Case Definition and Study Design

We conducted a retrospective review of the medical records of all POWV/DTV cases that were reported to the New York State Department of Health (NYSDOH) during 2004–2012. During this period, the NYSDOH initiated the use of diagnostic testing methods to detect this virus for all patients for whom arboviral testing was requested. We also reviewed published reports of 6 of these cases for additional details (10,11,13–15). POWV/DTV neuroinvasive infection was defined by using the 2011 United States surveillance case definition (16), which includes clinical criteria (fever of >38°C with any peripheral or central nervous system dysfunction documented by a physician and the absence of another more likely explanation) and >1 of the following:

• Isolation of POWV/DTV from, or detection of specific viral antigen or nucleic acid in, tissue, blood, cerebrospinal fluid (CSF), or another body fluid;

• A >4-fold change in POWV/DTV-specific quantitative antibody titers in paired serum samples;

• POWV/DTV-specific IgM in serum with confirmatory POWV/DTV-specific neutralizing antibodies in the same or a later serum specimen (POWV/DTV neutralizing antibody was considered specific if the titer was >4-fold higher than the corresponding neutralizing antibody titer to West Nile virus [WNV] or Saint Louis encephalitis virus [SLEV]);

• POWV/DTV-specific IgM antibodies in CSF and a negative result for other IgM antibodies in CSF for arboviruses endemic to the region where exposure occurred.

Demographic data were summarized by using descriptive statistics, mean and SD for continuous variables, and numbers and percentages for categorical variables. The study was approved by the Institutional Review Board at New York Medical College.

Laboratory Analyses

Serologic testing for POWV/DTV, performed at the NYSDOH Wadsworth Center in Albany, included a microsphere immunoassay to detect IgM and, separately, total antibodies (IgG + IgA + IgM) against recombinant DTV envelope protein in serum. The microsphere immunoassay also was used to detect IgM antibodies against recombinant DTV envelope protein in CSF. Recombinant DTV envelope protein was produced from the DTV-Ipswich strain, as described (17,18). Results were assessed as the ratio of the median fluorescence intensity (MFI) for 100 beads that reacted with the patient’s serum to the MFI of beads that reacted with a negative control serum specimen. The cutoff for a positive result was a value of 3 SDs above the mean MFI result based on a panel of serum specimens from healthy subjects.

Serum and CSF samples from 2 patients were also tested for POWV/DTV at the Centers for Disease Control and Prevention diagnostic and reference laboratory (Arboviral Diseases Branch, Fort Collins, CO, USA) by using an IgM antibody capture ELISA (MAC-ELISA) and IgG ELISA against POWV envelope protein (LB strain, Canada, 1958). A plaque reduction neutralization test (PRNT) against POWV (LB strain, Canada, 1958) (2) was performed at the Wadsworth Center and/or the Arboviral Diseases Branch by using BHK-21 cells. The antibody titer reported is the reciprocal of the dilution of serum that inhibited 90% of the test virus inoculum. PCR for POWV/DTV and genetic sequence analysis were done at the Wadsworth Center, as described (10,11). Serum samples submitted to the NYSDOH for arboviral screening were also tested for antibodies to WNV by using a MAC-ELISA and for antibodies to SLEV by using an indirect immunofluorescence assay; if these assays were positive, specific WNV and SLEV neutralizing antibodies were measured by using PRNT.

Demographic Characteristics

Figure 1

Figure 1. [[caption]]

Figure 2

Figure 2. [[caption]]

Fourteen cases of POWV/DTV encephalitis were identified in New York during 2004–2012; geographic locations by case-patient county of residence are shown in Figure 1. Ten (72%) case-patients were from Westchester, Putnam, or Dutchess Counties, which are located in the Lower Hudson Valley (LHV), a highly Lyme disease–endemic region. Three (21%) patients were <10 years of age; 10 (72%) were >60 years of age. The male-to-female ratio was 3:1. Eight (57%) patients had symptom onset in June, July, or August (Figure 2). Ten (72%) patients lived in a wooded area or reported outdoor exposure, and 8 (57%) had a pet (mainly a dog or cat). A tick bite was reported for 5 (36%) case-patients before illness onset, but only 3 remembered the exact date; incubation times from the tick bite to the onset of symptoms in these patients were 9, 11, and 32 days.

All case-patients were hospitalized. The mean ± SD time from onset of symptoms to hospitalization for 13 patients was 3.5 ± 1.9 (range 1–6) days. For the remaining patient, time from onset of symptoms to hospitalization was reported to be 2–3 weeks. Seven case-patients had been prescribed an oral antimicrobial drug (amoxicillin or doxycycline) within the few days preceding hospital admission: 4 for a febrile illness with a nonspecific macular or papular rash, 2 for erythema migrans (both lived in the LHV), and 1 for headaches and malaise following a tick bite.

Clinical and Radiographic Data

Fever (100%), generalized weakness (86%), and lethargy (72%) were the most commonly reported signs and symptoms (Table 1); 11 patients (79%) had a temperature >39°C at the time of hospitalization. Results of computed tomography scans of the brain were negative for acute findings for 13 (93%) case-patients (whether intravenous contrast was given is unknown). Results of magnetic resonance imaging (MRI) (with and without contrast for 4 patients, without contrast for 2, unknown for 8) showed acute abnormalities in 13 (93%) case-patients (mean ± SD time to MRI after hospitalization 2.9 ± 2.3 [range 1–10] days). Scattered T2-hyperintense foci predominantly affecting the gray matter and acute ischemic changes were the most common abnormalities found (Table 2). Among the 9 patients for whom electroencephalography (EEG) data were available, epileptic waveforms were documented for 3 (33%) and diffuse or severe slowing for 7 (78%) (mean ± SD time to EEG after hospitalization 9.3 ± 9.5 [range 2–30] days).

Laboratory Findings

The mean ± SD hemoglobin level for all case-patients was 12.3 ± 1.62 g/dL; 12 case-patients were anemic based on reference ranges for age and sex. Mean ± SD platelet count was 217 ± 86 × 10⁹/L (reference range 172–450 × 10⁹/L). With the exception of 1 patient who had chronic lymphocytic leukemia, the mean ± SD leukocyte count at admission was 9.36 ± 3.99 × 10⁹ cells/L (reference range 4.4–11.3 × 10⁹ cells/L). Aspartate and alanine aminotransferase enzyme levels were within reference ranges.

CSF analysis (mean ± SD time to initial lumbar puncture after hospitalization 3.3 ± 3.9 [range 0–15] days) showed a mean ± SD leukocyte count of 84 ± 88 (range 0–263) cells/mm³. Of the 13 patients who had CSF pleocytosis, 12 had a predominance of lymphocytes; 1 patient had a CSF leukocyte count of 68 with 65% neutrophils. Although CSF in 1 patient initially showed no leukocytes, repeat CSF testing 2 days later showed 891 leukocytes with 93% lymphocytes, 6% monocytes, and 1% neutrophils. The mean ± SD CSF glucose and protein levels were 86 ± 43 (range 39–179) mg/dL and 76.7 ± 28 (range 52–142) mg/dL, respectively. Results of Gram stain testing and bacterial culture of CSF were negative for all patients.

For 2 case-patients who were residents of Putnam County, diagnosis was confirmed by detection of a conserved region in the POWV nonstructural protein 5 gene by real-time reverse transcription PCR. A positive result was obtained from brain tissue (from an autopsy specimen) from 1 patient, collected on day 17 after hospitalization, and from a CSF sample from the other patient, obtained on day 9 after hospitalization. Genetic sequencing analysis demonstrated that both of these patients were infected with DTV (Table 3). These 2 patients were previously reported (10,11). For 5 other case-patients for whom CSF was available for molecular testing, viral RNA was not detected by the same method.

For 12 case-patients, the diagnosis of POWV/DTV infection was made by serologic testing alone. Of these, 11 had a positive test for IgM in serum to recombinant DTV envelope protein and evidence of neutralizing antibodies against POWV; case-patient 8 had a 4-fold increase in neutralizing antibodies against POWV between acute and convalescent serum samples, without detectable IgM. In total, 8 (67%) of the 12 patients had a >4-fold increase in POWV PRNT titers between acute and convalescent serum samples (Table 3). Four patients had weak-positive test results for serum antibodies against SLEV, but serum PRNT titers for this virus were negative for 3 patients and at least 4-fold lower than the PRNT titer for POWV/DTV in 1 patient. The 2 patients with erythema migrans had detectable antibodies to Borrelia burgdorferi, the causative agent of Lyme disease, by standard 2-tier testing.

Histologic examination of brain tissue was available for 2 patients, 1 from a brain biopsy and 1 from a specimen obtained at autopsy. In both cases, hematoxylin and eosin stained sections revealed a reactive gliosis, increased numbers of microglial cells, and necrotizing inflammation with a lymphocytic infiltrate, predominantly affecting the gray matter, consistent with acute meningoencephalitis. For 1 of the patients, who had confirmed DTV infection by molecular analysis, a detailed histopathologic analysis showed lymphocytic infiltrates, which in the leptomeninges and perivascular spaces contained predominantly CD4+ helper T cells and in the brain parenchyma predominantly CD8+ cytotoxic T cells (10).

Molecular testing for 12 viruses besides POWV/DTV, including enteroviruses, herpesviruses, and arboviruses, was performed on CSF to investigate other potential viral causes of encephalitis for the 14 patients (21). CSF real-time PCR testing showed Epstein-Barr virus infection (EBV) at 36.2 cycle threshold (C_t) (weak signal) and human herpesvirus 6 (HHV-6) at 29.62 C_t in case-patient 3 and herpes simplex virus 1 at 37.25 C_t (weak signal) in case-patient 11.

Hospital Course

All patients received intravenous antimicrobial drugs during the first week of hospitalization (ceftriaxone [92%], vancomycin [67%], ampicillin [42%], and acyclovir [86%]). Case-patient 13 received oral ribavirin and subcutaneous pegylated interferon α for 2 weeks (11). Of the 14 patients, 5 (36%) received intravenous corticosteroids during illness (dexamethasone 0.5–1.0 mg/kg/d for 10 days or methylprednisolone 500–1,000 mg/d for 3–5 days). Time of initiation of systemic corticosteroid therapy ranged from 1 to 45 days after hospitalization.

Twelve (86%) patients were admitted to the intensive care unit (ICU) during hospitalization (mean ± SD time from hospitalization to ICU admission [n = 11] 2 ± 1.6 [range 1–6] days). Seven patients (50%) required endotracheal intubation and mechanical ventilation (mean ± SD time from hospitalization to intubation 4.6 ± 1.8 [range 3–7] days), and 5 (36%) required tracheostomy and gastric feeding tube placement.

Outcomes

Patients were followed for a mean ± SD time of 66 ± 67 (range 10–240) days after the date of hospitalization. The mean ± SD hospital stay was 33 ± 22 (range 9–90) days. For patients admitted to the ICU for whom length of stay was available (n = 10), the mean ± SD ICU stay was 22 ± 17 (range 5–58) days.

Five (36%) patients died. One patient died in the hospital after withdrawal of life support according to the family’s wishes; 4 died after discharge from the hospital (mean ± SD time to death from the date of hospitalization, 116 ± 94 [range 13–240] days). Although none of the postdischarge deaths could be directly attributed to POWV/DTV infection, they did appear to be related to the severely impaired health status caused by the infection. All patients who died were >60 years of age. Both patients who had DTV infection confirmed by genetic sequencing died. All 5 patients treated with corticosteroids survived; 5 (71%) of the 7 patients who did not receive corticosteroids died. (Information on corticosteroid use was unavailable for 2 patients, both of whom survived.) All of the patients who were discharged from the hospital (n = 13) had neurologic deficits at the time of discharge (Table 4).