Moving the Needle on Clostridioides difficile Research in Children: A Major Step Forward but Challenges Remain

(See the Major Article by Wolf et al on pages 2581–8.)

“Not small adults” is the pediatrician’s mantra that reinforces that children are clinically distinct from adults and justifies the need for focused research in pediatric populations. Clostridioides (Clostridium) difficile infection (CDI) embodies this mantra. Infants are particularly distinct from adults as it pertains to C. difficile. Carriage with toxigenic C. difficile is identified in more than 40% of infants in the first year of life [1, 2]. However, for reasons not yet clearly delineated in humans, infants rarely, if ever, develop symptomatic CDI despite frequent detection of C. difficile and its toxins in stool [1, 2]. This unique aspect of C. difficile epidemiology prompted a viewpoint authored by a multinational group of CDI experts that challenged the regulatory requirement for investigation of C. difficile treatments in children younger than 2 years old. The authors reasoned that the lack of evidence of CDI in this age group argues against an unmet need for CDI therapy in this young patient population [3]. Not only are there ethical issues with investigating a drug in a population unlikely to be suffering from the disease of interest, doing so also introduces bias into clinical trials. Nonetheless, this regulatory requirement remains.

C. difficile carriage decreases over time, reaching its general population prevalence nadir of approximately 1–3% at approximately 2–3 years of age [1, 2]. However, C. difficile carriage is much higher in particular pediatric patient populations. For example, approximately 25% of hospitalized children [4], approximately 30–50% of children with cancer [5], and approximately 20% of children with inflammatory bowel disease (IBD) [6] have asymptomatic C. difficile carriage. Because other infectious and noninfectious diarrheal etiologies are common in these populations, and because the majority of US children’s hospitals use nucleic acid amplification tests (NAATs) that do not alone differentiate C. difficile carriage and infection [7], there is risk of misdiagnosis of CDI in children with asymptomatic C. difficile carriage [8]. Although toxin assays are generally more clinically predictive of CDI than NAATs, suboptimal sensitivity has limited their use in the clinical setting [8]. Thus, highly sensitive NAATs may avoid underdiagnosis of CDI and may have value in clinical practice, particularly at institutions with preagreed criteria for patient stool submission [1]. However, using NAATs alone to identify CDI cases for enrollment into clinical trials may lead to CDI misdiagnosis and introduce bias into clinical trials. Although toxin assays also may be positive in patients with C. difficile carriage and are not impervious to this NAAT limitation, it is generally considered that this occurs less frequently than in NAATs. Because of these challenges, clinical practice and clinical trials may require different CDI case ascertainment strategies in children, both in terms of diagnostic test selection and efforts to rule out alternative and/or concomitant diarrheal etiologies.

Recent guidance [1] from the Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America removed metronidazole as a recommended treatment option for nonsevere CDI in adults; vancomycin and fidaxomicin were strongly recommended as first-line therapies based on high quality of evidence. However, in children, metronidazole remains a recommended option for initial or first recurrence of nonsevere CDI, primarily because robust CDI treatment data were lacking in children.

A major step forward in addressing this evidence gap is taken with the publication of the SUNSHINE trial, the first phase 3 pediatric clinical trial of CDI treatment, in this issue of Clinical Infectious Diseases [9]. In this multicenter trial, 142 children with CDI were randomized (2:1) and completed 10 days of fidaxomicin or vancomycin. Children were included if they met clinical criteria for CDI (at least 3 unformed stools per day) and tested positive for C. difficile (either detection of toxins A/B or toxigenic C. difficile in stool). To avoid confounding from other potential diarrheal etiologies, a negative rotavirus test was required for young children, and children with IBD were excluded. Confirmed clinical response (CCR) was defined as initial clinical response at the end of therapy (EOT) with no further requirement for CDI therapy at 2 days after EOT. CCR frequency was not significantly different between the fidaxomicin and vancomycin groups (78% vs 71%). However, among those achieving CCR, those in the fidaxomicin group had a significantly lower frequency of CDI recurrence (12% vs 29%). Similarly, those in the fidaxomicin group had a significantly higher frequency of global cure (GC) (ie, CCR without CDI recurrence; 68% vs 50%).

Similar to the phase 3 fidaxomicin trial in adults [10], fidaxomicin was associated with a lower CDI recurrence rate in children. This is very important because it will lead to the availability of a new safe and effective CDI treatment option in children that is supported by rigorous clinical investigation. However, a deeper dive into the SUNSHINE trial data raises several important questions. Overall, CCR and GC frequency with both fidaxomicin and vancomycin was lower in children [9] than in adults [10] (Table 1). Of note, both trials used similar outcome definitions, permitted enrollment of subjects with 1 CDI episode in the prior 12 weeks, and excluded subjects with fulminant CDI; and neither trial excluded immunocompromised subjects. Fidaxomicin pharmacokinetics (high fecal concentrations and very low systemic absorption) were generally similar between children and adults, and although antimicrobial susceptibilities were not reported in the pediatric trial, a prior single-center pediatric study [11] reported C. difficile resistance to vancomycin and fidaxomicin occurred similarly infrequently as the adult fidaxomicin trial [10]. Thus, resistance to either drug is unlikely to account for the lower CCR in children.

Table 1.

Comparison of Frequency of Confirmed Clinical Response and Global Cure in Children (n = 142) and Adults (n = 548) Enrolled in a Phase 3 Trial of Fidaxomicin Versus Vancomycin (Per-Protocol Analyses)

	Fidaxomicin, %	Vancomycin, %
CCR
Children	78	71
Adults	92	90
GC
Children	68	50
Adults	78	67

Data from references [9] (children) and [10] (adults). Abbreviations: CCR, confirmed clinical response; GC, global cure.

The SUNSHINE trial enrolled children 0–2 years old, a group unlikely to suffer from CDI. Further, the adult trial limited enrollment to subjects whose stool was toxin positive. Although the SUNSHINE trial permitted enrollment of NAAT-positive children without confirmation of toxin positivity (only 46% were tested by a toxin assay), a subgroup analysis of toxin-positive children was performed post hoc. Upon a subgroup analysis of toxin-positive children age 2 years or older (ie, the subgroup with the highest CDI probability), CCR remained relatively low in both groups. There are several possible explanations for this interesting finding. Although it is possible that children, in general, may have a poorer initial response to therapy for bonafide CDI compared with adults, a biological explanation to support this is unknown and, in my opinion, not likely. More likely is that some children testing positive for C. difficile suffer from an alternative or concomitant diarrheal etiology, in either case limiting response to CDI therapy. Possible alternative diarrheal etiologies are supported by lower fidaxomicin CCR in children younger than 2 years old and reported use of medications for constipation among 11% of subjects. Possible concomitant diarrheal etiologies in immunocompromised children are supported by the improved CCR among nonimmunocompromised children. Another possibility is that some children may have had an analytical or diagnostic false-positive toxin test, supporting prior observations that false positives are not limited to NAATs [1]. Although compared with NAATs detection of toxin in stool is generally considered to be more clinically predictive of CDI, these data are primarily derived from adult studies [8], and frequent analytical false-positive toxin results have been previously reported in children at a single center [12]. It has been recently observed that infants with C. difficile carriage develop a humoral immune response following early-life exposure to toxigenic C. difficile [13]. It is conceivable that antitoxin antibodies that develop in infancy persist and protect against CDI later in childhood. If so, this could limit the clinical predictive value of toxin testing in children with antitoxin antibodies because toxin-positive children could remain asymptomatic.

Despite these important unanswered questions, the SUNSHINE trial is a major step forward in addressing the need for evidence-based CDI treatment options in children. However, to continue to move the needle for C. difficile research in children, a deeper understanding of the pathogenesis, clinical microbiology, and immunology of CDI in children is needed. This important knowledge will guide the design of future trials that optimize CDI case ascertainment in children and limit bias potentially related to enrollment of children with alternative and/or concomitant diarrheal etiologies into CDI clinical trials.

Notes

Acknowledgments. L. K. K. is supported by the National Institutes of Health (grant number K23 AI123525).

Potential conflicts of interest. L. K. K. is a scientific advisor for Synthetic Biologics. The author has submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

(See the Major Article by...)