Pediatric Enteric Diagnostic Stewardship: The Right Test in the Right Context

In this issue of Pediatrics, Cotter et al¹  offer valuable analysis of the utility of molecular (nucleic acid amplification) enteric pathogen diagnostic technology in hospitalized children. These tests return highly worthwhile information for some infections, but many of their results are prone to misinterpretation. However, diagnostic guidelines based on classic microbiologic techniques do not perform well either.²  Molecular enteric microbiology is, therefore, a model opportunity for diagnostic stewardship to maximize worth and minimize wasteful expense.

Cotter et al¹  find that only 3.7% of tested hospitalized children benefitted from molecular diagnosis in terms of decreased length of stay and/or initiation of appropriate treatment. These data reinforce previous findings: enteric diagnostic use increases briskly after molecular testing becomes available and the tests find many unexpected signals reflecting difficult-to-interpret and/or nonactionable microbes. The authors candidly acknowledge the study’s limitations: nongeneralizability (their findings do not apply to outpatients) and reliance on historic comparison data. The benefit they cite applies to hospitalized children, and such populations are probably enriched for severe bacterial infections for which rapid and accurate molecular diagnosis has the greatest value; one would expect different yields in emergency facilities and ambulatory settings. Nonetheless, Cotter et al¹  reflect real-world practice.

Molecular enteric testing has undeniable appeal. Information is returned rapidly and can occasionally prompt timely treatment, such as antibiotics for Shigella and Campylobacter or antibiotic avoidance and volume expansion for Shiga toxin-producing Escherichia coli.^3,4  Molecular enteric testing can be an ally of antibiotic stewardship. For example, in the Netherlands, adults with diarrhea treated without microbiologic testing were more likely to receive antimicrobial agents not recommended by clinical guidelines than those whose treatments were informed by stool testing.⁵  Axelrad et al⁶  and Cybulski et al⁷  similarly reported 4.7% and 16.5% decreases, respectively, in empirical antimicrobial use in children and adults with diarrhea. These data are relevant to pediatrics: antimicrobial agents are used in 10% of gastrointestinal infections among outpatient US children ages 1 to 19 years,⁸  which almost certainly exceeds the proportion of cases in which they are warranted.^9,10  Also, although the preponderance of childhood enteric infections is caused by viruses and resolve without etiologic certainty, establishing norovirus as a cause of severe vomiting and diarrhea might avert radiologic testing and hospital admission.

However, molecular enteric tests have drawbacks. They might prompt antibiotic use even when current data do not support treatment, such as certain diarrheagenic E coli infections or Clostridioides difficile infection in infants, because panels return these results without the provider requesting that the identified agents be sought. The Children’s Hospital Colorado microbiology laboratory wisely masked such results when positive. Also, molecular diagnostics increasingly supplant standard culture. Although diagnostic laboratory decisions to forego culture are not a drawback of the technology per se, the consequence is that public health laboratories lose their ability to type strains to detect or monitor outbreaks.¹¹  The abandonment of pathogen isolation by culture also prevents studies of pathogen characteristics. Finally, absence of molecular proof is not proof of pathogen absence: Campylobacter and Salmonella can evade molecular detection.^12,13  These performance issues stem from processes that underlie regulatory approval and marketing by using data from limited numbers of specimens that are highly likely to contain or not contain the target pathogen(s). Because pretest probabilities in the real world are considerably lower, after-market performance is invariably worse. It is also concerning that there is little postmarket surveillance of diagnostics, so if cultures are abandoned, comparisons cannot be made.

When should one test for bacterial enteric pathogens using molecular techniques? We encourage seeking selected bacterial pathogens in children with high pretest probability of having an actionable infection (diarrhea of <1 week duration, visibly bloody stools, illness accompanied by severe abdominal pain and/or fever), which is the standard we also apply when deciding to perform stool cultures. Note, however, that children infected with bona fide bacterial enteric pathogens are often afebrile,¹⁴  so absence of fever should not dissuade a physician from seeking a diagnosis if other illness characteristics are present.

Finally, we urge expanded annotation of results that are sent to the provider to offer perspective on questionably significant results in stool analyses, such as finding C difficile in infants or diarrheagenic E coli other than those that produce Shiga toxin in people of any age in North America. These organisms are likely not pathogens, and current literature does not support antibiotic treatment when present.

Data offered by Cotter et al¹  will accelerate conversations about diagnostic stewardship in children suspected of having enteric infections. The topic will compel considerations of yield, risks of over- and underdiagnosing infections and over- and undertreating patients on the basis of the results, and individual and aggregate costs of molecular enteric tests, which are currently expensive. Public health implications of early notification (good) and nonprovision of pathogens for typing (bad) should also be considered. We look forward to more stewardship studies of molecular enteric testing to maximize the value of this powerful new technology.