In neonates, C. difficile frequently colonizes the gastrointestinal tract without causing the disease, considering that colonization rates are reportedly 25–36% at 1 month of age [7]. Moreover, studies reported in the 1980s showed that the isolation rate of C. difficile was very high. Al-Jumaili et al. [10] found that the isolation rate of toxigenic C. difficile increased progressively with infant age, from 7% at birth to 100% by 26–35 days. They detected 66% (61/94) of toxigenic strains from neonates aged 1 to 35 days in the special care baby unit. Donta et al. [11] reported that C. difficile toxin was detected in the feces, 10.5% in normal newborn infants, and five times higher (55%) in neonates in the NICU [11]. Vaginal delivery and breastfeeding were associated with increased rates of toxin carriage. A study performed three decades later found that 45% of infants carried C. difficile, with 13% carrying a toxigenic isolate among healthy infants (age, 0–3 years). Toxigenic strains appeared after four months in 40% of the infants under evaluation [12].
Unlike previous reports, we did not find any toxigenic C. difficile in newborns and infants under 4 months of age. In this study, antibiotics were administered to all 13 patients in the NICU and 74% of patients under 4 months of age. In neonates, antibiotic administration has been reported to delay C. difficile colonization for at least two months [13]. This may explain why neither neonates nor infants of up to 3 months of age had detectable C. difficile in this study. The odds ratio, which was statistically significant in age group 0, including neonates, indicated that antibiotic usage does indeed delay C. difficile colonization (Table 2).
Larson et al. [14]surveyed three postnatal wards and reported a positivity rate of 2–52%, and their difference was statistically significant [14]. They also found epidemiological clusters in ward environments. They suspected a nosocomial spreading, which caused the high prevalence in previous studies. Hospitals systematically developed many infection control measures, such as hand hygiene and standard precautions, which may result in a lower acquisition rate in neonates. Rousseau et al. [12] divided the acquisition period into the neonatal phase (early) and infant stage (4–6 months, late). Our youngest toxigenic C. difficile tcdB-positive infant was 4 months old; therefore, the subject would have been included in the “late acquisition” group in Rousseau’s study. Late acquisition has been reported to be caused by modifications in the gut microbiota composition during a variable food trial.
Figure 1 shows that positivity is the highest in the age group of 1 year in both outpatients and inpatients, and since then, the positivity decreases continually. The high colonization rate of C. difficile in infants could result from the commensal microbiota in the pre-weaning period, which is dominated by Bifidobacterium spp. and Lactobacillus spp., which is more permissive to colonization [15]. After solid food intake, the microbiota is similar to that of adults dominated by Bacteroidetes and Firmicutes spp. According to a longitudinal observation of the gut microbiome analyzed by 16S rRNA gene sequencing from an infant, the introduction of solid food at around 4 months resulted in a huge change in the microbiome, and the diversity of intestinal microbiota was related to C. difficile disappearance [15]. During the observation period, C. difficile counts varied with fluctuations of more than 105 and eventually disappeared at 12 months. This may explain our first detection of C. difficile tcdB in a 4-month-old infant.
We observed that antibiotic usage within 30 days did not increase the positive rate of C. difficile infection (Table 1). The odds ratios of the age groups 1, 2–6 and total indicated that antibiotic usage is inversely related to the tcdB positive rate. Antibiotic use is a major risk factor for adult CDI, and research by Donta and Myers showed that C. difficile toxin could be found in 85% of infants after 14 days of exposure to antibiotics, even when the toxin was not detected during antibiotic therapy [11]. However, only 5% (3/53) were positive for C. difficile tcdB 14 days after antibiotic therapy in our study (data not shown). Similar to our results, found antibiotic exposure prior to C. difficile detection not different between positive patients and the overall population [12].
Considering the process by which toxigenic C. difficile is acquired in these age groups, our study suggests that multiple factors beyond antibiotic usage might affect the positive rate. Our study was based on a molecular method using fresh stool specimens to detect the C. difficile tcdB gene. Molecular testing has a higher sensitivity than other methods, which used cell culture with frozen stool samples. Although non-toxigenic C. difficile was not included in this study, toxic C. difficile detection is important in clinical settings. Although molecular testing alone is too sensitive and not specific for diagnosing CDI [8], it is an appropriate test to estimate the presence of low concentrations, not infection status by C. difficile overgrowth, and production of abundant toxin.
In addition, C. difficile tcdB was detected in 10.1% of patients with diarrhea in the population under study (0–18 years of age). In a previous study, C. difficile was cultured in 7.0% of patients with diarrhea and 14.8% of patients without diarrhea between 2 weeks and 16 years of age. Therefore, the C. difficile isolation rate in patients without diarrhea was more than 50% higher than that in patients with diarrhea among outpatients [16]. Another study showed no correlation between diarrhea and C. difficile colonization rates in infants [17]. Further, a group aged over 8 years had an infection rate of approximately 5%, which was similar to that of healthy adults [5]. Among children under 15 years of age, Kim et al. [18] reported that 15.6% of the group with diarrhea and 6.7% of the control group had C. difficile toxin by cytotoxicity neutralization assay, indicating a higher positive rate in those with diarrhea. In the group with diarrhea, the possibility of C. difficile infection should be considered for some positive patients.
We observed that the C. difficile tcdB-positive group was younger than the negative group. More tcdB-positives were found in outpatients than that of inpatients, and their length of stay was shorter than that of the negative group. Underlying diseases such as neoplasm, hematologic, respiratory, genitourinary disorder, and inflammatory bowel disease were not statistically related to tcdB positivity, which was contrary to adult CDI. Although severe infection sometimes occurs, most cases in this age group are asymptomatic [19]. It has been suggested that C. difficile cannot play a pathogenic role, since immature intestinal epithelial cells of neonates and infants lack receptors that enable the invasion of C. difficile toxin A in a study of rabbit ileum [20].
The pediatric patients included in this study were not entirely healthy, since they had diarrhea that required hospital visits, during which stool samples were collected and tested for the presence of diarrhea-causing pathogens. Therefore, our study included both patients with CDI and carriers with non-CDI diarrhea etiology. The limitation of our study is that patients with CDI and carriers cannot be accurately discriminated.
Clinically significant diarrhea can be defined as frequent passage of loose stool. When a child has clinically significant diarrhea with proven toxigenic C. difficile presence and without other gastroenteric pathogens, it can be CDI [8]. In clinical practice, even when a diarrheal patient with toxigenic C. difficile is positive, it is challenging to distinguish true C. difficile infection from colonization.
In our study, 22 patients were clinically diagnosed with CDI and treated with metronidazole or vancomycin, yet nine had no proven existence of C. difficile tcdB (data not shown). Moreover, 30.9% (30/97) of C. difficile tcdB-positive patients showed positive gastrointestinal pathogens at the same time. This is in accordance with another study which reported a simultaneous positive rate of > 50% with C. difficile [21]. We cannot define the remaining 70% as CDI because we did not exclude all other etiologies.
We noticed that C. difficile tcdB positivity was not affected by concomitant gastrointestinal pathogens. This result suggests that most C. difficile tcdB positive cases are more likely to be colonization and not CDI. The clinical factors that were known as risk factors for CDI, such as underlying disease, antibiotic exposure, and hospital administration, did not increase the CD positive rate in this study. CDI cases were certainly included, but the rate did not appear to be substantial. Therefore, we may cautiously draw a sketch of pediatric CD colonization with this positive rate rather than CDI.