Epidemiology, Diagnosis and Treatment of Clostridium Difficile Infection

Matteo Bassetti; Giovanni Villa; Davide Pecori; Alessandra Arzese; Mark Wilcox


Expert Rev Anti Infect Ther. 2012;10(12):1405-1423. 

In This Article


Since the beginning of the 20th century a continuous rise in the incidence of severe CDI has been observed in Canada,[7] in the USA[8] and in Europe.[9–11] In North America, there was a fivefold increase in the incidence in the whole population and an eightfold increase in the elderly.[7] In Europe, this continuous increase has been associated with outbreaks, first in the UK from 2003 to 2004, then in The Netherlands and Belgium from 2005, followed by France and other European countries.[12] Barbut et al. reported a mean incidence of nosocomial CDI in 23 European hospitals of 245 per 10,000 patient-days (minimum to maximum range: 0.1–7.1),[13] which is lower than the overall figure of 41 per 10,000 patient-days in a more recent study by Bauer et al.[14]

The most recent published data in the US report 336,600 hospitalizations that involved CDI, equivalent to almost one in 100 of all hospital stays, in 2009. CDI cases increased steadily during the past decade, but hospital stays have leveled off between 2008 and 2009.[201] During the last 10 years, other observed epidemiological changes include an increased rate of severe manifestations, such as septic shock, toxic megacolon and intestinal perforation. Furthermore, both an increase in treatment failure with metronidazole and several relapses were observed. Some authors reported an increased incidence in populations previously considered at low risk,[15,16] although this remains a contentious issue.[17]

The continent of Asia, as shown by three prospective studies, prevalence rates of C. difficile associated diarrhea (CDAD) in hospitalized patients developing acute diarrhea ranging from 11.1 to 26.6%.[18–20]

In 2005, an increase in CDI rates was observed in North American hospitals. Thanks to molecular analysis a new strain of C. difficile was identified as the cause of a large number of cases, and it was characterized as group BI by restriction endonuclease analysis. Other characterizations include one obtained by pulse-field gel electrophoresis, North American pulse-field type (NAP1); this isolate is also referred to as ribotype 027 by PCR ribotyping.[21] It is referred to C. difficile BI/NAP1/027 in this review, this epidemic strain is capable of producing in vitro higher levels of toxins A (TcdA) and B (TcdB).[22]

This hyperproduction of toxins is associated with accelerated kinetics and sustained production; the synthesis of toxins occurs in these strains during both the exponential and stationary growth phases, in contrast with common strains, in which toxins synthesis increases as bacteria enter the stationary phase.[23] Another postulated reason for the hypervirulence of C. difficile BI/NAP1/027 is the production of a binary toxin, called CDT, acting as a regulatory gene. The CDT binary toxin is found in approximately 6% of C. difficile isolates[24,25] and was discovered by Popoff et al.[26] The toxin consists of a binding component and an enzymatic component and displays an actin-specific ADP ribosyl transferase activity that leads to cytoskeleton disorganization. The genes encoding toxin A and toxin B and regulatory protein, are co-located on a locus called CdT.[27] Up-dated experimental evidence by molecular studies suggests that although the relative pathogenicity of most toxigenic and hyper-toxigenic genotypes is still unclear, it may be influenced by a PaLoc genetic variant. Indeed, comparative molecular studies of the PaLoc locus between several C. difficile strains isolated from various clinical conditions, including the BI/NAP1/027 strain found that the hypervirulent behavior of C. difficile isolates should be due to mutations in the negative regulator tcdC gene.[28] Additionally, it has been also shown that the 027 genotype, as well as genotype 078, which has been isolated from severe CDI and an equally considered hypervirulent C. difficile strain[29] appear to be genetically divergent from other strains.[183] Moreover, five additional unique genetic regions (a novel phage island, a two component regulatory system and transcriptional regulators) have been found in the BI/NAP1/027 C. difficile strain, absent in a nonepidemic O27 genotype, and in a C. difficile ribotype O12 genome reference strain, thus leading to the hypothesis that a hypervirulent C. difficile phenotype might be also related to expression of multiple genetic elements.[30]

A further characteristic of C. difficile BI/NAP1/027 is an increased sporulation rate in vitro in the absence or presence of non-chloride cleaning agents.[31] This may lead to a better survival and spread of the strain in the environment.

In contrast with historic strains of 027 C. difficile, the new hypervirulent strains are resistant to fluoroquinolones and erythromycin (MIC of >32 and 256 mg/l, respectively).[32,33] High levels of resistance to clindamycin have recently been described in Europe regarding C. difficile BI/NAP1/027; this strain is usually susceptible to standard therapy (metronidazole and vancomycin), even though there is a concern regarding a reduced susceptibility to metronidazole. Isolates with reduced susceptibility to metronidazole have been found to be transmitted between patients, but their clinical significance in terms of response to antibiotic treatment remains unclear.[34–36] There is evidence that sub-inhibitory concentrations of metronidazole, vancomycin and linezolid induce TcdA and TcdB gene trascription and toxin production.[37] Multiple antibiotics have been shown to promote C. difficile spore germination, vegetative cell growth and toxin production, including both older and newer fluoroquinolones[38] and cephalosporins.[39] Notably, using the same in vitro model approach, neither piperacillin–tazobactam nor tigecycline induce C. difficile toxin production.[40,41]

As mentioned above, cases of community-associated CDI (CA-CDI) are reported, albeit at a considerably lower rate than the hospitalized cases. In the USA, reported CA-CDI rates are 7.7 cases per 100,000 persons per year, of which 35% received no antibiotics within 42 days of C. difficile detection. Although still considered rare, CA-CDI can affect groups that were previously considered at low risk, such as children or pregnant women.[42]C. difficile BI/NAP1/027 infections are mostly described in hospitalized patients, but there is a recent evidence of community-acquired cases, especially in the surrounding community of a hospital where other cases were diagnosed.[43]

Possible community sources of CDI include soil, water, pets, animals used for food, meats and vegetables.[44] Even though there is no sufficient evidence that contamination of food may lead to clinical CDI in human beings, since rapid spread has been observed during recent years, a common vehicle such as food cannot be ruled out.[45] C. difficile can also be a cause of infection in animals, with similar mechanisms of pathogenicity.[46]

Wilcox et al. have recently published an accurate analysis on the current epidemiology in the UK.[47] An increase in the number of outbreaks of severe CDIs was observed in Europe, with more than 55,000 CDI cases only in England from 2007 to 2008.[202]

From 2007, due to this increase in incidence and mortality, all NHS hospitals in England were asked to report all cases of CDI. Reduction targets were established, and a C. difficile Ribotyping Network (CDRN) was created. Fecal samples were sent to CDRN reference laboratories when increased incidence, severity, recurrence, complications or mortality was observed.

In the 3 years of the analysis, ribotype 027 was the most frequently detected ribotype and more than 85% of the samples belonged to patients older than 65 years. These two characteristics, plus a severe disease (PMC, toxic megacolon and the need of abdominal surgery for CDI) and the exposure to at least two antibiotics, significantly correlated with an increased mortality.

Furthermore, from 2007 to 2010 a significant reduction in the incidence of CDI was observed (61% reduction of CDI cases in England),[203] coincident with the control of the epidemic C. difficile ribotype 027. This may reflect the success of control measures aimed at reducing cross infection; these measures were probably intensified in hospitals once ribotype 027 was isolated from the CDRN laboratories. The emergence of different and less-virulent ribotypes was also observed, and this may be a further explanation on the reduction of severe cases.[48] This decrease in the incidence of CDI caused by ribotype 027 was accompanied by a reduction of the incidence of serious complications, due to either the less frequent number of cases caused by the more virulent strain and better clinical vigilance and aggressive interventions.[47]

On this very last matter, interventions aimed at the reduction in the level of use of fluoroquinolones and cephalosporins, clindamycin and broad-spectrum penicillins through a strict antibiotic stewardship clearly demonstrated a reduction in the incidence of CDI cases,[49] also when associated with measures aimed at limiting the spread of the spores within the hospital environment, such as isolating or cohorting positive cases.[50]