Introduction

Clostridioides difficile (CD), formerly known as Clostridium difficile, is an obligate anaerobic, spore-forming, Gram-positive rod, first identified in 1935 as Bacillus difficilis in the fecal flora of healthy infants. It remained unrecognized as a cause of human infection until 1977 when it was identified as the predominate bacterial cause of antibiotic-associated diarrhea and pseudomembranous colitis (PMC) [1].

Until 2000, Clostridioides difficile infections (CDIs) were found only in sporadic cases; thereafter, there was an evolution toward an urgent public health threat worldwide with severe clinical presentations ultimately causing death [2].

The milestone event that showed the virulence of CD occurred in 2001, due to the newly emergent epidemic strain of CD, that later became the most prevalent strain, usually referred as ribotype (RT) 027, causing 30%–50% of CDI cases [3].

In 2005, the first outbreaks due to CD (RT 027) were reported in Europe, then followed by epidemics by the same RT [4, 5]. In these outbreaks, a higher death rate, poor response to metronidazole therapy and resistance to fluoroquinolones (levofloxacin, moxifloxacin) with anti-anaerobic activities have been observed. In 2012, the European Centre for Disease Prevention and Control (ECDC) reported that 48% of hospital gastrointestinal and 7.7% of all health care-associated infections were due to CDI, concluding that CD is the eighth most frequent pathogen causing nosocomial infections [6]. In 2016 EUCLID study (European, multicentre, prospective, biannual, point-prevalence study of CDI in hospitalized patients with diarrhea) was carried out in 482 hospitals from 20 European countries [7, 8]. The CDI rate during this period was 7.0 cases per 10,000 patient-days with great differences among countries; 138 different RTs were found and RT 027 was the most prevalent (18.5%). Compared to earlier findings, the most prevalent RTs were 001/072 and 014, while there was a threefold reduction in the prevalence of RT 078 and a shift in endemicity of RT 027 from UK to Germany and Eastern European countries [8]. A detailed analysis on the emerging RTs of CD in Europe can be found in the electronic supplemental material (EMS).

Interestingly, the EUCLID study also reported an underdiagnosis of CDI (ranged from 0% in Belgium, the Netherlands and Sweden to more than 60% in Bulgaria, Greece and Romania) because of suboptimal laboratory diagnostics, previous lack of consensus on optimal testing methodology, non-availability of typing and deficiency in uniformity of case definitions, clinical algorithms and recognition amongst clinicians of when to suspect CDI [7]. The total burden of CDI on the healthcare system is significantly underestimated because almost all studies are based on CDI in the hospital setting although it is becoming more prominent in the long-term care setting and in the community [9].

Considering this worrisome scenario, we wrote this expert opinion aiming to gather the available literature about the pathophysiology, the risk factors, diagnosis and treatment of CDI, giving a special insight into the critically ill setting. We searched Medline and Embase databases for studies published in English from inception until 01/09/2019 (as detailed in the ESM) including all the data from meta-analyses, randomized controlled trials (RCTs) and any recent interventional or observational studies considered relevant for the topic.

Pathophysiology

Transmission is via the fecal–oral route; when the ingested spores survive the acidic environment of the stomach and are exposed to bile acids, they germinate into vegetative bacteria in the small intestine.

The patients can remain asymptomatic, acting as a reservoir for contamination of the environment. Conversely, when the normal flora is disrupted, the microbe proliferates, attaches to the mucosa of the colon and produces the two toxins (toxin A, an enterotoxin, and toxin B, a cytotoxin, ten times more potent), generating severe inflammation and necrosis of the mucosal due to irreversible glycosylation of cell proteins [10].

Risk factors

Exposure to broad-spectrum antimicrobials is a known risk factor for CDI in all patients. Indeed, the antibiotics disrupt the gut microbiota by reducing the gut microbiome diversity and causing the depletion of “health-promoting” commensal bacteria that protect against CD colonization [11]. Clindamycin, cephalosporins, and fluoroquinolones are the most reported antibiotic classes associated with CDI. Also, carbapenems and trimethoprim/sulphonamides have been reported even though with lower power of association [12]. A significant protective effect of antibiotic stewardship programs on CDI incidence has been demonstrated [13], encouraging the clinicians to a cautious prescription of the broad-spectrum antimicrobials, choosing the appropriate dose and duration of treatment.

Recently, it has been shown that other drugs, such as nonsteroidal anti-inflammatory drugs (NSAIDs), are also linked to the development of CDI [14].

In the past years, several large controlled observational studies have identified an association between CDI and the use of proton pump inhibitors (PPIs), often used in ICU for the stress ulcer prophylaxis (SUP). PPIs increase the pH of the stomach that is barrier to ingested bacteria and prevents the proliferation of spores and the conversion to a vegetative form of CD. Moreover, they decrease the production of reactive oxygen species by neutrophils, which mediate the host defence against CD.

A recent meta-analysis involving 342,532 patients has shown a significant association between PPI use and increased risk of CDI (OR = 1.26, 95% confidence interval (CI), 1.12–1.39) [15]. Of note the use of PPI has been associated with increased risk of CDI occurrence even when compared to H2-receptor antagonist (H2-RA) (OR = 1.386; 95% CI 1.152–1.668) [16].

This association has also been reported in a retrospective study on critically ill patients receiving SUP, where the authors found a significantly higher incidence of CDI in patients treated with PPIs than with H2-RAs (6.7% vs 1.8%), and PPI use was found to be an independent risk factor for SUP-related CDI (OR = 3.3; 95% CI 1.5–7.1) [17].

However, it is remarkable that there are only three RCTs published on this topic and they all have failed to show any association between PPI use and CDI [18,19,20], even when pooled in a meta-analysis [21]. Therefore, further large RCTs are warranted to confirm the relationship between PPI use and risk of CDI.

Several other risk factors have been identified; they are widely explained in the electronic supplemental material (ESM) and summarized in Table 1.

Table 1 Clostridioides difficile risk factors*
Full size table

Infection control

CD colonization has been described as an independent risk factor for the subsequent development of CDI in hospital [22] and in intensive care unit (ICU) [23]. Although treatment of asymptomatic carriers is not recommended, they could play a role in the transmission [24], based on the contact with remaining spores in a contaminated environment.

Recently, a quasi-experimental controlled study using time series analysis has shown a possible benefit of active surveillance and contact precautions to reduce the rate of Hospital-acquired CDI (HA-CDI) [25]. However, the active surveillance and isolation of asymptomatic carriers is still not supported by sufficient evidence [26].

Conversely, there is evidence that the use of barrier precautions (gloves and gowns), the isolation of the infected patients and the use of chlorine-based chemical wipes are essential strategies to prevent transmission [2, 27, 28].

Noteworthy, most disinfectants commonly used to clean the hospital environment are not sporicidal, so the use sodium hypochlorite-based disinfectants and hand hygiene with soap and water should be implemented.

Last, a cornerstone strategy to reduce CDI burden is a strong antibiotic stewardship program. Indeed, it is essential to limit the use of antibiotics to mitigate the dysbiosis that predisposes to CD colonization. Interestingly, it has been demonstrated that up to 25% of antibiotic administration is not indicated, even in the ICU [29].

Diagnosis

According to the recent guidelines, the diagnosis of CDI relies on a compatible clinical picture (new-onset diarrhea, defined as a stool frequency of three or more loose stools per day, corresponding to Bristol stool chart types 5–7), the presence of individual risk factors in association with laboratory tests confirming the presence of toxigenic strain of CD or toxins in stools [2, 30, 31].

Various laboratory tests to diagnose CDI have been proposed. Those include culture, molecular testing (nucleic acid amplification test, NAAT) for the gene encoding toxins, detection of free toxins via enzyme-linked immunosorbent assay (ELISA) and antigen testing for glutamate dehydrogenase (GDH), a metabolic enzyme expressed by all CD strains.

The optimal approach for the diagnosis of CDI is still debated [32]. Recently, the European society of clinical microbiology and infectious disease (ESCMID) has recommended a two-step algorithm [30]. First, a screening test with high sensitivity and negative predictive value (NPV), e.g. NAAT or GDH tests. If positive, this test should be followed by a more specific test with high positive predictive value (PPV) to confirm the diagnosis, e.g. ELISA detecting free toxins. Importantly, in case the second test is negative, the patients should be clinically evaluated: they can be either truly infected (with toxin levels below the threshold of detection) or carriers of a toxigenic strain (Fig. 1).

Fig. 1
figure 1

Proposed diagnostic algorithm for patients with suspected Clostridioides difficile infection in ICU. ICU intensive care unit, NAAT nucleic acid amplification test, GDH glutamate dehydrogenase, ELISA enzyme-linked immunosorbent assay

Full size image

Noteworthy, in patients with high clinical suspicion of CDI, imaging can assist the diagnosis.

The most common CT finding is colonic wall thickening, even though it lacks specificity and can be present in other form of colitis. Other signs are dilatation, peri-colonic fat stranding, ascites, stratification of either two (“double-halo sign”) or three (“target sign”) layers of different attenuation, where the lower-attenuated layer represents the edematous submucosa (Fig. 2) [33].

Fig. 2
figure 2

CT scan showing hyper-enhancement of the bowel wall (arrow), diffuse mural thickening with submucosal edema (star) and ascites (with permission)

Full size image

Of note, CT findings frequently do not allow a reliable distinction between CDI and other diseases such as ischemic and radiation colitis, inflammatory bowel disease (IBD), colitis caused by cytomegalovirus or Escherichia coli [33].

A possible role of the flexible sigmoidoscopy with biopsy analysis (eFig. 1) has also been reported to rule out other possibilities, especially when the stool tests are negative or clinical status worsens [34].

Despite this procedure is not routinely part of the evaluation of CDI, the need for a diagnosis may outweigh the risk of perforation of the inflamed colon.

Therapy

Treatment of CDI depends on the severity of the disease. Nevertheless, to date, there are no universally accepted criteria or prospectively validated severity scores for CDI.

In 2014, the ESCMID defined “severe disease” as an episode of CDI with one or more specific signs and symptoms of severe colitis or a complicated course of disease, with significant systemic toxin effects and shock, resulting in need for ICU admission, colectomy or death. In this setting, white blood cells (WBCs) > 15,000 cell/mm3, decreased serum albumin (< 30 g/L) and rise in serum creatinine level (≥ 1.5 times the premorbid level) can be present [35].

More recently, the guidelines of the Infectious Diseases Society of America (IDSA) used criteria based on expert opinion, defining a severe disease only by the presence of WBCs > 15,000 cell/mm3 or serum creatinine level > 1.5 mg/day [26].

The latter guidelines recommend either vancomycin (125 mg orally four times daily) or fidaxomicin (200 mg twice daily) for an initial episode of non-severe CDI [26]. Oral vancomycin is concentrated in the gut lumen conferring to this antibiotic a superior pharmacokinetic property compared to metronidazole that is no longer recommended in the treatment of non-severe disease.

Fidaxomicin is an antibacterial agent that inhibits sporulation; it is more specific with less impact on the normal gut microbiome than vancomycin. A recent meta-analysis has compared the two antibiotics finding no differences in the cure rate, but patients treated with fidaxomicin had a significant lower recurrence rate. This could justify the use despite the higher cost [36]. Fidaxomicin has been already used among critically ill patients with efficacy comparable to that of general patients [37], even when crushed and administered via nasogastric tube (NGT).

In patients with severe disease, vancomycin doses can be increased up to 500 mg four times daily and intravenous metronidazole (dosage: 500 mg three times daily) should be added. In the ICU setting, treatment with the combination of oral vancomycin and intravenous metronidazole was found by a retrospective study to improve mortality rates [38].

To note, the use of tigecycline for the primary treatment of severe CDI has been reported; a recent retrospective cohort study has also demonstrated the superiority of this drug to the standard treatment [39]. Figure 3 summarizes all the therapeutic opportunities targeted on the severity of the disease with potential fields for future research.

Fig. 3
figure 3

Proposed algorithm and future research areas for the management of Clostridioides difficile infection in ICU. ICU intensive care unit, po: per orem (per os, orally), qds quater die sumendum (four times daily), bid: bis in die (twice daily), NPO nihil per orem (nothing through the mouth), WBCs white blood cells, CDI Clostridioides difficile infection, iv: intravenous

Full size image

Specific challenges in ICU

An important topic in the management of a patient with CDI is the discontinuation of the inciting antibiotics. It is recommended as an ancillary strategy for CDI, since the antibiotics may decrease the clinical response and increase the risk of recurrence [26]. Clearly, this may be challenging in the ICU setting, where 60% of the patients developing CDI have been shown having concomitant documented infections [40]. It seems appropriate once the results of the cultures are available to narrow the spectrum of the treatment avoiding the high-risk antibiotics discussed above.

A systematic review and meta-analysis of 22 studies including 80,835 critically ill patients has demonstrated a significantly longer ICU and hospital stay together with an increased mortality risk in patients with CDI compared to non-CDI patients with comparable morbidity score at ICU admission [41].

However, with an early treatment, ICU-acquired CDI is not independently associated with an increased mortality and does not impact significantly on the ICU length of stay [42]. In this scenario, the ICU physician should promptly treat the disease and in hospitals where the results of the tests are not available in a few hours, treatment should be started while awaiting the results.

In patients with negative results, the clinician should also consider repeat testing, well recommended in patients with worsening symptoms [26]. Despite the lack of any evidence, in critically ill patients with hemodynamic instability and without a definitive diagnosis, it would be appropriate to repeat the tests within a short timeframe (e.g. 48 h) even in patients with atypical presentations.

Critically ill patients can develop fulminant CDI (FCDI), characterized by hypotension, shock, ileus or megacolon. Ileus is defined by signs of severely disturbed bowel function such as vomiting and absence of stool with radiological signs of bowel distension. In this atypical presentation, when CDI is suspected, it is acceptable to collect a rectal swab as a substitute for the stool sample.

Interestingly, in this condition or in case of severe colitis, when medication cannot be given orally, vancomycin can also be administered per rectum or by endoscopic placement (carefully in case of toxic megacolon) [43]. However, to date there is no study directly comparing these forms of application to the oral application of vancomycin.

Toxic megacolon is characterized by radiological signs of distension of the colon (> 6 cm in transversal width of colon) associated with the signs of systemic toxicity such as hemodynamic instability, marked elevated WBCs (> 25,000/mm3), hypoalbuminemia and need of cardiorespiratory support.

In case of systemic toxicity, the patient should be urgently evaluated for surgical intervention [44, 45].

The surgical management can consist of subtotal colectomy with preservation of the rectum or of a diverting loop ileostomy, namely the creation of a loop ileostomy with colonic lavage followed by antegrade vancomycin enemas during the post-operative period. The latter is less invasive (usually with laparoscopic technique) and better tolerated by patients, but it is contraindicated in case of bowel perforation or ischemia.

Interestingly, a recent retrospective cohort study compared the two techniques and failed to show any difference in the outcome [46]. Further, well-designed studies are needed to understand if the diverting loop ileostomy may lead to improved outcomes as well as colon salvage and which patient can have the best benefit.

The need of life saving colectomy is required in up to 20% of the cases [47] and has a mortality rate above 45%. Colectomy is indicated in case of perforation of the colon or worsening clinical condition not responding to medical therapy [35]. Of note, hyperlactatemia is a marker of severity and surgery should be performed before lactate exceeds 5 mmol/L [35]. In patients who do not respond to medical treatment, timing is of paramount importance since early colectomy (within 48–72 h from the onset of systemic toxicity) is associated with a better outcome [44, 45].

Fecal microbiota transplantation: evidences and applications in the icu

Faecal microbiota transplantation (FMT) involves transfer of faecal material from a healthy donor to the gastrointestinal tract of the patient thereby restoring the gut microbiota (Fig. 4). It is clearly recognized as a highly effective therapy for recurrent CDI, as shown in several meta-analyses of RCTs [48, 49] and international guidelines [26, 35].

Fig. 4
figure 4

Clostridioides difficile infection on antibiotic-induced dysbiosis and fecal microbiota transplantation (FMT)

Full size image

Based on this relevant body of evidence, FMT may represent a useful treatment for patients with CDI in the ICU, for different reasons. First, CDI is the most common infectious cause of diarroea in the ICU, as 1–2% of all patients in the ICU develop this disorder [50] and has shown to increase length of stay, costs and mortality [51]. Therefore, FMT would be useful in all cases of multiple recurrences, when antibiotics have failed to cure the disease, with a relevant impact on health outcomes.

Moreover, FMT could be also used as a potential therapy for severe and complicated CDI. This condition is hardly manageable in clinical practice, as common standard treatment options are still not satisfactory. In randomized [52] and non-randomized [53, 54] clinical studies, FMT has achieved high (nearly 70–100%) cure rates of severe and complicated CDI, and in one study [53] it was associated with a decrease of the need for CDI-related surgery. Another reason to consider FMT as a potential therapy for ICU patients with CDI is that it has been shown to be highly safe, also in fragile and immunocompromised patients, at least in the short-term, while long-term safety data still lack [55].

Beyond CDI, FMT could represent a promising strategy to another relevant issue, that are the MDR infections. By restoring the healthy gut microbiota, FMT could promote the decrease of antibiotic resistance gene expression in the gut resistome of the patient. In a pilot clinical trial, FMT was able to fully eradicate 75% of antibiotic-resistant intestinal bacteria in patients with blood disorders [56].

Despite this new therapeutic option is promising in the reconstitution of the intestinal microbiota and in the treatment of the disease associated with dysbiosis, some issues must still be clarified to understand the feasibility of FMT in critically ill patients [57, 58]. These include the selection of the patients and donors, modes of preparation, routes of administration (NGT, enemas or endoscopy) and timing. Since antibiotics destroy the transplant, FMT should be probably performed after antibiotic therapy is discontinued. At present, more evidence is needed to determine the role of this exciting frontier in the care of critically ill patients.

Conclusions

CDI represents a redoubtable problem for the fragile ICU patients, in large part related to the use and abuse of antibiotics. Specific preventive strategies and therapies must be adopted to stem the phenomenon, often underestimated. The advancements of FMT could represent a highly interesting option in critically ill patients. Intensive care physicians should be aware of this possibility, although where and when to use it still needs to be clarified.

Unfortunately, most of the knowledge about CDI comes from studies performed outside the ICU. The authors look forward to the results of the upcoming study “diarrhoea interventions, consequences and epidemiology in the intensive care unit” (DICE-ICU) that will provide more insights into critically ill patients developing CDI [59].