Clostridium difficile PCR ribotypes 001 and 176 – the common denominator of C . difficile infection epidemiology in the Czech Republic , 2014

M Krutova 1 2 , J Matejkova 1 , EJ Kuijper 3 , P Drevinek 1 , O Nyc 1 , Czech Clostridium difficile study group 4 1. Department of Medical Microbiology, 2nd Faculty of Medicine, Charles University in Prague and University Hospital Motol, Czech Republic 2. DNA Laboratory, Department of Paediatric Neurology, 2nd Faculty of Medicine, Charles University in Prague and University Hospital Motol, Czech Republic 3. Department of Medical Microbiology, Leiden University Medical Centre, Leiden, the Netherlands 4. Members of the study group are listed at the end of the article


Introduction
Clostridium difficile is the most important bacterial cause of hospital-acquired diarrhoea.Two large studies have been carried out to map and update data on C. difficile infection (CDI) in Europe [1,2].CDI incidence showed an increasing trend: in the first study in 2008, the mean incidence in the participating countries was 4.1 cases per 10,000 patient bed-days [1], while in the second, in 2011-13, it was 7.0 CDI cases per 10, 000 patient bed-days in the countries involved [2].
Results of the 2008 study -a hospital-based survey involving 34 European countries -showed that the Czech Republic had a low incidence of CDI (1.1/10,000 patient bed-days), without the presence of C. difficile PCR ribotypes 027 and 176 [1].Spread of PCR ribotype 027 has been seen worldwide [3] and is known to be associated with hospital CDI outbreaks [4] and severe course of disease and increased mortality [5].
Ribotype 176 is closely related to 027 [6,7] and can be misidentified by commercial tests targeting a singlebase-pair deletion at nucleotide 117 in the C. diffici-letcdC gene [8].In 2009, shortly afterward the 2008 study, the occurrence of ribotype 176 was reported in certain areas of the country (Eastern Bohemia and Moravia) [9].This ribotype has persisted in the Czech Republic [8] and was also reported in Poland in 2008-13, which borders the country [10,11].
Results from the second study -involving 20 European countries -revealed an increasing CDI incidence rate in the Czech Republic (4.4 cases in 2011-12/10,000 patient bed-days and 6.2 cases/10,000 patient beddays in 2012-13) [2].This observation prompted us to determine the CDI incidence in 2014 in a number of hospitals distributed across the Czech Republic (n = 18) and to gain an insight into the prevailing C. difficile ribotypes.

Methods
A CDI case was defined as a hospitalised patient (more than two years-old) with both diarrhoea and laboratory confirmation of CDI by a positive test result for the presence of GDH and toxin A/B and/or the detection of a toxin-producing C. difficile strain using toxigenic culture or nucleic acid amplification test (NAAT) in the stool sample.
Testing for CDI was requested by the attending physician based on clinical symptoms indicating CDI (primarily diarrhoea).Hospital-associated and community-associated CDI cases were included in the analysis.
During 2014, hospital microbiology departments of the 18 selected hospitals were asked to send C. difficile isolates cultured from stool samples from hospitalised CDI patients to the Department of Medical Microbiology of the University Hospital Motol in Prague.
Mean CDI incidence and CDI testing frequency for all participating hospitals was calculated using the total number of admissions, total number of patient beddays, number of non-duplicated glutamate dehydrogenase (GDH) and toxin A/B positive tests performed in 2014, using information obtained from the participating hospitals.The hospitals also provided information about their CDI laboratory diagnostic algorithms.
PCR ribotyping based on capillary electrophoresis was performed according to the method described by Stubbs et al. [13].The results were compared with data in WEBRIBO, a web-based database containing a broad spectrum of uploaded capillary electrophoresisribotyping profiles [14], and profiles from an international capillary electrophoresis-ribotyping validation study [15].The diversity of ribotypes for each hospital was calculated using the Shannon index [16], for which a higher value is an indicator of greater diversity.
For MLVA, five regions with short tandem repeats were sequenced: A6Cd, B7Cd, C6Cd, G8Cd [17] and CDR60 [18], with a change of reverse primer for G8Cd, as described elsewhere [19].The number of tandem repeats was counted manually after software processing (Sequencing Analysis Software, Applied Biosystems).The sum of tandem repeat differences (STRD) in five loci determines the genetic relatedness of isolates.Minimum spanning trees were created using Bionumerics v5.1 (Applied Maths).A clonal complex was defined as an STRD ≤ 2, a genetically related cluster as an STRD ≥ 3 to ≤ 10 [17].

Participating hospitals
A total of 18 hospitals, covering the country's major regions, voluntarily participated in the survey: seven tertiary care institutions, 10 secondary care facilities and one specialised centre.The size of hospital is indicated by the number of beds in 2014 (Table 1).These 18 hospitals represented about 30% of hospital-bed capacity in the Czech Republic in 2013 [20] (2014 data unavailable).Their location is shown in Figure 1.

C. difficile infection testing algorithms
Four different CDI testing algorithms were used during the study period (Table 1).All hospitals in the study used the detection of GDH and toxins A/B as the first (screening) part of their testing algorithm: 14 used lateral flow immunoassay (LFIA), three used a chemiluminescent immunoassay (CLIA) and one a chromatographic immunoassay (CIA).
A total of 16 hospitals performed anaerobic culture of GDH-positive and toxin-positive or toxin-negative samples, but only two of these tested toxin production or detected the presence of genes for toxin production of isolated C. difficile strains (one by LFIA and one by toxin gene multiplex PCR).The remaining two hospitals, which did not routinely perform anaerobic culture, used PCR detection of the presence of C. difficile toxin Of the 18 hospitals, 10 used a commercial PCR test, eight for rapid diagnosis if requested by the physician.In total, 774 C. difficile isolates were available for further analysis in our study: 378 were from male patients (49%) and 396 from female patients (51%).The mean age was 68 years (SD: 20); the median was 72 years (range: 2-101).Of the 774 patients, 537 (69%) were aged 65 years or older.

PCR ribotypes of C. difficile isolates
Of the 774 C. difficile isolates, 737 (95%) belonged to 33 different ribotypes, and 37 (5%) were defined as new ribotypes, as their electrophoretic profiles differed from each other and did not match any in the WEBRIBO database.
MLVA of five variable-number tandem repeat loci was performed for the 225 isolates of ribotype 176 and 184 isolates of ribotype 001, and two minimum spanning trees were generated.
In total, 27 clonal complexes comprising 190 isolates (84%) were found in the minimum spanning tree of PCR ribotype 176 isolates (Table 2).For each clonal complex, the number of isolates/number of hospitals in which they were found are shown in parentheses: CC1 (52/10); CC2 (19/4); CC3 (19/1); CC4 (11/3); CC5 (10/3); CC6 (11/1); Each hospital is represented by a different colour (see key).The numbers in the circles represent the number of C. difficile PCR ribotype 176 isolates.If the number is greater than one, it represents the number of isolates with a sum of tandem repeat differences (STRD) = 0 (i.e.100% identical in five variable-number tandem repeat loci).The numbers on the lines represent the STRD between isolates.

Figure 4
Minimum spanning tree of Clostridium difficile PCR ribotype 001 isolates from 14 of 18 hospitals participating in survey of incidence of C. difficile infection, Czech Republic, 2014 (n = 184) MLVA showed an STRD ≥ 3 to ≤ 10 in 33 isolates and an STRD > 10 in three isolates (Figure 3).

Discussion
In 2008, three Czech tertiary care hospitals participated in the European C. difficile infection study (ECDIS) [1].In 2012-13, 10 Czech hospitals (nine tertiary care, three of which had participated in the 2008 study, and one secondary care) took part in the European, multicentre, prospective, biannual, point-prevalence study of CDI in hospitalised patients with diarrhoea (EUCLID) [2].Our current study, which involved 18 hospitals distributed across the Czech Republic (including seven tertiary, 10 secondary and one specialised healthcare facility) reflects better the CDI epidemiological situation in the country.Of these 18 hospitals, eight (seven tertiary care and one secondary care) also participated in EUCLID.
In the Czech Republic, it is mandatory to report cases of CDI to EPI-DAT, the Czech reporting system for infectious diseases, but CDI is reported as 'other bacterial intestinal infections'.Colonisation by C. difficile is not mandatorily reportable.An increasing incidence of other bacterial intestinal infections was observed, from 26.4 per 100,000 habitants in 2005 to 64.3 per 100,000 habitants in 2014 [21].As it is impossible to determine which of these infections are CDIs, however, CDI incidence data among hospitalised patients can only be derived from our study and the European studies mentioned above.
The results of our study showed a mean incidence of CDI per hospital of 6.1 cases per 10,000 patient beddays and 37.8 cases per 10,000 admissions.Compared with incidence data for the Czech Republic in the 2008 European study [1], the incidence of CDI in the country has dramatically increased.Our findings are similar to those of EUCLID, which reported an incidence rate of 6.2 CDI cases per 10,000 patient bed-days in 2012-13 for the Czech Republic [2].
Our study also showed that the mean reported testing frequency was 39.5 tests per 10,000 patient bed-days, which is 1.7 times less than the mean testing frequency reported in EUCLID (65.8 tests per 10,000 patient beddays) and almost three times less than the mean testing frequency reported for the United Kingdom (139 tests per 10,000 patient bed days) [2].This indicates that CDI in the Czech Republic is most likely underdiagnosed and highlights the need for improvement of clinical awareness and laboratory algorithms (by adding a confirmatory test for GDH positive and toxin A/B-negative stool samples from patients with clinical symptoms of CDI).
It should be noted that considerable variation in CDI incidence was seen between the 18 participating hospitals.The highest incidence seen in Hospital K is probably due to the fact that this hospital also had the highest testing frequency, as there were sufficient local financial sources for extensive CDI testing.The high incidence in this hospital had a considerable impact on the SD of the mean incidence for all 18 hospitals.Despite the high incidence, the number of isolates submitted during the study was small.The sending of strains was voluntary and this hospital was unable to send a representative number of isolates.
All 18 participating hospitals used GDH testing, as a recommended screening step [22].A total of 14 used a lateral flow immunoassay for a single GDH and toxin A/B test as the first step of their testing algorithm; the other four used these tests separately.The use of two separate tests is more economical because testing can be stopped when samples are GDH negative, as a GDH-negative result has a high predictive value for the absence of CDI [23].
Of the 18 hospitals, eight did not confirm toxin production in GDH-positive and toxin A/B-negative stool samples, although they performed anaerobic culture; thus their testing algorithm were suboptimal.[10,11].In 2013, the first sporadic case of an imported infection caused by PCR ribotype 027 was found in the Czech Republic [25].In the study presented here, we diagnosed two new CDI cases due to PCR ribotype 027 infection: one was a man in his 30s, the other a man in his 70s.
Whereas PCR ribotype 176 has been only reported from two countries (Czech Republic and Poland) [10,11,25], which neighbour each other, PCR ribotype 001 has been problematic for a long time in many European countries [1,2,26].It has dominated, as in Slovakia in 2012 [27], or has occurred together with PCR ribotype 027, as reported from Germany (Hesse region) in 2011-13 [28], the north-east of England in July 2009 to December 2010 [29] and Scotland in November 2007 to December 2009 [30].It has also occurred together with other PCR ribotypes, such as 014/020 and 126/078, in a single-day study in Spain [31].
MLVA of the two predominant ribotypes identified in our study revealed close genetic relatedness between isolates of each ribotype.The occurrence of 100%-identical (STRD = 0) PCR ribotype 176 isolates in 11 hospitals and PCR ribotype 001 isolates in seven hospitals, suggests clonal clusters within and between healthcare facilities, probably due to ineffective hospital infection control measures and transfer of patients between healthcare facilities who were in fact CDI cases but had not been diagnosed.This is supported by the observation of clonal complexes in tertiary and secondary hospitals in the same region.The question remains as to which specific molecular characteristics of PCR ribotypes 176 and 001 allow them to spread rapidly within healthcare facilities in contrast to the other less frequent PCR ribotypes identified in the study.
Antibiotic susceptibility testing of C. difficile isolates was not performed in this study but multiresistance of PCR ribotype 176 isolates [32,33], as well as PCR ribotype 001 isolates, has been reported [26,34].The results of a recently published European study on antibiotic resistance among prevalent C. difficile ribotypes showed the Czech Republic as a country with a high cumulative resistance score (4-5), calculated based on susceptibility to nine antimicrobials tested [26].
An important limitation of our study is the lack of clinical patient data.The Czech national reference centre for healthcare-associated infections is currently organising the implementation of CDI surveillance based on the recent CDI surveillance protocol from the European Centre for Disease Prevention and Control [35].The first national CDI incidence data, including clinical data on CDI patients and data on antibiotic susceptibility to metronidazole, vancomycin and moxifloxacin of C. difficile isolates, should be available in 2016 (CDI surveillance started in April 2016 in the Czech Republic).

Conclusion
The results of our study showed an unfavourable CDI epidemiological situation in the Czech Republic in 2014 caused by the occurrence of epidemic PCR ribotypes 176 and 001.The absence of national surveillance at that time, the low frequency of testing and variability in testing algorithms probably contributed to the spread of these PCR ribotypes.
A Czech standardised CDI testing protocol and the implementation of CDI surveillance in a large number of hospitals is urgently needed for monitoring, management and reduction of these infections in the Czech Republic.

Members of the Czech Clostridium difficile study group (in no particular order)
authors approved the final version of the manuscript and agreed with its submission.

Figure 1
Figure 1 Location of hospitals participating in survey of incidence of Clostridium difficile infection, Czech Republic, 2014 (n = 18)

Table 1
Characteristics of hospitals participating in survey of the incidence of difficile infection, CzechRepublic, 2014  (n = 18) [16]: anaerobic culture on selective media; CDI: Clostridium difficile infection; CIA: chromatographic immunoassay (two separate tests for GDH and toxins A/B); CLIA: chemiluminescent immunoassay (two separate tests for GDH and toxins A/B); GDH: glutamate dehydrogenase; LFIA: lateral flow immunoassay (simultaneous tests to detect GDH and toxins A/B; NAAT: nucleic acid amplification test; S: secondary care hospital; SC: specialised centre; SD: standard deviation; T: tertiary care hospital; TC: anaerobic culture on selective media followed by LFIA (Hospital B) or NAAT (Hospital L). a Calculated using the Shannon index[16].

Table 2
MLVA characteristics of Clostridium difficile PCR ribotype 176 isolates (n = 225) from 17 of 18 hospitals participating in survey of incidence of C. difficile infection, Czech Republic, 2014

Table 3
MLVA characteristics of Clostridium difficile PCR ribotype 001 isolates (n = 184) from 14 of 18 hospitals participating in survey of incidence of C. difficile infection, Czech Republic, 2014 a Sum of tandem repeat differences (STRD) = 0.