Introduction

Clostridium difficile is the most common infectious cause of healthcare-associated diarrhea and rivals methicillin-resistant Staphylococcus aureus as the most common bacterial cause of health care-associated infections [1•, 2]. The Centers for Disease Control and Prevention (CDC) estimates that in the United States, C. difficile infections cause 250,000 illnesses and 14,000 deaths annually [3••]. Associated medical costs impose a burden in excess of $1 billioneach year [3••]. As with most health care-associated infections (HAIs], strategies to identify, treat and prevent C. difficile infection require a multi-pronged effort that encompasses both acute and long-term care facilities. Supported by a comprehensive body of high-quality studies and guidelines that focus on C. difficile in hospitals [1•, 4•, 5•, 6], there is a growing body of literature addressing the additional challenges faced by long-term care facilities (LTCFs). Here, we discuss prevention and management of C. difficile infection in LTCFs, the majority of which are nursing homes.

Microbiology and Pathogenesis

C. difficile is a Gram-positive bacillus that forms spores capable of resisting an array of adverse conditions, including exposure to acidic conditions (pH <1), heat (10 minutes at up to 80 °C), dehydration, and alcohol-based hand sanitizers [7, 8]. In its spore form, C. difficile also resists most routine environmental cleaning agents and may last for months on surfaces [9]. Both patients and healthcare workers may acquire spores on their hands, unwittingly disseminating spores throughout their environment and leading to unintended ingestion of the spores. Exposure to C. difficile spores may go unnoticed by individuals with a healthy gut microbiome as the bacteria pass through the intestine without finding an ecological niche. The phenomenon, termed colonization resistance, is a form of host-defense that protects most individuals from enteric pathogens like C. difficile [10]. For people with a disrupted gut microbiome, which is most commonly due to a systemic antimicrobial, ingested spores germinate and grow to high concentrations in the intestinal tract with toxin production and spore formation. Similar to infections caused by other clostridial bacteria, the primary means through which C. difficile causes disease is through toxins. The toxins, TcdA and TcdB, translocate across epithelial cell membranes, and cause depolymerization of the cytoskeleton, which leads to cell death. Both toxins are involved in disease pathogenesis.

In 2003, several reports described a dramatic increase in C. difficile infection rates associated with increased disease fatality, particularly among older adults [11]. This change was caused by the emergence of a new C. difficile strain, characterized as toxinotype III, restriction endonuclease group BI, North American pulsed field gel electrophoresis type 1 (NAP1) and ribotype 027 [12, 13]. Frequently referred to as epidemic C. difficile, the BI/NAP1/027 strain has three distinct features that may help explain both its rapid spread and resulting increase in disease severity. First, it is resistant to fluoroquinolone antibiotics. In 2002, these became the most commonly prescribed antibiotic in the United States, which coincides with the emergence of the epidemic strain [14]. At least in the outpatient setting, fluoroquinolone prescriptions among adults and older adults in the US remained essentially unchanged from 2000 to 2010, raising the possibility of persistent selective pressure that favors the epidemic over the non-epidemic strain as one reason for persistent and widespread dissemination [15, 16]. Second, compared to most non-epidemic strains, BI/NAP1/027 strains have an 18-base pair deletion in tcdC, a gene that is a putative negative regulator of toxin production [17]. Some studies have demonstrated that the BI/NAP1/027 strain produces greater concentrations of toxins TcdA and TcdB in vitro than other strains [18]. However, a recent study found that BI/NAP1/027 strains exhibited robust toxin production, the amounts were not significantly different from those of non-BI/NAP1/027 strains tested [19]. Moreover, a recent study involving precise genetic manipulation demonstrated that an aberrant tcdC genotype did not result in increased toxin production [20]. Finally, the BI/NAP1/027 strain produces CDT, a binary toxin associated with more severe diarrhea, higher fatality rates and increased risk of recurrent disease [21, 22]. CDTb binds the cell surface and induces translocation, thus permitting CDTa access to cytosolic contents and promoting cell death through cytoskeletal depolymerization, acting upon different molecular targets than TcdA and TcdB [23].

Epidemiology of C. difficile Infection in LTCFs

Since the advent of the BI/NAP1/027 strain, rates of C. difficile infection steadily increased, such that by 2009, it was part of nearly 1 % of all hospital stays [24]. This percentage of hospital stays disproportionately involved older adults. In 2009, the rate of C. difficile infection-related hospitals stays for adults 65–84 years and ≥85 years was fourfold and tenfold greater, respectively, than for adults 45–64 years [24]. Hospitalized patients developing C. difficile infection are more likely to be discharged to an LTCF [25–27], yet we know relatively little about the burden of this disease within this vulnerable population.

There is evidence that the BI/NAP1/027 strain may be a common cause of infections in LTCF populations [28–30]. In a study of the epidemiology of C. difficile in multiple hospitals in the Chicago area, Black et al. found that 67 % of patients with C. difficile infection discharged to LTCFs were infected with BI/NAP1/027 strains [27]. Among hospitalized patients with C. difficile infection, Archbald-Pannone et al. reported that LTCF residents were significantly more likely to be infected with BI/NAP1/027 strains than those admitted from home [30]. Patients infected with BI/NAP1/027 strains had a higher 6-month mortality and greater inflammation based on fecal lactoferrin testing than those infected with non-epidemic strains [25].

Measuring the burden of C. difficile infection in LTCFs requires a standard set of clinical case definitions and surveillance methods that are applicable to that setting (Table 1). While the clinical case definitions are easily applicable across both inpatient and outpatient settings, the current surveillance definitions may be less relevant for estimating the disease burden among LTCFs. Specifically, Mylotte hypothesized that exposure to systemic antibiotics and to C. difficile spores often occurs in hospitals, with symptom onset in nursing homes shortly after hospital discharge [28]. Accordingly, he proposed subdividing the definition for health care facility (HCF)-onset, HCF-associated C. difficile infection into LTCF-onset, hospital-associated and LTCF-associated (see Table 1 for details). Using these definitions, Guerrero et al. reported that among 40 patients at a single Veterans Affairs Medical Center (VA) with HCF-onset, HCF-associated disease, 34 (85 %) met the criteria for LTCF-onset, hospital-associated C. difficile infection, while six (15 %) had LTCF-associated disease [29]. Taking his sample from four community nursing homes, Mylotte et al. reported similar outcomes, with 69 % of incident C. difficile infections developing within 30 days of admission [31]. Using a larger sample of eight diverse geographic areas, the CDC reported a nearly identical rate, with 67 % of people with nursing home-onset C. difficile infections having been discharged from a hospital in the previous 4 weeks [32].

Table 1 Surveillance definitions of C. difficile infection
Full size table

Employing an alternative approach, the CDC’s National Healthcare Surveillance Network (NHSN) uses proxy measure to estimate the burden of C. difficile infection [33]. Their definition, based solely on laboratory data, uses the number of positive C. difficile tests per 10,000 resident days, excluding positive tests from the same resident following a previous C. difficile-positive test within the previous 2 weeks. Among 30 acute care hospitals in New York State, comparison of C. difficile infections detected using the NHSN laboratory-based definition versus those identified using a clinical definition yielded >80 % agreement [34]. A study at a single VA LTCF found a similar rate of concordance. The NHSN laboratory-based definition detected 76 of 100 C. difficile infections identified using a clinical definition [35]. The most notable area of discordance was among residents admitted to the LTCF who were already diagnosed with and on therapy for C. difficile infection.

To date, the most comprehensive description of the burden of C. difficile infection in LTCFs comes from the Ohio Department of Public Health, which mandated reporting of healthcare-onset C. difficile infection. Based on data from 2006, Campbell et al. found that the overall rate for initial cases was lower in nursing homes compared to hospitals (1.7–2.9 vs. 6.4–7.9 cases/10,000 patient days, respectively) [36]. The absolute number of C. difficile infections in nursing homes, however, exceeded those in acute care by more than 50 % (11,200 vs. 7,000 cases, respectively). Furthermore, using even a very conservative definition of recurrent disease (within 6 months of an initial case), both the number (4,300 vs. 1,300 cases, respectively) and proportion (38 % vs. 23 %, respectively) of recurrent cases in nursing homes far exceeded those in hospitals.

LTCF residents include both traditional nursing home residents and patients receiving short-term rehabilitation or post-acute care. Limited data are available on the incidence of C. difficile infection among these different resident categories. However, it has been noted that those receiving short-term rehabilitation after hospitalization may be at particularly high risk for infection [24]. Laffan et al. reported that the incidence of C. difficile infection was much higher on rehabilitation and subacute (i.e., ventilator-dependent rehabilitation unit) wards of an LTCF than on a traditional nursing home ward in the same facility [37].

Risk Factors for C. difficile Infection in LTCF Residents

Among the general population, exposure to systemic antibiotics and advanced age are the two primary risk factors for C. difficile infection [4•, 38]. Others, reviewed in greater detail elsewhere, include suppression of gastric acid production, underlying disease severity and low albumin [12, 38–42]. Additionally, hospitalization is a risk factor for C. difficile infection, which reflects the combination of diminished health and exposure to antibiotics in a location with opportunity to acquire C. difficile spores from the environment and from health care workers [32, 43, 44]. Not surprisingly, residence in an LTCF is also a risk factor for C. difficile infection for similar reasons [32, 45].

Distinct to LTCFs, however, is the proportion of residents colonized with C. difficile. Reported rates of asymptomatic colonization among LTCF residents ranges from 5 to 51 %, far exceeding the 1–3 % rate reported among the general population [46–52]. In general, studies have found that the prevalence of asymptomatic colonization is higher among LTCF residents than among hospitalized patients. For example, Riggs et al. [44] found that 51 % of LTCF residents were asymptomatically colonized with toxigenic C. difficile, whereas a subsequent study in the same facility demonstrated that only 11 % of hospitalized patients were asymptomatic carriers of toxigenic strains [53]. Asymptomatic carriers shed C. difficile spores into their environment [50]. Furthermore, they also have spores on their skin, which are easily acquired on the hands of health care workers [50]. Given that nearly 80 % of LTCF residents require assistance with at least four of five activities of daily living, the risk for unwitting acquisition and dissemination of spores by health care workers is notable [54]. These findings help explain the high incidence (40–50 %) of initial C. difficile infections unrelated to recent hospitalizations reported at some LTCFs [55, 56].

Diagnosis

The diagnosis of C. difficile infection requires both clinical symptoms consistent with the diagnosis (diarrhea defined as ≥3 unformed stools in <24 hours) and a positive test for genes that encode for toxins, or for the toxins themselves (Table 1). Inappropriate testing of individuals with loose stools not meeting criteria for diarrhea or with diarrhea attributable to non-infectious causes (e.g., laxatives, viral gastroenteritis) may result in false-positive diagnoses of C. difficile infection if asymptomatic carriage of toxigenic strains is present. For example, there have been several reports of pseudo-outbreaks of C. difficile infection when stool specimens were submitted for testing during Norovirus outbreaks [57–59]. Given the high prevalence of asymptomatic carriage in LTCFs, education of nurses and physicians on appropriate testing is particularly important in this setting.

Efficient diagnostic testing for C. difficile infection is needed to minimize delays in initiation of isolation and treatment for confirmed cases, while also allowing rapid discontinuation of empirical therapy and isolation when testing is negative. However, delays in diagnosis are common in practice. At a large private hospital, the time between symptom onset to sampling and sampling to treatment was 2.24 (range 1–17 days) and 3.76 days (range 1–19 days), respectively [60]. In a VA hospital and attached LTCF, the average time between placing an order and obtaining a test result from the on-site laboratory was 1.8 days (range 0.2–10.6 days), with the time required for collection of stool specimens contributing to much of the delay [61]. An intervention focused on expediting stool sample collection and testing and reducing rejection of specimens was effective in significantly reducing the time from test order to diagnosis [50]. Notably, in a prior study conducted by the same institution at a time when the affiliated LTCF was separate from the hospital, the average time from onset of diarrhea to diagnosis of C. difficile infection was significantly longer in the LTCF than in the hospital (5 versus 2 days, respectively) [25]. Because many LTCFs use off-site laboratories, improving the timelines of diagnostic testing may be a particular challenge in this setting.

Given the delays inherent in the use of off-site laboratories, it is often necessary to consider empiric treatment for C. difficile infection in LTCF settings. Current practice guidelines recommend empiric treatment only for patients with suspected severe C. difficile [3••]. Empiric treatment of patients with suspected recurrence of infection is also reasonable, given the high likelihood of infection in the setting of typical symptoms recurring after discontinuation of therapy. If delays in testing are anticipated in LTCF settings, empiric treatment for residents with high clinical suspicion for C. difficile infection but mild to moderate symptoms may be reasonable, rather than waiting for test results. In this setting, the risks of adverse effects of treatment (e.g., adverse drug reactions, promotion of colonization by vancomycin-resistant enterococci) must be balanced against the risks of adverse outcomes due to delays in treatment.

Management

The treatment of C. difficile infection among LTCF residents is the same as treatment in the general adult population. It begins with supportive measures that include replacing fluid and electrolyte losses, avoiding anti-peristaltic agents, and, whenever possible, stopping the inciting antibiotic [4•, 5•]. Metronidazole is the first-line agent recommended for non-severe disease, while oral vancomycin is recommended for those with severe disease [3••]. Due to a significant drug–drug interaction resulting in INR elevation, metronidazole should be avoided in patients receiving warfarin or the INR should be closely monitored. Since the emergence of the BI/NAP1/027 strain, there have been increasing reports of metronidazole treatment failure. In a recent systematic review of the evidence, Vardakas et al. concluded that oral vancomycin offers some advantages over metronidazole, with fewer treatment failures (22 % vs. 14 %, respectively) and a slight reduction in the risk for recurrent disease (24 % vs. 27 %) [62]. For first recurrences, current guidelines recommend treatment with a second course of the agent used for the initial infection; for additional recurrences, a course of tapered and/or pulsed oral vancomycin is recommended [4•, 5•]. Two recent therapeutic advances, fidaxomicin and fecal microbiota transplant (FMT), have increased the array of evidence-based options available for treating C. difficile infection, particularly for reducing the risk for recurrent disease and treating patients with multiple recurrences.

In general, ~25 % of adults successfully treated for C. difficile infection will experience recurrent disease, though this may be notably higher among LTCF residents [36, 62]. Risk factors associated with recurrent infection include previous recurrences, increasing age and exposure to additional antimicrobials (other than those used to treat C. difficile infections) [63–65]. Molecular typing shows that ~50 % of recurrent C. difficile infections are caused by a new strain [66, 67]. These findings suggest that vulnerability to recurrent disease may in part reflect failure to recover colonization resistance. To study this, Abujamel et al. collected serial stool samples from hospitalized patients during and following treatment for C. difficile infection and tested if the samples inhibited or supported C. difficile growth [68]. They found that most patients required 3 weeks following completion of either metronidazole or oral vancomycin for their fecal microbiota to recovery sufficiently to reestablish colonization resistance against C. difficile.

Accordingly, to minimize the risk for recurrent disease, an ideal therapy for C. difficile infection should favor more rapid restoration of the gut microbiota. This appears to be the advantage that fidaxomicin offers over oral vancomycin for treating initial C. difficile infections caused by strains other than BI/NAP1/027 and for first recurrences [69, 70]. Fidaxomicin is a novel macrocyclic antibiotic approved by the Food and Drug Administration (FDA) for the treatment of C. difficile infection in 2011. Compared to vancomycin, it appears to have little effect upon the major bacterial phylogenetic clusters that comprise a significant portion of human fecal microbiota, including those from Clostridium clusters IV and XIVa and the Bifidobacteriaceae family [71]. The disadvantage of fidaxomicin is its substantial cost. A 10-day course costs $2,800 dollars, compared with just $250 dollars for oral vancomycin compounded from a 1 gm dose of the intravenous formulation. Fidaxomicin may offer some overall cost-benefit by reducing expenses associated with recurrent disease, though this remains controversial [72, 73].

FMT may hold the most promise for treatment of both initial and recurrent disease. First described over 30 years ago, FMT uses feces from a healthy donor to instill and restore a healthy fecal microbiota to patients with active C. difficile infection [74, 75]. Aesthetic considerations aside, FMT seems to be an effective and safe treatment, curing a majority of recurrent C. difficile infections with one to two treatments [76••, 77, 78]. Even among a brief case series of ambulatory adults 80 years and older, FMT led to symptom resolution in eight of ten cases described [79]. Studies evaluating the fecal microbiome of people with recurrent C. difficile infection reveal an overall lack of microbial diversity [52, 76••]. Two weeks following FMT, the recipients showed an increase in the diversity of their microbiome, specifically with recovery of species from the Bacteroidetes family and from Clostridium clusters IV and XIVa, and overall patterns indistinguishable from the donor sample [76••]. A cost-effectiveness analysis that compared treatment of recurrent C. difficile infection with metronidazole, oral vancomycin, fidaxomicin and FMT found that FMT was the most cost-effective strategy [80]. Interestingly, the same authors report that if FMT is not feasible, oral vancomycin is the preferred alternative.

Prevention

Efforts to prevent C. difficile infection include reducing patients’ vulnerability to infection as well as stringent efforts to prevent exposure to spores through infection control and environmental decontamination.

Antimicrobial Stewardship

Among the many risk factors for C. difficile infection, the most readily modifiable is antibiotic exposure. This is especially important in LTCFs where antibiotics account for 40 % of prescriptions [81]. An alarming 25–75 % of those prescriptions are either inappropriate or unnecessary [82, 83]. In LTCFs, one of the most common reasons residents receive antimicrobials is for concerns of a urinary tract infection (UTI). Rojanapan et al. reported that, compared to remainder of nursing home population, residents in two nursing homes who were prescribed antibiotics for a UTI that did not fulfill the McGeer criteria were eight times as likely to develop C. difficile infection in the 3 months following treatment [84]. Reducing antimicrobial use also reduces C. difficile infection rates. Through a remarkable effort, the Scottish Government supported the development of a national antimicrobial stewardship plan, with a specific goal to reduce C. difficile infections in older adults [85]. Between 2008 and 2010, the rates of C. difficile infection/1,000 bed-days among patients aged ≥65 years were more than halved. At a VA LTCF, an infectious disease consult service achieved a 30 % reduction in antibiotic use, which correlated with a significant decrease in the rate of positive C. difficile tests [86, 87]. The resources necessary to support these types of intervention are not available to most LTCFs, and as the Scottish program suggests, may require a concerted national effort. Developing effective strategies to reduce antimicrobial use at the level of LTCFs remains a challenge and area of intense interest [88].

Infection Control

Current guidelines for prevention of C. difficile infections focus on the acute care setting [1•]. Potential strategies to adapt hospital-based recommendations for preventing C. difficile infection in LTCFs are detailed in Table 2. Because patients with C. difficile infection are considered the major source for transmission, basic measures to be implemented in all facilities focus on reducing the risk for transmission from symptomatic patients. These basic measures include placement of infected patients in contact precautions, in a private room if available, until diarrhea resolves and disinfection of their rooms and portable equipment after patient discharge, preferably with a sporicidal agent such as sodium hypochlorite, has occurred [1•]. If basic measures are unsuccessful in preventing C. difficile transmission, adherence to basic practices should be assessed prior to addition of other control strategies. Unfortunately, adherence to basic measures is often suboptimal. If implementation of basic measures has been optimized, several special measures can be considered in addition to basic measures [1•]. These special measures include placement of patients with suspected C. difficile infection preemptively in contact precautions, extending the duration of contact precautions until discharge, and interventions to improve environmental disinfection (e.g., daily disinfection of high-touch surfaces).

Table 2 Potential strategies to adapt recommendations to prevent Clostridium difficile infections in acute care facilities to long-term care facilities
Full size table

Although infection control measures are similar in hospitals and LTCFs, the LTCF setting offers several unique challenges for prevention of pathogen transmission. First, nursing homes are the long-term home of many residents and the need to prevent transmission of C. difficile must be balanced with the goal to provide a home-like environment. Second, LTCFs often lack sufficient private rooms to provide single room isolation. Third, many LTCFs have shared bathrooms, rehabilitation facilities, and dining and recreation areas. Fourth, many LTCF residents have dementia or other chronic conditions that compromise their ability to adhere to basic standards of hygiene and to comply with contact precautions. Fifth, the staff in LTCFs may have less training in infection control and less experience with C. difficile infection. Sixth, special approaches such as extending the duration of contact precautions may be much less feasible in LTCFs than in hospitals because the length of stay is much longer. Jinno et al. found that asymptomatic carriage with shedding of spores was common during the month after treatment of C. difficile infection, but noted that a majority of patients with recent infection in a VA facility were cared for in a long-term care setting [89]. Finally, as noted previously, many LTCFs do not have on-site laboratory services, and thus may experience significant delays in diagnosis of C. difficile infection.

Vaccination

A systemic antibody response to C. difficile toxins provides protection against development of acute diarrhea and against recurrence [90, 91]. Based upon these findings, development of an effective vaccine to prevent C. difficile infection has been an active area of clinical investigation. One candidate vaccine is now in Phase 3 trials and others are currently under development.

Conclusion

Age, comorbid illnesses, frequent antibiotic exposure and dependence on health care workers, in the setting of communal living, all serve to increase the risk of LTCF residents becoming colonized or infected with C. difficile. While the primary goal for treating C. difficile infection is symptom resolution, an important secondary goal is to reduce the risk of recurrent disease by using therapies that promote rapid restoration of a healthy gut microbiota capable of colonization resistance. Vaccines that promote robust antibody production against TcdA and/or Tcd B may be an effective long-term strategy to reduce the burden of C. difficile in older adults. Until then, the mainstays of prevention will continue to be the reduction of unnecessary antibiotic exposure and improvement of infection control measures.