Current research on hand hygiene

The focus on bulk reduction in microbial load is apparent in the majority of studies of hygiene—even those conducted by clinical microbiologists [11–14]. The concepts of hygiene and sterilization are often conflated, which is perhaps unsurprising given the history of hospital sanitation practices, which seek to remove all microbes from the environment [8]. There is a logical link between bulk reduction in microbial load and reduction in pathogen spread; however, relatively few studies go beyond cleaning and link hygiene directly to health outcomes, and many of these are specifically concerned with nosocomial infection [12].

Hand hygiene research has focused largely on hospital settings and the spread of nosocomial infection (reviewed in [2, 12]), in part due to the history of the field but moreover because of recognition of the increased risk of infection in places where potentially contagious and immunocompromised people are gathered. Where work on hand hygiene has taken place outside of hospital settings, it has focused on other areas with high risk of pathogen transmission, such as childcare facilities [15–17] or food handling situations [18], or has been undertaken in combination with efforts to improve community environmental hygiene, such as improved sanitation infrastructure in developing countries (reviewed in [19]).

Aiello and Larson found only 53 studies published between 1980 and 2001, out of thousands of studies matching their search criteria, that explicitly linked hygiene to health outcomes outside of healthcare settings [19]. Studies linking hygiene intervention to health demonstrate the effectiveness of handwashing at reducing the risk of diarrhetic disease [20, 21] and upper respiratory infection [21, 22]. Reduction in the rates of handwashing in response to fears of lead contamination have been suggested as a factor contributing to a recent Shigella outbreak in Flint, Michigan [23].

The lack of a clear and consistent definition of hygiene has led to confusion in the scientific literature. One example of this confusion is apparent across the hygiene literature related to hand drying. Though previously ignored [24], hand drying is a crucial aspect of hand hygiene because of the significant role that residual moisture plays in the transfer of microbes between surfaces [25–28]. The different methods of hand drying have varying hygienic advantages and concerns. For example, drying with paper towels is the method recommended for healthcare workers by both the Centers for Disease Control and Prevention [29] and the WHO [2], due in large part to bulk bacterial count data indicating that paper towels are effective at removing transient surface bacteria [24, 30–35]. Nevertheless, it is possible that waste paper towels may serve as a bacterial reservoir that may facilitate the spread of disease [36, 37]. Newer alternatives to paper towels such as jet air dryers (e.g., the Dyson Airblade™) are marketed as designed with a high-efficiency particulate air (HEPA) filter built into the airflow system, which reduces the risk of redistribution of airborne microbes to the hands [13]. However, there is concern about the propensity of such rapid air movement to aerosolize microbes from the users’ hands or the surrounding environment [14, 34, 37–39].

Much of the existing work on hand drying has examined the hygienic efficacy of various methods. What is meant by “hygiene” in any given study, however, is often left unstated and is generally inconsistent between studies from different research groups [40] but usually is measured by change in microbial load [13, 34], dispersal of microbes from the hands or some proxy thereof [38, 39], and/or effectiveness of drying [13, 36, 37]. Utilizing a definition of hygiene that explicitly relies on reduction in disease spread and takes into account microbial community dynamics would allow future experiments to adequately address the possible hygienic concerns of paper towel bacterial reservoirs or microbes aerosolized by dryers.

Skin habitat and microbial diversity

The human skin may be open to colonization from the environment, but it is thought to be a strong selective filter, largely unsuitable for most microbes to permanently reside [50]. The three major skin habitats (sebaceous, dry, moist), and the gradations of environmental conditions within and between them, largely determine the bacterial community living at a particular skin site [51, 52]. Skin bacterial communities, therefore, appear to have generally predictable biogeographic patterns. The normal/healthy skin microbiota is composed of a limited number of types of bacterial species (mainly Gram-positive species) [50, 52–55]. Dry regions, such as the forearm and palm, are often the richest in bacterial diversity, are generally less restricted in membership, and are more susceptible to temporal variability, while the sebaceous sites are generally poorer in bacterial diversity and dominated by Propionibacterium acnes, presumably due to high sebaceous gland activity that may result in more exclusivity [52, 53, 55, 56].

While there is some similarity in microbiota of similar body habitats and across individuals, it is abundantly clear that not all skin communities are alike [52, 53, 55, 57, 58]. There is some evidence of persistent community members at specific skin sites, with a few taxa being detected from the same site for up to a year [57, 59]. Nevertheless, it is difficult to define a core microbiota for a given anatomic site on the skin [60]. One review listed the most “common” human skin bacterial residents as: Staphylococcus, Corynebacterium, Propionibacterium, Micrococcus, Streptococcus, Brevibacterium, Acinetobacterium, and Pseudomonas [61]. Many of these (e.g., Staphylococcus aureus and Staphylococcus epidermidis) also have the potential to become multidrug-resistant pathogens, emphasizing the insidious nature of classifying microorganisms as one ecological mode (e.g., commensal vs. pathogenic) and the need for a conceptual framework taking into account the existing ecological continua.

The mutualist–pathogen continuum

Much of the scientific literature related to the human skin microbiota utilizes two different sets of vocabulary: resident and transient microbes (often used in the clinical literature; e.g., [59]) vs. commensal and pathogenic microbes (often used in the ecological literature; e.g., [62]). However, these terms are rarely defined and are frequently conflated; residents are often assumed to be commensal, and transients are often assumed to be potential pathogens.

Historically, resident microbes were thought of as those that were stable on the human skin and were difficult to remove, whereas transient microbes were thought to be acquired by contact and could be easily removed from the skin [63]. This notion of resident/transient microbes has continued for decades and has morphed into the assumption that resident microbes are those that commonly reside on skin whereas transient microbes are viewed as contaminants [64].

Commensal microbes on the human skin are regarded as those that are not typically associated with disease [61]. However, the ecological definition of commensalism refers to the condition where only one organism receives benefit and the other organism suffers no harm [61]. The use of the term “commensal” to describe non-harmful microbes on the skin is suggestive that only the microbe is receiving benefit from living on the skin’s surface and no benefit is provided to the human host. However, this definition is misleading because there is growing evidence that microbes once thought to be commensal may actually be involved in host defense, which would suggest a mutualistic, rather than a commensalistic, relationship [61]. In comparison, a pathogenic microbe on the skin is one that causes harm to the host. There are, however, many microbes that are associated with disease which exist as normal members of the skin microbiota in healthy individuals. Taking a lesson from ecology, it is necessary to consider the host, the microbe, and the environmental context to assess the relationship of any particular microbe to a disease state [65]. When the ecological relationship between host and microbe is unclear (that is, when it is impossible to say if a given microbe is acting as a mutualist, a commensal, or a pathogen), we prefer the term symbiont (literally, “together living”), which does not imply an ecological mode.

Both of these dichotomies represent continua, which are related but orthogonal to each other. The idea of a mutualist–pathogen continuum has been successfully applied in the plant microbial ecology literature for decades [66, 67]. This continuum represents a position in niche space, which can change through alterations to microbial or host genetics, environmental conditions, and community context [53, 61, 68, 69]. The resident–transient continuum represents a temporal dimension and is defined by the length of time that a given microbe is associated with its host—though we must consider the effect of the limits of detection with current techniques [57]. It is important to recognize that resident does not necessarily equate to commensal nor does transient always mean pathogenic.

It is possible for the human skin to have mutualistic, commensal, and pathogenic microbes as part of its resident “core” microbiota; a single microbial species may be all of these things. For example, the bacterium S. epidermidis is commonly found on the human skin and is generally regarded as commensal [61], although it can occasionally act as an opportunistic pathogen [53, 68] or a protective mutualist [70]. Recent evidence suggests that S. aureus, which has been typically thought of as a pathogenic microbe, is commonly present on the healthy skin, specifically in the nasal area [71]. Following this logic, it is likely that many of the transient microbes that people are exposed to in the environment are not only non-pathogenic, but in fact could be beneficial to the skin microbiota.

Microbial ecology of the skin

There is an emerging appreciation of the microbial ecology of the skin. Community ecology seeks to understand what factors determine the presence, abundance, and diversity of species in a community [72]. Island biogeography theory [73], in particular, allows us to conceptualize each person as an island: a patch of habitat that must emerge and assemble its communities by the fundamental processes of community ecology. The interactions between skin microbial communities and the host make understanding the ecological factors contributing to microbial communities particularly important. Multiple ecological factors interact to determine the species composition in a given ecological community; dispersal (Fig. 2a) and environmental filtering (Fig. 2b) are the two factors most relevant to the discussion of hygiene.

Dispersal (Fig. 2a) of commensal or mutualistic organisms may be particularly relevant to human health. Studies that have examined the transmission of human-associated microorganisms have almost exclusively focused on pathogenic microbes in healthcare settings [72]. The transmission of other (i.e., non-pathogenic) members of the skin microbiota is poorly understood, including the roles of a number of factors (e.g., diversity, interspecies interactions, host factors, environmental factors) on the ease of microbial transfer and subsequent colonization. Transmission via direct contact with other individuals, or indirectly with fomites or water droplets, introduces transient microbes that could alter the ecological dynamics of the skin microbiota [72].

All persons are dispersers of their microbiota, though dispersal rates vary within and among people. Organisms living on the skin have the potential to be dispersed as a result of normal desquamation (i.e., shedding or peeling of the outermost layer of the skin) [74]. Individuals emit a personalized microbial cloud that likely impacts both cohabitants and the microbiota of the built (indoor) environment itself [75]. While research on whether resident microorganisms can be transferred among individuals is nascent, it is hypothesized that delivery method at birth (vaginal vs. Cesarean section) affects initial skin microbial communities of infants [76] (but see [77]).

Environmental filtering (Fig. 2b) of dispersed microbes functions primarily through differences in skin habitat (e.g., dry vs. sebaceous sites). Interactions between microbial populations may be part of the “filtering” of the environment; thus, priority effects [78, 79] and factors related to the established microbiota can be considered part of the environment to newly dispersed microbes. There is evidence that host factors vary in the ability to promote bacterial colonization and that this varies by skin site [55]. The role of host/microbe feedbacks (Fig. 2d) in determining environmental selective pressures may also influence the outcomes of potential dispersal events. There is some evidence that microbial communities may be transferred between people or their environments [80, 81].

Invasion ecology focuses on perturbations of established communities and attempts to understand the factors that allow invasion by exogenous species [82, 83]. Applied to the skin microbiota, disturbance (e.g., hand hygiene practices) may be a major factor in alterations of the skin microbiota through invasion; protective mutualisms (Fig. 2c) may be disrupted or eliminated, allowing invaders to colonize. While the microbial ecosystem of the skin is relatively stable in the face of continual desquamation of the skin surface and frequent perturbations [58], disruptions by antibiotics, handwashing, or cosmetic application may alter the microbial community, enabling invasion of pathogenic microbes or a shift in dominance leading to dysbiosis [72]. Skin disturbance can predispose the host to a number of cutaneous infections and inflammatory conditions [61]. Despite the multitude of studies emphasizing the benefits of personal hygiene on reducing disease transmission by removing or reducing transient microorganisms, the effects of personal hygiene (e.g., handwashing) on the resident microbiota are not well studied [46]. Different behavioral habits (e.g., frequency and duration of washes, product used) likely account for at least some of the variation in microbial community structure and membership observed in human studies [72]. The frequency and magnitude of such disturbances likely facilitate the invasion of potentially undesirable, pathogenic species [84, 85].

A conceptual framework for understanding the interactions between the skin microbiota, the human host and environment, and the impact on human health must take into account all of these ecological factors [72, 84]. Significant and potentially harmful alterations of the skin microbial community structure may occur as a result of several factors: dispersal of non-resident microbes to the host microbiota, disturbance regimes (e.g., handwashing practices), local and regional environmental factors (i.e., environmental selective filters on source and sink populations, such as host skin condition and indoor settings), and the genetics and demographic characteristics of the host, which also provide selective filtering [72].

Skin microbiota and host immunity

The resident microbiota has evolved in conjunction with the human host and is thought to be important to the maintenance of healthy ‘normal’ skin function. Generally, the resident microbiota have a positive effect on human health through protective mutualism (Fig. 2c); it is only when the host becomes compromised that the resident microbiota displays pathogenic potential [86, 87]. As the skin is our body’s interface with the outside world, it must act to both prevent colonization by pathogens and tolerate or encourage the presence of potentially protective bacteria. The skin is a complex immunological organ with both innate and adaptive immune cells, including multiple dendritic and T cell subsets: antimicrobial peptides, proinflammatory cytokines, and chemokines that are secreted by keratinocytes to support an immune response [88–93]. While pathways related to infection response are relatively well understood [56, 61, 87], the mechanisms by which commensal or transient bacteria are tolerated by the cutaneous immune system are only beginning to be understood [93].

Host–microbe feedbacks (Fig. 2d), modulated through the host immune system, have been recently demonstrated and are likely to play critical roles in maintaining healthy host/microbiota relationships. For example, S. epidermidis has been shown to produce antimicrobial peptides and may modulate the host immune response [94]; S. epidermidis and Corynebacterium spp. are capable of reversing or preventing the successful colonization and proliferation of S. aureus in the human nares [47], such that removal of S. epidermidis may be harmful to the host through increased colonization of opportunistic pathogens [61]. Additionally, some studies have shown that microbial exposure—particularly early in life—may have long-term implications for health and immunity (known as the hygiene hypothesis [95, 96], but see [97]). Skin dysbiosis has been linked to many skin disorders, including acne vulgaris, psoriasis, and atopic dermatitis [98–106]. Investigation of the potential of microbiota transplants and probiotic skin treatments for these diseases are underway [53]. Thus, the skin microbiota is likely of direct benefit to the host and only rarely exhibits pathogenicity.

Microbial ecology applied to hand hygiene

The hands harbor greater bacterial diversity and are more temporally dynamic than other body sites [107]. More than 150 bacterial species have been recovered from human hands; these species primarily belong to the phyla Firmicutes, Actinobacteria, Proteobacteria, and Bacteroidetes [54, 107]. This increased diversity on human hands compared to other skin sites may be a result of the exposure of the hands to consistently varying external environments. Like other skin sites, there is a high degree of interpersonal variation in the hand microbiota; a minority of taxa (13%) are shared between the hands of any two individuals, and the two hands of a single person may share only a slightly larger fraction, though those communities appear somewhat stable [54] (but see [46], which suggests that resident communities are quite similar between dominant and non-dominant hands). Despite the evidence that certain bacterial species remain present on the hands over time, their relative abundances are variable [57]. Microbial communities on people’s hands are significantly affected by host factors, including sex, relatedness, living quarters, hand hygiene, and even pet ownership [54, 107, 108].

Hand hygiene is still regarded as the most important practice to prevent the transmission of microbes and minimize the spread of disease [109]. However, compliance with hand hygiene practices in healthcare settings is generally low, with mean baseline rates ranging from 5 to 89% [2]; typical rates may be no better than 40% [8, 109]. Current understanding of the effects of hand hygiene in healthcare settings largely stems from cultivation-based methods focusing on identification of pathogenic microbes. These clinical studies have historically been performed during periods of infectious outbreak in hospital settings with the assumption that bacteria on the skin are pathogenic contaminants [110]. Even with the growing use of high-throughput sequencing, there have been few cultivation-independent studies that have investigated the direct effects of hand hygiene or product use on the hand microbiota [46, 107, 111]. There is great potential to further our understanding of the human hand microbiota by utilizing an ecological perspective in healthcare settings, where hygiene practices are vital. Despite this current gap in knowledge, we are still able to draw preliminary conclusions about hand hygiene and its effect on the skin microbiota from cultivation-based studies and the few cultivation-independent studies that have looked at this relationship indirectly.

In cultivation-based studies, the length of direct patient contact is positively correlated with bacterial counts [109], and surface area and time of contact significantly affect the abundance of bacteria present on the hands of healthcare workers [112]. Older work has shown that soap and water handwashing is effective at removal of patient-acquired microbes [113], and more recent studies have shown alcohol-based handrubs to be as effective [109, 114] or even superior to soap and water [64]. There is also an interaction between skin health and the effect of hand hygiene that may be of concern: increased handwashing may increase the amount of microbes on the hands due to worsening skin health [46, 115]. Additionally, moisture level has a significant effect on cross-contamination rates [25–28]. However, these studies examined bacterial load on the hands and failed to address the identities of the species that were affected by hand hygiene practices—identity matters when most members of the microbiota are commensal or even potentially mutualistic.

Cultivation-independent studies show some similar trends: hand hygiene apparently reduced bacterial diversity on healthcare workers’ hands, though pathogen carriage on the hands was also observed [111]. Time since last handwashing was significantly correlated with changes in bacterial community composition but did not affect bacterial diversity [54]. This result could suggest that the microbiota present on the hands quickly reestablishes itself post-handwashing, or that few bacterial taxa are removed during the handwashing process [54].

In order for human-to-human microbial transmission to occur in a healthcare setting, the microbe must be capable of surviving for a period of time before it is transferred to another person or surface [116]. One study that looked at the transmission of Klebsiella spp. among healthcare workers in an intensive care unit found that only a slight contact with patients was needed to transfer the microbe to healthcare workers and that Klebsiella spp. could survive on dry hands for up to 150 min [117]. Another study found that the transmission potential of microbes to and from the hands and sterile fabrics was highly species-dependent, suggesting that hygienic practices may play a more vital role in transmission prevention of certain microbes over others [114].

From the available predominantly culture-dependent data, we can support that transmission via the hands is common and often related to microbial load and that variation in moisture levels affects transmission efficiency. However, since the majority of studies examined microbial load without taking taxonomic identities and ecological interactions into account, more studies are needed to quantify the role of interactions with the resident hand microbiota in the transmission of potentially pathogenic microorganisms.