- ED —
- emergency department
- HSV —
- herpes simplex virus
- LP —
- lumbar puncture
- PCR —
- polymerase chain reaction
In this issue of Pediatrics, Cruz et al1 address the following clinical question: what is the probability that an infant <60 days of age presenting to the emergency department (ED) who receives a lumbar puncture (LP) for concern of meningitis has a herpes simplex virus (HSV) infection? To answer this question, Cruz et al1 conducted a retrospective observational study and report on the incidence of HSV in infants <60 days old who had cerebrospinal fluid bacterial culture testing performed at 1 of 23 North American, tertiary-care, pediatric EDs over a 9-year period. There were 112 infants who had HSV identified, which amounts to 0.42%, or 42 of 10 000 (95% confidence interval, 35–51 of 10 000) infants evaluated. This estimate serves as the pretest probability of HSV infection in infants presenting to the ED who receive an LP for meningitis. On average, EDs saw ∼1 infant with HSV infection per year. Of neonatal HSV cases, 89.3% occurred in the first 6 weeks, with a median age of presentation of 14 days.
Cruz et al1 also bring attention to an important issue on the state of affairs of hospital management of HSV in young infants: there is substantial variation in HSV testing and acyclovir administration across EDs. HSV testing ranged from 14% to 72%, and empirical acyclovir administration ranged from 4% to 53% across EDs. Furthermore, 16% of infants who were subsequently diagnosed with HSV infection did not receive empirical acyclovir administration.
There are several important strengths of this study. Studies in which researchers report on rare conditions can only provide us with precise, meaningful estimates with large, multicenter collaborations in which researchers review data over years. Cruz et al1 reviewed ∼25 000 encounters of infants and relevant laboratory data across the 23 EDs. The retrospective and multicenter study design made this a relatively efficient study. The investigators used hospital charts for case verification and outcome measurement versus administrative data and billing codes. This reduces the chances of selection bias or bias from misclassification of outcomes. The study by Cruz et al1 also provides data on incidence that are clinically meaningful for ED practitioners. Interestingly, large population-based studies on the incidence and care patterns for neonatal HSV in the United States are missing. One recent report provides international estimates of neonatal HSV incidence at the level of different countries; however, mathematical modeling rather than actual cases is used to generate estimates.2
The timely diagnosis of infants with neonatal HSV is a challenge for the individual clinician and the health care system. Neonatal HSV infection carries a high mortality and morbidity rate, and early antiviral treatment is associated with better outcomes. This can be used to argue strongly in favor of having a low threshold to assess for it. On the other hand, neonatal HSV is a rare infection, as Cruz et al1 and others have shown, so the overwhelming majority of infants who are assessed (and presumably treated) for it will not actually have an HSV infection. Furthermore, neonatal HSV infection has multiple presentations, including without rash and/or fever, and has considerable overlap with the initial presentation of bacterial and other viral infections.3,4 Validated clinical prediction rules are lacking (and likely will not be developed because of the low incidence of this infection) as are point-of-care diagnostics that can rapidly guide management. So, it is no surprise that there is debate even among experts around the best approach to HSV testing and empirical acyclovir administration.5,6 Cruz et al1 have estimated that 237 infants who were <60 days old and received an LP in the ED would need to be treated empirically with acyclovir so that all infants who are diagnosed with HSV receive acyclovir at the initial encounter. However, with safety as a motive, the empirical treatment of all such infants with acyclovir exposes thousands of infants across the country to the potential toxicities of acyclovir, iatrogenic harms of a longer hospital stay while waiting for viral test results, and increased hospital costs. Cruz et al1 do not report on harms either experienced by infants on empirical acyclovir treatment or differences in harm rates among centers in relation to their rates of empirical testing and treatment. The number needed to harm is an important balancing measure to consider.
How do we improve care in the ED and hospital so that we identify and initiate empirical treatment in those infants who are at the highest risk of HSV infection and reduce the exposure and subsequent harms of testing and acyclovir treatment to those infants who are at the lowest risk? To begin with, we cannot rely solely on the education of clinicians on the diagnosis of neonatal HSV infection and the harms of overtesting and treatment. Pathways or protocols need to be implemented that can be used to guide clinicians on when testing and empirical treatment are appropriate. Those who create these pathways will have to rely on published case series and expert opinion to define the presentation characteristics of the at-risk group. Moreover, timely diagnostics, such as rapid HSV polymerase chain reaction (PCR) testing, need to be incorporated into these pathways when feasible, and these diagnostics must be assessed in this population (ie, we must assess neonates with HSV and not simply extrapolate results from adult populations).7 In addition to these recommendations, all infants who are assessed for neonatal HSV need to have the full workup performed on them. This includes cerebrospinal fluid and blood testing by PCR for HSV DNA detection as well as surface and lesion swabs for PCR and possibly viral cultures.8 For example, in the Cruz et al1 study, skin swabs for PCR or cultures were the most likely to have positive results, yet more than half of the infants with a central nervous system disease did not have skin specimens obtained.
Most importantly, centers need to collect outcome data after the implementation of their pathways (to understand diagnostic outcomes, harms, and costs for the entire denominator of infants) so that we can learn from their strategies. Researchers in recent publications describe such efforts and early outcomes.9,10 Cruz et al1 provide substantial data that underscore these paths forward and are to be commended on their detailed efforts to answer some of these important questions.
- Accepted November 2, 2017.
- Address correspondence to Sanjay Mahant, MD, Department of Paediatrics, Hospital for Sick Children, 555 University Ave, Toronto, ON, Canada M5G1X8. E-mail:
Opinions expressed in these commentaries are those of the authors and not necessarily those of the American Academy of Pediatrics or its Committees.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
COMPANION PAPER: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/peds.2017-1688.
- Cruz AT,
- Freedman SB,
- Kulik DM, et al
- Van TT,
- Mongkolrattanothai K,
- Arevalo M,
- Lustestica M,
- Dien Bard J
- American Academy of Pediatrics
- Brower L,
- Schondelmeyer A,
- Wilson P,
- Shah SS
- Copyright © 2018 by the American Academy of Pediatrics