Enhanced surveillance for WN virus human disease in New York and New Jersey during 2000 was instituted to facilitate timely reporting of viral meningoencephalitis and to ensure rapid and accurate laboratory testing for WN virus. In NYC, encephalitis and viral meningitis are reportable conditions. From May to October 2000, the following measures were implemented by the NYC Department of Health to supplement existing passive surveillance: 1) Enhanced passive surveillance--to encourage physician reporting citywide, information on WN virus reporting and testing procedures was widely disseminated to the medical community through invited presentations by departmental medical staff, an agency publication mailed to >65,000 health-care providers, and biweekly broadcast facsimile and e-mail alerts to all NYC hospitals; 2) Hospital-based active physician surveillance-- neurologists, infectious disease consultants, intensive-care physicians, and chief medical residents at 18 sentinel NYC hospitals were called every 2 weeks to ascertain potential cases meeting clinical criteria; 3) Hospital-based active laboratory surveillance--laboratories at 12 sentinel NYC hospitals submitted all cerebrospinal fluid (CSF) specimens suggestive of a viral cause for WN virus testing at the NYC health department (CSF with negative Gram stain and culture with either a CSF leukocyte count >5/mm³ or protein >40 mg/dL). A special unit was established within the Communicable Disease Program of the NY City Department of Health to implement these additional surveillance activities. This unit ensured that all suspected cases were prioritized for next-business-day specimen collection and transportation to the city's Public Health Laboratories for WN virus testing.

In New Jersey, bacterial meningitis and encephalitis are reportable to the New Jersey Department of Health and Senior Services (NJDHSS). Active, hospital-based surveillance by infection control practitioners targeted patients admitted with the diagnoses of aseptic meningitis or encephalitis in 42 hospitals in six northern counties. Passive surveillance was enhanced through the distribution of reporting protocols, surveillance criteria, and WN virus educational materials to the medical community. Medical providers were reminded to notify the NJDHSS of suspected cases. NJDHSS conducted weekly follow-up with physicians and infection control practitioners to review the status of pending cases.

A patient was considered to have a confirmed WN case if he or she was hospitalized with an illness associated with neurologic manifestations consistent with meningitis or encephalitis, and had laboratory confirmation of WN infection. The four laboratory confirmation criteria used for WN infection, established by CDC (1), are as follows: 1) isolation of WN virus from, or demonstration of WN viral antigen or genomic sequences in tissue, blood, CSF, or other body fluid; 2) demonstration of immunoglobulin M (IgM) antibody to WN virus in CSF by IgM-capture enzyme immunoassay (EIA); 3) >4-fold serial change in plaque-reduction neutralizing antibody titer (PRNT) to WN virus in paired, appropriately timed serum or CSF samples; and 4) demonstration of both WN virus-specific IgM (by EIA) and IgG (screened by EIA and confirmed by PRNT) antibody in a single serum specimen.

Patients were classified into three clinical categories: meningitis, if they had fever plus headache, stiff neck, or photophobia; encephalitis, if they had altered mental status or other cortical signs; or meningoencephalitis, if they met both criteria. The categories of encephalitis and meningoencephalitis were combined as "any encephalitis" in some analyses. All syndromes required abnormal CSF findings consistent with a viral cause (CSF with negative Gram stain and culture with either a CSF leukocyte count >5/mm³ or protein >40 mg/dL).

IgM-capture EIA was performed at either the NY City Public Health Laboratories or the New Jersey Public Health and Environmental Laboratory; confirmation of positive results by PRNT was performed by CDC or the NY State Department of Health. Viral neutralization testing followed CDC protocol (R. Lanciotti, pers. commun.). Real-time Taqman polymerase chain reaction (PCR) was performed at CDC. Medical chart reviews and patient interviews were completed on all patients with positive tests for WN virus by IgM-capture EIA.

Supplementary medical chart reviews by the physician authors were completed in November-December 2000 after confirmation of initial results. Information abstracted included demographics, symptoms, chronology of illness, admission diagnosis, clinical findings, coexisting illness, laboratory findings, hospital course, diagnostic procedures, complications, level of neurologic involvement, discharge condition, and diagnoses. If a symptom was not specifically mentioned or a physical finding was not noted in the medical record, it was considered to be absent. Patient addresses were geocoded and mapped to compare the geography of the 2000 and 1999 epidemics. Descriptive statistics and Fisher's exact p values were calculated with SPSS (SPSS Chicago, version 10.0) and Epi Info (CDC, Atlanta, version 6.04b). Mapping was done in ArcView (ERSI, Redlands, CA, version 3.2).

Nineteen hospitalized adults were diagnosed with WN virus infection in the New York metropolitan area in 2000, a decline of 68% from 1999. The epicenter was located in Staten Island, approximately 20 miles west and south of the 1999 epicenter in northern Queens. Most patients had a febrile illness associated with meningeal signs, altered mental status, or both. The median age of hospitalized patients was lower in 2000 (63 vs. 71 years), and the proportion with encephalitis decreased from 63% to 58% (p=0.1). Gastrointestinal complaints were common, and severe motor weakness was reported less frequently than in 1999 (16% in 2000 vs. 27% in 1999). In 1999, seven deaths were caused by WN virus; in 2000 two were. The case-fatality rates for the 2 years do not differ statistically (11.9% in 1999, 10.5% in 2000, p=0.6).

Routine laboratory findings were nonspecific. CSF findings were consistent with a nonbacterial inflammatory process. Mild hyponatremia was found in eight patients. The syndrome of inappropriate secretion of antidiuretic hormone has been described in viral meningitis, St. Louis encephalitis, and WN virus (3-5). Two of eight patients with hyponatremia had other reasons for this finding (one with suspected dehydration and the other with hyperglycemia), and information on the use of antihypertensive medications, including diuretics, was not collected nor were urine osmolalities measured. A possible association of WN infection with this syndrome cannot be determined from this case series and requires further investigation.

Reasons for the differences seen in the number of human cases over the two epidemic years are speculative. Aggressive mosquito and larval control activities, particularly on Staten Island, may have reduced the infected mosquito population enough to diminish WN virus transmission to humans. Increased immunity in the resident avian population may have also prevented the re-establishment of an enzootic amplification cycle sufficient to cause significant human disease in Queens, the epicenter of the 1999 outbreak. Evidence from an avian serosurvey conducted after the 1999 epidemic supports this theory, since 51% of birds captured in Queens and 2% of those in Staten Island tested positive for WN virus (6).

WN virus infection in hospitalized cases in 2000 occurred over a 9-week period from mid-July to mid-September. A greater proportion of cases occurring outside Staten Island were recognized toward the end of this interval, which may relate to differences in vector ecology or control measures used. The timing of the 2000 epidemic curve closely resembles that of the recent outbreak in Romania (7) and preceded the 1999 New York epidemic curve by 2 weeks. The earlier onset in 2000 may have resulted from enhanced surveillance efforts that were not in place before the 1999 outbreak was recognized. In the 1996 Romanian outbreak, the predilection for WN virus to cause severe disease with increasing age and the frequency of gastrointestinal complaints were similar to findings in this series. Most encephalitis cases were in persons >50 years of age; vomiting occurred in 63% and diarrhea in 12% of cases (8). The propensity of WN virus to affect the elderly more seriously has been seen with other flaviviruses, most notably St. Louis encephalitis (9). The common contributing factors of age, hypertension, and previous brain insult may relate to a decline in the integrity of the blood brain barrier and facilitated access of WN virus to the central nervous system.

The interpretation of the findings of this case series is limited because of the small number of cases. Only hospitalized patients were included, and most WN virus infections are subclinical. Two additional nonhospitalized WN fever cases, one in Connecticut and one in New Jersey, were detected through surveillance and were not included in this case series (10). Focusing on the most severely ill obscures the true spectrum of WN illness. A 1999 serosurvey in Queens, New York, estimates that for every hospitalized case of WN virus infection there were 24 mild febrile and 110 subclinical illnesses (F. Mostashari, pers. commun.). Surveillance in 2000 focused on adults with aseptic meningitis or encephalitis; patients <18 years old were only included if they had encephalitis. The active laboratory surveillance component, however, included patients <18 years old. Four hundred fifty-three CSF specimens were received through active laboratory surveillance; 13 (3%) were from children <18 years. No positive results in children were found.

Another limitation was that the data described were abstracted from medical records that varied greatly in their completeness and legibility. The frequency of missing, missed, and omitted information was approximately 5%-10%. For some analyses, clinical information not located in the medical record was coded as negative, possibly introducing bias that could have produced spurious or hidden real associations.

WN virus appears to have established an enzootic cycle in the northeast United States with positive avian or mosquito findings extending from New Hampshire to North Carolina (11). Clinicians practicing along the eastern seaboard should consider this diagnosis when evaluating febrile patients during the summer months with neurologic complaints, especially those with a gastrointestinal prodrome or muscle weakness.