The study, approved by the University of Colorado Multiple Institutional Review Board, was conducted with all CSF clinical samples from pediatric patients submitted for herpes simplex virus (HSV) PCR to the Clinical Virology Laboratory at the University of Colorado from December 1998 through February 2000. When multiple specimens were submitted for one patient, only the first one was tested. Specimens positive for other microorganisms were not excluded. Peripheral blood specimens from these patients were not available to study. Information regarding demographics, clinical manifestations, diagnostic studies, management, discharge diagnosis, and outcome was gathered by retrospective chart review for patients seen at The Children's Hospital, Denver. CNS diagnoses and classification of seizures were based on the assessments of the primary treating physicians.

Infectious and postinfectious encephalitis were defined as the presence of encephalopathy or focal neurologic abnormalities, an abnormal CSF profile but negative CSF microbiologic studies, and a history or serologic result consistent with a current or preceding acute infectious illness. CSF pleocytosis was defined as >25 leukocytes x 10⁶/L for preterm neonates, >22 leukocytes x 10⁶/L for term neonates, and >7 leukocytes x 10⁶/L for all other patients. Infections not involving the CNS were classified as other infections.

HHV-6 PCR was performed (23) with the following primers and probes (24): 5´ AAG CTT GCA CAA TGC CAA AAA ACA G (17627–17603), 5´ AAC TGT CTG ACT GGC AAA AAC TTT T (17405–17429), and 5´ AAC TGT CTG GCA AAA ACT TTT (17516–17492). DNA was extracted from 50-µL aliquots of CSF previously stored at –70°C with Chelex purification matrix (Bio-Rad Laboratories, Inc, Hercules, CA). PCR was carried out in 50-µL mixtures containing 20 µL of extracted DNA, 1.25 units of PfuI (Stratagene, La Jolla, CA), 200 µmol/L of each of four deoxynucleoside triphosphates, and 0.5 µmol/L of each primer in PfuI buffer (Stratagene). The samples were amplified in duplicate for 45 cycles. The amplified DNA was separated according to molecular weight by using 3% agarose gel electrophoresis and transferred to nylon membranes. The identity of the DNA band (223 bp) was confirmed by detection with a digoxigenin-conjugated probe, antidigoxigenin antibody conjugated to alkaline phosphatase (Boehringer Mannheim Biochemicals, Indianopolis, IN), and chemiluminescent substrate (Tropix, Bedford, MA). Each patient sample was run in duplicate. Each assay included two negative controls (water) and two positive controls (100 genomic copies of HHV-6 DNA/reaction tube). The tests were considered valid if the controls yielded the expected results, and a specimen was considered positive if both replicates tested positive. Specimens with discordant replicates were reanalyzed in an independent run and considered positive if >50% of replicates gave positive results. The ability to detect HHV-6 DNA after at least four freeze-thaw cycles remained intact.

The analytical sensitivity of HHV-6 PCR was 8–10 copies of genomic DNA per reaction tube, as determined by serial dilutions of a sample with known DNA copy number. To rule out the presence of inhibitors in the extracted DNA, 20 HSV-negative CSF samples were spiked with the equivalent of 2 HSV DNA copies per PCR reaction tube. All samples yielded a positive result for HSV, whereas the unspiked specimens remained HSV-negative. The ability to detect HHV-6 strains from different geographic regions was confirmed by using 14 HHV-6–containing specimens obtained from diverse laboratories with recognized expertise in diagnosing HHV-6 infections as well as our laboratory. No cross-reactivity occurred with other DNA viruses, including other herpesviruses, parvovirus, adenovirus, and polyomavirus, thereby establishing the specificity of the HHV-6 PCR.

HHV-7 PCR was performed by using the methods described for HHV-6 with modified sample preparation. The primers were 5´ TAT CCC AGC TGT TTT CAT ATA GTA AC and 5´ GCC TTG CGG TAG CAC TAG ATT TTT TG, and the probe was 5´ AGA ATT CTG TAC CCA TGG GCA CAT TTG TAC (25; GenBank accession no. L03525). The PCR product was 186 bp long. The assay included two negative controls (water) and two positive controls (100 genomic copies of HHV-7 DNA per reaction tube).

The sensitivity of HHV-7 PCR was 8–10 copies of DNA per reaction tube, as determined by detecting HHV-7 DNA from samples extracted from infected cells. Specimens were prepared by boiling aliquots of nonhemorrhagic CSF for 55 s. If CSF contained blood, DNA was extracted with Qiagen DNA extraction kit (Qiagen, Valencia, CA). Testing with control viral DNA indicated that boiling did not affect the sensitivity of the PCR. The presence of inhibitors was ruled out by spiking 20 HSV-negative CSF samples with the equivalent of two HSV DNA copies per reaction tube and showing that all spiked samples became HSV-positive, whereas unspiked samples remained negative. The ability to detect HHV-7 DNA after at least three freeze-thaw cycles remained intact.

CSF samples obtained from 245 patients were tested for HHV-6 and HHV-7 DNA by PCR. Clinical data were available for the 218 patients hospitalized at The Children's Hospital, Denver (Table 1). The median age was 43 days; 68% were <2 years of age; 13% were 2–6 years, and 19% >6 years. Fever occurred in 56% of patients, rash in 11%, and seizures in 25%. CNS disorder diagnoses at discharge included meningitis in 61 patients (enteroviral, 21; bacterial, 4; aseptic/indeterminate etiology, 36), encephalitis in 21 patients (HSV, 4; acute disseminated encephalomyelitis, 4; infectious/postinfectious, 13), febrile seizures in 7 patients, and seizure disorder in 29 patients. A non-CNS disease had been diagnosed in the remaining 100 patients, which included a diagnosis of other infections in 60. Nineteen patients were immunocompromised (leukemia, 6; posttransplant, 3; premature, 3; HIV, 2; others, 5), of whom 1 had a febrile seizure, 1 had encephalitis, 2 had meningitis, and 5 had other infections. CSF pleocytosis was found in 90 (43%) of 209 patients for whom data were available.

PCR identified HHV-6 DNA in CSF of 3 of 245 patients (Table 2). Patient 1, a 5-day-old full-term male infant of a diabetic mother, had a generalized seizure on the day he was born. His seizure was attributed to hypoglycemia because his concurrent glucose level was 0 mmol/L. On day 5 of life, a fever of 38.3ºC developed for which he underwent a complete workup for sepsis. Results of a physical examination were normal with no signs or symptoms of neurologic dysfunction. His blood count was remarkable for a leukocyte count of 8,600 x 10⁶/L (46% neutrophils and 21% bands) and a platelet count of 40,000 x 10⁶/L. His CSF showed 40 leukocytes x 10⁶/L (20% neutrophils, 9% bands, 31% lymphocytes, 33% monocytes, and 6% macrophages), 4,110 erythrocytes x 10⁶/L, protein 169 g x 10^-2/L, and glucose 2.3 mmol/L. CSF HSV PCR, viral culture, and bacterial cultures of CSF and blood and urine were negative. Diagnosis at discharge was aseptic meningitis. The patient recovered spontaneously without sequelae.

Patient 2 was a 38-day-old full-term female infant who had a fever of 39.2ºC and irritability. The physical examination showed a few vesicular lesions on the tonsillar pillar as well as a diffuse, erythematous, macular blanching rash. CSF studies found 19 leukocytes x 10⁶/L (34% lymphocytes, 12% monocytes, and 55% macrophages), 1 erythrocyte x 10⁶/L, protein 48 g x 10^-2/L, and glucose 2.3 mmol/L. PCRs of CSF for HSV and enterovirus and bacterial cultures of CSF, blood, and urine were negative. She recovered spontaneously without sequalae. Her discharge diagnosis was aseptic meningitis.

Patient 3 was a male infant of 24 weeks' gestation with respiratory distress syndrome. A blood culture was taken on day 10 of life because of worsening respiratory distress; results were positive for Candida. A lumbar puncture was performed when the blood grew Candida. The patient did not have any neurologic symptoms or signs. CSF had 1 leukocyte x 10⁶/L, 1,780 erthrocytes x 10⁶/L, protein 105 g x 10^-2/L, and glucose 8.1 mmol/L. CSF studies, including PCRs for HSV and enterovirus, as well as bacterial, viral, and fungal cultures, were negative. He subsequently died of the complications of prematurity and candidal sepsis, which were unrelated to the CNS.

All 3 were <2 months of age, and 2 had CSF pleocytosis. Overall, HHV-6 DNA was detected in 2 (2.2%) of 90 specimens from patients with pleocytosis and in 2 (3.3%) of 61 patients with a discharge diagnosis of meningitis. HHV-6 DNA was not found in any patients who had encephalitis, febrile seizures, or a seizure disorder. None of the 19 immunocompromised patients had HHV-6 DNA in CSF.

HHV-7 DNA was not found in any of the 245 CSF samples tested by PCR. Spiking random samples with HHV-7 DNA generated positive PCR results.