|Home | About | Journals | Submit | Contact Us | Français|
The etiology of nosocomial infections, the frequency of contaminated hands with the different nosocomial pathogens, and the role of health care workers' hands during outbreaks suggest that a hand hygiene preparation should at least have activity against bacteria, yeasts, and coated viruses. The importance of efficacy in choosing the right hand hygiene product is reflected in the new Centers for Disease Control and Prevention guideline on hand hygiene (J. M. Boyce and D. Pittet, Morb. Mortal. Wkly. Rep. 51:1-45, 2002). The best antimicrobial efficacy can be achieved with ethanol (60 to 85%), isopropanol (60 to 80%), and n-propanol (60 to 80%). The activity is broad and immediate. Ethanol at high concentrations (e.g., 95%) is the most effective treatment against naked viruses, whereas n-propanol seems to be more effective against the resident bacterial flora. The combination of alcohols may have a synergistic effect. The antimicrobial efficacy of chlorhexidine (2 to 4%) and triclosan (1 to 2%) is both lower and slower. Additionally, both agents have a risk of bacterial resistance, which is higher for chlorhexidine than triclosan. Their activity is often supported by the mechanical removal of pathogens during hand washing. Taking the antimicrobial efficacy and the mechanical removal together, they are still less effective than the alcohols. Plain soap and water has the lowest efficacy of all. In the new Centers for Disease Control and Prevention guideline, promotion of alcohol-based hand rubs containing various emollients instead of irritating soaps and detergents is one strategy to reduce skin damage, dryness, and irritation. Irritant contact dermatitis is highest with preparations containing 4% chlorhexidine gluconate, less frequent with nonantimicrobial soaps and preparations containing lower concentrations of chlorhexidine gluconate, and lowest with well-formulated alcohol-based hand rubs containing emollients and other skin conditioners. Too few published data from comparative trials are available to reliably rank triclosan. Personnel should be reminded that it is neither necessary nor recommended to routinely wash hands after each application of an alcohol-based hand rub. Long-lasting improvement of compliance with hand hygiene protocols can be successful if an effective and accessible alcohol-based hand rub with a proven dermal tolerance and an excellent user acceptability is supplied, accompanied by education of health care workers and promotion of the use of the product.
Nosocomial infections (NIs) remain a major global concern. Approximately 2 million NIs occur annually in the United States (232). Overall national prevalence rates have been described as ranging between 3.5 and 9.9% (160), but they vary significantly between departments, patient groups, types of surgical procedures, and the use of indwelling medical devices, etc. (20, 162). The most common NIs are urinary tract infections, lower respiratory tract infections, surgical-site infections, and primary septicemia (27, 159, 528, 532). They lead to additional days of treatment (146, 232, 411, 431, 605), increase the risk of death (27, 157), and increase treatment costs (217, 232, 234, 414, 431, 440, 460, 489, 605). The overall financial burden incurred by NIs has been estimated to be $4.5 billion per year in the United States alone (232). Approximately one-third of all NIs are regarded as preventable (193).
In 2002, a new Centers for Disease Control and Prevention (CDC) guideline for hand hygiene in health care settings, entitled Recommendations of the Healthcare Infection Control Practices Advisory Committee and the HICPAC/SHEA/APIC/IDSA Hand Hygiene Task Force, was published (71). It provides health care workers with a review of data on hand washing and hand antisepsis in health care settings and provides specific recommendations to promote improved hand hygiene practices and reduce the transmission of pathogenic microorganisms to patients and personnel in health care settings. As a clinical guideline, its chief aim is to reduce the incidence of NIs by providing detailed recommendations on two main aspects of hand hygiene: (i) choice of the most appropriate agents for hand hygiene in terms of efficacy and dermal tolerance and (ii) different strategies to improve compliance in hand hygiene, including hand hygiene practices among health care workers, behavioral theories, and methods for reducing adverse effects of agents. Our review is intended to support the CDC guideline by presenting specific additional aspects of the various agents, such as a broader evaluation of the in vitro and in vivo efficacy in various test models and their mode of action, resistance potential, and effect on compliance in hand hygiene.
Hand hygiene has been considered to be the most important tool in NI control (403, 462) ever since Semmelweis observed its immense effect on the incidence of childbed fever (473). Health care workers have three opportunities for the postcontamination treatment of hands: (i) the social hand wash, which is the cleaning of hands with plain, nonmedicated bar or liquid soap and water for removal of dirt, soil, and various organic substances; (ii) the hygienic (Europe) or antiseptic (United States) hand wash, which is the cleaning of hands with antimicrobial or medicated soap and water (“scrub”); most antimicrobial soaps contain a single active agent and are usually available as liquid preparations; and (iii) the hygienic hand disinfection (Europe), which normally consists of the application of an alcohol-based hand rub into dry hands without water.
For the preoperative treatment of hands two options are available: (i) the surgical hand wash (Europe) or surgical hand scrub (United States) which is the cleaning of hands with antimicrobial soap and water; and (ii) the surgical hand disinfection (Europe), which is the application of an alcohol-based hand rub into dry hands without water.
Three main types of preparations can be used for the different procedures of hand hygiene. (i) The first is plain, nonmedicated soap (social hand wash). (ii) The second is medicated soap (antiseptic and surgical hand wash). The most commonly used agent is chlorhexidine, usually at a concentration of 4 or 2%. Triclosan can also be found in medicated soaps, usually at a concentration of 1%. Hexachlorophene has now been banned worldwide because of its high rate of dermal absorption and subsequent toxic effects, especially among newborns (84, 98). Levels of 0.1 to 0.6 ppm in blood were found among health care workers who regularly used a 3% hexachlorophene preparation for hand washing (323). These findings speak strongly against the topical use of this active agent. The Food and Drug Administration classifies this agent as not being generally recognized as safe and effective for use as an antiseptic hand wash (21). Hexachlorophene is therefore not included in this review. Other active agents such as povidone iodine have rarely been used for the postcontamination treatment of hands and therefore are also not addressed in this review. (iii) The final type is the alcohol-based hand rub (hygienic and surgical hand disinfection). This is a leave-on preparation and this applied to the skin without the use of water.
In addition, non-alcohol-based waterless antiseptic agents are available for use by health care workers. Some of these contain quaternary ammonium-type compounds. They were not discussed in the CDC hand hygiene guideline because there was insufficient evidence at the time to promote their use; therefore, they are not further evaluated here.
This review provides an in-depth comparison of the several options for hand hygiene, with the aim of further supporting the CDC guideline on hand hygiene.
Three principal types of skin flora have been described. The resident and transient flora were already distinguished in 1938 (447, 470). In addition, the infectious flora was described, with species such as Staphylococcus aureus or beta-hemolytic streptococci, which are frequently isolated from abscesses, whitlows, paronychia, or infected eczema (475).
The resident flora consists of permanent inhabitants of the skin. They are found mainly on the surface of the skin and under the superficial cells of the stratum corneum (379). These bacteria are not regarded as pathogens on intact skin but may cause infections in sterile body cavities, in the eyes, or on nonintact skin (292). Resident skin bacteria survive longer on intact skin than do gram-negative transient species (325). The protective function of the resident flora, so-called colonization resistance, has been demonstrated in various in vitro and in vivo studies. Its purpose is twofold: microbial antagonism and the competition for nutrients in the ecosystem (12). Nevertheless, the interactions between bacteria and fungi on the skin are still inadequately understood. Many such interactions have been demonstrated experimentally. Their contribution—which is thought to be a major mechanism of preventing the adherence of pathogens—to the stability of the dermal ecosystem, however, remains unclear (375).
The dominant species is Staphylococcus epidermidis, which is found on almost every hand (311, 454, 522). The incidence of oxacillin resistance among isolates of S. epidermidis is up to 64.3% (311) and is higher among health care workers who have direct contact with patients than in those who do not (522). Other regular residents are Staphylococcus hominis and other coagulase-negative staphylococci, followed by coryneform bacteria such as propionibacteria, corynebacteria, dermabacteria, and micrococci (137, 315, 401). Among fungi, the most important genus of the resident skin flora is Pityrosporum (Malassezia) (201). Viruses are usually not resident on the skin but can proliferate within the living epidermis, where they may induce pathological changes (361).
Total counts of bacteria on the hands of medical staff have ranged from 3.9 × 104 to 4.6 × 106 (294, 309, 338, 447). Their number increases with the duration of clinical activities, on average by 16 cells per min (438). Some clinical situations are associated with a higher bacterial load on the hands of health care workers: direct contact with patients, respiratory tract care, contact with body fluids, and after being interrupted while caring for a patient (438). In general, however, it is difficult to clearly assign a specific risk of hand contamination to certain patient care activities. Nurses can contaminate their hands with 100 to 1,000 CFU of Klebsiella spp. during “clean activities” (81), while 10 to 600 CFU/ml can be found on nurses' hands after touching the groins of patients heavily contaminated with Proteus mirabilis (129). In intensive care units (ICU), the number of direct contacts between the hands of the health care workers and the patients is particularly high, leading to a higher risk of NI (148).
The transient skin flora consists of bacteria, fungi, and viruses that may be found on the skin only at times (447). They usually do not multiply on the skin, but they survive and occasionally multiply and cause disease (15). They may come from patients or inanimate surfaces. Between 4 and 16% of the hand surface is exposed by a single direct contact, and after 12 direct contacts, up to 40% of the hand surface may have been touched (74). The transmissibility of transient bacteria depends on the species, the number of bacteria on the hand, their survival on skin, and the dermal water content (230, 344, 418).
In addition, there is the temporary resident skin flora, which persists and multiplies for a limited period on the skin. The definition is more or less identical to that of transient skin flora, because the duration of residence on human skin is uncertain and variable but never permanent (5). In addition, the temporary resident skin flora often includes nosocomial bacteria and fungi (5, 201, 399, 400).
S. aureus is the most common gram-positive bacterium causing NIs (353, 533). Its frequency among all pathogens in NIs varies between 11.1 and 17.2% (265, 484, 493, 583). Methicillin resistance in S. aureus (MRSA) is increasing worldwide (113, 503, 578), leading not only to NIs but recently also to community-acquired infection. In 139 ICUs in Germany, 14.3% of all 1,535 NIs due to S. aureus have been caused by MRSA. This proportion is highest for urinary tract infections (26.4%), followed by primary septicemia (23.3%), and lower respiratory tract infection (12.9%) (161). The most common type of NI caused by S. aureus is the surgical-site infection (245, 259, 422).
Enterococcus spp. are isolated in up to 14.8% of patients with NI (484). The most common species are Enterococcus faecium and E. faecalis (385), which frequently cause urinary tract infections (533). The emergence of vancomycin resistance among enterococci (VRE) has led to an increased recognition of cross-transmission of VRE, including the role of health care workers' hands (29, 347).
Coagulase-negative staphylococci, such as S. epidermidis, mainly cause catheter-associated primary bloodstream infections. In ICUs, approximately one-third of all blood culture isolates from patients with nosocomial bloodstream infections were found to be coagulase-negative staphylococci (463, 533).
Colonization of health care workers' hands with S. aureus has been described to range between 10.5 and 78.3% (Table (Table1).1). Up to 24,000,000 cells can be found per hand (33). The colonization rate with S. aureus was higher among doctors (36%) than among nurses (18%), as was the bacterial density of S. aureus on the hands (21 and 5%, respectively, with more than 1,000 CFU per hand) (101). The carrier rate may be up to 28% if the health care worker contacts patients with an atopic dermatitis which is colonized by S. aureus (608, 609). MRSA has been isolated from the hands of up to 16.9% of health care workers. VRE can be found on the hands of up to 41% of health care workers (Table (Table11).
Hand carriage of pathogens such as S. aureus, MRSA, or S. epidermidis has repeatedly been associated with different types of NI (Table (Table2)2) (212, 455). The analysis of outbreaks revealed that dermatitis on the hands of health care workers was a risk factor for colonization or for inadequate hand hygiene, resulting in various types of NI (Table (Table22).
Transmissibility of VRE has also been demonstrated. The hands and gloves of 44 health care workers were sampled after care of VRE-positive patients. Gloves were VRE positive for 17 of 44 healthcare workers, and hands were positive for 5 of 44, even though they had worn gloves (553). One health care worker was even VRE positive on the hands although the culture from the glove was negative (553).
S. aureus can survive on hands for at least 150 min; VRE survives on hands or gloves for up to 60 min (Table (Table1).1). On inanimate surfaces, S. aureus and MRSA may survive for 7 months, with wild strains surviving longer than laboratory strains (Table (Table1).1). VRE may survive on surfaces for 4 months. The long survival on surfaces, together with the relatively short survival on hands, suggests that contaminated surfaces may well be the source of transient colonization despite negative hand cultures.
Escherichia coli is the most common gram-negative bacterium, causing mainly urinary tract infections (265, 463). Pseudomonas aeruginosa is also very common, chiefly causing lower respiratory tract infections (265, 463). In the majority of cases, both types of infection are device associated (364, 463, 531) and are often found among patients in ICUs (260). Manual handling of devices such as urinary catheters, ventilation equipment, and suction tubes emphasizes the importance of the hands of health care workers in possible cross-transmission of gram-negative bacteria. Overall, gram-negative bacteria are found in up to 64% of all NIs (463).
Colonization rates of gram-negative bacteria on the hands of health care workers have been described as ranging from 21 to 86.1% (Table (Table1),1), with the highest rate being found in ICUs (271). The number of gram-negative bacteria per hand may be as large as 13,000,000 cells (33). The colonization may be long-lasting (302). Even in nursing homes, a rate of 76% has been described for nurses hands (610). Colonization with gram-negative bacteria is influenced by various factors. For example, it is higher before patient contact than after the work shift (187). Hands with artificial fingernails harbor gram-negative bacteria more often than those without (207). Higher colonization rates with gram-negative bacteria also occur during periods of higher ambient temperature and high air humidity (358).
Different species of gram-negative bacteria exhibit different colonization rates. For instance, the colonization rate is 3 to 15% for Acinetobacter baumannii, 1.3 to 25% for Pseudomonas spp., and 15.4 to 24% for Serratia marcescens (Table (Table1).1). Klebsiella spp. were found on the hands of 17% of the ICU staff sampled, with up to 10,000 bacteria per hand (81). Artificial fingernails have been associated with a higher risk for colonization with P. aeruginosa (144).
Transient hand carriage of various gram-negative bacterial species has quite often been suspected to be responsible for cross-transmission during outbreaks resulting in various types of NI (155, 426, 514, 571). Most reports of cross-transmission of specific gram-negative bacteria come from critical-care areas, such as neonatal ICUs and burn units. Contaminated hands (Table (Table1),1), brushes, contaminated plain soap, and contaminated antiseptic soap have been associated with various types of NI, which were quite often caused by S. marcescens (Table (Table22).
Most gram-negative bacteria survive on the hands for 1 h or more. Survival on inanimate surfaces has been reported to be different for the different gram-negative species, with most of them surviving for many months (Table (Table1).1). In general, gram-negative bacteria survive for longer on inanimate surfaces than on human skin (151).
The main spore-forming bacterium causing NIs is Clostridium difficile. It is estimated that between 15 and 55% of all cases of nosocomial antibiotic-associated diarrhea are caused by C. difficile (40, 374, 567, 613). Patients with diarrhea caused by C. difficile have on average 3.6 additional hospital days attributable to the NI, which in the United States costs approximately $3,669 per case or $1.1 billion per year (289). The overall mortality is 15% (381). Extraintestinal manifestations are very uncommon (≤1%) (156). Patients can be contaminated from, for instance, the hands of hospital personnel and from inanimate surfaces (40).
In one study, the hands of 59% of 35 health care workers were C. difficile positive after direct contact with culture-positive patients. Colonization was found mainly in the subungual area (43%), on the fingertips (37%), on the palm (37%), and under rings (20%) (362). In another study, 14% of 73 health care worker were culture positive for C. difficile on their hands. The presence of C. difficile on the hands correlated with the density of environmental contamination (491). During a third outbreak, caused by Bacillus cereus in a neonatal ICU, 11 (37%) of 30 fingerprints from health care workers were positive for Bacillus spp. (569).
Transmission of C. difficile in an endemic setting on a general medical ward has been shown to occur in 21% of patients, with 37% of them suffering from diarrhea (362). An outbreak of necrotizing enterocolitis among neonates was associated with clostridial hand carriage in four of seven health care workers (173). Another spore-forming bacterium has been described as well: B. cereus was transmitted to the umbilicus in 49% of newborns on a maternity ward; the hands of 15% of the health care workers were found to be culture positive (62).
Vegetative cells of C. difficile can survive for at least 24 h on inanimate surfaces, and spores survive for up to 5 months (Table (Table11).
Fungi are less commonly found than bacteria as the causative agent of NIs, but their frequency and importance are increasing (216, 502, 527). In Germany and New Zealand, 6% of all NIs were caused by fungi (397, 484). In Spain, the overall rate was found to be 2.4% in 1990 and 3.2% in 1999, indicating a higher clinical relevance for NIs in the more recent study (26). In the United States, an increase in isolation of yeasts from 7.6 to 10.6% has been noted over a period of 10 years in patients with NIs (593). The most important fungus with respect to NIs is Candida albicans. Fungi may cause septicemia, urinary tract infections, or surgical-site infections (463, 500). Device-associated bloodstream infections caused by Candida spp. have become more common among critically ill patients in the last decades (89, 128, 163, 342); the contribution of non-albicans Candida spp. is increasingly significant (216). It has also been reported that 21% of all urinary tract infections among ICU patients are caused by C. albicans (463).
In an ICU, 67 (46%) of the hands of 146 health care workers were colonized with a yeast. The most common species were Candida and Rhodotorula spp. Respiratory therapists were found to have the highest colonization rate (69%) (221). In another study of nurses and other hospital staff, 75% of the nurses and 81% of the other hospital staff were colonized with a yeast (541). In a long-term-care facility, 41% of 42 health care workers were found to have Candida spp. on their hands (378). Yeasts quite often also colonize artificial fingernails (207). Acquisition of C. albicans on the hands of health care workers immediately after attending systemically infected patients was reported to occur in 2 of 17 nurses (79).
Only a few studies are found in the literature which demonstrate the role of hands in cross-transmission (Table (Table2),2), sometimes despite negative hand cultures (572). The analysis of an outbreak revealed that caring for a patient who is colonized with Candida parapsilosis can lead to positive hand cultures and finally to severe infections or colonization among patients (501). The transmissibility of yeasts from hand to hand is high (Table (Table33).
On fingertips, only 20% of viable cells of C. albicans and C. parapsilosis remain detectable after 1 h (79, 564). Candida spp. can survive on surfaces for up to 150 days (452, 564). During this period of survival, most yeast cells die within the first few minutes (452).
Viruses account for approximately 5% of all NIs. On pediatric wards, the proportion is higher at 23% (6). Five main groups of viruses have been identified with respect to their nosocomial transmission: blood-borne viruses (e.g., hepatitis B virus [HBV], hepatitis C virus [HCV], and human immunodeficiency virus [HIV]), respiratory route viruses (e.g., respiratory syncytial virus [RSV], influenza virus, rhinovirus, coronavirus, and adenovirus), fecal-oral route viruses (e.g., rotavirus, small round structured viruses [noroviruses], enteroviruses, and hepatitis A virus [HAV]), herpesviruses obtained from direct contact with skin, mucous membranes, or wounds (e.g., herpes simplex viruses, varicella zoster virus, cytomegalovirus, and Epstein-Barr virus), and exotic viruses such as viral hemorrhagic fever viruses (Ebola virus, Marburg virus, Lassa fever virus, and Congo Crimean hemorrhagic fever virus) and rabies virus (8). The fingers, especially the pads and tips, are the most likely areas to come into contact with viruses while touching infected people and their bodily substances as well as other contaminated materials (499, 576).
The risk of direct contact with blood and thereby with blood-borne viruses is variable. In general, it must be assumed that a health care worker wears protective gloves if contact with blood is expected. However, there are still clinical situations in which contamination with blood is unexpected. Health care workers in invasive radiology have blood contact in 3% of clinical activities, surgeons have blood contact in 50%, and midwives have blood contact in 71% (48). Surgical gloves should protect from direct contact with blood, but perforations are found on average in 17% of gloves, which correlates with the detection of blood under surgical gloves in 13% of surgeons (392). Perforations in most gloves (83%) remain undetected by the surgeon (557). Up to 82.5% of protective gloves have invisible perforations (276). In an acute viremic state, HBV may be present in blood at a concentration of 5 × 108 IU per ml of blood (623). A 1-μl volume of blood, which is hardly visible on a hand, may still contain 500 IU of HBV. For HCV, a concentration of 104 to 107 IU was found in blood (105). Virus detection on the hands has been investigated in a few studies. In a dialysis unit, 23.8% of samples obtained from health care workers' hands were positive for HCV RNA after treatment of HCV-positive patients despite the use of standard precautions, whereas the rate was 8% after treatment of HCV-negative patients (11).
Viruses from the respiratory tract are often found on hands, e.g., rhinoviruses in up to 65% from persons with a common cold (191, 457). Adenovirus has been found on the hands of healthcare workers during outbreaks of keratoconjunctivitis (380) and was isolated from the hands of 46% of patients with epidemic keratoconjunctivitis (35), which emphasizes the potential of virus transfer to hospital personnel through casual hand contact. No data were available regarding the detection of severe acute respiratory syndrome (SARS) virus on hands during the outbreaks in Asia and Canada in 2003.
Rotaviruses can be found on the hands in up to 78.6% of individuals sampled (Table (Table1)1) and also on surfaces with frequent hand contact, e.g., TV sets, toys, and patient charts (9). At the peak of a bout of rotavirus gastroenteritis, every gram of feces may contain more than 107 to 108 infectious viral particles (590).
Cytomegalovirus has been isolated from the hands of day care workers (224), but exotic viruses such as hemorrhagic fever viruses have to date not been detected on health care workers' hands.
Hands play a major role especially in the transmission of blood-borne, fecal, and respiratory tract viruses. The transmission of some viruses from the hands of health care workers has been described (Table (Table2).2). In addition, transient hand carriage is associated with the transmission of many viruses, such as rhinovirus (99, 191), RSV (194, 488), astrovirus (136), and cytomegalovirus (109). For the SARS virus, a similar correlation has been described, since hand hygiene was found to be the second most effective measure to prevent cross-transmission of the SARS virus in a hospital (510). Most viruses are easily transmitted from hand to hand, food, or surfaces (Table (Table33).
Persistance of viruses on the hands has been investigated mainly for fecal and respiratory tract viruses. Artificial contamination of hands with HAV led to an immediate-recovery rate of 70.5% (59). HAV persisted for several hours on human hands (354, 355). With poliovirus, the immediate-recovery rate was 22% but the whole inoculum was recovered after 150 min, indicating an almost complete persistence of poliovirus on hands (505). Rotavirus has been described as persisting on hands for up to 260 min, with 57% recovery after 20 min, 42.6% recovery after 60 min, and 7.1% recovery after 260 min (22). It can be transferred from contaminated hands to clean hands, with 6.6% of the viral contamination transferred 20 min after contamination (Table (Table3),3), and 2.8% of the viral contamination transferred 60 min after contamination (22). Rotavirus has been described to persist better on hands than rhinovirus or parainfluenzavirus (24).
Many enveloped viruses such as influenza virus, parainfluenza virus (Table (Table1),1), and cytomegalovirus (139) may survive on the hands for 10 to 15 min or even up to 2 h (herpes simplex virus type 1 [Table [Table1]).1]). Adenoviruses have been described to persist on human skin for many hours (499).
Only a few studies of the persistence of viruses on surfaces have been performed. Rotavirus and HAV can persist for up to 60 days (Table (Table1)1) depending on the room temperature, air humidity, and type of surface (495). HIV remains infective on surfaces for up to 7 days, depending on the inoculum and the type of preparation (cell-associated virus or cell-free virus). HIV obtained from clinical specimens remains infective for a few days (568). Influenza A virus may persist on steel for up to 48 h; on other materials, such as paper or handkerchiefs, the virus persists for up to 12 h (46). Rhinovirus may persist for up to 7 days (Table (Table11).
The new CDC guideline on hand hygiene does not suggest a specific minimum spectrum of antimicrobial activity of a suitable hand hygiene agent (71). However, it can be derived from the etiology of NIs as well as the data on the skin flora of the hands of health care workers and their role in the transmission of nosocomial pathogens (Table (Table4).4). A procedure for the postcontamination treatment of hands must have at least bactericidal, fungicidal (yeasts), and virucidal (coated viruses) activity.
The spectrum of activity can be substantiated in suspension tests (474). In principle, suspension tests are suitable to substantiate the spectrum of antimicrobial activity (474). The suggested activity against coated viruses is based on the frequent contamination of health care workers' hands with blood during routine patient care and thereby possibly with blood-borne viruses, such as HCV or HIV, where neither patients nor health care workers can be protected by vaccination. The contamination of hands with blood may not be visible but may still be infective with HCV or HIV for the health care worker or the next patient (123). That is why activity against coated viruses should be included in the minimum spectrum of activity for an active agent for hand hygiene. Uncoated viruses, however, are usually spread from patients with infective gastroenteritis (e.g., caused by noroviruses or rotaviruses), upper and lower respiratory tract infections, or keratoconjunctivitis (e.g., caused by adenoviruses). These infections often have typical and visible symptoms. The activity against uncoated viruses can be restricted to a specific clinical area, e.g., in ophthalmology (adenovirus), pediatrics (rotavirus), or oncology (parvovirus) or to outbreaks of specific infectious diseases caused by uncoated viruses. Additional activity against the whole spectrum of fungi (including molds), mycobacteria, and bacterial spores may be relevant in special patient care situations (e.g., in bone-marrow transplant units) or during outbreaks. A procedure for the preoperative treatment of hands should be at least bactericidally and fungicidally (yeasts) effective, since the hands of most health care workers' hands carry yeasts and since surgical-site infections have also been associated with hand carriage of yeasts during an outbreak.
Normally, nonmedicated soaps are detergent-based products. Those based on esterified fatty acids and sodium or potassium hydroxide are less skin compatible (Table (Table5).5). They are available in various forms including bar soaps, tissue, leaflet, and liquid preparations. This cleaning activity can be attributed to the detergent properties.
Nonmedicated soaps do not contain any active ingredient with an antimicrobial activity apart from preservatives. That is why in vitro data on the antimicrobial activity of nonmedicated soap rarely exist. The first experiments with soft alkaline soap were carried out by Robert Koch. He found out that multiplication of the vegetative cells of Bacillus anthracis was completely (dilution of 1:1,000) or partly (dilution of 1:5,000) inhibited (273). A more recent study described a fungistatic effect of a tenside-based soap at dilutions between 1:64 and 1:1,000 against Trichosporon cutaneum. C. albicans. Trichophyton rubrum. Trichophyton schönleinii. Microsporum audouinii, and Microsporum canis (277). With one plain soap, even limited fungicidal activity was described and largely explained by the presence of preservatives (603).
The use of plain soap and water reduces the numbers of microorganisms and viruses by mechanical removal of loosely adherent microorganisms from the hands. Many studies are available which address the reduction of the transient hand flora. The most common type of artificial contamination of hands for test purposes in the United States is S. marcescens (21), whereas E. coli is the main contaminant used in Europe (115). Regarding the transient flora, a reduction between 0.5 and 2.8 log10 units can be found within 1 min for E. coli (Table (Table6).6). Other types of artificial contamination have been used as well, such as VRE, rotavirus, Klebsiella spp., or spores of Bacillus atrophaeus. A simple hand wash still leads to a mean reduction of up to 2.4 log10 units within 1 min (Table (Table7).7). There is basically no effect on resident hand flora after a 2-min hand wash; after a 5-min hand wash, a reduction of 0.4 log10 unit was found, and after 3 h of wearing gloves, no reduction at all was observed (Table (Table88).
The effect of a social hand wash “in real life” has also been studied. Among 224 healthy homemakers, a single hand wash had little impact on microbial counts (mean log counts before hand wash, 5.72 ± 0.99; mean log counts after hand wash, 5.69 ± 1.04) (307). In a study with 11 volunteers who washed their hands for 15 s with water alone 24 times per day for a total of 5 days, a slight increase of the bacterial counts was observed (mean log bacterial counts: prewash, 4.91 ± 0.46; postwash, 5.12 ± 0.44); when bar soap was used, a similar result was found (mean log bacterial counts: prewash, 4.81 ± 0.46; postwash, 5.07 ± 0.47) (299). Other authors, too, have found paradoxical increases in bacterial counts on the skin after hand washing with plain soap (299, 371, 611). In contrast, another study showed that a 5-min hand wash with regular bar soap reduced the resident hand flora by 0.33 log10 units (326). The use of a nonmedicated soap by a surgical nurse for the preoperative treatment of hands even led to eight cases of surgical-site infection after cardiac surgery, which underscores the limited efficacy of nonmedicated soap (226).
Some studies have examined only microorganisms that are left on the hands after a hand wash. Washing hands with soap and water has been described to be ineffective in eliminating adenovirus from the culture-positive hands of a physician and patients, indicating that mechanical removal was incomplete (235). Transient gram-negative bacteria remained on the hands of health care workers in 10 of 10 cases despite five successive hand washes with soap and water (187). Furthermore, transmission of gram-negative bacteria from hands has been shown to occur 11 of 12 cases when a simple hand wash is carried out (129).
One risk of using soap and water is the contamination of hands by the washing process per se. This has been reported for P. aeruginosa (143). A possible source is the sink itself, when splashes of contaminated water come in contact with the hand of the health care worker (119). The reason is that the microorganisms are not killed during the hand wash but only removed and distributed in the immediate surroundings of the person, including the clothes. Nonmedicated soaps may also become contaminated and lead to colonization of the hands of personnel and to NIs, e.g., with S. marcescens (492) or Serratia liquefaciens (183).
Although the data involving nonmedicated soap suggest that a simple hand wash has some effect on the transient hand flora, it must be borne in mind that, in reality, a simple hand wash often does not last longer than 10 s (121, 145, 176, 177, 180, 300, 334, 450, 552).
Each hand wash detrimentally alters the water-lipid layer of the superficial skin, resulting in a loss of various protective agents such as amino acids and antimicrobial protective factors. Regeneration of the protective film may be insufficient if many hand washes are carried out in a row. This may lead to damage of the barrier function of the stratum corneum by inhingement of intercellular putty substances. The transepidermal water loss (TEWL) increases, and the skin becomes more permeable for toxic agents. At the same time, the superficial skin cells dry out, resulting in dehiscence of the stratum corneum, initially on the microscopic level and in due course on the macroscopic level (280).
The incidence with which simple soaps and detergents affect the condition of the skin of health care workers' hands varies considerably (407). For years, natural soaps that have high pH values were thought to be more irritating to the skin than synthetic detergents with neutral or acidic pHs. However, subsequent studies have found that pH is less important than other product characteristics as a cause of skin irritation (200). In some studies, plain soaps have caused less skin irritation than synthetic detergents, while in others, plain bar soap caused greater skin irritation than did a synthetic antimicrobial-containing detergent (299, 565). Synthetic detergents also vary in their propensity to cause skin irritation (200, 407). The incidence of detergent-related-irritant contact dermatitis is affected by various factors: the concentration of the compound, the type of detergent (anionic, cationic, amphoteric, or nonionic) and its quantity, the refattening, the vehicle, the time of exposure, and area exposed (50, 133, 283, 565). For example, it has been shown in vivo that higher concentrations of sodium lauryl sulfate (a detergent) caused greater skin irritation than lower concentrations did (133). In addition, anionic detergents are known to cause greater skin irritation than amphoteric or nonionic detergents (565).
Another factor is the temperature of the water that is used for the hand wash. Hot water leads to greater skin irritation, as reflected by in vivo measurements of TEWL and in vitro measurements of the penetration of detergent through the skin (50, 133, 405). This is explained by an increased penetration of detergents into the epidermis (405). In addition, scaling of the skin is greater when hands are washed with hot water (50). Only skin hydration does not appear to be affected by higher water temperatures (50, 405).
Frequent hand washing induces irritative contact dermatitis (ICD) and dry skin (70, 275, 525, 611), which may become colonized with nosocomial pathogens. ICD can be found in 18.3% of nursing staff in hospitals and is a major occupational health concern (523). A single hand wash already significantly reduces the dermal sebum content; the reduction lasts for 1 h. Skin hydration drops at the same time (280). If hands are washed four times within 1 h, the skin does not recover to its normal state within this period (337). In a study with 52 volunteers who washed their hands 24 times per day for a total of 5 days, a significant increase of the TEWL was observed, indicating that the skin barrier function is impaired (299). The prevalence of ICD caused by hand washing with antimicrobial soaps (detergents) is related to the factors listed above (540). The hardness of water may also affect the incidence of ICD due to frequent hand washing (591).
In summary, plain soap has basically no antimicrobial activity. A simple hand wash can reduce transient bacteria by 0.5 to 3 log10 units but has no real effect on the resident hand flora. The dermal tolerance is rather poor (Table (Table99).
Chlorhexidine is a cationic biguanide (485) and was first established as an antimicrobial agent in 1954 (104). It exists as acetate (diacetate), gluconate, and hydrochloride salts (485). Chlorhexidine gluconate is commonly used either at 0.5 to 0.75% in aqueous solution or in some detergent preparations or at 2 to 4% in other detergent preparations (327, 328). Its activity is greatly reduced in the presence of organic matter (485), natural corks (321), and hand creams containing anionic emulsifying agents (586). Inactivation of chlorhexidine may result in contamination of solutions containing 0.1% chlorhexidine, e.g., with Pseudomonas spp. (78).
The main target is the bacterial cytoplasmic membrane (360, 464). After chlorhexidine has caused extensive damage to the cytoplasmic inner membrane, precipitation or coagulation of protein and nucleic acids occurs (487). Damage also occurs to the outer membrane in gram-negative bacteria and the cell wall in gram-positive cells (131, 142, 227, 228, 236). Chlorhexidine also damages the cytoplasmic membrane of yeasts (588) and prevents the outgrowth, but not the germination, of bacterial spores (511). If chlorhexidine is hydrolyzed, small amounts of carcinogenic para-chloraniline may develop (87); this chemical has been found even in manufactured chlorhexidine solutions (274). At temperatures above 70°C, chlorhexidine is not stable and may degrade to para-chloraniline (171). An upper limit for para-chloraniline has been set in the British Pharmacopoeia at 0.25 mg per 100 mg of chlorhexidine (17).
The antimicrobial activity of chlorhexidine is dependent on its concentration. At lower concentrations, chlorhexidine has a bacteriostatic effect against most gram-positive bacteria (e.g., at 1 μg/ml), many gram-negative bacteria (e.g., at 2 to 2.5 μg/ml) (100, 195), and bacterial spores (513). At chlorhexidine concentrations of 20 μg/ml or more, a bactericidal effect can be expected as well as activity against yeasts (487). The actual effective concentration against Burkholderia cepacia and S. aureus varies with different supplements from 0.004 to 0.4% (factor 100), and the actual killing time also varies with different supplements (phenylethanol or edetate disodium) from <15 min to >360 min (465). In most studies, concentrations for rapid inactivation are well in excess of MICs, e.g., for S. aureus (103), E. coli. Vibrio cholerae (237), and yeasts (214). When used in a liquid soap, chlorhexidine usually has a concentration of 4% and exhibits a bactericidal activity against various gram-negative (130) and gram-positive (249) bacteria. In some comparative studies using suspension tests chlorhexidine (4%) was found to be less effective against MRSA than against methicillin-susceptible S. aureus, which has raised concerns about the suitability of the active agent in the prevention of transmission of MRSA (93, 192, 249). This concern has been confirmed with enterococci. Against Enterococcus species and VRE, chlorhexidine (4%) was found to be essentially ineffective in suspension tests if neutralization of residual activity is excluded (247). In a comparison with a nonmedicated hand wash product, a chlorhexidine-based scrub yielded a lower reduction of different antibiotic-resistant test bacteria such as MRSA, VRE, or high-level gentamicin-resistant enterococci (175). Chlorhexidine has no sporicidal activity (513). The data on mycobactericidal activity are not unambiguous but do indicate the relevance of a threshold concentration of chlorhexidine. In one report, 4% chlorhexidine was described as having very good activity against Mycobacterium smegmatis (reduction of >6 log10 units within 1 min) (54), whereas another study with Mycobacterium tuberculosis suggested a low activity of 4% chlorhexidine (reduction of <3 log10 units within 1 min) (55). Chlorhexidine at 1.5% did not reveal sufficient activity against Mycobacterium bovis (56), and chlorhexidine at 0.5% had no activity against Mycobacterium avium. Mycobacterium kansasii, or M. tuberculosis within 120 min (466).
Against dermatophytes such as Trichophyton mentagrophytes, chlorhexidine (1.5%) has been described as having no activity (56).
Antiviral activity has been described as good against most enveloped viruses, such as HIV, cytomegalovirus, influenza virus, RSV, and herpes simplex virus (284, 441), but the virucidal activity of chlorhexidine against naked viruses such as rotavirus, adenovirus, or enteroviruses is low (391, 498).
In comparison to other active agents, chlorhexidine has been described to be less effective in vitro against various nosocomial pathogens than is benzalkonium chloride or povidone iodine (517).
Overall, chlorhexidine seems to have good residual activity (13, 34, 305, 328, 423, 468, 476), but the residual activity must be assessed with caution. It may be false positive due to insufficient neutralization of chlorhexidine in the test design, leading to bacteriostatic concentrations beyond the actual exposure time. Significant difficulties in effective neutralization in in vitro tests have been described, and may yield false-positive activity data for this active agent (246, 516, 517, 600). In addition, the clinical benefit of such a residual effect has never been shown.
A 1-min hand wash with soap containing 4% chlorhexidine has been reported to lead to a mean reduction of E. coli of 3.08 log10 units on artificially contaminated hands (478). In a study with 52 volunteers who washed their hands 24 times per day for a total of 5 days, a significant decrease in the number of resident skin bacteria was observed with a 4% chlorhexidine liquid soap (mean reduction of 0.76 log10 unit) compared to nonmedicated bar soap (mean increase of 0.21 log10 unit) and a povidone-iodine soap (mean reduction of 0.32 log10 unit) (299). Under practical conditions with hands artificially contaminated by MRSA, chlorhexidine-based liquid soap was equally effective as simple soap (188, 220). A similar result was reported after contamination of hands with S. aureus (577). A reduction of 2.1 to 3 log10 units was found on hands contaminated with Klebsiella spp. after a 20-s hand wash with a soap based on 4% chlorhexidine (81). If hands were contaminated with rotavirus and treated with chlorhexidine soap for 10 s, the number of test viruses was reduced by 86.9%, which was significantly lower than the reductions achieved with 70% ethanol (99.8%) and 70% isopropanol (99.8%) (23). Treatment with 4% chlorhexidine soap for 30 s on hands contaminated with rotavirus leads to a similar effect of only 0.27 to 0.5 log10 unit (47). Under practical conditions and in terms of removal rate from hands, the efficacy against bacterial spores (e.g., B. atrophaeus) of an antiseptic liquid soap based on chlorhexidine was similar to that of nonmedicated soap, indicating that within 10 s or 60 s, chlorhexidine does not exhibit a significant sporicidal activity (57, 594). The effect of 4% chlorhexidine on the resident hand flora was found to be a reduction of between 0.35 and 2.29 log10 units, depending on the application time (Table (Table88).
The in-use studies yield a heterogeneous picture of the efficacy of chlorhexidine. One of the first studies with chlorhexidine was performed in 1955. A hand cream containing 1% chlorhexidine was rubbed into dry hands and led to a substantial reduction is the number of resident skin bacteria after 30 min (386). In another clinical study, 74 health care workers evaluated plain soap and a liquid soap based on 4% chlorhexidine over 4 months in a neurosurgical unit and a vascular surgery ward. Overall hand contamination was found to be significantly lower after the use of plain soap (mean number of CFU, 125) than after the use of chlorhexidine (mean number of CFU, 150) (343). A hand wash with 4% chlorhexidine was reported to be more effective on the total bacterial count under clinical conditions than was a 1% triclosan hand wash (140). In a prospective crossover study over 4 months with plain soap and a 4% chlorhexidine soap among health care workers in two surgical units, plain soap was found to be significantly more effective than chlorhexidine in reducing bacterial counts from the hands of health care workers (343). After contamination of hands with Klebsiella spp., a 98% reduction was described in 19 of 23 experiments in which a soap based on 4% chlorhexidine was used (81); this is an almost 2 log10 unit reduction. Chlorhexidine failed to eliminate MRSA from the hands (140). In contrast, gram-negative bacteria were more likely to be eliminated after the use of chlorhexidine (140, 357, 573, 580). The mean resident flora of the hands of surgeons was reduced by a 3-min application of 4% chlorhexidine from 3.5 log10 units (preoperatively) to 3.15 log10 units (postoperatively) in operations lasting less than 2 h. It has been shown that for operations lasting more than 3 h, 4% chlorhexidine was unable to keep the resident skin bacteria below the baseline value (4.5 preoperatively and 5.2 postoperatively) (76).
The definition of chlorhexidine resistance is often based on a report from 1982 in which the MICs of chlorhexidine for 317 clinical isolates of P. aeruginosa were analyzed, leading to the suggestion that resistance to chlorhexidine should be reported if the MIC is ≥50 mg/liter (390).
Resistance to chlorhexidine among gram-positive bacterial species is rather uncommon. Among Streptococcus and Enterococcus species, no chlorhexidine resistance has been demonstrated (42, 231). However, gram-negative bacteria, such as E. coli (389), Proteus mirabilis (100, 536), Providencia stuartii (227, 228, 554), P. aeruginosa (390, 556), P. cepacia (348), and S. marcescens (291), have frequently been reported to be resistant to chlorhexidine. The frequency of resistance for the different species is variable. A total of 84.6% of clinical isolates of P. mirabilis must be considered resistant to chlorhexidine (536). Among other gram-negative bacteria, the rate is lower (42, 195). C. albicans was found to have a resistance rate of 10.5% (42, 231).
Acquired resistance to chlorhexidine has been reported to occur in S. aureus (249) and among many gram-negative bacteria (37, 38, 434) which were isolated after recurrent bladder washouts using 600 mg of chlorhexidine per liter (537, 538) or after addition of chlorhexidine to catheter bags for paraplegic patients (584). Some of the isolates were highly resistant, with chlorhexidine MICs of ≥500 mg/liter (538). The chlorhexidine resistance is quite clearly linked to hospital isolates only. A selection of 196 environmental gram-negative isolates did not reveal a resistance to chlorhexidine (147).
High chlorhexidine MICs correlate with poor reduction in the number of test bacteria in suspension tests, which highlights the potential hazard (555). The MIC may be as high as 1600 μg/ml and correlates well with a slow and insufficient bacterial reduction in suspension tests, as shown with strains of Providencia (539). The resistance may be single (83), but cross-resistance to other anti-infective agents can also occur. Among isolates of P. aeruginosa from industry and hospitals, an association between resistance to antibiotics and chlorhexidine has been described (290). The potential for cross-resistance between antiseptic agents and antibiotics must be given careful consideration (443). Various nonfermenting gram-negative bacteria which were isolated from blood cultures of oncology patients were inactivated only with >500 mg of chlorhexidine per liter (210).
Different mechanisms of resistance have been found. The acquired resistance is probably linked to the inner (227) or the outer (551) membrane of bacterial cells, the cell surface (131), or the cell wall (549). It may also be explained by the presence of plasmids which code for chlorhexidine resistance (269) and may therefore be transfered to other bacterial species (486, 619). A change in lipid content or a reduced adsorption of the antiseptic can be excluded as the main mechanism of resistance, as shown with isolates from urinary tract infections caused by P. mirabilis (554) and S. marcescens (410).
Recurrent exposure of bacteria to chlorhexidine may lead to adaptation and may enhance their resistance. This phenomenen was shown with S. marcescens. One example involves repeated exposure to various contact lens solutions containing between 0.001 and 0.006% chlorhexidine, which enabled S. marcescens to multiply in the disinfectant solution (154). Repeated exposure of P. aeruginosa to 5 mg of chlorhexidine per liter was shown to increase the MIC from <10 to 70 mg/liter within 6 days (556). A similar result was reported with Pseudomonas stutzeri, which became resistant (MIC, 50 mg/liter) after 12 days of exposure to chlorhexidine (550). Even with Streptococcus sanguis, a clear increase of the chlorhexidine MIC during permanent chlorhexidine exposure was observed (601). In general, higher exposures to chlorhexidine in hospitals were reported to be associated with higher rates of resistance (67). Recently, some isolates of P. aeruginosa. K. pneumoniae, and A. baumannii isolated from soap dispensers were reported to multiply in a 1:2 dilution of a 2% chlorhexidine liquid soap; ATCC strains of K. pneumoniae and A. baumannii multiplied only at higher dilutions (73). The latter report highlights the potential danger for the hospital.
Resistance to chlorhexidine may even result in nosocomial infections. Occasional outbreaks of NIs have been traced to contaminated solutions of chlorhexidine (345). There is one report that a 0.5% chlorhexidine solution which was used to disinfect plastic clamps for Hickman lines and was handled by health care workers who transmitted the adapted bacteria to intravenous lines led to 12 cases of bacteremia with three fatalities (357). In another outbreak, contamination of a disinfectant solution with Burkholderia multivorans led to nine cases of surgical site infection (45). Especially when chlorhexidine resistance is endemic in gram-negative bacteria, the use of chlorhexidine-based hand antiseptics may lead to an increase of NIs by the chlorhexidine-resistant species (100).
Chlorhexidine gluconate is among the most common antiseptics causing ICD (540). However, the frequency of hand dermatitis associated with chlorhexidine-containing detergents is concentration dependent; products containing 4% chlorhexidine cause dermatitis much more frequently than do those containing lower concentrations (540). However, even preparations with the same concentration of chlorhexidine (4%) may cause skin irritation at different frequencies (398, 508). The differences are presumably due to other components of the various formulations. The relatively large number of reports of dermatitis related to chlorhexidine gluconate was partly explained by the fact that it was one of the most widely used antiseptics. In a survey of over 400 nurses working in several hospitals, detergents containing chlorhexidine were reported to cause skin damage less frequently than was nonantimicrobial soap or other detergents containing antimicrobial agents (298). In one 5-day prospective clinical trial, a detergent containing 4% chlorhexidine gluconate caused less irritation than did plain bar soap (300). Nonetheless, dry skin may occur with repeated exposure to preparations containing 4% chlorhexidine gluconate (339, 398).
The potential for contact allergy has been studied as well. Among eczema patients, 5.4% were found to have a positive skin reaction after a single patch test with 1% chlorhexidine, indicating the presence of an allergic contact dermatitis. Repeated exposure resulted in a sensitization rate of ca. 50% (310). In another study, 15 (2.5%) of 551 patients showed a strong and obviously relevant skin reaction in a single patch test with 1% chlorhexidine (415). Although these studies were carried out with patients and not with health care workers, the results nevertheless indicate the potential for sensitization and allergic contact dermatitis during frequent use. Allergic reactions to the use of detergents containing chlorhexidine gluconate on intact skin have been reported and can be severe, including dyspnea and anaphylactic shock (30, 92, 124, 138, 158, 270, 409, 425, 430, 468, 526, 563). Some cases of contact urticaria have also occurred as a result of chlorhexidine use (141, 617).
In summary, chlorhexidine (2 to 4%) has good activity against most vegetative bacteria, yeasts, and enveloped viruses but limited activity against mycobacteria, dermatophytes, and naked viruses. It has a moderate potential for acquired bacterial resistance. A hand wash with a chlorhexidine-based soap can reduce the number of transient bacteria by 2.1 to 3 log10 units; the effect on the resident hand flora is smaller, with a mean reduction between 0.35 and 2.29 log10 units. The dermal tolerance is rather poor, and anaphylactic reactions have been reported (Table (Table99).
Triclosan is one of many phenol derivatives (diphenoxyethyl ether) which have been used as a group of active agents since 1815, when coal tar was used for disinfection (222). Ever since, many different derivatives, such as thymol, cresol, and hexachlorophene, have been isolated and synthesized. Some of them have been used in antiseptic soaps for health care workers. Triclosan was introduced in 1965 and has been marketed as cloxifenol, Irgasan CH 3565, and Irgasan DP 300. It has very good stability (585) and resists diluted acid and alkali (453). The commonly used concentration in antiseptic soaps is 1%.
The mode of action of triclosan was identified some years ago. For decades, it has been assumed that triclosan attacks the bacterial cytoplasmic membrane (372, 458). Since 1998, we have known that it blocks lipid synthesis by inhibition of the enzyme enoyl-acyl carrier protein reductase, which plays an essential role in lipid synthesis (367). Mutation and overexpression of the fabI gene—which encodes the enoyl-acyl carrier protein reductase—are able to abolish the blockage of lipid synthesis caused by triclosan (205, 312). The fabI gene was first found in E. coli (366) and was subsequently also found in various other bacterial species such as P. aeruginosa (215), S. aureus (203, 520), and M. smegmatis (365). Some other bacteria, such as Bacillus subtilis, contain orthologous enoyl-acyl carrier protein reductases, namely those encoded by fabI and fabK, which are not inhibited by triclosan (204, 206). A genetic sequence coding for broad-spectrum resistance to triclosan has been identified (239).
The identification of the specific mode of action has raised concerns about the development of resistance to triclosan (313, 366, 506). A recent study has shown that this concern is valid. Strains of P. aeruginosa were exposed to triclosan and subsequently developed multiresistance to various antibiotics, including ciprofloxacin (86). Particular care should be taken in the use of triclosan in ICUs, where P. aeruginosa is the most common nosocomial pathogen, causing lower respiratory tract infection (260).
In vitro, triclosan exhibits a bacteriostatic effect at lower concentrations (575); at higher concentrations, it has bactericidal activity (560). The activity of triclosan is greater against gram-positive organisms than against gram-negative bacteria, particularly P. aeruginosa (238). MIC of triclosan generally range between 0.025 and 4 mg/liter among isolates of S. aureus and MRSA (94, 459, 543). The fungicidal activity of triclosan is good and includes yeasts and dermatophytes (459).
For artificially contaminated hands, a 1-min hand wash with 0.1% triclosan has been shown to reduce the number of test bacteria by 2.8 log10 units (475), which is essentially identical to the results obtained with nonmedicated soap (257). A soap based on 1% triclosan was found to reduce the resident hand flora within 5 min by 0.6 log10 unit (305). A 2% concentration yielded no major difference at 0.8 log10 unit (49). If hands were contaminated with rotavirus and treated with 2% triclosan for 30 s, the number of test viruses was reduced by 2.1 log10 units (47). On the resident hand flora, 1 or 2% triclosan has only a small effect, showing a mean reduction between 0.29 and 0.8 log10 unit within 5 min (Table (Table88).
In comparison to plain soap, at 0.2% triclosan does not further reduce bacterial counts on the hands (295). Under clinical conditions, a hand wash with 1% triclosan was reported to be less effective on the total bacterial count than a 4% chlorhexidine hand wash (140). Triclosan was able to eliminate MRSA from the hands (140). In contrast, gram-negative bacteria were less likely to be eliminated after the use of triclosan (140).
One S. aureus isolate for which the triclosan MIC is >6,400 mg/liter has been described (494). Some isolates of gram-negative bacteria have been found with triclosan MICs of >100 mg/liter as well (459). This high resistance was not transferable and was probably chromosomal (494). Exposure of S. aureus to 0.01% triclosan over 28 days did not result in a change of the triclosan MIC (543). Using S. epidermidis in a similar test, however, resulted in an increase of the MIC from 2.5 to 20 mg/liter, indicating a high potential for adaptation of the bacterium (545). Exposure of P. aeruginosa to 25 mg of triclosan per liter yielded multiresistant mutants which exhibited resistance to triclosan (MIC, >128 mg/liter) and some antibiotics, e.g., tetracycline (MIC, >256 mg/liter), trimethoprim (MIC, >1,024 mg/liter), and erythromycin (MIC, >1,024 mg/liter) (86).
An antiseptic hand wash preparation based on 1% triclosan was found to be contaminated with S. marcescens in an operating theater and a surgical ICU (43). This involved 4 (17%) of 23 bottles and 5 (28%) of 18 wall dispensers, but no association with a higher rate of NIs was found (43).
The widespread use of triclosan in antibacterial household products such as liquid soaps is cause for concern that selection for bacteria with an intrinsic resistance to triclosan may be occurring (314). Triclosan can be found in 76% of antibacterial liquid soaps in the United States (424), which has led to the recommendation that it should not be used in consumer products (547). It is therefore not surprising that highly resistant bacteria were detected in compost, water, and soil (369). Two species, Pseudomonas putida and Alcaligenes xylosoxidans, were even capable of metabolizing triclosan and thereby of actively “digesting” the active agent (369).
Detergents containing less than 2% triclosan are generally well tolerated. In one laboratory-based study of surgical hand disinfectants, a detergent containing 1% triclosan caused fewer subjective skin problems than did formulations containing an iodophor, 70% ethanol plus 0.5% chlorhexidine gluconate, or 4% chlorhexidine gluconate (305). Allergic reactions to triclosan-based handwash products are uncommon (616).
In summary, triclosan (1 to 2%) has good activity against vegetative bacteria and yeasts but limited activity against mycobacteria and dermatophytes. The activity against viruses is unknown. Triclosan has a low potential for acquired bacterial resistance. A hand wash with a triclosan-based soap can reduce the number of transient bacteria by 2.8 log10 units; the effect on the resident hand flora is lower, yielding a mean reduction between 0.29 and 0.8 log10 unit. The dermal tolerance is rather poor (Table (Table99).
The general antimicrobial activity of alcohols has been described to increase with the length of the carbon chain and reaches a maximum at six carbon atoms (548). Solubility in water has led to a preference for ethanol and the two propanols. Alcohols have a nonspecific mode of action, consisting mainly of denaturation and coagulation of proteins (241). Cells are lysed (229, 428), and the cellular metabolism is disrupted (360).
Ethanol is a well-known antimicrobial agent, which was first recommended for the treatment of hands in 1888 (473). The antimicrobial activity of isopropanol (equivalent to propan-2-ol) and n-propanol (equivalent to propan-1-ol) was first investigated in 1904 (612). Many studies followed and supported the use of the two propanols for hand disinfection (52, 85, 322, 395).
Both the alkyl chain length and branching affect the antimicrobial activity (562). The following ranking regarding the bactericidal activity has been generally established: n-propanol > isopropanol > ethanol (95, 476, 548). The bactericidal activity is also higher at 30 to 40°C than at 20 to 30°C (561). In terms of virucidal activity, ethanol is superior to the propanols.
(a) Ethanol. Ethanol has a strong immediate bactericidal activity (297) that is observed at 30% and higher concentrations (383, 444, 448, 449). Against S. aureus. E. faecium, or P. aeruginosa, its bactericidal activity seems to be slightly higher, at 80% than at 95% (110). According to the tentative final monograph for health care antiseptic products, ethanol is considered to be generally effective at between 60 and 95% (21). The spectrum of bactericidal activity of ethanol is broad (198).
Ethanol is also effective against various mycobacteria. Ethanol at 95% killed M. tuberculosis in sputum within 15 s, 70% ethanol required a contact time of 30 s, and 50% ethanol required 60 s (524), which was also required against M. smegmatis (54). Similar results were obtained with 70% ethanol and M. tuberculosis (55). For Mycobacterium terrae, the surrogate test strain for M. tuberculosis, a log10 reduction of >4 was found with 85% ethanol within 30 s (258). Very good activity was also shown with 70% ethanol against M. bovis (56).
The spectrum of virucidal activity is largely dependent on the concentration of ethanol. Higher concentrations of ethanol (e.g., 95%) generally have better virucidal activity than do lower concentrations, such as 60 to 80%, especially against naked viruses (127, 244, 534). A hand rub based on 95% ethanol has been described to have broad virucidal activity within 2 min, even against the most common nonenveloped viruses such as poliovirus and adenovirus (19). A gel based on 85% ethanol was still effective with a reduction factor (RF) of >4 against poliovirus within 3 min and against adenovirus within 2 min (258). Most naked viruses such as poliovirus (258, 262, 268, 535, 566), astroviruses (288), feline calicivirus (164), rotaviruses (258, 288), and echoviruses (287, 288) are inactivated by ethanol as well. HAV may be the only virus which is not fully inactivated; however, a higher RF of 3.2 was found with 95% ethanol whereas the RF was only 1.8 with 80% ethanol (615). Preparations containing less than 85% ethanol are usually less effective against viruses (570), although they may reveal sufficient activity within 10 min against various nonenveloped viruses such as adenovirus, poliovirus, echovirus, or Coxsackie virus (268). Under variable test conditions and at different exposure times, ethanol has broad general activity against the enveloped viruses, such as vaccinia virus (61, 184, 185, 268), influenza A virus (185, 268), togaviruses (77), Newcastle disease virus (97), HIV (346, 529), HBV (68, 272), and herpes simplex viruses (268).
Ethanol is known to have virtually no sporicidal activity (56, 165). This was first described over a century ago (135, 199, 395, 461). A pseudo-outbreak was reported due to contamination of ethanol with spores of B. cereus. The ethanol was used in the hospital pharmacy for preparation of skin antiseptics without spore filtration (219). Another report described contamination of 70% ethanol with spores of Clostridium perfringens, which was eliminated by addition of 0.27% hydrogen peroxide over 24 h (602).
(b) Isopropanol. The bactericidal activity of isopropanol begins at a concentration of 30% (445) and increases with increasing concentration but is lower again at 90% (544). It is similar to the bactericidal activity of n-propanol (612). In suspension tests, a hand rub based on propanols (total of 75%, wt/wt) had a comprehensive bactericidal activity against 13 gram-positive species, 18 gram-negative species, and 14 emerging pathogens within 30 s. Test bacteria included both ATCC strains and clinical isolates (248). Variations of the test conditions (e.g., with organic load) usually have no effect on the overall result in suspension tests (253). A tuberculocidal activity was found with isopropanol between 50 and 70% (150). The virucidal activity against naked viruses is limited and usually does not include enteroviruses such as astrovirus or echovirus (287, 288). If the exposure time is extended, sufficient activity against some nonenveloped viruses—such as echovirus (90% isopropanol for 10 min), feline calicivirus (50 to 70% isopropanol for ≥3 min), or adenovirus (50% isopropanol for 10 min)—can be achieved (164, 268). Isopropanol alone has no sporicidal activity, as shown with spores of B. subtilis and Clostridium novyi (445).
(c) n-Propanol. As early as 1904, n-propanol was described as an alcohol with a very strong bactericidal effect (548, 612) starting at a concentration of 30% (250). Compared to isopropanol, the activity against feline calicivirus seems to be better (164). In general, however, the antimicrobial activity of n-propanol is thought to be similar to that of isopropanol (475).
(a) Ethanol. On hands artificially contaminated with E. coli, ethanol at concentrations between 70 and 80% caused a reduction in the number of test organisms of between 3.8 and 4.5 log10 units within 60 s (475-477), and 1.96 log10 units within 10 s (23). Significant differences may be observed among alcohol-based gels. Up to an ethanol concentration of 70%, gels have been described to be significantly less effective than the reference hand disinfection (282, 432). A preparation with 85% ethanol, however, was found to be as effective as the reference hand disinfection, with 3 ml within 30 s (258).
Other types of artificial contamination of hands have only rarely been tested. Using S. aureus, a 30-s application of 70% ethanol achieved a 2.6 or 3.7 log10 unit reduction (34, 318). A similar result was found with 79% ethanol against Micrococcus luteus (mean RF, 3.2 after 30 s) (174). If hands were contaminated with rotavirus and treated with 70% ethanol for 10 s, the number of test virus was reduced by 2.05 log10 units (23). A longer application time of 30 s revealed a similar reduction of 2.72 log10 units (47). Low ethanol concentrations, e.g., 70 or 62%, did not even achieve a 1 log10-unit reduction of HAV on contaminated hands (355) but achieved a 2.9 to 4.2 log10-unit reduction within 20 s against adenovirus, rhinovirus, and rotavirus (496). Contamination with poliovirus was reduced by only 1.6 log10 units within 10 s by use of 70% ethanol (534). A solution of 80% ethanol reduced the carriage of poliovirus on fingers by only 0.4 log10 unit within 30 s (106). A higher concentration of ethanol (95%) reduced different naked viruses, such as adenovirus (RF, >2.3), poliovirus (RF, between 0.7 and 2.5), and coxsackievirus (RF, 2.9), significantly better on the hands (504). Against feline calicivirus, a sufficient efficacy (RF ≥ 3.83) was observed with 70% ethanol within 1 min without an organic load (164). Experiments with 5% fecal test suspension as the organic load, however, demonstrated a lowered efficacy of ethanol. Within 30 s, ethanol at 70% revealed a mean log10 reduction between 1.27 and 1.56 (244) and ethanol at 95% was more effective (mean RF between 1.63 and 2.17) (244).
The lack of sporicidal efficacy has been recently confirmed under practical conditions of hand contamination, using spores of B. atrophaeus, a surrogate for B. anthracis (594).
The effect on the resident hand flora depends on the ethanol concentration and the application time. A reduction between 1.0 and 1.5 log10 units has been found with ethanol at 70 and 80% within 2 min; higher concentrations (80 and 85%) and longer application times led to mean reductions between 2.1 and 2.5 log10 units (Table (Table88).
Comparison to antimicrobial soaps or nonmedicated soaps usually reveals the superior efficacy of ethanol on the resident hand flora or on artificial contamination of hands with E. coli or S. marcescens (32, 34, 66, 80, 267, 318, 377, 406, 419, 476). To date, there is only one study with a 2-min application time, yielding the opposite result (319). Other test models have been investigated as well. Compared to washing hands with plain soap, a 30-s hand disinfection using 70% ethanol was significantly more effective in reducing the transfer of Staphyloccocus saprophyticus (344). The higher bactericidal efficacy of ethanol than of antimicrobial soaps is even more pronounced in the presence of blood (296, 297).
Comparison to other alcohols reveals only minor differences. Using S. marcescens as a test organism, 70% ethanol with 0.5% chlorhexidine was described to be more effective under practical conditions than was 70% isopropanol, which may be explained by the different type of alcohol, the additional chlorhexidine, or both (14).
(b) Isopropanol. Isopropanol (60%) has been chosen as the reference agent for testing the efficacy of hygienic hand disinfection in European standard EN 1500 (116). With the reference treatment on hands which were artificially contaminated with E. coli and treated with two 3-ml doses for a total of 60 s, a mean reduction of 4.6 log10 units was achieved (256, 257). In other studies, similar results of 4.0 to 4.4 log10 units within 60 s were found (472, 475, 480, 482). The reduction with 70% isopropanol after 10 s, however, is 2.15 log10 units (23). In contrast, a gel based on 60% isopropanol was found to be significantly less effective than three liquid rinses against three test bacteria at 15 and 30 s (110). Using bacteria other than E. coli to artificially contaminate hands, similar mean reductions were found after 30 s in S. aureus (mean RF, 6.36), E. faecalis (mean RF, 6.07), and P. aeruginosa (mean RF, 6.81) (110). After 15 s, mean RFs were only marginally lower in S. aureus (mean RF, 5.90), E. faecalis (mean RF, 5.03), and P. aeruginosa (mean RF, 6.05) (110). If hands were contaminated with rotavirus and treated with 70% isopropanol for 10 s, the number of test viruses was reduced by 99.8% (RF, 2.7). The number of E. coli cells is reduced to a similar extent (99%; RF, 2.0) (23). A similar result was obtained when hands were contaminated with rotavirus and treated with 70% isopropanol for 30 s. The number of test viruses was reduced by 3.1 log10 units (47). Contamination with poliovirus was reduced only by 0.8 log10 units within 10 s after use of 70% isopropanol (534). The efficacy against feline calicivirus is also quite low, with a mean reduction of 0.76 log10 unit (90% isopropanol) or 2.15 log10 units (70% isopropanol) within 30 s (164).
Isopropanol at 60 and 70% has a rather low efficacy against the resident hand flora within 2 min (RF, between 0.7 and 1.2). With longer application times (3 and 5 min) and higher concentrations of isopropanol (80 and 90%), the mean reduction of the resident hand flora is between 1.5 and 2.4 (Table (Table88).
Comparison of isopropanol with nonmedicated soaps and antimicrobial soaps reveals the better efficacy of isopropanol, both on the resident hand flora (129, 316) and on hands which were artificially contaminated (33, 44, 472), with only one study showing discrepant results (306).
(c) n-Propanol. On hands which were artificially contaminated with E. coli, n-propanol at 100, 60, or 50% reduced the number of test bacteria within 1 min by 5.8, 5.5, or 5.0 log10 units, respectively (482, 604). Lower concentrations, e.g., 40%, still reduce the test bacteria by 4.3 log10 unit within 1 min (475). The efficacy against feline calicivirus seems to be quite good, with a mean RF between 1.9 (80% n-propanol) and ≥ 4.13 (50% n-propanol) within 30 s (164). Against the resident hand flora, 60% n-propanol is quite effective, with a mean reduction of 1.1 after 1 min and of 2.05 to 2.9 after 5 min (Table (Table8).8). A combination of isopropanol (45%) with n-propanol (30%) is significantly more efficacious than n-propanol (60%) on the resident hand flora in two studies; yielding a mean RF of 4.61 versus 2.9 in one study (240) and a mean RF of 1.45 versus 0.83 in the other (341).
(a) Ethanol. During an outbreak of gentamicin-resistant Klebsiella aerogenes, a health care worker was found to carry the strain on her hand. K. aerogenes was still detectable on two occasions after use of 95% ethanol for hand disinfection. The nurse continued to carry the strain for almost 4 weeks on her hand (82). Especially among health care workers wearing artificial fingernails, ethanol (60%) was found to be more effective in the removal of nosocomial pathogens than was an antimicrobial soap (368).
(b) Isopropanol. Under clinical conditions, a combination of isopropanol, n-propanol, and mecetronium etilsulfate was found to be significantly more effective than a chlorhexidine-based liquid soap (168). Isopropanol at 60% was found to have a better bactericidal efficacy on the resident hand flora than do antiseptic soaps based on chlorhexidine or triclosan (382). The higher bactericidal efficacy of isopropanol compared to antimicrobial soaps is even more pronounced in the presence of blood (296, 297).
Isopropanol at 60 to 70% was found to be necessary for removal of aerobic gram-negative bacteria from hands, whereas a simple hand wash with soap was inadequate (125). Transmission of gram-negative bacteria was also significantly better interrupted by propanol than by a social hand wash following brief contact with a heavily contaminated patient source (129).
(c) n-Propanol. Comparisons of n-propanol with nonmedicated soaps and antimicrobial soaps consistently reveal the greater efficacy of n-propanol on hands which were artificially contaminated (33, 480, 482). Comparison between n-propanol and isopropanol reveals a slightly greater efficacy of n-propanol (33). The efficacy of 60% n-propanol was found to be similar to that of 90% isopropanol on the resident bacteria (483).
No acquired resistance to ethanol, isopropanol, or n-propanol has been reported to date.
Alcohols are considered to be among the safest antiseptics available and generally have no toxic effect on human skin (332). One of the first studies was carried out in 1923 and found that isopropanol had no noticeable harmful effect on human skin (181). This has been confirmed in a repetitive occlusive patch test with n-propanol at various concentrations (333). In addition, different formulations based on various alcohols were tested on intact skin for 6 days and 4 weeks and were well tolerated (279). The skin barrier remains intact, dermal hydration does not change significantly, and the dermal sebum content remains unchanged (279). A similar result was found in a repetitive occlusive patch test with an ethanol-based hand gel (255) and a propanol-based hand rub (254). Even on preirritated skin, the potential for irritation by commonly used alcohols is very low (333). Repeated exposure to alcohol or a moderately formulated product can cause or maintain skin dryness and irritation (108, 197, 475). Ethanol is less cytotoxic (278) and may be less irritating than n-propanol or isopropanol (108, 281, 423). Adding 1 to 3% glycerol, humectants, emollients, or other skin-conditioning agents can reduce or eliminate the drying effects of alcohol (34, 182, 306, 328, 396, 408, 481, 587).
Various studies have addressed the question whether alcohol-based hand rubs have a dermal tolerance that is similar to or better than that of nonmedicated or antimicrobial soaps. Several prospective trials have demonstrated that alcohol-based hand rubs containing emollients may cause significantly less skin dryness and irritation than washing hands with liquid detergents (70, 303, 304, 378, 611). For example, a prospective, randomized clinical trial with crossover design was conducted with nurses working on several hospital wards in order to compare hand washing with a nonantimicrobial liquid detergent and hand disinfection using a commercially available alcohol hand gel. The condition of the skin of nurses' hands was determined at the beginning, midpoint, and end of each phase of the trial by using participants' self-assessment, visual assessment by an observer, and objective assessment of skin dryness via measurements of the electrical capacitance of the skin on the dorsal surface of the hands. Self-assessments and visual assessments by the observer both found that skin irritation and dryness occurred significantly less often when nurses routinely used the alcohol-based hand gel between attending to patients, and electrical skin capacitance readings demonstrated that skin dryness occurred significantly less often when the alcohol hand gel was used (70). A questionnaire study conducted at the end of the trial found that more than 85% of nurses felt that the alcohol hand rub caused less skin dryness than did washing with soap and water and that they would be willing to use the product routinely for hand hygiene (69). In another study of 77 operating-room staff who used either an alcohol-based hand rub or an antiseptic liquid soap for surgical hand disinfection, skin dryness and skin irritation decreased significantly in the group using the alcohol rub whereas they both increased in the group using soap (416). In another clinical trial, nurses were randomly assigned to use either a nonantimicrobial liquid detergent or an alcohol-based hand rinse, and skin tolerance was studied by using a combination of self-assessments, evaluations by a dermatologist, and measurements of TEWL. Self-assessments and those of the dermatologist found that the alcohol hand rinse was tolerated significantly better than the liquid detergent (611). There was no significant difference in TEWL readings with the two regimens. In a prospective, randomized trial conducted with ICU personnel, the effects on skin condition of a detergent containing 2% chlorhexidine were compared to those of an alcohol-based hand rub. Both the skin scaling scores and self-assessments found that the alcohol-based hand rub was tolerated better than the detergent containing 2% chlorhexidine (303). In a similar randomized, prospective trial in a neonatal ICU, the alcohol-based hand rinse regimen was tolerated significantly better than a detergent containing 2% chlorhexidine (301).
In a prospective intervention trial designed to study the impact of introducing an alcohol hand rinse on hand hygiene compliance among health care workers, dermatologist-assessed skin dryness and irritation revealed that the alcohol hand rinse was tolerated better than the traditional antiseptic hand-washing preparation (167). Measurements of skin hydration improved (although not significantly) after the alcohol hand rinse was introduced. Other clinical studies have also shown that alcohol-based hand rubs are tolerated well by health care workers (351). Furthermore, in a laboratory-based study of hand disinfection which compiled observations by an expert, self-assessments, and TEWL measurements, an alcohol-based hand rub caused less skin irritation than did a detergent containing 2% chlorhexidine (186).
Another trial based only on self-assessments to determine the impact on skin condition of an alcohol hand rub versus a detergent containing 4% chlorhexidine gluconate also found that the alcohol-based product was better tolerated (384).
In health care facilities where hand washing with plain soap or antimicrobial soap and water has been the rule, switching (particularly in the winter) to an alcohol-based hand rub may cause some personnel to complain of burning or stinging of the skin when applying alcohol. This is usually due to the presence of underlying, detergent-associated ICD among personnel (252). Skin that has been damaged by preexisting exposure to detergents may be more susceptible to irritation by alcohols than are nondamaged skin areas (333). As the skin condition improves with continued use of alcohol-based hand rubs, the burning and stinging associated with applying alcohol invariably disappears.
Allergic contact dermatitis or contact urticaria syndrome induced by exposure to alcohol-based hand rubs occurs rarely (88), and the cause is not clear. For example, surveillance at a large hospital where a commercial alcohol hand rub has been used for more than 10 years has not identified a single case of well-documented allergy to the product (606). In the few observed cases, however, it remains unclear whether the allergic reaction to the product is caused by the ethanol or by any of the auxiliary agents of the formulation (88). When reactions do occur, they may be caused by hypersensitivity to the alcohol itself, to aldehyde metabolites, or to some other additive (413). Allergic reactions to ethanol or isopropanol have been reported, are extremely rare (413), and depend on the chemical purity of the tested alcohol. Other ingredients in alcohol-based hand rubs that could be responsible for allergic reactions include fragrances, stearyl or isostearyl alcohol, benzyl alcohol, myristyl alcohol, phenoxyethanol, propylene glycol, parabens, and benzalkonium chloride (28, 107, 152, 189, 413, 442, 620).
In summary, ethanol (60 to 85%), isopropanol (60 to 80%) and n-propanol (60 to 80%) have very good activity against vegetative bacteria, mycobacteria, yeasts, dermatophytes, and enveloped viruses. Ethanol is more effective against naked viruses than are isopropanol and n-propanol. None of the alcohols has a potential for acquired bacterial resistance. Hand disinfection with an alcohol-based hand rub can reduce transient bacteria by 2.6 to 6.8 log10 units, but the effect on the resident hand flora is lower, with a mean reduction between 1.5 and 2.9 log10 units. The dermal tolerance is good (Table (Table99).
Compared with no hand washing at all, a simple hand wash reduces the transmission of nosocomial pathogens. Enforcement of a simple hand wash together with other infection control measures on a neonatal ICU led to a significant reduction of rectal colonization with VRE among newborns (40.2 versus 7%) (340). The simple hand wash has also been shown to be effective after direct contact with contaminated objects and before meals for prevention of infectious enteritis caused by Salmonella enterica servar Enteritidis (149). A similar effect on the incidence of diarrhea has been reported from India, although the hand wash had no effect on diarrhea caused by rotaviruses (512). One study exists which shows an effect even on the transmission of respiratory tract viruses. More frequent hand washing by health care workers in combination with cohorting of patients with respiratory tract infections caused by RSV has been found to reduce the nosocomial spread of RSV (225). Although these studies indicate that hand washing can reduce the transmission of nosocomial pathogens, especially during outbreak investigations involving multiple control measures, it is impossible to determine the individual contribution made by hand hygiene in preventing transmission.
Overall, wet hands have been described to significantly increase the risk of cross-transmission, indicating that hands should always be thoroughly dried (373).
Only a few studies were available which examined the impact of antimicrobial soaps on NIs. One study looked at the colonization and infection rate with Klebsiella spp. in an ICU. The annual rate was reduced from 22% in 1972 and 22.6% in 1973 to 15.5% in 1974, which was explained mainly by the introduction of a chlorhexidine-based liquid soap (81). In another study, NIs were less frequent when personnel performed antiseptic hand washing instead a simple hand wash (339). Antiseptic hand washing was also associated with lower NI rates in some ICUs but not in others (349). Some investigators have found that nosocomial acquisition of MRSA was reduced when the antimicrobial soap used for hand washing was changed (597, 621).
The use of alcohol-based hand rubs in regular patient care and its promotion over the years resulted in an increase of compliance in hand hygiene from 48 to 66% and a decreased in the rate of NIs from 16.9 to 9.9% at the same time. This is a significant decrease, of 41.1% in the NI rate (439). A comparative study of ICUs was carried out to determine the efficacy of a chlorhexidine-based soap (4%) and an isopropanol-based hand rub (60%) with the optional use of bland soap in reducing NIs. Washing hands with the chlorhexidine-based soap resulted in a lower rate of NIs, but the difference was not significant. However, it has been stated that this study does not indicate which of the two hand hygiene treatments is superior in ICUs. The personnel used much smaller volumes of isopropanol than of chlorhexidine and washed their hands more often than they used the hand rub (117). The data should therefore be regarded as resulting from a comparison between a social hand wash and chlorhexidine rather than a comparison between isopropanol and chlorhexidine (170). On a single ward in a 498-bed acute-care facility, use of an alcohol-based hand preparation over a 10-month period resulted in a 36% decrease in the incidence of two indicator NIs (urinary tract infections and surgical-site infections), expressed as the infection rate per 1,000 patient-days (211). In another study with an ethanol-based hand gel, the incidence of C. difficile-associated diarrhea decreased from 11.5 to 9.5 cases per 1,000 admissions within 1 year, but the difference was not significant (172). At the same time, the incidence of hospital-acquired MRSA decreased from 50 to 39% (172). Introduction of an ethanol-based hand disinfectant in a neonatal ICU significantly reduced cross-transmission of K. pneumoniae within 3 months from 21.5 to 4.7 cases of nosocomial colonization per 1,000 patient-days (75). Cross-transmission of E. faecium and C. albicans decreased as well, while rates for E. coli. Enterobacter agglomerans, and E. faecalis remained low and almost unchanged (75). The use of a virucidal hand rub based on 95% ethanol was part of an effective outbreak management of gastroenteritis caused by norovirus which involved 63 patients and health care workers (263).
Compliance with hand hygiene practices is known to be low. Compliance rates have been described to vary between 16 and 81% (437), with an overall average of 40% (71). One of the main goals of the new CDC guideline on hand hygiene is to provide evidence-based recommendations for improvement of compliance with hand hygiene (71). It is known that strategies to improve compliance with hand hygiene practices should be multimodal and multidisciplinary (435). Many individual parameters with a proven effect on hand hygiene compliance, however, have been identified in the new CDC guideline (71). These are efficacy, dermal tolerance, accessibility, time required for the procedure education, and personal perception; they are discussed below.
Hospital personnel should be provided with efficacious hand hygiene products, such as alcohol-based hand rubs. A change of the hand hygiene agent has been described to be particularly beneficial in institutions or hospital wards with a high workload and a high demand for hand hygiene (71).
Hand-washing agents are known to cause irritation and dryness, resulting in lower compliance rates (71). Hand hygiene products should have a low irritancy potential, particularly when these products are used multiple times per shift. A change to alcohol-based hand rubs should be made with great care, especially during winter, when hand skin is more irritable (252). Provision of skin care products may help (71, 437). However, they should not impair the efficacy of agents applied to the hands (71).
Easy access to a fast-acting hand hygiene agents should be viewed as the main tool of the strategy (71, 435, 437). Hand hygiene should be made possible, easy, and convenient. In areas with high workload, alcohol-based hand rubs should be made available at the entrance to the patient's room or at the bedside, in other convenient locations, or in individual pocket-sized bottles to be carried by health care workers (71).
Insufficient time to carry out the procedure, e.g., caused by high workload or understaffing, is associated with poor compliance (71). The time required for nurses to leave a patient's bedside, to to a sink, and wash and dry their hands before attending the next patient is a deterrent to a high compliance rate (71). A hand wash may take 62 s, whereas only one-fourth that time is required to use an alcohol-based hand rub placed at the bedside (579).
Ongoing education and promotion of hand hygiene should accompany the introduction of alcohol-based hand rubs in order to achieve long-lasting improvement in hand hygiene practices. Educational elements should include topics such as the rationale for hand hygiene, indications for hand hygiene, techniques of hand hygiene, methods to maintain hand skin health, and the correct use of gloves (71).
The smell, consistency (“feel”), and color may be important characteristics of the hand hygiene preparation that can influence the compliance rate by affecting the personal perception of those who use it (71). Differences in the acceptability of various agents have been described (258, 279, 349).
As well as the above parameters, the choice of the agent and the contents of a preparation may well have an impact on the compliance rate (242, 349, 481). The choice of a hand hygiene agent has been described to be one of many factors contributing to a strategy to successfully promote hand hygiene in hospitals (436). Baseline compliance rates in different departments vary between 30 and 63% for any soap, including plain and medicated soap, and sometimes even with the optional use of alcohol-based hand rubs if no intervention is done (Table (Table10).10). Several other studies have been conducted to measure the effect of various interventions on compliance rates with hand hygiene practices. In many studies the agent for hand hygiene remained unchanged. A higher compliance rate could be achieved by educational and training. In other studies, introduction of an alcohol-based hand rub or gel was accompanied by an educational and motivational campaign. Compliance rates could also be increased, often to a higher rate compared with the rate associated with no change of the hand hygiene agent (Table (Table11).11). Introduction of ethanol-based hand rubs sometimes revealed lower compliance rates and sometimes revealed higher compliance rates, with a trend toward the higher rates. The acceptability of the preparation and the role model function of physicians apparently have considerable influence. Preparations based on isopropanol or a combination of isopropanol and n-propanol revealed consistently higher compliance rates if education and promotion are carried out during introduction of the preparation and if the preparation has a superior dermal tolerance (Table (Table11).11). A 25% increase of compliance with hand hygiene is possible (64) with the right choice of agent (which should have an excellent dermal tolerance and a high acceptability among users) and with an intensive educational and promotional campaign. These data very much support the recommendation of the CDC guideline to choose hand hygiene products with a low irritance potential and with a maximum acceptability by health care workers (71). The acceptability includes an assessment of the feel, fragrance, and subjective skin tolerance of the product (71). In this respect, well-formulated preparations based on propanol have been shown to have a better acceptability in terms of skin tolerance and skin dryness (279).
The social hand wash has only a few indications in hospitals and community medicine (243): mechanical cleaning when hands are visibly soiled with blood or other body fluids, before eating and after using the restroom, and if contamination of hands with bacterial spores is suspected (71). In these clinical situations, the simple hand wash reveals the best results compared with other possible hand treatments.
In the CDC guideline, a hygienic hand disinfection with an alcohol-based hand rub is the preferred treatment to be carried out after patient care activities that could lead to contamination of the hands of the health care workers, e.g., after contact with the patient's intact skin, body fluids or excretions, mucous membranes, nonintact skin, and wound dressings (if hands are not visibly soiled), when moving from a contaminated body site to a clean body site, after contact with environmental surfaces in the immediate vicinity of patients, and after glove removal (71). Hands should also be treated before having direct contact with patients, before donning sterile gloves when inserting devices such as vascular lines, indwelling urinary catheters, or peripheral vascular catheters (71). Hand hygiene is also indicated after using the restroom in cases of diarrhea and after blowing the nose in cases of an upper respiratory tract infection (18). The use of antimicrobial soaps in all these situations will probably be less effective in preventing cross-transmission of nosocomial pathogens and also has the risk of inducing occupational ICD.