Primary Versus Nonprimary West Nile Virus Infection: A Cohort Study

BACKGROUND Since 2001, we have observed patients with a clinical picture consistent with West Nile virus (WNV) infection, which was defined as nonprimary infection (NPI) owing to the presence of highly elevated serum immunoglobulin G antibody titers with a high avidity index (≥ 55%), absent or low titers of serum and cerebrospinal fluid (CSF) immunoglobulin M, and occasionally positive results of WNV-specific real-time reverse-transcription polymerase chain reaction analysis of CSF and/or blood specimens. METHODS We investigated 124 patients with a diagnosis of primary WNV infection (PI) or NPI during 2005-2007 at Sheba Medical Center (Tel-Hashomer, Israel). Logistic regression was used to evaluate the association of variables with PI and NPI and with in-hospital mortality. RESULTS A total of 68 and 50 patients with PI and NPI, respectively were included; 6 patients had incomplete data. In multivariate models, NPI was significantly associated with underlying psychiatric disorders (adjusted odds ratio [aOR], 13.73; 95% confidence interval [CI], 2.28-82.56; P = .004), hospitalization during winter and spring (aOR, 8.82; 95% CI, 1.59-48.87; P = .013), and fever (aOR, 0.61; 95% CI, .39-.95; P = .031). In-hospital mortality was significantly associated with NPI (aOR, 3.86; 95% CI, 1.12-13.28; P = .032) and a higher Charlson comorbidity index (aOR, 1.37; 95% CI, 1.03-1.83; P = .032). CONCLUSIONS The possibility that NPI may be an emerging clinical entity with a high mortality rate must be considered seriously.