As with antibiotics and other chemotherapeutic drugs, acquired resistance to antiseptics and disinfectants can arise by either mutation or the acquisition of genetic material in the form of plasmids or transposons. It is important to note that “resistance” as a term can often be used loosely and in many cases must be interpreted with some prudence. This is particularly true with MIC analysis. Unlike antibiotics, “resistance,” or an increase in the MIC of a biocide, does not necessarily correlate with therapeutic failure. An increase in an antibiotic MIC can may have significant consequences, often indicating that the target organism is unaffected by its antimicrobial action. Increased biocide MICs due to acquired mechanisms have also been reported and in some case misinterpreted as indicating resistance. It is important that issues including the pleiotropic action of most biocides, bactericidal activity, concentrations used in products, direct product application, formulation effects, etc., be considered in evaluating the clinical implications of these reports.
Plasmids and bacterial resistance to antiseptics and disinfectants.
Chopra (
82,
83) examined the role of plasmids in encoding resistance (or increased tolerance) to antiseptics and disinfectants; this topic was considered further by Russell (
413). It was concluded that apart from certain specific examples such as silver, other metals, and organomercurials, plasmids were not normally responsible for the elevated levels of antiseptic or disinfectant resistance associated with certain species or strains. Since then, however, there have been numerous reports linking the presence of plasmids in bacteria with increased tolerance to chlorhexidine, QACs, and triclosan, as well as to diamidines, acridines and ethidium bromide, and the topic was reconsidered (
83,
423,
427) (Table
9).
Plasmid-encoded resistance to antiseptics and disinfectants had at one time been most extensively investigated with mercurials (both inorganic and organic), silver compounds, and other cations and anions. Mercurials are no longer used as disinfectants, but phenylmercuric salts and thiomersal are still used as preservatives in some types of pharmaceutical products (
226). Resistance to mercury is plasmid borne, inducible, and may be transferred by conjugation or transduction. Inorganic mercury (Hg
2+) and organomercury resistance is a common property of clinical isolates of
S. aureus containing penicillinase plasmids (
110). Plasmids conferring resistance to mercurials are either narrow spectrum, specifying resistance to Hg
2+ and to some organomercurials, or broad-spectrum, with resistance to the above compounds and to additional organomercurials (
331). Silver salts are still used as topical antimicrobial agents (
50,
443). Plasmid-encoded resistance to silver has been found in
Pseudomonas stutzeri (
192), members of the
Enterobacteriaceae (
479,
480,
511), and
Citrobacter spp. (
511). The mechanism of resistance has yet to be elucidated fully but may be associated with silver accumulation (
152,
511).
(i) Plasmid-mediated antiseptic and disinfectant resistance in gram-negative bacteria.
Occasional reports have examined the possible role of plasmids in the resistance of gram-negative bacteria to antiseptics and disinfectants. Sutton and Jacoby (
498) observed that plasmid RP1 did not significantly alter the resistance of
P. aeruginosa to QACs, chlorhexidine, iodine, or chlorinated phenols, although increased resistance to hexachlorophene was observed. This compound has a much greater effect on gram-positive than gram-negative bacteria, so that it is difficult to assess the significance of this finding. Transformation of this plasmid (which encodes resistance to carbenicillin, tetracycline, and neomycin and kanamycin) into
E. coli or
P. aeruginosa did not increase the sensitivity of these organisms to a range of antiseptics (
5).
Strains of
Providencia stuartii may be highly tolerant to Hg
2+, cationic disinfectants (such as chlorhexidine and QACs), and antibiotics (
496). No evidence has been presented to show that there is a plasmid-linked association between antibiotic resistance and disinfectant resistance in these organisms, pseudomonads, or
Proteus spp. (
492). High levels of disinfectant resistance have been reported in other hospital isolates (
195), although no clear-cut role for plasmid-specified resistance has emerged (
102,
250,
348,
373,
518). High levels of tolerance to chlorhexidine and QACs (
311) may be intrinsic or may have resulted from mutation. It has been proposed (
492,
496) that the extensive usage of these cationic agents could be responsible for the selection of antiseptic-disinfectant-, and antibiotic-resistant strains; however, there is little evidence to support this conclusion. All of these studies demonstrated that it was difficult to transfer chlorhexidine or QAC resistance under normal conditions and that plasmid-mediated resistance to these chemicals in gram-negative bacteria was an unlikely event. By contrast, plasmid R124 alters the OmpF outer membrane protein in
E. coli, and cells containing this plasmid are more resistant to a QAC (cetrimide) and to other agents (
406).
Bacterial resistance mechanisms to formaldehyde and industrial biocides may be plasmid encoded (
71,
193). Alterations in the cell surface (outer membrane proteins [
19,
246]) and formaldehyde dehydrogenase (
247,
269) are considered to be responsible (
202). In addition, the so-called TOM plasmid encodes enzymes for toluene and phenol degradation in
B. cepacia (
476).
(ii) Plasmid-mediated antiseptic and disinfectant resistance in staphylococci.
Methicillin-resistant
S. aureus (MRSA) strains are a major cause of sepsis in hospitals throughout the world, although not all strains have increased virulence. Many can be referred to as “epidemic” MRSA because of the ease with which they can spread (
91,
295,
317). Patients at particularly high risk are those who are debilitated or immunocompromised or who have open sores.
It has been known for several years that some antiseptics and disinfectants are, on the basis of MICs, somewhat less inhibitory to
S. aureus strains that contain a plasmid carrying genes encoding resistance to the aminoglycoside antibiotic gentamicin (Table
10). These biocidal agents include chlorhexidine, diamidines, and QACs, together with ethidium bromide and acridines (
8,
238,
289,
368,
423,
427,
463). According to Mycock (
346), MRSA strains showed a remarkable increase in tolerance (at least 5,000-fold) to povidone-iodine. However, there was no decrease in susceptibility of antibiotic-resistant strains to phenolics (phenol, cresol, and chlorocresol) or to the preservatives known as parabens (
8).
Tennent et al. (
505) proposed that increased resistances to cetyltrimethylammonium bromide (CTAB) and to propamidine isethionate were linked and that these cationic agents may be acting as a selective pressure for the retention of plasmids encoding resistance to them. The potential clinical significance of this finding remains to be determined.
Staphylococci are the only bacteria in which the genetic aspects of plasmid-mediated antiseptic and disinfectant resistant mechanisms have been described (
466). In
S. aureus, these mechanisms are encoded by at least three separate multidrug resistance determinants (Tables
10 and
11). Increased antiseptic MICs have been reported to be widespread among MRSA strains and to be specified by two gene families (
qacAB and
qacCD) of determinants (
188,
280,
281,
288,
289,
363-365,
367,
506). The
qacAB family of genes (Table
11) encodes proton-dependant export proteins that develop significant homology to other energy-dependent transporters such as the tetracycline transporters found in various strains of tetracycline-resistant bacteria (
405). The
qacAgene is present predominantly on the pSK1 family of multiresistance plasmids but is also likely to be present on the chromosome of clinical
S. aureus strains as an integrated family plasmid or part thereof. The
qacB gene is detected on large heavy-metal resistance plasmids. The
qacC and
qacD genes encode identical phenotypes and show restriction site homology; the
qacC gene may have evolved from
qacD(
288).
Interesting studies by Reverdy et al. (
395,
396), Dussau et al. (
129) and Behr et al. (
31) demonstrated a relationship between increased
S. aureus MICs to the β-lactam oxacillin and four antiseptics (chlorhexidine, benzalkonium chloride, hexamine, and acriflavine). A gene encoding multidrug resistance was not found in susceptible strains but was present in 70% of
S. aureus strains for which the MICs of all four of these antiseptics were increased and in 45% of the remaining strains resistant to at least one of these antiseptics (
31). Genes encoding increased QAC tolerance may be widespread in food-associated staphylococcal species (
203). Some 40% of isolates of coagulase-negative staphylococci (CNS) contain both
qacA and
qacC genes, with a possible selective advantage in possessing both as opposed to
qacAonly (
281). Furthermore, there is growing evidence that
S. aureus and CNS have a common pool of resistance determinants.
Triclosan is used in surgical scrubs, soaps, and deodorants. It is active against staphylococci and less active against most gram-negative organisms, especially
P. aeruginosa, probably by virtue of a permeability barrier (
428). Low-level transferable resistance to triclosan was reported in MRSA strains (
88,
90); however, no further work on these organisms has been described. According to Sasatsu et al. (
465), the MICs of triclosan against sensitive and resistant
S. aureusstrains were 100 and >6,400 μg/ml, respectively; these results were disputed because these concentrations are well in excess of the solubility of triclosan (
515), which is practically insoluble in water. Sasatsu et al. (
464) described a high-level resistant strain of
S. aureus for which the MICs of chlorhexidine, CTAB, and butylparaben were the same as for a low-level resistant strain. Furthermore, the MIC quoted for methylparaben comfortably exceeds its aqueous solubility. Most of these studies with sensitive and “resistant” strains involved the use of MIC evaluations (for example, Table
6). A few investigations examined the bactericidal effects of antiseptics. Cookson et al. (
89) pointed out that curing of resistance plasmids produced a fall in MICs but not a consistent decrease in the lethal activity of chlorhexidine. There is a poor correlation between MIC and the rate of bactericidal action of chlorhexidine (
88,
89,
319) and triclosan (
90,
319). McDonnell et al. (
318,
319) have described methicillin-susceptible
S. aureus (MSSA) and MRSA strains with increased triclosan MICs (up to 1.6 μg/ml) but showed that the MBCs for these strains were identical; these results were not surprising, considering that biocides (unlike antibiotics) have multiple cellular targets. Irizarry et al. (
229) compared the susceptibility of MRSA and MSSA strains to CPC and chlorhexidine by both MIC and bactericidal testing methods. However, the conclusion of this study that MRSA strains were more resistant warrants additional comments. On the basis of rather high actual MICs, MRSA strains were some four times more resistant to chlorhexidine and five times more resistant to a QAC (CPC) than were MSSA strains. At a concentration in broth of 2 μg of CPC/ml, two MRSA strains grew normally with a threefold increase in viable numbers over a 4-h test period whereas an MSSA strain showed a 97% decrease in viability. From this, the authors concluded that it was reasonable to speculate that the residual amounts of antiseptics and disinfectants found in the hospital environment could contribute to the selection and maintenance of multiresistant MRSA strains. Irizarry et al. (
229) also concluded that MRSA strains are less susceptible than MSSA strains to both chronic and acute exposures to antiseptics and disinfectants. However, their results with 4 μg of CPC/ml show no such pattern: at this higher concentration, the turbidities (and viability) of the two MRSA and one MSSA strains decreased at very similar rates (if anything, one MRSA strain appeared to be affected to a slightly greater extent that the MSSA strain). Furthermore, the authors stated that chlorhexidine gave similar results to CPC. It is therefore difficult to see how Irizarry et al. arrived at their highly selective conclusions.
Plasmid-mediated efflux pumps are particularly important mechanisms of resistance to many antibiotics (
85), metals (
349), and cationic disinfectants and antiseptics such as QACs, chlorhexidine, diamidines, and acridines, as well as to ethidium bromide (
239,
289,
324-336,
363-368). Recombinant
S. aureusplasmids transferred into
E. coli are responsible for conferring increased MICs of cationic agents to the gram-negative organism (
505,
544). Midgley (
324,
325) demonstrated that a plasmid-borne, ethidium resistance determinant from
S. aureus cloned in
E. coli encodes resistance to ethidium bromide and to QACs, which are expelled from the cells. A similar efflux system is present in
Enterococcus hirae (
326).
Sasatsu et al. (
463) showed that duplication of
ebr is responsible for resistance to ethidium bromide and to some antiseptics. Later, Sasatsu et al. (
466) examined the origin of
ebr (now known to be identical to
qacCD) in
S. aureus;
ebr was found in antibiotic-resistant and -sensitive strains of
S. aureus, CNS, and enterococcal strains. The nucleotide sequences of the amplified DNA fragment from sensitive and resistant strains were identical, and it was proposed that in antiseptic-resistant cells there was an increase in the copy number of the
ebr(
qacCD) gene whose normal function was to remove toxic substances from normal cells of staphylococci and enterococci.
Based on DNA homology, it was proposed that
qacA and related genes carrying resistance determinants evolved from preexisting genes responsible for normal cellular transport systems (
405) and that the antiseptic resistance genes evolved before the introduction and use of topical antimicrobial products and other antiseptics and disinfectants (
288,
289,
365,
367,
368,
405).
Baquero et al. (
23) have pointed out that for antibiotics, the presence of a specific resistance mechanism frequently contributes to the long-term selection of resistant variants under in vivo conditions. Whether low-level resistance to cationic antiseptics, e.g., chlorhexidine, QACs, can likewise provide a selective advantage on staphylococci carrying
qac genes remains to be elucidated. The evidence is currently contentious and inconclusive.
Mutational resistance to antiseptics and disinfectants.
Chromosomal mutation to antibiotics has been recognized for decades. By contrast, fewer studies have been performed to determine whether mutation confers resistance to antiseptics and disinfectants. It was, however, demonstrated over 40 years ago (
77,
78) that
S. marcescens, normally inhibited by QACs at <100 μg/ml, could adapt to grow in 100,000 μg of a QAC per ml. The resistant and sensitive cells had different surface characteristics (electrophoretic mobilities), but resistance could be lost when the cells were grown on QAC-free media. One problem associated with QACs and chlorhexidine is the turbidity produced in liquid culture media above a certain concentration (interaction with agar also occurs), which could undoubtedly interfere with the determination of growth. This observation is reinforced by the findings presented by Nicoletti et al. (
354).
Prince et al. (
383) reported that resistance to chlorhexidine could be induced in some organisms but not in others. For example,
P. mirabilis and
S. marcescens displayed 128- and 258-fold increases, respectively, in resistance to chlorhexidine, whereas it was not possible to develop resistance to chlorhexidine in
Salmonella enteritidis. The resistant strains did not show altered biochemical properties of changed virulence for mice, and some strains were resistant to the QAC benzalkonium chloride. Moreover, resistance to chlorhexidine was stable in
S. marcescens but not in
P. mirabilis. Despite extensive experimentation with a variety of procedures, Fitzgerald et al. (
148) were unable to develop stable chlorhexidine resistance in
E. coli or
S. aureus. Similar observations were made by Cookson et al. (
89), who worked with MRSA and other strains of
S. aureus, and by McDonnell et al. (
319), who worked with MRSA and enterococci. Recently, stable chlorhexidine resistance was developed in
P. stutzeri (
502); these strains showed various levels of increased tolerance to QACs, triclosan, and some antibiotics, probably as a result of a nonspecific alteration of the cell envelope (
452). The adaptation and growth of
S. marcescens in contact lens disinfectants containing chlorhexidine, with cross-resistance to a QAC, have been described previously (
166).
Chloroxylenol-resistant strains of
P. aeruginosa were isolated by repeated exposure in media containing gradually increasing concentrations of the phenolic, but the resistance was unstable (
432). The adaptation of
P. aeruginosato QACs is a well-known phenomenon (
1,
2,
240). Resistance to amphoteric surfactants has also been observed, and, interestingly, cross-resistance to chlorhexidine has been noted (
240). This implies that the mechanism of such resistance is nonspecific and that it involves cellular changes that modify the response of organisms to unrelated biocidal agents. Outer membrane modification is an obvious factor and has indeed been found with QAC-resistant and amphoteric compound-resistant
P. aeruginosa (
240) and with chlorhexidine-resistant
S. marcescens(
166). Such changes involve fatty acid profiles and, perhaps more importantly, outer membrane proteins. It is also pertinent to note here the recent findings of Langsrud and Sundheim (
274). In this study, resistance of
P. fluorescens to QACs was reduced when EDTA was present with the QAC (although the lethal effect was mitigated after the cells were grown in medium containing QAC and EDTA); similar results have been found with laboratory-generated
E. coli mutants for which the MICs of triclosan were increased (
318). EDTA has long been known (
175,
410) to produce changes in the outer membrane of gram-negative bacteria, especially pseudomonads. Thus, it appears that, again, the development of resistance is associated with changes in the cell envelope, thereby limiting uptake of antiseptics and disinfectants.
Hospital (as for other environmental) isolates of gram-negative bacteria are invariably less sensitive to disinfectants than are laboratory strains (
196,
209,
279,
286,
492). Since plasmid-mediated transfer has apparently been ruled out (see above), selection and mutation could play an important role in the presence of these isolates. Subinhibitory antibiotic concentrations may cause subtle changes in the bacterial outer structure, thereby stimulating cell-to-cell contact (
109); it remains to be tested if residual concentrations of antiseptics or disinfectants in clinical situations could produce the same effect.
Another insusceptibility mechanism has been put forward, in this instance to explain acridine resistance. It has been proposed (
270,
351) that proflavine-sensitive and -resistant cells are equally permeable to the acridine but that resistant cells possessed the ability to expel the bound dye. This is an important point and one that has been reinforced by more recent studies that demonstrate the significance of efflux in resistance of bacteria to antibiotics (
284,
330,
355). Furthermore, multidrug resistance (MDR) is a serious problem in enteric and other gram-negative bacteria. MDR is a term used to describe resistance mechanisms used by genes that form part of the normal cell genome (
168). These genes are activated by induction or mutation caused by some types of stress, and because they are distributed ubiquitously, genetic transfer is not needed. Although such systems are most important in the context of antibiotic resistance, George (
168) provides several examples of MDR systems in which an operon or gene is associated with changes in antiseptic or disinfectant susceptibility; e.g., (i) mutations at an
acrlocus in the Acr system render
E. coli more sensitive to hydrophobic antibiotics, dyes, and detergents; (ii) the
robA gene is responsible for overexpression in
E. coli of the RobA protein that confers multiple antibiotic and heavy-metal resistance (interestingly, Ag
+ may be effluxed [
350]); and (iii) the MarA protein controls a set of genes (
mar and
soxRS regulons) that confer resistance not only to several antibiotics but also to superoxide-generating agents. Moken et al. (
333) have found that low concentrations of pine oil (used as a disinfectant) could select for
E. coli mutants that overexpressed MarA and demonstrated low levels of cross-resistance to antibiotics. Deletion of the
mar or
acrAB locus (the latter encodes a PMF-dependant efflux pump) increased the susceptibility of
E. coli to pine oil; deletion of
acrAB, but not of
mar, increased the susceptibility of
E. coli to chloroxylenol and to a QAC. In addition, the
E. coli MdfA (multidrug transporter) protein was recently identified and confers greater tolerance to both antibiotics and a QAC (benzalkonium) (
132). The significance of these and other MDR systems in bacterial susceptibility to antiseptics and disinfectants, in particular the issue of cross-resistance with antibiotics, must be studied further. At present, it is difficult to translate these laboratory findings to actual clinical use, and some studies have demonstrated that antibiotic-resistant bacteria are not significantly more resistant to the lethal (or bactericidal) effects of antiseptic and disinfectants than are antibiotic-sensitive strains (
11,
88,
89,
319).