Detection of Clostridioides difficile toxin B gene: benefits of identifying gastrointestinal pathogens by mPCR assay in the diagnosis of diarrhea in pediatric patients.

BACKGROUND: In the pediatric population, severe Clostridioides difficile infection (CDI) sometimes occurs, but most cases are asymptomatic. The asymptomatic carriage rate in pediatric populations is reportedly higher than in the adult population. It is difficult to diagnose CDI, even if C. difficile is detected in children with diarrhea. This study aimed to evaluate the positivity rate of toxigenic C. difficile in the pediatric population with diarrhea.
METHODS: We collected and retrospectively analyzed gastrointestinal pathogen multiplex PCR results of 960 patients to estimate the positivity rate of toxigenic C. difficile in pediatric populations aged between 0 and 18 years.
RESULTS: The overall rate of C. difficile toxin B positivity was 10.1% in the stool samples. The positivity rate peaked in 1-year-old infants (29/153, 19.0%) and continually decreased thereafter. The positivity rate we observed was lower than the rates described in the literature. Remarkably, no C. difficile was detected in neonates. Antibiotic usage was inversely related to the positivity rate, especially in infants < 2 years of age. The odds ratio of antibiotics was 0.44 (95% confidence interval (CI) 0.28-0.68; P < 0.001). The presence of concomitant gastrointestinal pathogens was not associated with toxigenic C. difficile positivity.
CONCLUSIONS: Even though toxigenic C. difficile infection is neither an important nor a common cause of pediatric diarrhea, children can spread it to adults at risk of developing CDI. The pediatric population can act as hidden reservoirs for pathogenic strains in the community.

Background

Clostridioides difficile, formerly known as Clostridium difficile, is a spore-forming, obligate anaerobic, gram-positive bacillus acquired either from the environment or through the fecal-oral route [1]. It is known to cause a wide range of symptoms, from mild diarrhea to severe life-threatening complications such as toxic megacolon [2]. The major virulence factors of C. difficile are large clostridial toxins, toxin A and toxin B, which are encoded by tcdA and tcdB [3]. C. difficile infection (CDI) mainly occurs in healthcare-associated cases and adults. However, over the last decade, CDI has emerged as an important community-associated infection both in adults and children [4]. Approximately 4–5% of non-hospitalized healthy adults carry the pathogen in their intestinal flora [5], and varying positive rates of up to 70% have been reported in healthy newborns [6]. In children, the carrying capacity decreases with age, reaching adult levels of approximately 5% by the age of 2 [7]. According to the guidelines for pediatric consideration, because of the high prevalence of asymptomatic carriage of toxigenic C. difficile, testing for CDI should not be routinely performed in children under 12 months of age with diarrhea [1, 8]. If they have rare and severe symptoms of pseudomembranous colitis, toxic megacolon, or clinically significant diarrhea, C. difficile testing should be performed. In children aged between 1 and 3 years, a diagnostic workup for other diarrheal causes should be performed first, while C. difficile testing can be considered at later stages. Due to the unclear role of C. difficile in children with diarrhea, there are few reports on the positivity of toxigenic C. difficile in the pediatric population. Therefore, this study focuses on multiplex PCR (mPCR), which is increasingly being applied to detect gastrointestinal pathogens and provides additional information on the C. difficile toxin B gene (tcdB) in pediatric patients. In this study, we estimated the positivity rate of C. difficile tcdB and interpreted the results through electronic medical record review.

Methods

We reviewed 960 non-duplicated stool mPCR (Seeplex Diarrhea-B1/2 and V ACE Detection, Seegen, Korea) results obtained from pediatric patients up to the age of 18 [9] collected over 39 months (October 2014–December 2017) and submitted to a tertiary referral hospital in Seoul. The mPCR included C. difficile tcdB and 13 other diarrhea-causing pathogens (Salmonella spp., Shigella spp., Vibrio spp., Campylobacter spp., Escherichia coli O157:H7, Clostridium perfringens, Yersinia enterocolitica, Aeromonas spp., Verocytotoxin-producing E. coli, rotavirus group A, norovirus, astrovirus, and enteric adenovirus). The electronic medical records of the patients were reviewed to acquire information regarding age, length of hospital stay, underlying diseases (malignant neoplasm, hematology/immunology, endocrinology, cardiovascular, respiratory, digestive, inflammatory bowel diseases, and genitourinary disorder), previous history of antibiotics, and clinical diagnosis of C. difficile enterocolitis during the entire period of hospitalization with or without metronidazole or oral vancomycin treatment.

Unpaired t-test or Mann-Whitney U test was used for continuous data. Pearson’s chi-squared test or Fisher’s exact test was used for categorical data. The odds ratios of the antibiotic-treated versus naive groups were calculated. All statistical analyses were performed using MedCalc Statistical Software version 18 (MedCalc Software bvba, Ostend, Belgium; http://www.medcalc.org; 2018).

Results

B positivity andB negativity

The overall positivity of C. difficile tcdB, as determined by mPCR, was 10.1% (97/960). Patients with positive results were younger (median age, 1.6 years) than those with negative results (median age, 3.8 years) (P <0.01) (Table 1).

Table 1

Baseline characteristics of 960 pediatric patients with diarrhea based on Clostridioides difficile toxin B (tcdB) detected by multiplex PCR

	tcdB positive (N = 97)	tcdB negative (N = 863)	P
Age, median (95% CIa)	1.6 (1.4–2.2)	3.8 (3.3–4.5)	< 0.01
Age group (N)
0 (219)
Inpatient	12	140	0.09
Outpatient	11	56
Subtotal	23	196
1 (153)
Inpatient	18	95	0.16
Outpatient	11	29
Subtotal	29	124
2–6 (265)
Inpatient	11	204	< 0.01
Outpatient	15	61
Subtotal	26	239
7–12 (171)
Inpatient	7	137	0.42
Outpatient	3	34
Subtotal	10	161
13–18 (152)
Inpatient	9	125	0.36
Outpatient	0	27
Subtotal	9	143
Total (960)
Inpatient	57	674	< 0.01
Outpatient	40	189
Subtotal	97	863
Sex			0.64
Male	54	502
Female	43	361
Length of stay at testing, median (95% CI)	5 (3.6–6.0)	4 (4.0–5.0)	0.807
30-day mortality	2	10	0.346
Other gastrointestinal pathogens
Detected	30	295	0.52
Not detected	67	568

aConfidence interval (CI)

No tcdB was detected in neonates (0/13). They were admitted to the neonatal intensive care unit (NICU) and administered antibiotics. Their mean length of stay was 5.0 days from the day of testing. No other diarrheal pathogens were detected in these neonates. While the youngest tcdB-positive infant was a 4-month-old, the tcdB positivity rate among infants aged 1 month to 1 year was 11.2% (23/206).

The tcdB positivity peaked at 1 year of age (29/153, 19.0%) and was inversely correlated with age. In children aged 2–6 years, the positivity rate dropped to 9.8%, and this incidence decreased in the group aged 7–12 years (5.8%) and 13–18 years (5.9%) (Fig. 1). The tcdB positivity rates were higher in outpatients than in inpatients, except in the 13–18 years group (P < 0.001).

Fig. 1

Clostridioides difficile toxin B positivity detected by multiplex PCR of indicated age groups

Among inpatients, the difference in hospital length of stay between tcdB positivity and tcdB negativity was not statistically significant. Sex and 30-day mortality were not related to tcdB positivity. None of the underlying diseases were related to tcdB positivity. We categorized the underlying diseases into eight groups, and none of the odds ratios in each group reached statistical significance (data not shown).

Presence of concomitant gastrointestinal pathogens

Other gastrointestinal pathogens were detected in the stool samples from 325 patients (33.8%) using mPCR. Clostridium perfringens 32.6%, norovirus 20.9%, Campylobacter spp. 14.5%, and Salmonella spp. 10.2% were detected. Among them, 30 (9.2%) had C. difficile and other pathogens simultaneously. In 635 patients, no proven etiology of diarrhea was detected. Among them, 67 patients (9.0%) were tcdB-positive. Altogether, the presence of concomitant gastrointestinal pathogens did not affect the tcdB positivity rate (P = 0.52).

Antibiotic exposure

Antibiotic exposure did not increase tcdB positivity. The odds ratio of antibiotics in the antibiotic-treated group (N = 541) compared to that in the antibiotic-naïve group (N = 419) was 0.44 (95% confidence interval–CI: 0.28–0.67; P < 0.001). Interestingly, when we stratified the groups by age, the tcdB positivity was inversely proportional to antibiotic exposure in those under 7 years of age. The odds ratios of the groups over 7 years of age were not statistically significant (Table 2).

Table 2

Odds ratio of antibiotics exposed group compared to antibiotics naïve group stratified by age

Age (years)	Odds ratio	95% CI	P value
0	0.24	0.08–0.66	<0.01
1	0.35	0.14–0.79	0.01
2–6	0.38	0.17–0.87	0.02
7–12	1.35	0.34–5.41	0.67
13–18	1.23	0.32–4.78	0.76
Total	0.44	0.28–0.68	<0.01

aConfidence interval (CI)

Diagnosis and treatment of C. difficile infection

A review of the electronic medical records revealed that a total of 22 patients (22/960, 2.3%) were clinically diagnosed with CDI and treated with metronidazole or oral vancomycin, but nine had no proven existence of C. difficile tcdB (data not shown).

Discussion

In neonates, C. difficile frequently colonizes the gastrointestinal tract without causing disease since colonization rates are reportedly 25–36% at 1 month of age [7]. 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. 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 2 months [11]. 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. [12] surveyed three postnatal wards and reported a positivity rate of 2–52%, and their difference was statistically significant. They also found epidemiological clusters in ward environments. They suspected a nosocomial spreading, which caused the high prevalence in previous studies. Hospitals systematically develop many infection control measures, such as hand hygiene and standard precautions, which may result in a lower acquisition rate in neonates. Rousseau et al. [13] classified 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 was reportedly caused by modifications in the gut microbiota composition during a variable food trial.

The high colonization rate of C. difficile in infants could result from the commensal microbiota in the pre-weaning period, dominated by Bifidobacterium spp. and Lactobacillus spp., which are more permissive to colonization [14]. 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 16 S 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 [14]. 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 positivity rate of C. difficile (Table 1). The odds ratios of the age groups 1, 2–6, and total indicated that antibiotic usage is inversely related to the tcdB positivity rate. Antibiotic use is a major risk factor for adult CDI, and research by Donta and Myers conducted in 1980 using cell culture neutralization assay (CCNA) 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 [15]. However, in our study, only 5% (3/53) were positive for C. difficile tcdB 14 days after antibiotic therapy (data not shown). Similar to our results, in a study reported in 2012 using toxigenic C. difficile culture method, antibiotic exposure prior to C. difficile detection did not cause a difference between positive patients and the overall population [13]. Although the method used in this study showed higher sensitivity than cytotoxic neutralization analysis (CCNA), the positive rate was lower. In a study using CCNA [15], vaginal delivery and breastfeeding were related to high positive rates, with differences in environmental factors thought to be the cause.

Considering how toxigenic C. difficile is acquired in these age groups, our study suggests that multiple factors beyond antibiotic usage might affect the positivity rate. Our study was based on a molecular method using fresh stool specimens to detect the C. difficile tcdB gene. Molecular testing, which uses cell culture with frozen stool samples, has a higher sensitivity than other methods. Although non-toxigenic C. difficile was not included in this study, detecting toxic C. difficile 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 detected in culture samples of 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%, 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 through cytotoxicity neutralization assay, indicating a higher positivity rate in diarrhea patients. In the group with diarrhea, the possibility of C. difficile infection should be considered for some positive patients.

In our results, the clinical characteristics of the tcdB-positive group were different from those of adult CDI. Patients in the C. difficile tcdB-positive group were younger than those in the negative group. tcdB-positivity was found more in outpatients than in inpatients, and length of stay was shorter in this group than in the tcdB negative group. Underlying diseases such as neoplasm, hematologic, respiratory, genitourinary disorder, and inflammatory bowel disease were not statistically related to tcdB positivity, contrary to adult CDI.

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, tcdB positivity in this study included both patients with CDI and carriers with non-CDI diarrhea etiology. A limitation of our study is that not all tcdB positivity implies colonization. In addition, this study is a retrospective study; there is a possibility that the records were inaccurate, and some data might have been omitted.

In our study, 22 patients were clinically diagnosed with CDI and treated with metronidazole or vancomycin, yet nine of them 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 simultaneously. This is in accordance with another study which reported a simultaneous positive rate of > 50% with C. difficile [19]. We cannot define the remaining 70% as CDI because we could 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 known as risk factors for CDI, such as underlying disease, antibiotic exposure, and hospital administration, did not increase the CD positivity 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 positivity rate rather than CDI.

Conclusions

Clostridioides difficile is thought to be a hospital-associated infectious pathogen. The acquisition in the community seems to have prevailed due to improvements in individual hygiene levels and hospital infection control, at least in the pediatric population. This study demonstrated lower C. difficile positivity in the pediatric population than previously reported. Although toxigenic C. difficile infection is neither an important nor a common cause of pediatric diarrhea, children can spread it to adults at risk of developing CDI. Therefore, children can act as hidden reservoirs for pathogenic strains in the community. Monitoring of toxigenic C. difficile positivity in the pediatric population should be approached as an infection control measure, as well as individual diagnosis.

In stool samples of patients with diarrhea aged 0–18 years, C. difficile toxin B positivity by gastrointestinal pathogen mPCR was 10.1%. The presence of concomitant gastrointestinal pathogens was not associated with toxigenic C. difficile positivity; therefore, it is difficult to determine whether C. difficile is the cause of diarrhea in the pediatric population. The risk of transmission of CDI along with the difficulty of diagnosis should be noted.