Clostridium difficile colonization in healthy adults: transient colonization and correlation with enterococcal colonization

Clostridium difficile is the principal pathogen that causes antibiotic-associated diarrhoea and colitis (Lyerly et al., 1988). It has been well documented that this organism spreads nosocomially and causes hospital outbreaks of C. difficile-associated diarrhoea (CDAD) in various clinical settings (Cartmill et al., 1994; Samore et al., 1996). In the hospital setting, infected and colonized patients and contaminated environments have been implicated as potential sources of C. difficile (Gerding et al., 1986). C. difficile colonization of asymptomatic humans has been studied by various groups (Nakamura et al., 1981; Viscidi et al., 1981; Wilson et al., 1982; Kobayashi, 1983; Aronsson et al., 1985). Recently, we reported the colonization and transmission of C. difficile in healthy individuals in Japan. We also demonstrated that C. difficile could be reisolated from one-third of C. difficile-positive subjects after a half-year interval; half of these latter subjects were found to harbour a different strain on the second isolation (Kato et al., 2001).

The aim of the present study was to further investigate the colonization status of C. difficile in healthy individuals. Moreover, intestinal microflora of C. difficile-positive and -negative subjects was compared, in order to analyse various host-related factors that may affect C. difficile colonization.

Subjects.
In total, 139 healthy adults (age range, 1965 years; median, 22 years) from two study groups were examined by stool culture for intestinal colonization by C. difficile (Table 1). The study groups were different from those in our previous study (Kato et al., 2001). None of the subjects had had diarrhoea or had been administered antimicrobial agents for at least 4 weeks before examination. One group, which consisted of 85 individuals, was a class of university students and the other, which comprised 54 individuals, was a group of employees at a company. For the student group, all subjects were examined three times at intervals of 3 months. Subjects who were C. difficile-positive in at least one of the three examinations were examined once more, i.e. four times in total. All subjects in the employee group were examined four times at intervals of 3 months. Examination of intestinal microflora was carried out for seven C. difficile-positive subjects in the student group (Table 2; S-1, S-2, S-3, S-4, S-5, S-6 and S-7) and nine age-matched, C. difficile-negative subjects that were also from the student group.

Table 1. Asymptomatic intestinal colonization by C. difficile in healthy subjects

Table 2. Types of C. difficile isolates from healthy subjects All 137 isolates, representing a maximum of five (15) colonies from each specimen, were examined. NT, Non-typable by PFGE, due to DNA degradation; -, C. difficile-negative.

Bacterial isolation and identification.
Stool specimens were collected and frozen at -80 °C until they were used for isolation of C. difficile. In examinations of intestinal microflora, stool specimens were collected immediately after defecation and were put into an anaerobic jar with an AnaeroPack (Mitsubishi Gas Chemical Company). Culture was started within 1 day of collection. For isolation of C. difficile, stool specimens were homogenized with an equal volume of ethanol for spore selection and were cultured on cycloserine/cefoxitin/mannitol agar (CCMA) (Kato et al., 2001). To investigate possible colonization of an individual by multiple strains, colonies (a maximum of five) were isolated randomly from the primary culture plate for each specimen and subcultured on CCMA. C. difficile was identified as described previously (Kato et al., 1998). Examinations of intestinal microflora were performed by using the following selective and non-selective media: VLM (total anaerobes; Becton Dickinson) (Morotomi et al., 1981), VLM-KV (Bacteroides; Becton Dickinson) (Morotomi et al., 1981), MTP (bifidobacteria; Eikennkagaku) (Morotomi et al., 1981), CW (lecithinase-positive clostridia; Nikkenseibutsu), LBS (lactobacilli; Becton Dickinson), TSA (total aerobes; Nikkenseibutsu), DHL (Enterobacteriaceae; Nikkenseibutsu), COBA (enterococci; Becton Dickinson) (Petts, 1984) and #110 (staphylococci; Nikkenseibutsu). Stool samples were homogenized and diluted serially (tenfold) with anaerobic dilution buffer (1.7 mM KH₂PO₄, 1.3 mM K₂HPO₄, 7.7 mM NaCl, 1.7 mM (NH₄)₂SO₄, 0.2 mM MgSO₄, 0.2 mM CaCl₂, 28 mM Na₂CO₃, 3 mM L-Cysteine.HCl, 0.0001 % resazurin). A 50 µl aliquot of diluted samples from 10^-1 to 10^-8 was spread on each plate and incubated under either anaerobic or aerobic conditions. A 500 µl aliquot of diluted samples from 10^-6 to 10^-9 was spread on MLV and MLV-KV. The number of c.f.u. [log c.f.u. (g stool)^-1] was calculated after incubation. The MannWhitney U test was used for statistical analysis.

Typing of C. difficile isolates.
The toxin gene type of the isolates was determined by a previously described PCR assay system (Kato et al., 2001). In this system, C. difficile strains were classified as toxin A-positive, toxin B-positive (A⁺B⁺), toxin A-negative, toxin B-positive (A^-B⁺) or toxin A-negative, toxin B-negative (A^-B^-). Typing analysis by using PCR ribotyping and PFGE was performed as described previously (Kato et al., 2001). Isolates with patterns that differed by one or more major bands were assigned to different PCR ribotypes; differences in faint bands were ignored. Major PFGE types were defined by differences in more than three fragments; these major types were subtyped further according to the criteria described by Tenover et al. (1995).

Asymptomatic colonization and persistence in healthy adults
In both groups, C. difficile-positive individuals were identified at all examinations (which were performed three times in the student group and four times in the employee group) (Table 1). Colonization rates ranged from 2.4 to 13.0 %. In general, C. difficile was isolated from 34 (7.1 %) of 478 samples. In the student group, seven subjects who were C. difficile-positive at any of the first, second or third examinations were examined one additional time. In the fourth examination, four subjects were found to be positive for C. difficile.

In the student group, the number of subjects from whom C. difficile was isolated once, twice, three times or four times was two (subjects S-6 and S-7, 2.4 %), two (S-4 and S-5, 2.4 %), one (S-3, 1.2 %) and two (S-1 and S-2, 2.4 %), respectively (Table 2). In the employee group, the number of subjects from whom C. difficile was isolated once, twice, three times or four times was eight (C-4 to C-11, 15.0 %), one (C-3, 1.9 %), one (C-2, 1.9 %) and one (C-1, 1.9 %), respectively. Overall, among the 18 C. difficile-positive subjects, the number from whom C. difficile was isolated once, twice, three times or four times was 10 (55.6 %), three (16.7 %), two (11.1 %) and three (16.7 %), respectively.

A maximum of five colonies was picked from each C. difficile-positive specimen; these colonies were examined by toxin gene typing and PCR ribotyping. All isolates from each specimen gave an identical toxin gene type and PCR ribotype in all 34 specimens. Therefore, one isolate from each of the 34 specimens was analysed further by PFGE. All 34 isolates were resolved into 19 ribotypes (Table 2). Four isolates were not typable by PFGE, due to DNA degradation during sample processing. The 30 PFGE-typable isolates were classified into 17 major types. Among the four non-typable isolates, two PCR ribotypes were identified. All isolates with the same PFGE type showed the same PCR ribotype; in other words, one PFGE type corresponded to one PCR ribotype. In the student group, different subjects were colonized by different PCR ribotype/PFGE types, whereas the same PCR ribotype/PFGE types of C. difficile were isolated from different subjects in the employee group: ntt/R115 was isolated from subjects C-1, C-2 and C-7 and yb51/R426 was isolated from subjects C-3, C-9 and C-11.

Continuous colonization by the same PCR ribotype/PFGE type was observed in three subjects: S-1 (isolated four times), S-3 (three times) and C-2 (three times) (Table 2 and Fig. 1). On the other hand, the PCR ribotype/PFGE type changed in the other subjects (S-2, S-4, S-5, C-1 and C-3). It is of note that in subjects S-2 and C-1, from whom C. difficile was isolated four times, the same PCR ribotype/PFGE type was observed on the second and third examinations, respectively.

(57K): Fig. 1. Representative PCR ribotypes and PFGE patterns of C. difficile isolates. M, Standard 100 bp DNA ladders (PCR ribotype) and chromosomal DNA from Saccharomyces cerevisiae (PFGE patterns) as molecular size standards; lanes 14, isolates from S-1 of the student group; lanes 58, isolates from C-1 of the employee group.

Regarding the toxin gene type, 22 (65 %) and 12 (35 %) of the 34 subjects were colonized by A⁺B⁺ and A^-B^- C. difficile, respectively.

Composition of the intestinal microflora in C. difficile-positive and -negative subjects
Composition of faecal microflora was compared in seven C. difficile-positive (aged 2023) and nine C. difficile- negative (aged 2023) subjects in the student group (Table 3). The age distribution of the subjects was adjusted in order to compare the two groups. C. difficile-positive subjects were significantly more densely colonized by enterococci (P < 0.05) than C. difficile-negative subjects. Moreover, it was shown that all subjects who were found to be C. difficile-positive three or four times were colonized by a concentration of 10⁷ enterococci (g stool)^-1 or more, in comparison with the nine C. difficile-negative subjects, among whom only one was colonized by a level of 10⁷ enterococci (g stool)^-1. The other eight subjects in that group were colonized by fewer than 10⁷ enterococci (g stool)^-1 (Table 4). Percentage colonization and viable counts of other bacteria tested were similar in C. difficile-positive and -negative subjects.

Table 3. Composition of faecal microflora in C. difficile-positive and -negative subjects

Table 4. Viable counts of enterococci in the faeces of C. difficile-positive and -negative subjects

In previous reports of intestinal colonization of healthy adults by C. difficile, colonization rates ranged from 0 to 17.5 % (Nakamura et al., 1981; Viscidi et al., 1981; Wilson et al., 1982; Kobayashi, 1983; Aronsson et al., 1985). Fekety & Shah (1993) reported that the faeces of about 5 % of healthy adults are colonized by toxigenic C. difficile. In our previous report (Kato et al., 2001), colonization rates of toxigenic and non-toxigenic C. difficile ranged from 4.2 to 15.3 % in the surveyed groups; 4.5 % of healthy adults were colonized by toxigenic C. difficile. In this study, colonization rates of toxigenic and non-toxigenic C. difficile ranged from 2.4 to 13.0 % in seven inspections (Table 1), and that of toxigenic C. difficile was 4.0 % overall. Colonization rates in the present study were similar to those in previous reports.

In our previous paper, we reported the possibility that cross-transmission of C. difficile can occur not only in nosocomial settings, but also among healthy adults in community settings (Kato et al., 2001). In this study, C. difficile strains with the same PCR ribotype/PFGE type were isolated from different subjects in the employee group. These results suggest that cross-transmission of C. difficile may be relatively common among healthy individuals. However, there is a possibility that spread was due to a common source of C. difficile in the work environment. Further studies are required to clarify this point.

In our previous report, C. difficile was reisolated from 32 % of C. difficile-positive subjects after an interval of 6 months; 50 % of these individuals harboured a different strain on the second investigation (Kato et al., 2001). Among the 18 C. difficile-positive subjects in the present study, the number of subjects from whom C. difficile was isolated once was 10 (55.6 %), and the number of subjects from whom C. difficile was isolated twice, three times or four times was eight (44.4 %) (Table 2). Among these eight subjects, only three (37.5 %) were colonized continuously by the same strains. These results demonstrate that colonization of healthy individuals by C. difficile is transient in many cases. However, the present study indicates clearly that there are healthy individuals who are continuously colonized by C. difficile, although this is rare. In all three subjects who were C. difficile-positive four times (S-1, S-2 and C-1), the C. difficile isolates from each subject were different from each other and only one of the three subjects (S-1) harboured the same type continuously. These findings suggest that factors other than bacterial properties of C. difficile strains may play an important role in continuous colonization by C. difficile. Considering that all five subjects (S-1, S-2, S-3, C-1 and C-2) from whom C. difficile was isolated three or four times harboured organisms of the same type on at least two consecutive examinations, the subjects themselves may have contaminated their environment, leading in turn to repeated infections from the environment.

To examine possible host-related factors that may affect colonization by C. difficile, we examined the hosts microflora and found that the number of enterococci in faeces was significantly higher among C. difficile-positive subjects. In addition, we recently encountered two additional students who were colonized persistently by C. difficile, who were also colonized by a level of enterococci that exceeded 10⁸ (g stool)^-1 (data not shown). Hopkins & Macfarlane (2002) also demonstrated that levels of enterococci were higher in the faeces of CDAD patients, compared to healthy adults. Interestingly, colonization by C. difficile occurs at high frequency in infants (Stark et al., 1982), who are generally colonized densely by enterococci (Mitsuoka et al., 1974). Taken together, these data suggest that dense colonization of the intestine by enterococci may be associated with C. difficile colonization.

Interactions between C. difficile and intestinal flora have been discussed in the literature. Growth and colonization of C. difficile was shown to be inhibited by the following factors of other intestinal flora: depletion of amino acids, decrease in pH, presence of volatile fatty acids (Yamamoto-Osaki et al., 1994) and competition for association with mucosal surfaces, mucin and N-acetylglucosamine (Borriello, 1990). In this study, we examined whether or not the culture supernatants of enterococci have an effect on the rate of germination of C. difficile spores (the germination frequency of which was found to be quite low; Nakamura et al., 1985) and on the binding of C. difficile to cultured cells; however, in neither of these cases were any effects observed (data not shown). The reason for dense colonization of C. difficile-positive individuals by enterococci remains unknown; further studies will be required to account for these findings.

We would like to thank T. Muraki of the Ishikawa Yakult Company for sample collection. This work was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology, Japan, and by the Yakult Bio-Science Foundation, Japan.