i(ad) ad(ad)
A(A) S(A)
***
a(a)
S, synergistic (ZFIC 5 0.75); ad, additive (ZFIC = 1); i, indifferent (1 < ZFIC < 2); A, antagonistic (CFIC 2 2). Normal characters : antimicrobial interaction with Staph. aureus; Italic characters : antimicrobial interaction with E . colt. Key as for Table 1. ***, Untested.
398 J.-L. PONS E r A L .
E. coli. A major discordance (synergy with one method,
0.25
0.50
0.75
I
mC or PEA (FIC)
Fig. 2 Isobolograms of the associations of metacresol (mC, 0) or phenylethyl alcohol (PEA,). with Amonyl671 SB@(Amo SB). Test strain: Staphylococcus aureus CIP 53154
Table 2. Interpretation is based upon C F I C values, which were calculated by comparison with the MICs determined concomitantly. Results show a fair correlation between the methods in most of the 49 combinations tested with Staph. aureus and
antagonism with the other one) was noted in 15 cases among 98 tests. An additivity-antagonism discordance was noted in five cases. Antimicrobial interactions may differ according to the test strain. T h e BZA-CC association was antagonistic with Staph. aureus, whereas it was additive or indifferent with E. coli. Associations of CNX or BZA with T R I led to opposite results with the two test strains. Another typical example is given with the PEA-Amo SB association. T h e following trends are worth noting in the checkerboard study : associations including CHX as antimicrobial were often synergistic, especially in the case of CHXquaternary ammonium compound associations. In the study of preservatives, the combination of PEA with phenolic compounds (mC, CC) led to a synergistic effect against both test strains. Considering excipients, it may be noted that TRI-antimicrobial agent associations were often synergistic, whereas Amo BA-antimicrobial agent associations were antagonistic except with KAT; in addition IPA potentiated the activity of antiseptics and preservatives against E. coli. Isobolograms (graphs with points recording equivalent antibacterial effect) provide an additional information to compare two agents whose combinations with the same excipient are synergistic; Fig. 2 shows that synergistic combinations mC-Amo SB and PEA-Amo SB are easily differentiated. As the M I C is a partial result of inhibitory activity for a single compound (see above), the isobologram of FICs of an association is inadequate to quantify relative inhibition at various concentrations. Figure 3 presents a threedimensional isobologram of the Amo BA-CC association, including growth intensity as a third parameter. This type of graph may be particularly helpful for compounds such as surface-active agents which do not give clear-cut M I C endpoints when they are tested singly. DISCUSSION
Fig. 3 Three-dimensional isobologram of the association of
parachlorometacresol (CC) with Amonyl380 BA@(Amo BA). Test strain: Staphylococcus aureus CIP 53154. Growth intensity is expressed as a percentage of growth in drug-free control cultures
T h e aim of the present study was to evaluate antimicrobial interactions between numerous associations with methods that appear well adapted to large scale studies. T h e MICs of individually tested compounds were determined as a first step, but they do not give any predictive information about their activity in complex formulations. I n addition, our results point out that doseresponse curves are necessary to describe the bacteriostatic effect of nontypical antimicrobials such as surface-active agents. M I C determinations, however, are essential as an initial step in the study of combinations and must be verified when testing combinations to avoid misinterpretations of synergistic or antagonistic effects. They are also helpful in the
I N VITRO INTERACTIONS I N ANTISEPTICS
initial choice of the various components of an association because antiseptics require a broad spectrum of activity that may be achieved with more than one antimicrobial agent. Antimicrobial interactions in the associations were evaluated by two different methods whose results showed a fair correlation. In most cases, the discordances were observed for combinations of antimicrobial agents with surface-active agents. It may be that the serial dilutions in the wells of a microplate, in the checkerboard method, are not well adapted to ‘scrubbing’ mixtures. However, the method of Berenbaum (1978) tests combinations of compounds at the same concentrations, whereas the checkerboard method allows combinations of various concentrations of two components. The first method may be considered as a simple and powerful tool to eliminate antagonistic associations in the initial step of a formulation study. Further studies are then necessary with the checkerboard titration method to elaborate a synergistic association. Isobolograms may then serve in discriminating various synergistic combinations. Berenbaum (1978) pointed out that his method was of value to study combinations of any number of agents. It would be perhaps well advised to complete the results by recording growth intensity, as we proposed with the checkerboard titration method. Synergistic combinations of CHX with BAK or C E T have already been reported (Delmotte et al. 1972), and are widely used in antiseptic formulations. Our work confirms that the associations of quaternary ammonium compounds, which are generally known to damage the cytoplasmic membrane of the bacterial cell, with compounds exhibiting other sites of action (CHX, PEA) may be synergistic. Previous studies had also demonstrated that CHX, which acts on the permeability barriers, may enhance the penetration and the intracellular activity of antibiotics such as gentamycin or neomycin on bacterial protein synthesis (Barnham & Kerby 1980). Associations between agents acting on the same biological targets may be merely additive. The associations between antiseptics and preservatives were not antagonistic except in one case (BZA-CC with Staph. aureus). These observations agree with the results of Hugbo (1976) who demonstrated synergistic bactericidal effects between BAK and phenolic compounds. Herein, associations of PEA and phenolic compounds showed a synergistic bacteriostatic activity. They had been previously considered as either additive (Richards 8i McBride 1971) or synergistic (Hugbo 1976) in tests of bactericidal activity. Scrub formulations, especially needed in surgical hand disinfection, include surface-active agents as excipient. Amo BA, Amo SB (amphoteric surface-active agents) and T R I (a non-ionic surface-active agent) were chosen as test excipients because they were physically compatible with the cationic antiseptics included in this work. Our results confirm that physical compatibility may be insufficient to
399
maintain the antibacterial activity of an antiseptic, as evidenced by the antagonistic associations of Amo BA. Furthermore, the exploration of combinations of various concentrations of both components is essential and requires the use of the checkerboard titration method. Isobolograms showing growth intensity may be proposed in the study of interactions involving excipients that often fail to provide a clear-cut end-point of antibacterial activity. T h e two methods chosen in this study are powerful tools to evaluate interactions within topical antimicrobial formulations. They allow large scale studies with other strains representative of the cutaneous microflora.
REFERENCES AL-NAJJAR, A. & QUESNEL, L.B. (1979) Synergism between chlorhexidine and polymyxins against Pseudomonas aeruginosa. Journal of Applied Bacteriology 47, 469-476. A N O N . (1973) Analysis of in vitro antibacterial activities of the combination of trimethoprim and sulfamethoxazole on clinical isolates in Japan (Japan cooperative bacteriological study group for co-trimoxazole). Journal of Infectious Diseases 128 (suppl.), 502-507. A N O N . (1980) Oflcial Methods of Analysis, 13th edn. Washington, DC: Association of Official Analytical Chemists (AOAC). A N O N . (1981) Richtlinienfir die Priifng chemischer Desinfektionsmittel. (Deutsche Gesellschaft fur Hygiene und Mikrobiologie ; DGHM). Stuttgart: G. Fischer Verlag. A N O N. (1987) Prkparations antiseptiques. Pharmacopie Francaise, 10th edn. Additif no. 5. Paris: Adrapharm. ANON. (1989) Antiseptiques et diinfectants. Recueil de Normes francaises, 2nd edn. Versions anglaise et franpise. Paris: Association Franpise de Normalisation (AFNOR). BARNHAM M., & K E R B YJ, . (1980) Antibacterial activity of combinations of chlorhexidine with neomycin and gentamycin. Journal of Hospital Infiction 1, 77-81. BERENBAUM, M.C. (1978) A method for testing for synergy with any number of agents. Journal of Infectious Diseases 137, 122-130. BRUMFITT,W., DIXSON, S. & H A M I L T O N - M I L L E R , J .H . T. (1985) Resistance to antiseptics in methicillin and gentamycin resistant Staphylococcus aureus. Lancet 1, 1442-1443. C O M B E R K.R., , BASKER,M.J., O S B O R N EC.D. , & S U T H E R L A NR.D ,(1977) Synergy between ticarcillin and tobramycin against Pseudomonas aeruginosa and Enterobacteriaceae in vitro and in vivo. Antimicrobial Agents and Chemotherapy 11, 956-964. C R E M I E U XA., , GUIRRAUD-DAURIA HC . , & BENDJELLOUL, D. (1983a) Influence des tensio-actifs non ioniques et ampholytes sur I’activite bactkricide de I’hexachIoroph&-ie, du glutaraldkhyde et de la chlorhexidine. Journal de Pharmacie de Belgique 38,22-26. C R E M I E U XA, , , G U I R R A U D - D A U R I A HC . , & BENDJELLOUL,D . (1983b) Intefierence de deux surfactants anioniques sur I’activitk bactkricide de quelques antiseptiques. Journal de Pharmacie de Belgique 38, 101-104.
400 J.-L. P O N S E T A L .
D E L M O T T EA, , , B E U M E RJ, . , C O T T O N ,E., D E K E Y S E R D E L M O T T EN, . , M I L L E T , M . , VAN D E N ABBEELE, K . G . & YOURASSOWSKY, E. (1972) Etude sur la sensibilite du bacille pyocyanique (Pseudomonas aeruginosa) aux antiseptiques et antibiotiques. IV. Synergie ou antagonisme entre antiseptiques. Therapie 28, 445455. HUGBOP , . G . (1976) Additivity and synergism in vitro as displayed by mixtures of some commonly employed antibacterial preservatives. Canadian Journal of Pharmaceutical Sciences 11, 17-20. L O W B U R YE., J . L . (1976) Prophylaxis and treatment for infection of burns. British Journal of Hospital Medicine 15, 566472. Q U E S N E L , L.B., A L - N A J J A R , A.R. & BUDDHAV U D H I K R A I , P. (1978) Synergism between chlorhexidine and sulphadiazine. Journal of Applied Bacteriology 45, 397405.
REYBROUCK G,. (1986) Unification of the testing of disinfectants in Europe. Zentralblatt f i r Bakteriologie, Parasitenkunde, Infektionskrankenheiten und Hygiene, 1, Abteilung, Originale Reihe B 182, 485498. R I C H A R D SR, . M . E . & M C B R I D ER.J. , (1971) Phenylethanol enhancement of preservatives used in ophthalmic preparations. Journal of Pharmacy and Pharmacology 23 (suppl.), 141-146. S I M O N E T T IN , . , D ’ A U R I A , F . D . , S T R I P P O LV. I, & L U C H E T T I ,G . (1988) Itraconazole: increased activity by chlorhexidine. Drugs Under Experimental and Clinical Research 14, 19-23. T U R N EAR. ,, H A W K E YP, . H . & P E D L E RS.J. , (1983) Computer assisted determination of antibiotic susceptibility in photometer read microtiter plate. Journal of Clinical Microbiology 18,996-998.
