In Defense of Empiric Acyclovir Therapy in Certain Neonates
Article Outline
Abbreviations: HSV, Herpes simplex virus
In this issue of The Journal, Caviness et al1 report their attempt to estimate the prevalence of herpes simplex virus (HSV) infection among neonates hospitalized after evaluation in an urban Houston emergency department. In an accompanying editorial, Kimberlin,2 a renowned herpes virus expert, synthesizes the clinical applicability of the findings and summarizes the problems associated with empiric use of acyclovir and salient case features that would lead him to treat infants empirically.
See related editorial, p 155 and related article, p 164
Approaches among pediatric infectious diseases experts on empiric use of acyclovir in neonates is so varied as to leave recommending bodies with weasel words such as “Some experts
….” The situation is reminiscent of recommendations for use of dexamethasone along with antibiotics for treatment for presumed bacterial meningitis in older infants and children. In the case of meningitis, there are real data from prospective, randomized placebo-controlled trials, for which applicability can be debated. Smart people who are experts in pediatric infectious diseases, looking at the comparatively robust dexamethasone data, calculate the same benefit-to-risk ratio and yet assign weight to one or the other side of the equation differently—leaving consensus in the lurch. It is not a surprise, then, that “experts differ” at the tipping point for use of acyclovir for possible HSV therapy in neonates. We are grateful to Caviness et al1 for bringing data to the table and Kimberlin2 for framing the conversation. A real debate would require more data. We are at a stage of opinion.
The following is a single subspecialist's “take” on the data of Caviness et al, and an opinion.1 The first part regards incidence rates of HSV infection derived from an administrative database in which there are uncertainties and assumptions about both denominators and numerators. Appropriate to the goal of capturing all possible presentations of HSV infection and a robust sample size, investigators included all infants ≤28 days of age who were admitted to the hospital for any reason. Because only 20% of infants had either fever or hypothermia, the study cannot be characterized as, nor incidence figures applied to, “rule-out-sepsis” admissions. Only 50% of infants had a blood culture performed, and only 30% had any viral test performed. Investigators presumed that all cases of HSV infection would be captured because of severity of disease—although most of the study population had no HSV diagnostic test performed. In addition, it is not clear whether some infants received empiric acyclovir therapy, and no post-discharge follow-up was performed.
Relatively few neonates evaluated in clinical practice or in any study population have HSV infection. There is tyranny in small numbers. It is almost impossible to know accurate attack rate or to dissect the critically relevant case scenarios to be able to calculate positive predictive values, negative predictive values and numbers-needed-to-treat. The study by Caviness et al1 included 5817 neonates, over 5 years, yet netted only 10 cases of proven HSV infection. With all limitations of the administrative data method and caveats of inclusion criteria, the incidence of HSV infection (0.2%) was statistically (and arguably, clinically) not different from that of bacterial meningitis (0.4%). Consider the lengths to which clinicians go to diagnose and treat possible bacterial meningitis empirically pending exclusion by results of definitive tests. Why not HSV?
Kimberlin2 considers empiric evaluation and therapy for neonatal HSV warranted in any of the following settings: (1) There is clear index of suspicion of HSV (ie, skin vesicles, seizures, marked elevation of hepatic aminotransaminases); (2) A sepsis-like picture (including hypothermia) is present; and (3) The infant appears more ill than would be expected for another suspect diagnosis. Empiric therapy also might be warranted if cerebrospinal fluid mononuclear cell pleocytosis is found, outside of the enteroviral season.2
At the bedside, Kimberlin2 and Long would treat almost identically. Practicing in a facility with a primary isolation virus laboratory, polymerase chain reaction tests, and rapid antigen detection tests for multiple viruses performed routinely in infants hospitalized for acute illnesses, and desiring that doctors at the bedside who are not pediatric infectious diseases specialists make decisions, we cast a broader net. The Long approach supports evaluation and empiric acyclovir therapy for Kimberlin's2 scenarios 1, 2 and 3, plus for infants <21 days of age (ie, eliminating 27% of the unselected neonatal population of Caviness et al,1 and in practice eliminating a higher percentage of neonates admitted with any clinical syndrome possibly caused by HSV). The Long approach also would include scenarios 4, 5, and 6: cerebrospinal fluid mononuclear cell pleocytosis regardless of season, fever (≥38.0° C) without another clear diagnosis (eg, staphylococcal pustulosis, bloody diarrhea, lobar pneumonia, cellulitis/lymphadenitis, respiratory virus antigen-positive illness if result available immediately), and persistent or recurrent erythema or purulence/crusting at a former scalp electrode site. Although inclusion of infants <21 days old with fever may appear “extravagant” to some, most community-acquired viral illnesses in young infants such as respiratory syncytial virus, influenza. and enterovirus3, 4, 5 occur after 21 days when maternal antibodies wane and exposures accrue.
In this writer's experience and review of the literature, except for the rare severe central nervous system presentation of obtundation and seizures in the fourth week of life (in which case empiric therapy would always be given), the clinical presentation of perinatally acquired HSV occurs before 21 days of age. All 10 cases in the report of Caviness et al1 were <21 days of age. In fact, 9 of 10 infants were 8 to 14 days of age, leading some experts (personal communication, Michael Brady, Columbus Children's Hospital) to recommend empiric therapy for fever alone only in infants ≤14 days of age. Neither the Kimberlin2 nor Long scheme would have missed the 10 cases of Caviness et al,1 and Long undoubtedly would have given more empiric acyclovir therapy. But 10 cases are only 10 cases; 7 infants had skin lesions and the other 3 had hepatic necrosis. Neonatal HSV is “rare” in Birmingham (Kimberlin2) and “uncommon” (2-4 cases/year) in North Philadelphia (Long), frankly apparent in Houston but frequently more subtle in North Philadelphia. There may be real geographical differences—related to maternal age and prevalence of genital HSV, maternal prior herpes virus exposures, obstetric practices, access to care, and more. The opportunity to begin treatment before severe visceral or central nervous system manifestations appear is the driving force. With careful attention to hydration and rapid viral diagnostics/culture methods and alternative bacterial diagnoses leading to <48-hour acyclovir therapy for most HSV-uninfected infants, this writer is unaware of adverse effects of acyclovir. Squeamishness about discussing HSV with parents should not enter the decision-making process. HSV frequently is a subclinical infectious disease in parents but a devastating infection in neonates for which there is lifesaving therapy. HSV is not a judgment of parental behavior.
There are no right or wrong answers. There are individual opinions, flavored by knowledge of frequency and presentation of neonatal HSV disease in the community in which we serve, testing resources available, and personal level of comfort with exclusion of HSV infection and risks of treating uninfected infants with acyclovir and not treating infected infants early.
References
- The prevalence of neonatal herpes simplex virus infection compared with serious bacterial illness in hospitalized neonates. J Pediatr. 2008;153:164–169
- . When should you initiate acyclovir therapy in a neonate?. J Pediatr. 2008;153:155–156
- Diagnosis of enterovirus infection in the first 2 months of life by real-time polymerase chain reaction. Clin Infect Dis. 2003;37:1–6
- Diagnosis and outcomes of enterovirus infections in young infants. Pediatr Infect Dis J. 2005;24:546–550
- A prospective study of enterovirus infection in Thai infants presenting as sepsis. Asian Biomed. 2007;1:59–65
PII: S0022-3476(08)00381-8
doi:10.1016/j.jpeds.2008.04.071
© 2008 Mosby, Inc. All rights reserved.
Refers to article:
- When Should You Initiate Acyclovir Therapy in a Neonate?
- The Prevalence of Neonatal Herpes Simplex Virus Infection Compared with Serious Bacterial Illness in Hospitalized Neonates , 04 April 2008
