Herpes encephalitis is a rare but devastating infection in premature infants. We report a 29 week gestation infant with severe intrauterine cutaneous and central nervous system herpes accompanied by hyperleukocytosis. Leukemoid reactions are not uncommon in this population, but the association of herpes encephalitis and a leukemoid reaction or hyperleukocytosis has not been reported previously.
Neonatal herpes simplex virus (HSV) encephalitis is associated with high morbidity and mortality even with prompt treatment.1 The most common presentation of neonatal HSV infection occurs when a woman delivers vaginally during a primary herpes outbreak. In such a case the newborns's symptoms usually appear at 2 to 3 weeks of life. Presenting symptoms include skin lesions (vesicles, pustules and/or papules), lethargy, fever, conjunctivitis, seizures, apnea and respiratory distress. HSV infection in premature infants may differ from that of term infants with a high incidence of respiratory distress and disseminated disease in one series, suggesting that viral respiratory cultures and serum transaminases may be helpful in diagnosis.2 In one series of 56 neonatal HSV infections, 50% of the infants were born prematurely and 70% of the mothers were asymptomatic.3 Intrauterine infection before the onset of labor and rupture of membranes is less common and usually consists of a triad of skin, eye and neurological findings present at birth with a poor prognosis.4, 5
Neonatal leukemoid reactions are not uncommon in ill neonates with prevalence estimates up to 15%.6 Historically leukemoid reactions have been defined as a leukocyte count >50 × 109 l−1 with no evidence of malignancy.7 A more recent definition is an absolute neutrophil count (ANC)>10 s.d.s above the mean for gestational age (that is, ANC>30 000 in the first 60 h of life or ANC>15 000 from 60 h to 28 days of life).8 Hyperleukocytosis is often defined as total leukocyte count >100 × 109 l−1. To our knowledge, neither leukemoid reaction nor hyperleukocytosis has been previously reported in intrauterine or neonatal herpes infection.
The patient’s mother was a primigravida woman with unremarkable maternal labs and history aside from a Group-B streptococcus urinary tract infection at 14 weeks gestation. There was no maternal history of genital herpes. The mother presented 4 days before delivery with preterm labor and received indomethacin for 2 days followed by magnesium sulfate, two doses of betamethasone, and prophylactic ampicillin and azithromycin. Her white blood cell (WBC) count on admission was 17.7 × 109 l−1 and decreased to 13.0 × 109 l−1 the day before delivery. She had no uterine tenderness or fever and no observed genital lesions. Rupture of membranes occurred 10 min before spontaneous vaginal delivery; the amniotic fluid was not foul-smelling. The patient was born at 29 weeks gestation with a birth weight of 1460 g. The baby was apneic at birth requiring intubation. Apgar scores were 1, 2, 2, and 7 at 1, 5, 10, and 15 min, respectively. The initial physical examination revealed a vesicular rash on the hands, feet, lower abdomen, axilla and nose (Supplementary Material Figure 1). Lesions were unroofed and swabbed for viral cultures and the patient was treated with intravenous ampicillin, gentamicin, and acyclovir. The initial WBC count was 84.7 × 109 l−1 (manual differential: 32% neutrophils, 9% bands, 35% lymphocytes, 4% metamyelocytes, 7% myelocytes and 10% promyelocytes, ANC 34 727). This was repeated 2 h later with an increase in WBC to 116.7 × 109 l−1 (manual differential: 31% polys, 12% bands, 20% lymphocytes, 9% metamyelocytes, 11% myelocytes and 15% promyelocytes, ANC 50 181). A double volume exchange transfusion was performed for hyperleukocytosis. Further laboratory evaluations are summarized in Table 1.
On day 2, a head ultrasound revealed diffuse hemorrhagic encephalomalacia and cerebellar hypoplasia (Supplementary Material Figure 2). Brain magnetic resonance imaging showed lissencepaly, diffuse hemorrhagic encephalomalacia, cortical ischemia and cerebellar hypoplasia (Supplementary Material Figure 3). An ophthalmologic exam revealed no evidence of chorioretinitis. On day 3, the HSV PCR result from the cerebrospinal fluid was reported positive. After extensive discussions about the poor prognosis for neurodevelopment, the parents requested that the baby be allowed a natural death. The patient was extubated and held by the parents; she died ∼5 h later.
The placenta weighted 370 g (>90th percentile for gestational age) and measured 15.5 × 17.2 × 1.7 cm. The placental membranes were yellow-tan and translucent. The umbilical cord was 26.2 cm in length × 1.1 cm in diameter, contained three vessels and inserted centrally. The fetal surface was tan-pink to dusky blue with no vesicular lesions. The maternal surface was red-brown, spongy with intact cotyledons. Microscopic examination revealed evidence of chorioamnionitis and funisitis with neutrophilic and lymphocytic infiltration and areas of necrosis (Supplementary Material Figure 4). No intranuclear inclusions were identified and immunohistochemistry stains for HSV 1 and HSV 2 were negative. The parents declined the request for a post-mortem examination of the infant.
A total of 11 days after the birth of the baby, the mother was re-examined by her obstetrician and a lesion was noted on the left labium majus; a culture was obtained that was positive for HSV-2.
This patient represents a devastating presentation of intrauterine herpes: an asymptomatic primary infection in the mother with ascending infection resulting in fetal encephalitis before rupture of membranes. Based on the severe white matter damage, it is likely that fetal infection began several weeks before delivery; the positive cerebrospinal fluid PCR and the extensive skin lesions most likely represent a recurrent infection. The dogma that amniotic fluid is generally sterile before rupture of fetal membranes has recently been questioned. Non-culture molecular technologies have demonstrated bacteria in 70% of fetal membranes at the time of elective Caesarean section at term.9 These and similar findings in premature pregnancies suggest that ascending invasion of the amniotic fluid from the vagina is common.10
Intrauterine HSV infection can occur because of either transplacental or ascending infection. The former is rare and would be supported by evidence of placentitis with intranuclear inclusions and HSV identifiable within the placental tissue, whereas the latter is more common and results in chorioamnionitis and funisitis with HSV identifiable in the fetal membranes. Placentitis, chorioamnionitis and funisitis have been reported in intrauterine disseminated HSV infection with intact membranes.11 The mild chorioamnionitis, lack of placental intranuclear inclusions and negative placental staining for HSV in this case suggest an ascending infection.
The unusual aspect of this patient's presentation was the hyperleukocytosis. The first case series of intrauterine HSV infections does not include a description of the WBC count.4 Case series of neonatal herpes encephalitis and disseminated disease have described normal or decreased total WBC counts. This mother received antenatal steroids that have been proposed as being associated with leukemoid reactions, however, this possible association has not been confirmed in subsequent case series and mechanistic studies.6, 8, 12 Given the initial elevation of the maternal WBC count and the microscopic analysis of the placenta, bacterial chorioamnionitis is a possible cause of the infant's high WBC count; however, all bacterial cultures were negative and there were no toxic granulations or vacuolization noted on the infant's peripheral blood smear. Chromosomal analysis was performed to rule out the possibility of leukemoid reaction related to mosaic trisomy 21 (result: 46XX).
Emergent leukapheresis or exchange transfusion has been advocated for hyperleukocytosis accompanied by respiratory distress or cardiovascular instability.13, 14, 15 Exchange transfusion is not recommended for the neonate with asymptomatic leukemoid reaction. A recent review of 1200 infants with birth weight <2500 g revealed 17 cases of leukemoid reaction with peak WBC count at 9 days of age. Neonates with leukemoid reaction had a 4-fold increase in sepsis, 20-fold increase in intraventricular hemorrhage, 54-fold increase in bronchopulmonary dysplasia and 6-fold increase in mortality compared with controls.16 The decision to proceed with double volume exchange transfusion in this patient was based on concerns of high blood viscosity contributing to the risk of neurological injury. This baby was at considerable risk of a poor neurodevelopmental outcome because of prematurity, low 10 min Apgar score and the prominent vesicular rash at birth. The added potential risk of hyperviscosity was felt to be of greater concern than the risks of the procedure.
The value of measuring total serum immunoglobulin (Ig) M from neonatal or cord blood has been questioned as a screen for congenital infection.17 The IgM response of the fetus to intrauterine HSV infection is unknown. In cases of perinatally acquired HSV, IgM levels rise 2 to 4 weeks after initial infection.18 It is noteworthy that the total IgM level was low in this premature infant in spite of an active and devastating intrauterine infection. It would have been helpful to have HSV2 IgM and IgG titers on the infant to assess whether the infant was able to produce antibiodies. The positive IgM but equivocal IgG titers for HSV2 in the mother is similar to a case report of intrauterine HSV2 infection wherein the primary infection was known to occur at 19 weeks gestation (elevated maternal IgM and normal IgG at that time) with a recurrence of infection at delivery 8 weeks later (minimal maternal elevations of IgM and IgG then).11
We present a case report of a premature infant with severe intrauterine herpes simplex encephalitis/encephalomalacia and hyperleukocytosis. We postulate that the elevated WBC count was triggered by the extensive herpetic central nervous system destruction, though an antenatal steroid effect or undiagnosed bacterial infection cannot be completely ruled out.
We thank John W Bishop MD, Director of Anatomic Pathology, UC Davis Medical Center, for review of the placental pathology and special staining. This work was financially supported by MU: NIH HD059127 (NICHD). This publication was also made possible by Grant number UL1 RR024146 from the National Center for Research Resources (NCRR), a component of the NIH.
About this article
Supplementary Information accompanies the paper on the Journal of Perinatology website (http://www.nature.com/jp)
The Indian Journal of Pediatrics (2014)