Adaptive bacterial response to low level chlorhexidine exposure and its implications for hand hygiene

Chlorhexidine digluconate (CHG) is commonly used in healthcare, e.g. in skin antiseptics, antimicrobial soaps, alcohol-based hand rubs and oral or wound antiseptics. Aim of the literature review was to evaluate the potential of bacteria to adapt to low level CHG exposure. A maximum 4fold MIC increase to CHG was found after low level exposure in most of the 71 evaluated bacterial species. A strong adaptive mostly stable MIC change was described in strains or isolates of the healthcare-associated species E. coli, S. marcescens and P. aeruginosa (up to 500fold, 128fold or 32fold, respectively). The highest MIC values after adaptation were 2,048 mg/l (S. marcescens) and 1,024 mg/l (P. aeruginosa). A new resistance to tetracycline, gentamicin, meropeneme or triclosan was found in some adapted isolates. In E. coli horizontal gene transfer was induced (sulfonamide resistance by conjugation), pointing out an additional risk of sublethal CHG. The use of CHG in patient care - but also all other settings such as consumer products and households - should therefore be critically assessed and restricted to indications with a proven health benefit or justifiable public health benefits. Additional CHG has no health benefit when used in alcohol-based hand rubs and is not recommended by the WHO. For routine hand washing of soiled hands the use of plain soap is sufficient, CHG in soaps has no health benefit. In surgical hand antisepsis alcohol-based hand rubs should be preferred to CHG soaps. Implementation of these principles will help to reduce avoidable selection pressure.


INTRODUCTION
Chlorhexidine digluconate (CHG) is a commonly used antiseptic agent in human healthcare and veterinary medicine, mainly used for hand hygiene (e.g. at 2% -4% as the only active agent in antiseptic soaps or at 0.5% or 1% as an additional active agent in alcohol-based hand rubs), in alcohol-based skin antiseptics at 2% and in mouth rinse solutions at 0.12% -0.2% [1]. The widespread CHG use in various types of applications has probably lead to an increase of acquired bacterial resistances, mainly in Gram-negative species such as Pseudomonas aeruginosa (minimal inhibitory concentration (MIC) of up to 800 mg/l), Serratia marcescens (MIC of up to 400 mg/l) or Klebsiella pneumoniae (MIC of up to 256 mg/l) [1]. In some types of applications such as skin antiseptics CHG has been shown to reduce healthcare associated infections, e.g. catheter-associated bloodstream infections [2]. Recent evidence also suggests a contribution to the prevention of surgical site infections [3] although the single effect of CHG for this application is still under controversial debate [4][5][6].
Its widespread use in hand hygiene by healthcare workers in many countries suggests to look specifically at all possible applications in this area. The WHO has published a recommendation on hand hygiene for healthcare in 2009 with the aim to reduce healthcare-associated infections [7]. Three types of applications can be distinguished. The use of alcohol-based hand rubs is recommended on clean hands in five specific clinical situations: before touching a patient, before clean or aseptic procedures, after body fluid exposure, after touching a patient and after touching patient surroundings [7,8]. Hand washing with either plain soap or antiseptic soap and water is recommended for visibly soiled hands or in case of contamination with spore-forming bacteria such as Clostridium difficile [7]. The third type of application is in the surgical theater. Healthcare workers should decontaminate their hands prior to donning sterile surgical gloves with either alcohol-based hand rubs (surgical hand disinfection) or with antimicrobial soaps (surgical scrubbing) [7].
In the past years there is an increasing concern on the development of resistance not only to antibiotics but also to antiseptic agents which are essential to limit the spread of multidrug-resistant pathogens in healthcare [9,10]. Some antiseptic agents are more likely than others to cause a bacterial tolerance or even resistance [11]. Aim of the review is therefore to evaluate the potential of CHG to cause an adaptive bacterial response during exposure to sublethal concentrations and to propose reasonable implications for the use of CHG in hand hygiene.

General remark
The magnitude of any adaptive response to CHG is expressed as an MIC change and assigned to one of the following three categories: No adaptive response (no MIC increase), weak adaptive response (MIC increase ≤ 4fold) and strong adaptive response (MIC increase > 4fold). For some bacterial species two or more studies were found resulting in data from various isolates or strains. That is why some bacterial species can be found in two or three categories depending on the results obtained with the various isolates or strains of the same species.  Table 1).

Adaptive bacterial response in Gram-negative species
A strong but unstable MIC change (> 4fold) was found in isolates or strains of four species (Burkholderia cepacia, E. coli, Salmonella enteritidis, Salmonella Typhimurium). A strong and stable MIC change (> 4fold) was described for isolates or strains of seven species (E. coli, K. pneumoniae, P. aeruginosa, Salmonella Virchow, Salmonella spp., S. marcescens, Stenotrophomonas maltophilia). In isolates or strains of six species (Acinetobacter baylyi, Acinetobacter proteolyticus, E. coli, Pseudomonas spp., Ralstonia spp., S. marcescens) the adaptive response was strong but its stability was not described.
Cross resistance to various antibiotics such as tetracycline, gentamicin or meropeneme was found in some isolates of Bacterioides fragilis, B. cepacia complex and Salmonella spp.. In addition, a lower susceptibility to other biocidal agents was described for E. coli and S. Virchow to triclosan, for A. baylyi to hydrogen peroxide and for S. Typhimurium to benzalkonium chloride (BAC).
Other adaptive changes include a significant upregulation of efflux pump genes in B. fragilis and B. cepacia complex. Horizontal gene transfer (sulfonamide resistance by conjugation) was induced in E. coli. VanA-type vancomycin resistance gene expression was increased vanA Enterococcus faecium (≥ 10fold increase of vanHAX encoding). Enhanced biofilm formation was described for K. pneumoniae and S. marcescens, adherence to polyethylene was increased in S. marcescens. Biofilm formation was decreased in B. cepacia. → Cross resistance to tetracycline, gentamicin, meropeneme or triclosan was found in some isolates.

Adaptive bacterial response in Gram-positive species
→ Horizontal gene transfer (sulfonamide resistance by conjugation) was induced in E. coli.
→ The use of CHG in patient care and other settings such as consumer products and households should be restricted to indications with a proven health benefit. Cross-adaptation* to BAC (2fold -100fold; 2 strains), triclosan (4fold; 1 strain) and didecyldimethyl-ammonium bromide (16fold; 1 strain); crossresistance* to cefotaxime and ceftazidime (2 strains each), sulfamethoxazole, ampicillin and tetracycline (1 strain each) [43] OPEN ACCESS | www.microbialcell.com   Table 2). Some maximum MIC values are above the proposed epidemiological cut-off value (8 mg/l for S. aureus) and some below (64 mg/l for E. faecalis) to determine CHG resistance in Gram-positive bacterial species [12]. *spiral gradient endpoint method; **disc diffusion method; ***broth microdilution; ****macrodilution method OPEN ACCESS | www.microbialcell.com Cross tolerance to various antibiotics such as tetracycline, gentamicin or meropeneme could be found in some isolates of S. aureus. In E. faecium (vancomycin-resistant enterococcus; VRE) a more than 10fold vanA up-regulation was detected as well as reduced daptomycin susceptibility. An increase in biofilm formation was described in S. epidermidis.

DISCUSSION
The strongest adaptation to low level CHG exposure was found in common nosocomial pathogens such as E. coli (up to 500fold MIC increase), S. marcescens (up to 128fold MIC increase), P. aeruginosa (up to 32fold MIC increase) and K. pneumoniae (up to 16fold MIC increase). After sublethal exposure the highest MIC values were also found in common nosocomial pathogens such as S. marcescens (2,048 mg/l), P. aeruginosa (1,024 mg/l), K. pneumoniae (> 512 mg/l) and E. coli (500 mg/l), It is probably no coincidence that these pathogens are among those species considered to have extreme or even pan resistance to antibiotics [13].
Low level CHG exposure also reduced the susceptibility to selected antibiotics in Burkholderia spp. or Salmonella spp. In Burkholderia spp. an up-regulation of transporter and efflux pump genes was found. Efflux pumps are often not agent-specific and may well result in resistance to other biocidal agents or antibiotics [1]. A quite alarming finding was that horizontal gene transfer was induced in E. coli by low level CHG exposure enabling the faster spread of resistance genes within the bacterial community.
Some mechanisms of the adaptive response have been described. Increased expression of efflux pumps is recognized as a mechanism of antibiotic and biocide resistance. The pumps may have limited or broad substrates, the socalled multiple drug resistance pumps [14]. The multiple antibiotic resistance (mar) locus and mar regulon in E. coli and other members of the enterobacteriaceae is a paradigm for a generalized response locus leading to increased expression of efflux pumps. One such pump, the AcrAB pump, extrudes biocides such as triclosan, chlorhexidine and quaternary ammonium compounds as well as multiple antibiotics [14]. In P. aeruginosa, a number of multidrug efflux pumps export a broad range of substrates [14]. In C. jejuni and C. coli active efflux was identified in adapted strains. In addition, the outer membrane protein profiles had changed, along with morphological changes [15]. In K. pneumoniae CHG adaptation was associated with mutations in the two-component regulator phoPQ and a putative Tet repressor gene (smvR) adjacent to the major facilitator superfamily (MFS) efflux pump gene, smvA [16]. And in Salmonella spp. a defense network was described that involved multiple cell targets including those associated with the synthesis and modification of the cell wall, the SOS response, virulence, and a shift in cellular metabolism toward anoxic pathways. In addition, results indicated that CHG tolerance was associated with more extensive modifications of the same cellular processes involved in this proposed network, as well as a divergent defense response involving the up-regulation of additional targets such as the flagellar apparatus and an altered cellular phosphate metabolism [17].
A major limitation of this review is that most of the data were obtained in laboratories under defined conditions. The findings are certainly suitable to describe the potential for adaptation to CHG. But it is less clear if or how the findings are transferred to the clinic. In 2002 Block et al. described that the MIC for CHG was higher among clinical isolates when more CHG was used for any type of application [18]. A similar correlation between CHG usage and MIC values was described in 2018 with S. aureus [19]. Lindford et al. described an outbreak by MDR A. baumannii in a burn unit. One of the measures to finally control the outbreak was to reduce moist low-concentration CHG dressings on burn wounds [20]. And yet the clinical impact of an elevated MIC value remains under controversial debate [21]. In hand hygiene it is known that a low bactericidal effect of CHG on the skin can only be achieved in the presence of small volumes of water, the water released by the skin as transepidermal water loss does not seem to be sufficient [22]. If the water realised by the skin is sufficient to allow adaptive changes of the bacterial species on the skin is currently not know. And yet, the triclosan tale strongly suggested that "a chemical that constantly stresses bacteria to adapt, and behaviour that promotes antibiotic resistance needs to be stopped immediately when the benefits are null" [10]. CHG is obviously such a chemical that constantly stresses bacteria to adapt. Even if the clinical impact of isolates or strains with elevated MIC values cannot finally be evaluated at the moment it seems justified restricting the use of CHG to applications where health benefits are associated with its use.

Alcohol-based hand rubs
In alcohol-based hand rubs with additional CHG used for hygienic hand disinfection there is no sound evidence for an additional effect of CHG in vitro [23]. There is also no evidence on the prevention of any type of healthcareassociated infection by the additional CHG in hand rubs. But there are obvious risks such as acquired bacterial resistance, anaphylactic reactions or skin irritation [24]. Its use in the immediate patient environment may therefore contribute to the selection pressure especially when the CHG concentration is sublethal [20]. Additional biocidal agents in alcohol-based hand rubs such as CHG are not recommended by the WHO [7].
The same applies to hand rubs used for surgical hand disinfection [24]. For surgical hand disinfection additional biocidal agents such as CHG are not recommended because they do not contribute to the prevention of surgical site infections [3,25]. Replacing hand rubs with additional CHG by hand rubs without CHG will help to reduce avoidable CHG selection pressure. They should, however, have an equivalent efficacy, dermal tolerance and user acceptability [26].

Antimicrobial soaps
Another simple option to reduce CHG selection pressure is to ban CHG soaps in healthcare for regular hand washing. Based on the WHO recommendation for hand hygiene from 2009 hand washing is recommended to wash hands when they are visibly soiled. The use of plain soap, however, is adequate, there is no health benefit for antimicrobial soaps [7]. Another possible use of antimicrobial soaps is prior to surgery. Surgical scrubbing usually lasts for 6-10 min of scrubbing time and consumes between 5 and 20 l water per scrub [27][28][29]. Surgical scrub products may only be effective with additional post-scrub water-based CHG treatments of the hands which pose an additional contamination and selection pressure risk [30,31]. Alcoholbased hand rubs with an appropriate concentration of alcohol(s) have a stronger effect on the resident hand flora, require typically 1.5 min for application, cause less skin irritation [32] and do not pose any selection pressure to bacterial species due to their volatility [33,34].

CONCLUSION
Overall, the evidence on the adaptive potential of various pathogens to low level CHG exposure strongly suggests to critically review the use of CHG in patient care and to eliminate it in all applications where no health benefit has been shown or is realistically expectable. In addition, studies deemed suitable for this review were also included. Publications were included and results were extracted from them when they provided original data on any type of adaptive response to the exposure of bacteria to sublethal concentrations of CHG, corresponding changes of MICs (CHG, antibiotics, and other biocidal agents), survival in CHG solutions, efflux pump activity, gene expression or biofilm formation. Articles were excluded when they described only data on fungi, outbreaks, pseudo-outbreaks or infections caused by contaminated CHG products or solutions, only biochemical changes, an adaptive effect with other chlorhexidine salts or when a CHG solution or product was used for disinfection during an outbreak but without being the suspected or proven source. Reviews were also excluded and screened for any original information within the scope of the review.

METHODS
The susceptibility of isolates or strains to CHG is described as the minimum inhibitory concentration (MIC value). In most studies it was described as a single value and is presented as such unless stated otherwise. The magnitude of any adaptive response to CHG is expressed as an MIC change and assigned to one of the following three categories: no adaptive response (no MIC increase), weak adaptive response (MIC increase ≤ 4fold) and strong adaptive response (MIC increase > 4fold).

SUPPLEMENTAL MATERIAL
All supplemental data for this article are available online at www.microbialcell.com. *broth microdilution; **macrodilution method