New Test, Old Dilemma: Distinguishing Viral From Bacterial Infections

Résumé fourni par la source

Differentiating viral from bacterial infections is one of the greatest challenges in pediatrics. As our awareness of the harms associated with antibiotic use in children has increased, the tension between the desire to avoid unnecessary antibiotics and the imperative to identify and treat invasive bacterial infections, including bacteremia and meningitis, in a timely fashion has escalated. Countless algorithms, guidelines, recommendations, and tests have been developed to support clinical decision-making around antibiotic use. Yet, we still struggle.In this issue of Pediatrics, Kalmovich et al report on the use of a host-protein biomarker test intended to distinguish viral from bacterial infections at urgent care centers.1 The test, MeMed BV (MMBV), is a rapid immunoassay that measures a viral-induced protein called tumor necrosis factor–related apoptosis-inducing ligand, interferon-gamma inducible protein of 10 kDA and C-reactive protein. MMBV incorporates these results into a score from 0 to 100, with scores of 0 to less than 35 indicating viral infection, 35 to 65 being considered equivocal, and 66 to 100 being consistent with bacterial infection. In prior validations, MMBV has been reported to have high sensitivity and specificity as well as high positive and negative predictive values for bacterial infections as compared with gold standards that include additional diagnostic tests and adjudication by experienced pediatricians.2,3 In this study, the authors sought to assess how this test impacts real-world decision-making by urgent care physicians.MMBV was implemented at 10 outpatient urgent care centers in Israel in conjunction with education about when to order the test and how to interpret its results. Clinicians were asked about their likelihood to refer to the emergency department (ED) and/or prescribe antibiotics at the time of ordering MMBV and, at the end of the visit, whether the test results influenced their decisions. Investigators evaluated antibiotic prescribing at the index visit, referral to the ED, and antibiotic prescription or hospital admission within 7 days of discharge.In more than 2000 encounters, MMBV was consistent with bacterial infection in 20%, equivocal in 12%, and consistent with viral infection in 69%. Across multiple comparisons, physician ED referral and antibiotic prescribing rates generally aligned with test results, without measurable harm. There was no difference in outcomes based on whether clinician prescribing aligned with MMBV results.This study was designed to understand how MMBV might impact real-world decision-making. Although MMBV use coincided with safe reductions in ED referrals, further comparative studies are needed to evaluate how it actually impacts outcomes in children, such as overall antibiotic use, antibiotic-associated adverse events, and hospitalizations. A recent randomized controlled trial in adults with lower respiratory tract infections (LRTIs) in whom clinicians were considering prescribing antibiotics did find that antibiotic prescribing was less frequent in the group with access to MMBV.4 Additional research is needed to understand whether the same will hold true in pediatric populations—and in which clinical scenarios—because adult LRTI is a very different, and much narrower, population than all children presenting to urgent care.Most children seen in urgent care (and in this study) have acute respiratory tract infections (ARTIs). These are most commonly viral, but there are 4 ARTIs for which antibiotics are sometimes indicated—acute otitis media, acute sinusitis, streptococcal pharyngitis, and pneumonia—for which evidence-based guidelines already exist to guide diagnosis.5–8 Future studies should evaluate whether incorporating MMBV into standard practice guided by high-quality evidence-based recommendations for these specific diagnoses improves outcomes. Other diagnostics have not always cleared that bar. For example, procalcitonin raised interest as a stewardship tool when studies in adults showed that it reduced antibiotic durations for pneumonia and sepsis.9 However, on closer inspection, the treatment durations in the control groups have often been longer than evidence-based guidelines recommend, suggesting that if clinicians followed evidence-based recommendations and used the shortest effective antibiotic courses, benefit of procalcitonin may be limited.10 Indeed, several studies in children have shown that procalcitonin does not significantly reduce antibiotic use, particularly in settings with active stewardship programs, and is not cost-effective.11,12Implementation strategies heavily influence the utility of diagnostic testing. In the present study, alignment between MMBV and clinician prescribing was 78%, but clinicians still prescribed antibiotics to 20% of children whose MMBV indicated viral infection. As the authors acknowledge, there may be many reasons for this, including parental pressure or other conflicting testing, but the benefit of any test will be limited if clinicians fail to act on its results. The approach to implementation can be as important as the test itself, as demonstrated by rapid blood culture identification tests, which decrease unnecessary antibiotics for gram-positive infections but typically only if implemented in conjunction with clear oversight and guidance from a stewardship program.13 Further work is necessary to identify best practices in implementation to maximize clinical impact of MMBV.Beyond optimizing uptake of novel diagnostic tests, it is crucial to prevent their use in scenarios wherein they offer limited value or may cause potential harm. In the present study, physicians were instructed that MMBV has not been studied in children with symptoms for more than 7 days, gastrointestinal tract infections, or a number of comorbidities, including immunosuppression, and were discouraged from using it for suspected tonsilitis, urinary tract infections, and skin infections. Nevertheless, 28% of the patients in whom MMBV was performed fell outside of the advised use cohort. At best, inappropriate test use generates unnecessary costs, but, at worst, it can lead to diagnostic error. Respiratory viral tests have faced similar overuse challenges. Their clinical utility lies in early identification of viruses that might benefit from treatment, such as influenza and RSV, or when identification of a virus might lead a clinician not to prescribe antibiotics when they otherwise would have. In reality, they are often used in patients with low likelihood of bacterial infection or when positive results do not dissuade clinicians from using antibiotics, adding cost without reducing antibiotic use. Their use has skyrocketed without added value, now becoming a frequent target of deimplementation efforts.14,15 Instead of addressing overuse after overly broad application, future efforts should focus on strategies to optimize MMBV use in appropriate cohorts, potentially using electronic medical record decision support to restrict use to validated scenarios.The authors argue that a test that helps support decision-making is valuable in this age of high decision burden and decision fatigue in urgent care centers and EDs. However, of the 80% of visits wherein MMBV influenced providers, in 64%, it supported the existing plan and only changed management in 16%. The question is whether, in an already crowded field of diagnostic tests, adding another one that rarely changes management will help relieve the burden of clinical decision-making or further muddy an increasingly complicated clinical picture while adding costs and (potentially) time. To ensure that MMBV adds value will require successful demonstration of its benefit compared with existing evidence-based diagnostic strategies and, if found to be useful, identification of implementation strategies that can optimize application in appropriate clinical scenarios and limit unvalidated use.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.